ll:llN:l:l OlVlOd lVNOllVNll:llNI

. .

The cover inset rem inds us

that amidst rapid adva nces
in the world o f technology
and information, traditional
agriculture continues.
It is the artful blending of the
best of tradition with
advanced technologies that
allows CIP to have maximum
impact on our changing world.
Photo by

M . lwanaga

Impact on a
Changing World
Program Report 1997-98

International Potato Center

International Potato Center

Apartado 1558
Lima 12, Peru
Cip@ cgiar.org
http ://www.cipototo.org
Impact on a Changing World
Program Report 1997-98
ISSN 0256-6311
September, 1999, 2.25 M
Copyright International Potato Center 1999
The International Potato Center. 1999.
Impact on a Changing World Program Report 1997-98. Lima, Peru. 458 p.

Contents
Impact on a Changing World

Program Overview
Wanda Collins, Deputy Director General for Research
The Role of Regions in Cl P's Research
CIP's Research Project Portfolio
Feature Article: Incorporating Pove rty in Priority Setting: CIP's 1998-2000
Medium Term Plan
T. Walker and M. Co/lion

7
11
15
21

Research on Potato
Characterization of Phytophthora infestans Populations in Peru
W. Perez, S. Gamboa, M. Coca, R. Raymundo, R. Hijmans, and R. Nelson
Genetic Diversity among Isolates of Phytophthora infestans from Various Hosts
in Ecuador
L.}. Erselius, H.R . Hohl, M.E. Ordonez, PJ. Oyarzun, F j arrin, A. Velasco,
M .P. Ramon, and G.A. Forbes
He .::: Specificity of Phytophthora infestans on Tomato and Potato in Uganda
and Kenya
L.J. Erselius, M.). Vega -Sanchez, A.M. Rodriqu ez 0. Bastidas H.R . Hohl,
P.S. Ojiambo 1 }. Mukalazi, T. Vermeulen, WE. Fry, and G.A. Forbes
Genotype by Environment Reaction of Potato to the Late Blight Pathogen
G.A. Forbes
Identification of QTLs for Late Blight Resistance in a Cross Between 5. phureja
and a Di haploid 5. tuberosum and Association with a Plant Defense Gene
M. Ghislain, B. Trog nitz 1 R. Nelson, Ma. def R. Herrera, L. Portal, M. Ori/lo,
and F Trognitz
Parasitic Fitness and Temperature Response of New Lineages of Phytophthora
infestans from Peru
J.L. Andrade Piedra, G.A. Forbes, WE. Fry, and R. Nelson
Estimating the Global Severity of Potato Late Blight with a GIS-Linked
Disease Forecaster
R.}. Hijmans, G.A. Forbes, and T.S. Walker
Implementing IPM for Late Blight in the Andes
R. Torrez,}. Tenorio, C. Valencia, R. Orrego, 0. Ortiz, R. Nelson,
and G. Thie le
Understanding Farmers' Responses to Late Blight: Evidence from Peru, Bolivia,
Ecuador, and Uganda
0. Ortiz, P. Winters, H. Fano, G. Thiele, S. Guaman, R. Torres, V Barrera,
}. Unda,}. Hakiza
Sensitive Detection of Ralstonia solanacearum in Latently Infected Potato Tubers
and Soil by Postenrichment ELISA
S. Priou, L. Gutarra, H. Fe rnandez, and P Aley
1

31

39

49

57

67

77

83
91

1 01

111

An Improved Method for Fighting Bacterial Wilt: NCM-ELISA Detection Kit

and Training Materials
Survey of the Durability of Extreme Resistance to PVY Derived from So/anum
tuberosum subsp. andigena
E. Mih ovilovic h, L.F Sa lazar, F Saguma, and M . W. Bo nierba le
The Leafminer Fly in Potato: Plant Reaction and Natural Enemies as Natural
Mortality Factors
F Cis neros and N. M uj ica
Taxonomy and Bionomics of the Andean Potato Weevil Complex: Premnotrypes
spp. and Related Genera
}. A lcaza r and F. Cisneros
Population Dynamics of the Andean Potato Tuber Moth, Symmetrischema
tangolias (Gyen), in Three Agroecosystems in Peru
M . Palacios, }. Teno rio, M . Ve ra, F Zevallos, and A. Lagnao ui
Assessing Bt-Transformed Potatoes for Potato Tuber Moth, Phthorimaea
operculella (Zeller), Management
V Cane do, }. Be navides, A. Go lmirzaie, F, Cisneros, M. Ghisla in, and
A. Lagnao ui
Participatory Needs and Opportunity Assessment for Potato IPM Development
Planning: The Case of Indonesia
E. va n de Flie rt, Wars ito, and A. Lagnaou i
Interaction of Maternal Clones of Potato (Solanum tuberosum) with Pollinator
Clones of S. phureja for Induced Parthenogenesis
M O. Upadh ya an d R. Cabe llo
A Proposed Solanum tuberosum subsp. andigena Core Collection
Z. Hu am an, R. Ortiz, an d R. Gomez
Rustic Seedbeds: A Bridge Between Formal and Traditional Potato Seed Systems
in Bolivia
G. Aguirre, }. Ca lderon, 0 . Buitrago, V Iriarte, }. Ra m os, }. Blajos,
G. Thie le, and A. Deva ux
Intensification of Potato Production in Rice-Based Cropping Systems: A Rapid
Rural Appraisal in West Bengal
S.K. Ba rdhan Roy, T. Wa lke r, VS . Khatan a, N. K. Sa ha, V S. Ve rm a,
M .S. Kadian, A.}. Have rko rt, and W. Bo wen
On-farm Profitability of TPS Utilization Technologies
A. Chil ver, T. Wa lker, V Khatana, H. Fan o, R. Suh e rm an, and A. Rizk
Globalization Takes Root: Potato Trade in Latin America
CJ Scott and L. Maldo nado
The Economics of Generating International Public Goods from Investing in
Potato Plant Breeding
T. W alker and K. Fug li e

1 22

123

129

14 1

153

1 61

17 1

1 79
1 85

195

205

2 13
221

229

Research on Sweetpotato
The Causes and Control of Virus Diseases of Sweetpotato in Developing
Countries: Is Sweetpotato Virus Disease the Main Problem?
E.E. Ca rey, R. W. Gibson, S. Fu entes, M. Machmud, R.0.M. M wanga,
G. Turyamuree ba, L. Z hang, 0. Ma, F Abo EL -Abba s, R. EL -Bedewy, and
L.F Salazar.
Economic Impact of Virus-Free Sweetpotato Planting Material in Shandong
Province, China
K. O. Fug lie, L. Zh ang, L. Salaza r, and T. Wa lke r

24 1

249

Biological and Selective Control of the Sweetpotato Whitefly, Bemisia tabaci

(Gennadius) (Hom.: Aleyrodidae)
F. Cisneros and N. M uj ica
An Efficient System of Embryogenic Suspension Cultures and Plant Regeneration
in Sweetpotato
Q.C. Liu, H. Z hai, O.H. Lu, Y Wa ng, and D.P. Z hang
Expression of Soybean Proteinase Inhibitor in Sweetpotato
C. Cipriani, 0 . Mic haud, F Brun elle, A. Colmirzaie, and O. P. Z hang
Starch Content and Properties of 106 Sweetpotato Clones from the World
Germplasm Collection held at CIP, Peru
C. Brabet, 0 . Reynoso, 0. Du four, C. Mestres, }. Arredondo, and C Sco tt
Potential of Sweetpotato in Reducing Vitamin A Deficiency in Africa
V Ha genimana, L. M. K'osambo, and E. E. Carey
Development of a Sweetpotato-Based Instant Weaning Food for Poorly
Nourished Children Six Months to Three Years Old
N . Espinola, H. Creed- Ka nashiro, ME. Ugaz, M. van Hal, and C. Scott
AFLP Assessment of Sweetpotato Genetic Diversity in four Tropical
American Regions
O.P. Z hang, M . Chislain, Z. Huam an, J.C. Cerva ntes, and E.E. Ca rey
Introduction and Use of Exotic Germplasm in the Chinese Sweetpotato
Breeding Program
O .F Ma, H. M. Li, and J.C. Mo k
Farmer Maintenance of Sweetpotato Diversity in Asia: Dominant Cultivars and
Implications for In Situ Conservation
C. Prain and 0. Ca m p i/an
Sweetpotato for the New Millennium: Trends in Production and Utilization in
Developing Countries
C .}. Scott and L. Maldo nado
Sweetpotato in Ugandan Food Systems: Enhancing Food Security and
Alleviating Poverty
C .}. Sco tt, }. Otieno, S.B. Ferris, A. K. Muganga, and L. Ma ldona do

255

265
271

279
287

295

303

311

31 7

329

33 7

Research on Potato and Sweetpotato

In Vitro Conservation of Potato and Sweetpotato Germplasm
A. C o lmirza ie and }. To ledo
Global Proj~ctions for Potato and Sweetpotato to the Year 2020
C .}. Scott, M. Rosegrant, C. Ring ler, and L. Maldonado
The role of Potato and Sweetpotato in Disaster Relief: The Case of Rwandan
Refugees in South Kivu, Democratic Republic of the Congo (ex-Zaire), 1994-96
M. Tanganik, P. Ph ezo, P. T. Ewell, N. B. Lu ta ladio, and C .}. Sco tt

351
357

365

Research on Natural Resource Management in the Andes

Estimating Frost Risk in Potato Production on the Andean Altiplano Using
Interpolated Climate Data
R.J. Hijm ans
Simulating the Response of Potato to Applied Nitrogen
W. Bo wen, H. Cabrera, V. Ba rrera, and G. Baigorria
Se lecting Optimum Ranges of Technological Alternatives by Using Response
Surface Designs in Systems Analysis
C. Leon-Ve larde and R.A. Q uiroz

373
381

387

Mapping Aquatic and Semiarid Vegetation in the Andean Altiplano Using

Multichannel Radar Imagery
R. Quiroz and S. Saatch i
A Process-Based Model (WEPP) for Simulating Soil Erosion in the Andes
W. Bowen, C. Baigorria, V. Barrera,}. Cordova, P Muck, and R. Pastor
Ecoregional Research: A Vision from the Andes

395
40 3
40 9

Research on Andean Roots and Tubers

Reappraisal of Edible Canna as a High-Value Starch Crop in Vietnam
M. Herma nn, N. K. Quyn h, 0. Peters
Compositional Diversity of the Yacon Storage Root
M. He rma nn, I. Freire, C. Pazos

415

Training in 1997-98
Selected Publications from 1997-98
Staff in 1997-98
Acronyms and Abbreviations

433
439
447
455

42 5

Program
Overview
Impact on a Changing World

Poverty eradication, increasing food productivity, and conservation of natural resources

continue to be the primary goals of CIP's research . The last few years have brought
revolutionary changes in the way our researchers can work towards those goals. Biological
and information sciences have expanded explosively, providing new tools and new
methods, both for conducting research and for managing and using the wealth of information that results . A heightened awareness of the wisdom and value of involving end-users
early in the process of technology development has evolved. The resulting interactions
increase the applicability and acceptance of our work, and help us to unde1stand the
sociological and economical context for appropriate technology development.
CIP has reacted to the changing world of research by redesigning its research program.
We continue to concentrate on potatoes and sweetpotatoes, two of the world's most
important food crops; management of natural resources, particularly in the high mountain
areas of the world; and on the conservation and preservation of a group of underutilized
but potentially valuable Andean root and tuber crops. Research is now organized in 17
well-designed projects that operate through a team approach and in close conjunction
with partners in advanced research institutes, in national programs of developing countries, and in the field with farmer collaborators.
Research management is decentralized with most decisions made at the project level,
increasing our flexibility to reach project goals. At one end of the spectrum, the new
project structure allows us to take advantage of rapidly developing scientific and information innovations to shorten the technology development process. At the other end of the
spectrum, we bring our research processes closer to the field and to farmers through
participatory methods that involve them directly in the discovery process. These diverse
partners provide us a wide range of expertise from lab to land and help us to use our
I imited resources to the best advantage. A mutual environment of cooperation and trust is
the result. Project-level decision-making, guidance from the newly formed Program
Management Team, and dialogue with our many research partners have created a synergy
that greatly enhances CIP's research.
CIP's 1997-98 Program Report presents the first research progress repmts from the 17
newly created projects. Forty-six research reports are grouped into the broad categories of
Potato, Sweetpotato, Potato and Sweetpotato, Natural Resource Management in the Andes,
and Andean Roots and Tubers. A full listing of the 17 projects contributing to this Program
Report follows this overview.
In addition, we present a feature article "Incorporating Poverty in Priority Setting: Cl P's

1998-2000 Medium Term Plan" by T. Walker and M. Coll ion. This article describes the

CIP Program Report 1997-98

process by which C l P's proj ects were chosen based on a pr io rity-setti ng mechanism that
focuses o n th e poverty alleviat io n poten ti al of each pro ject.
Th e fol low ing description s give a sketch of th e research a1ticles w ithin thi s repo rt.

Potato
Research in potatoes du ring 1997-98 emphasized finding a solution to the increasing late
blight prob lem. Late blight is the wor ld ' s most costly pl ant disease in terms of lost production, c hem ica l inpu ts, and environmental co nta m ination. It is parti cul arl y devastating to
smal l, resource-poor farme rs. C IP 's researc h fe atures an integ rated appro ach: from highly
technical mo lecu lar geneti cs to help in g farme rs learn in th e field about recogn iz in g and
manag in g the dis ease; and from wo rkin g wi th th e microscop ic path oge n wh ich ca uses th e
disease to breed ing new resistant cul tivars. The integrated, fu ll spectrum approac h is
necessary to provide sol utio ns that farmers understand and use. Resea rch reports co ver
pathogen bio logy and eco logy, geneti cs of res istance to late blight, breeding for resistance ,
and integrated manageme nt of the di sease. Progress in using molecular genetics, soc ioeco nomi c perspect ives, and geog raph ic inform at io n systems, and part icip atory research
meth ods are repo rted an d lay the fo und at io n fo r bette r contro l of late b li gh t w ithin the next
few yea rs .
W h i le co ntrol l ing the late b li ght disease beca me a ve ry
hi gh pr iority in C IP beca use of the mag nitude and urgenc y of
th e probl em, resea rch progress w ith other serious co nstraints
facing small , poor producers is also cove red in this vo lum e,
in c ludin g:
Deve lopm ent of a sensit ive an d affordable techno logy to
detect Ra lsto nia solanacearum, th e o rga nism w hi c h
causes bacteria l wi lt in soil and tubers.
Taxonomic , popu lat io n, and behavioral informat io n
cr itica l for in tegrated management of three of the most
se riou s potato pests in th e wor ld: A nd ea n potato weevi l,
potato tuber moth , an d leafm in er fly.
Establishm en t of a co re co ll ection of potato ge neti c
resources so th ey ca n be used more effic ientl y.
Us ing potato in intens ified rice-wheat produ ctio ns for
di versi ty.
Eco nomic studi es of potato research cove rin g true potato
utilization tec hn o log ies, global iza ti o n of mark ets, and
in vestin g in potato breeding.

Improving Bacterial Wilt detection

(see pag e 122).

Sweetpotato
Sweetpotato cont inu es to be an im porta nt foo d crop in the wor ld , and in c reasing ly so in
Africa after th e reduc tion in cassava production due to disease outbreaks. C IP's h ig hest
priority fo r sweetpotato rese arc h is in deve lopin g and mak in g ava il ab le c lo nes w ith the
potent ial for yie ldin g h igh leve ls of dry matter per hectare. Hi gh dry matter is esse ntial in
every area of th e develo ping wor ld for sweetpotato to reach its fu l l potential for utilizati o n,
w hether it is for an im al feed, hum an consum pti o n, o r in dustria l proc ess in g.

Prog rnmOverview

Preliminary re su lts indicate that rapid progress is being

made in in c reas ing dry matter. In this Program Report,
progress is reported in a numb er of areas of direct importance to the introdu ct ion and utilization of new high dry
matter c lon es. Research in cludes significant progress in the
causes and controls of v iru s diseases and utilization to
reduce nutritional deficits such as Vitamin A, especially
among women and children. Trends in the production and
utilization ~f sweetpotato in developing countries are
reported as wel I as spec ifi c uses in food systems. Much of
the research repo rted in this vo lum e is focused on learning
more about sweetpota to germp las m. Studies report molecu lar assessme nt of germp lasm for di versity, survey in g
germp lasm for food quality traits , and determining how
farme rs help to maintain diversity in the field. Knowledge of
th e diversity hidden in sweetpotato germp lasm is essential
for maximizing its potenti al use.

Natural Resource Management in the Andes

Harvesting in situ sweetpotato

germplasmin the Phillippines
(related story begins on page 317).

Th e Andean ecoregion is a particularly vu ln erabl e ecosystem for agr icu lture. CIP's natural
resource management program focuses on providing sound sc ient ifi c, technical , and
economic bases for policy and techno logy decisions. Resea rch reported here focuses o n
assess in g the risks associated w ith production in these fragi le agricu ltural land s through the
development and use of advanced technological modeling approaches. The development
and app li cat ion of tools and methods to enhance natural resea rch management research
capacity, including the use of microwave remote sensing, G IS, and multichannel radar
im age ry, is also reported. These too ls and methods are app li ed at different spatial sca les,
from the field level to the landscape and regional leve ls.
Natural resource management in the Andes exemplifies an eco reg ional approach to land use ma na gement for
sustainab le agriculture. This ecoregiona l approach,
wh ich uses the tools and methods mentioned above for
research and development of management options, is
implemented in the Andes through the Consortium for
the Sustainable Development of the Andean Ecoregion
(CON DESAN). CONDESAN is also the Andean
Ecoregion partner of the Global Mountain Program, a
more exte nsive res ea rch program for the high mountain
areas of the world, convened and led by CIP. The Global
Mountain Program represents the CG IAR's contr ibuti on
to meeting the objectives of the UNCED acco rd for
sustainab le env ironmenta l development of mountain
areas and involves other international agricu ltural
research ce nters within the CG IAR conducting similar
research in the African Highlands and the Himalayas.

CON DESAN works in the Andes, one of

the most challenging agroecosystems
on earth (see page 409).

CIP P10g10rn Repo111997-98

Andean Roots and Tubers

A sma ll but important part of Cl P' s researc h
program is devoted to a group of
und er utili zed hi gh potent ial Andean root
and tuber cro ps. Th e focus is on p1eserv ing
the genet ic d ive rsity throug h both in- situ
and ex-situ co nservat ion method s, and by
in creas ing th e demand and utili zati on of
th e crops . Resea rch res ults are in c lud ed
wh ich exp lore th e pote nti al utili zat ion of
two of th ese crops: ca nn a and yaco n.

The first step in canna noodle production (see page 415) .

Cl P' s co ntinuin g goa l is to achieve its

obj ecti ves of co ntributin g to pove rty eradication through the deve lopme nt of envi ro nmentall y pos iti ve product ion tec hnol og ies th at increase productivity and sustain abl e li ve lihoods for deve lopin g co untry farmers. Th e resea 1c h 1esults repo 1ted in this editi o n of the
CIP Prog ram Report bring us close r to th at goa l and ass ure that CIP co ntinu es to have a
stron g and meas urabl e impact on o ur changing wo rl d.

Wanda Collins
Deputy Directo r Gen eral
for Research

l0

Progrom Overview

The Role of Regions

in CIP's Research
Imp act on a Changing World

Research through Partnership:

CIP's Regional and Liaison Offices
Increase Impact
The potential impact of CIP research is magnified by the posting of scientists in five
regional and nine liai son offices* throughout the developi ng world. Working side by
side w ith developing cou ntry research partners and keeping a users' perspective in
mind enab les new technologies to be user driven and quickly adopted, decreasing the
time lag for impact to be seen, felt, and measured. It also allows priority setting to be a
dynamic and participatory process reflecting local needs and constraints.
Potatoes and sweetpotatoes are grow n in many deve lopin g cou ntri es of the
world. Ecosystems, production systems, and food systems vary tremendo usly,
howeve r, makin g the adaptation of tec hnology to local conditions both necessary
and va lu ab le. Additionally, in its natural resource management work, CIP researc hers work on a scal e that encompasses a wide range of site spec ifi c ities. Characterizin g these systems and spec ificiti es and understanding underlyin g mec hanisms is a
necessary part of technology development and impl eme ntin g change . However,
the impact of new management technologies for this reso urce base will only be
rea li zed through workin g w ith communities and loca l age nci es who can bring
about the large-sca le and inter-related changes necessary to protect and conserve
these very frag il e produ ction zones. Thi s would not be possibl e w ithout reg ionaland li aison office-based researche rs.
CIP's decentralized organ ization enab les res earc hers to develop technologi es
w here th ey w ill be most readily used, and with th e people w ho w ill most quickly
adopt them . Increasi ngly, thi s includes farmers, co mmunities, and loca l non governmenta l orga nization s. Howeve r, intercha nge of tec hnologies also occurs across
regions. The new team-oriented project structure creates an environment for rapid
spillover to sha re successes.

* The Vietnam

liaison office opened in mid-1999, and does not appear in th e accompanying map.

CIP Program Report 199/.98

11

The importance of CIP's reg ional and liaiso n office location s is obvious in this
Program Report. More than half of th e papers report on 1esea rc h conducted w here
CIP has regional and counuy offices, and many loca l collaborators are includ ed as
autho1s.
Location and contact in fo rmation for all CIP offices are in c luded below .

Cl P Headquarters
Lim a, Peru
E-mail: cip@cg iar.org
Website: www. cipotato.org

Latin America and the Caribbean

(LAC) Regional Office
Lim a, Peru
Fern ando Ezeta, Reg ional Representati ve
E-mail: cip@c ipa.org .pe

LAC Liaison Offices

Ecuador (Quito)
Charles Crissman , Liaison Scientist
E-mail: cip-quito@cg iar.org
Bolivia (Cochabamba)
Graham Thiele, And ea n Potato Proj ect,
hosted by Fund ac i6n PROI N PA
E-mai I: g. th iele@cg iar.org

Sub-Saharan Africa (SSA) Regional

Office
Nairobi, Kenya
Peter Ew ell , Region al Representat ive
E-mail: cip -nbo@cgiar.org

SSA Liaison Offices

Cameroon (Bamenda)
Joseph Koi , Liaison Office Manager
E-mai I: cip-bamenda@cgi ar.org
Uganda (Kampala)
J. 0. Oti eno
E-mail: jot ieno@imul.com

South and West Asia (SWA)

Regional Office
New Delhi, India
Sarath llan ga ntileke, Region al
Representati ve
E-mail: cip -de lhi @cg iar.org

12

Role of Regions

SWA Liaison Offices

Pakistan (Islamabad)
Oscar Hidalgo, Project Leader
E-mail: o.hidal go@cgiar. org
Nepal (Lalitpur)
Deepak N. Ojha , Team Leader/ Nepal
E-mai I: potato@pps.wlink .com.np

East and Southeast Asia and the

Pacific (ESEAP) Regional Office
Bogor, Indonesia
Gordo n Prain , Reg ional Representati ve
E-mai l: cip-bogor@cgiar.org
Webs ite: www2.bonet.co .id/ cip
ESEAP Liaison Offices
China (Beijing)
Yi Wan g, Liaison Scienti st
E-mail: cip-chin a@cgiar.org
Philippines (Manila)
Di nd o Carn pi Ian, Liai son Scienti st
E-mail: c ip-m anila@cg iai-.o rg
Vietnam (Hanoi)
Dai Peters, Li aison Sci entist
E-mail: cip-h anoi@fpt.vn

Central and Eastern Europe,

Transcaucasia and Central Asia
(ECA) Regional Office
Berlin, Germany
Peter Schmiedich e, Coordinator
E-mai I: p.schm ied iche@cg iar.org

CI P Regional and Liaison Offices

Central and Eastern Europe,

Trans-Caucasia and Central Asia

(ECA)
South and West Asia

(SWA)

Sub-Saharan Africa

(SSA)

Latin America and

the Caribbean

East and South Asia

and the Pacific

(ESEAP)

(LAC)

CIP Program Report 1997-98

l3

CIP's Research
Project Portfolio
Impact o n a Chan g ing Wo rld

1 Integrated control of late blight (R. Nelson)

Late Blight is CIP's highest priority at the moment and this project is highly focu sed on
deve loping, adapting, and integrating technologies for th e management of the world's
worst agricultural crop disease. Th e project integrates biotechnologica l tools for pathogen
epidem iology and developing disease resistance with disease management and control in
farmers' fields through th e Farmer Field Schoo l concept.
Pathogen population bio logy and eco logy (G. Forbes)
Deve lopme nt and app li cat ion of methods for ph enotyp ic characterizat io n of res ista nce
(G. Forb es)
Develo pm ent of tool s: maps, mark ers , ge nes, and databases (M . Bonierba le)
Mark er-ass isted genetic anal ys is of res ista nce (M. Ghis lain )
Diploid pre-breeding for LB resistan ce (B. Trogn itz)
Intermediate breedin g fo r LB re sista nce at the tetrapl o id leve l (J. Land eo)
Gen et ic eng ineerin g for res ista nce to late blight (M. Ghis lain )
Analysi s of genotype by environm ent in te ract ion s (M. Bonierbal e)
Uti I izat ion of imp roved population s for late blight res istance (J. Land eo)
Deve lopme nt of component techn o log ies for integ rated management of late blight
(G. Fo rb es)
Farm er part ic ipatory resea rch and impl ementation fo r in teg rated late b li ght manage ment
(R. Nelson)

2 Integrated control of bacterial wilt (E. Chujoy)

Bacterial wilt is the second worst potato disease in th e world. As resistance is difficult to
develop and maintain, CJP 's program concentrates on understanding and diagnosing the
presence of bacteria in th e soil and in tubers. This is used to improve site management,
seed systems, and utiliza tion of ava ilable resistant material. Decreased transmission of the
disease is a key ta1get in control.

Charac teri za tion, detect ion, and eco logy of P solanacea rum (S. Priou)
Integ rated managem ent of bacteri al wilt of potato in Southeast A sia (E . Chu joy)
Integrated co ntrol of potato bacteri al w ilt in East Africa (E . Chu j oy)
Integ rated management of bacteri al w il t of potato in Ch ina (Y. W ang)
Deve lopme nt of components of bacte rial wi lt contro l and IDM (S . Priou )
Integrated control of di seases of potato: bacterial w ilt in South W est Asia (0. Hid algo)

CIP Progrnm Report 1997-98

15

3 Control of potato viruses (M. Querci)

Viruses cause serious losses in potato and impede movement of seed and genetic resources. The project concentrates on sensitive, low-cost detection and epidemiological
factors affecting spread. Genetic engineering is used to identify, clone, and transfer genes
of re lated species to potato for resistance; this is combined with traditional methods of
breeding for maximum efficiency in incorporating resistance to a range of viruses.
Studies on potato ph ytoplasmas and their vectors (L. Salazar)
Deve lopment of sero logical techniques for potato virus diagnosis (L. Salaza r)
Molecular tec hniques for detect ion and control of vi ru ses and viro id s in potato
(M. Querci)
Mo lecu lar cha racter izat ion of genetic resistan ce to v iruses and use of natural genes for
transgenic res istance (M. Querc i)
Breeding for virus resi sta nce (M. Bonierbale)
Adaptation and utilization of advanced virus-resistant clones and proge nitors
(M. Bonierbale)
In -v itro eradication of potato v iru ses and multipli cat ion of materials for distribution
(A. Golmirzaie)

4 Integrated management of potato pests (A. Lagnaoui)

Key pests are potato tuber moth (a rapidly growing threa t), Andean potato weevil,
leafminer flies, and whiteflies. The project objective is to combine management practices
into locally adapted packages to reduce chemical pesticide use and increase overall
benefits for farmers. Components in clude biological, cultura l, and resistance aspects of
control.

Management of potato tuber moth complex (A . Lagnaoui)

Management of leafminer fly and other foliag e insects (F. Cisneros)
Management of And ea n potato weevi l and other so i ls pests (F. Cisneros)
Management of in sect vecto rs (A. Lagnaoui)
Ma nagement of potato nematodes (M. Canto)
Entomopathogens as integrated pest ma nagement compo nents (F. Cisn eros)

5 Propagation of clonal potato planting materials (U. Jayasinghe)

In creasing efficiency and effectiveness of both informal and formal seed systems is the
target of this project. Varietal introdu ction and diffusion is dependent on the inform al
system, but it must be linked with the forma l and it must emphasize high quality planting
material. Th e project accomplishes this with farm er training and establishment of pilot
seed systems.
Impact of inn ovations in link s betwee n formal and informal seed systems in Ecuador,
Kenya, and the Philippines (C. Cr issman)
Strengthen in g research and production of seed potato in Bolivia (A. Deva ux)
Seed multiplication and varieta l diffusion throu gh the development of a farmer-based
seed system (H. Kidane Mariam)
D iffu sion and production of quality seed throu gh in forma l participatory seed systems in
Bangladesh and Sri Lanka (S. ll angant il eke)

16

ProjectPortfolio

 Research and technical assistance for the improvement of seed technologies and the
development of seed programs (0. Hidalgo)
Seed multiplication and varietal diffusion of new potato varieties through the farmerbased seed system in the Philippines, and Vietnam (U. Jayasinghe)
Seed degeneration studies in Southern Peru (L. Salazar)

6 Sexual potato propagation (True Potato Seed) (M. Upadhya)

True potato seed (TPS) has growing potential in specific areas of the world where the more
traditional production systems fail. CIP concentrates on the researchable areas of improving parental performance in hybrids and in certain constraints (/ate blight resistance,
earliness, seed set, etc.). Backstopping is being done by local organizations (p rivate sector,
NCOs, and NARs) in efforts to commercialize TPS systems and thus underpin the developing of small industries.
Breeding TPS parental lines : production and evaluation of hybrid TPS families
(M. Upadhya)
Breeding TPS parental lines: production and evaluation of hybrid TPS families
(S. llangantileke)
Evaluation of hybrid TPS families for adaptation and yield (R. El Bedewy,
S. llangantil eke, G. Prain, and M. Upadhya)
Physiological studies on TPS quality postharvest handlin g and storage (N. Pallais)
Evaluation of parental lines and hybrid TPS families fm reaction to late blight
(B . Trognitz)
Evaluation of parental lines and hybrid TPS families for total glycoalkaloids
(M. Upadhya)

7 Global sector commodity analysis and impact assessment for potato

(T. Walker)
The lack of consistent and reliable statistics covering the agronomic, economic, social, and
environmental aspects of new and improved technologies in potato decreases the ability to
docum ent effectiveness and guide investment. This project is providing that information
and determining rates of returns on CIP research. Price and production databases are
established for constant reference and priority setting. Commodity analysis is improving
domestic potato marketing and international trade prospects for deve loping countries.
Impact assessment (T. Walker)
Global characterization (T. Walker)
Characterizing consumption, markets, and trade (G. Scott)

8 Control of sweetpotato viruses (L. Salazar)

Although sweetpotato viruses cause significant production losses, especially in subSaharan Africa, there is still a lack of research in this area. This project is developing
methods of detection and control for the virus complex known as WBV It also seeks to
identify resistance to this virus complex.
Identification and characterization of sweetpotato viruses and search for resistance
(L. Salazar)
Eradication of sweetpotato viruses (A. Golmirzaie)

CIP Program Report 1997-98

17

 Evaluati on of v irus d iseases (L. Sa lazar)

Test of sweetpotato culti va rs resistant to sweetpotato v irus d isease (L. Salaza r)
Pro pagatio n of clean pl anting materi al (N. Pallais)

9 Integrated management of sweetpotato pests (F. Cisneros)

Integrated pes t managem ent for sweetpota to dep ends on good know ledge of the biology
and in cidence of the ta1get pests. /PM components are then de veloped and eva luated
aga in st site-spec ific ecologica l, eco nomical, and social factors . Pilot units and Farmer Fie ld
Schoo ls (FFS) are used as th e mechan isms for testing and implem enting /PM components.

Institutionaliz in g sweetpotato FFS in Indonesia (E. va n de Fli ert)

In stitut ionaliz ing sweetpotato FF S in Vietnam (E. va n de Fliert)
ICM for sweetpotato in U ga nda (N . Smit)
Management of sweetpo tato pests in Latin Ame ri ca (F. Ci sneros)

10 Postharvest utilization of sweetpotato (G. Scott)

This project studies a range of techno logies and technology adoptions to impro ve the
sustainable li velihoods of rural poo r through di versification and expansion of sweetpotato
use. In the va rious countries, th e main beneficiaries are wom en and children and poor
househo lds. The project goa l is to improve nutrition and incom e and reduce poverty.

Expandin g utilizati o n of sweetpo tato starc h (C. Oates)

Faci litatin g co nsumption of sweetpotato flo u r (V. Hage nim ana)
In creasin g use of swee tpotato for li vestock fe ed (D. Pete rs)
N ew uses for fresh sweetpotato (C. Scott)

11 Breeding for high dry matter in sweetpotato (D.P. Zhang)

This project pro vides th e raw material for increases in both fresh and processed use of
sweetpota to. Ory matter, the essential component in both types of use, has been increased
and gen etic material made ava ilab le. Th e ne w clones are being ada pted to low-input
subsistence sys tems in ta1geted en vironments to feed into the developing m arkets.

Sweetpotato breedin g and ge rmpl asm eva lua tion fo r SSA (E. Carey)
Breedin g hi gh dry m atte r swee tp ota to adaptable to So uth eas t As ia ( l.C. Mok)
Breeding dual purpo se and early maturing variet ies for SW As ia (T.R. Da ya l)
lntrogress ion of exoti c germp lasm into the Chin ese bree din g program (Y. Wa ng)
Cer mpla sm eva luati o n for ab ioti c stresses (N. Pall ais)
Cermp lasm enhan ceme nt thro ugh population improveme nt and biotech nology (0 .P.
Zhang)

12 Global sector commodity analysis and impact assessment for sweetpotato

(G. Scott)
Com m odity analysis and impact assessment data are difficult to loca te and to generate.
This project is providing mu ch-needed information through country case studies, lo ca l and
regional literature review, and data to develop the analyses necessary to study constraints

18

Project P01tfolio

as well as the potential for developing country production and postharvest sweetpotato
systems.
Global Characterization (G. Scott)
Impact studies (J. Otieno)
Projections/database (G. Scott)

13 Potato production in rice-wheat systems (T. Walker)

The assessment of opportunity for potato to fit into the cool dry season cropping systems
following rice in the rainy season has been completed. Potential is good for increasing the
efficiency of this cropping ecology through the addition of potato. Current work concentrates on diagnosing the constraints to and increasing the productivity of the full potatorice-based system. Ecosystem resource management is particularly critical, and the project
is developing models to provide natural resource management options.
Productivity and sustainability of the rice -potato-rice cropping system (T. Walker)
Characterizing the feasibility and desirability of intensifying potato production in cerealbased cropping systems (W. Bowen)

14 Sustainable land use in the Andes (R. Quiroz)

The Andean ecoregion is a particularly vulnerable ecosystem for sustainable agriculture.
The project is providing a sound scientific, technical, and economic base for both policy
and technology recommendations. Innovative tools for ecoregional research are developed through the integration of process-based crop growth models, remote sensing,
economic decision models, and geographic information systems (CIS). Such integrated
tools are also applicable to other global mountain ecosystems.
Characterization of land use in the Andean ecoregion (R. Hijmans)
Methodology development for ecoregional research in mountain environments
(W. Bowen)
Land use intensification and natural resource management in the Andean ecoregion
(C.L. Velarde)
Policy intervention in the high Andes (R. Estrada)
Linking research and development activities with CON DESAN and GMP (J. Posner)

15 Conservation and characterization of potato genetic resources (Z. Huaman)

This project conserves the most comprehensive collection of wild and cultivated potatoes
in the world. Characterization (both molecular and morphological) is the key objective,
along with database development and availability. Significant activities focus on identifying key desirable traits and distributing healthy material throughout the world for utilization in potato improvement projects. The project provides key input into Cf P's own
breeding efforts.
Conservation of potato genetic resources (Z. Huaman)
Molecular and agronomic evaluation of potato genetic resources (M. Bonierbale)
In vitro culture.conservation and cryopreservation of potato cultivars (A. Golmirzaie)

CIPProgromReportl99798

19

16 Conservation and characterization of sweetpotato genetic resources

(M. Hermann)
This project conserves the most comprehensive collection of wild and cultivated lpomoea
species (including cultivated sweetpotato). Ne w and more efficient methods of conservation are being studied (e.g., cryopreservation). The collection is being characterized for a
core collection to make it more accessible to breeders. Virus eradication is a key component. Desirable traits are being identified for more efficient use of the collection.
Conservation of sweetpotato genetic resources (Z. Huaman)
DNA fingerprinting for the selection of a core sweetpotato collection (D. Zhang)
In vitro conservation and cr yopreservation of sweetpotato genetic resources
(A. Golmirzaie)

17 Conservation and characterization of Andean root and tuber crops

(M. Holle)
The activities of this project assist national programs in the Andes in rationalizing strategies
in conservation of Andean roots and tubers (ARTCJ. Nine genera are included for collection, studYt and conservation of biodiversity. The potential for ARTC use on a wider basis is
also studied through market and consumption patterns, and identification of poorly
documented demand. Healthy planting material is produced for farmers . This project is
perhaps the only one with a significant effort toward de veloping virus identification and
eradication procedures for these important, underutilized crops.

Conservation methods (C. Arbizu)

ARTC diversity: Characterization and classification of ARTC variability (M. Hermann)
ARTC biotic constraints: Clean planting material (viruses emphasized ) (L. Salazar)
Product development of ARTC for specific characters (e.g., starch) and sites to increase
competiti veness (S. Salas)
Links with CONDESAN partners (M. Holle)

20

P10ject Portfolio

Incorporating Poverty in Priority Setting:

CIP's 1998-2000 Medium Term Plan
T. Walker 1 and M. Collion 2

A lleviating poverty is increasingly regarded

as an importan t objective of public-sector
ag ri cultural resea rch in deve loping co un tri es . In deed, the mi ss io n stateme nt of the
Co nsultative G ro up o n In ternational
Agri cultural Resea rc h (CG IAR) has evo lved
over the last 15 yea rs from an em ph as is on
" imp rov in g the economi c we ll- being of
low- in come people" to "co ntributin g to
poverty erad icatio n. "
Th e rol e of agricu ltu ral res ea rch in
in creas in g productivity and in priming the
pump for eco nomic growth is we ll known,
but agr icu ltural re sea rc h is also viewed as a
blun t in strum ent to improve in equ ality.
Poverty is often assoc iated w ith in eq uality,
but the two are not th e same. A stro ng
pos itive relat ions hip betwee n eco nom ic
gro wt h and reduction in absolute pove rty
has been docume nted in comprehensive
empirical studi es of th e World Ba nk
(Deining er an d Squire, 1996). But the same
resea rc h does not f in d a stat ist ica ll y signifi ca nt relat io nship betwee n eco nomi c growth
and in equa lity. Therefore, wh il e we are
confident that ag ri cu ltural resea rch wi ll
co ntribute to pove rty all ev iat io n in deve lopin g co untri es, it would be too mu ch to
expect that ag ri c ultural research w ill be a
ve hic le for red istribu tin g assets and in come
from th e ri c h to th e poor.
A lm ost always, agricultura l resea rch that
in creases the productivity of food crops w ill
favor th e poor of deve lopi ng co untri es.
Th ose be low the pove rt y lin e w ill reap a
substanti al share of th e net benefits of
tec hn o log ica l change. But quantifying the

1 CIP, Limo , Peru.

2 World Bank, Washington, D.C. , USA.

size of such benefits is a co mp lex empiri ca l

und ertak in g. To w hat exte nt do th e rura l
poor produce, co nsume, and market
potatoes and sweetpotatoes? To what
extent do the urban poor co nsum e potatoes
and sweetpotatoes? To w hat extent is the
economy open or c losed? To w hat exte nt
are project techno log ies labor saving o r
labor usin g? To what exte nt do pol ic ies and
in stitut ions rein force in ce ntives for th e
adoption of potato and sweetpotato tec hnol ogies? These are usua lly on ly a subset
of the questions that need to be answered
before the effects of agr icu ltural researc h o n
th e poor can be quantified w ith some
degree of precision.
In this arti c le, we add ress onl y o ne
dim ensio n of the 1e lat io nship between
agricultural resea rch and poverty: does
adju sting fo r th e spatial d istribution of
pove rty lead to substantial changes in the
project rankings in priority setting? In o ur
application, pro j ect rankings adjusted for
po verty depend heav il y o n est im ates of
poverty in Chin a, th e largest grow in g potato
and sweetpo tato co untry in the wo rld .
The Geography of Poverty and
CIP's Projects

In prior ity setting for its 1998-2 000 M edium-Term Plan, CIP sc ientists and management eval uated 15 projects in a conventional format of eco nom ic proj ect appra isa l
(CIP, 1997). Project benefits depend ed on
th e size of the recommendation dom ain (in
target and sp ill-ove r co untries), probabi li ties
of tec hnologica l success, per hecta re
ben efits of ado p ted tec hnology, and ce ilin g
levels of adoptio n in 20 15 (Wa lker and
Collion , 1997). Estimates for these four

21

pove rty in th e ea rl y 1990s ranged from

86% in Rwa nda and Bhutan to 4% in So uth
Korea . For the deve lop in g co untri es as a
w ho le, th e we ighted average in cidence of
pove rty was 31%, about three persons in
ten.

parameters we re eli c ited from scien ti sts in

group discuss ions.
Ca lc ulatin g the we ighted average
in c idence of poverty for each of th e
projects was the first step in assess in g th e
pote nti al leve rage each project ha s on
pove rty. A hea dcount ind ex is the most
co mmon way to measure abso lute or
" thres hold " poverty. It shows th e propo rti on of th e pop ul at ion fa llin g below a
pove rty lin e that reflects a minimum
stand ard o f li v ing or consumption . In th e
TAC data base that was used fo r prior ity
sett in g (Grysee ls et al. , 1996), th e pove rty
lin e is rou ghl y equi va lent to a stand ard of
li ving at $1 per day in purchas in g power
par ity in come that adju sts for di ffe rences in
pri ce leve ls across countri es . D ata o n the
population of the rural and urban poor
we re prese nted fo r 11 8 deve lo pin g countries. Th e estimated nati o nal in c id ence of

Th ese nation al poverty estim ates we re

co mbin ed w ith reg ional estimates in China
and Ind ia to generate project-spec ific mea n
est imates we ighted by th e geog raphi c
di stribution of benefits (Tabl e 1). Most of th e
proj ects are characte rized by a pove rt y
intensity that exceeds the we ighted globa l
ave rage of 31 % . Fo ur projects fa ll bel ow
thi s thres ho ld. Three pertain to
sweetpotato, and all are geog raph ica ll y
concentrated in Ch ina . In general , projec ts
w ith benefit s more heavil y co nce ntrated in
China rank at th e low er end of th e sca le o n
pove rty intensity.

Table 1. The overage poverty rating of Cl P's projects by measure.

Poverty measure

Project

Head-count index (%) 0

Potato Sustainable Intensive Cropping (PSIC)

Sweetpotato Planting Material (S PM)
Potato Seed Systems (PSS)
NRM in Tropical Mountains (NRM)
Sweetpotato IPM (SIPM)
Potato Utilization (PU)
IPM Potato Insects (PIPM)
ARTC Postharvest (ARTC)
Potato Late Blight (PLB)
True Potato Seed (TPS)
Potato Viruses (PV)
Sweetpotato Dry Matter and Ada ptation (SDM)
Potato Bacterial Wilt (PBW)
Sweetpototo Product Development (SPD)
Sweetpotato Virus (SV)
0

Original

Revised'

Original

Revised'

52.0
50.4
47.6
44.2
43.4
40.7
39.7
35.1
34.3
32.8
31.4
29.6
25.8
22.5
16.7

52.0
50.4
47.6
44.2
43.4
40.7
39.7
42.6
34.3
32.8
44.3
41.2
38.0
37.4
28.6

0.787
0.644
0.587
0.560
0.585
0.694
0.367
0.452
0.426
0.466
0.326
0.372
0.321
0.272
0.197

0.787
0.644
0.587
0.560
0.585
0.694
0.367
0.561
0.548
0.572
0.514
0.540
0.498
0.489
0.370

Percentageof people that foll below thepoverty line in countries and provinces where the project is expected to have results.

b Poverty

measure that combinesa poverty line with on estimate of income inequality (TAC, 1996).

' Auth ors' revisions based on recent estimates of poverty in China.

22

TAC projectedb

FeatureArlicle

Poverty in China
The conseq uences of poverty in Ch in a
for the project rankin gs warrant more
discussion. First, the use of a national
average in TAC's own priority setti ng for the
CG IAR as a w hole hides a great deal of
interreg ional var iation in poverty (TAC,
1996). For examp le, potatoes in China are
mainly produced in the poorer interi or
mountainous provin ces distant from the
ric her coast (Fi gure 1). Weighting growing
area by provincial esti mates of poverty
shows that potatoes are cul tivated in
regions that are two to three ord ers of
magnitude poorer than comparable regions
where rice is produced.
Low growth and adverse distributional
effects have dampen ed growth in the
poorer inland prov in ces (C hen and
Ravallion, 1996). In four provinces in
sou th ern China, rel iable hou seho ld survey
data on rura l liv in g sta ndards suggested a
range in the incidence of rural poverty of
from 3% to 5% (depend in g on the method
used) in prosperous coas tal Guangdong to
27 % to 42% in poor inland Guizhou.
These interregional differences appear to be
sharper among rural rather th an urban
households (World Bank, 1993). In part,
these differences refl ect competitive
adva ntages of the coasta l provin ces.
Government policies have also co ntributed
to these differences. Seve ral coastal zones
receive preferential treatment for foreign
investment. The government also enforces
a reg istration system that inc reases the cost
of interregional mi gratio n. Geographic
poverty traps are becoming more ev id ent.
Seco nd , the national headcount index
for Chin a looked surpri si ngly low compared
with est im ates for other deve lop in g co untries. In descending order of the inciden ce
of poverty in the TAC database, China
occupies position 116, w ith mean est im ated
poverty lower th an that of all othe r de ve loping co untri es exc ept Oman and South
Korea. Estimated poverty is greate r in many
upp er-m idd le-incom e eco nomi es, such as
Malaysia, Uruguay, and Cyprus , than in

China. As an extreme exam ple , per-p erso n

income ad j usted for purchasin g power
parity was abo ut eight times hi gher in highin come Singapore th an in Chin a, yet 15%
of th e population of Sin gapore was est imated to be poor compared with 11 % for
China.
Th e World Bank has revised upwa rd
estimates of poverty in China (T he Economi st, 1996). The use of this more recent
estimate, wh ich we ca ll the rev ised
headcount index, results in a tighter
grouping of projects than the use of the
lower est im ate. For most projects, the
mean incidence of poverty I ies between
37% and 47 % (Tabl e 1).
Ideal ly, a poverty meas ure shou ld also
convey in format ion on the severity of
poverty, and a class of " high er-orde r"
poverty measures is preferred in the
literature (Ra va llion , 1994a). In the TAC
priority-setting exerc ise, an " in clusive"
hi ghe 1-mder measure of poverty was used
(Rava llion, 1994b). Th e proportion of poor
below a threshold poverty lin e was not the
so le foc us in measuring soc ial we lfa re.
Som e notion of th e sever ity of poverty was
approximated by the deg ree of incom e
in equal ity in the TA C exe rci se . Th e poverty
w eight is also based on a projection of
in come in purchasing power parity to 2010.
In come in equal ity was assumed to be
positively re lated to abso lute poverty. In
other words, even if base purchasing power
parity in come and proj ected economic
growth were ass um ed to be the same for
Brazil and Bangladesh, the estimated TAC
poverty we ight would be larger fo r Brazil
than for Bangladesh because of Brazil 's
greater in come inequ ality. In practice,
differences in inequ ality ac ro ss cou ntri es
are not th at important in co ntributin g to th e
national variation in TAC est im ated poverty
we ights, w hi ch are influ enced heavily by
the var iat ion in base purchasing power
parity incom e and proj ected growth rates .
Like the headcount index, we use two
versio ns of the TAC poverty measure: th e

CIP Program Repo111997-98

23

N
_p.

900

!00

uo

130

120

>
r;

'"

Nei

40 0

Mot~ol

. ,'

:Il~'f~1 ~

~~

- 20

I
~o o

(XJ O

~~~~~~~~~

Figure 1. The geography of potato area and poverty in China.

100

110

120

130

"or ig in al," based on a purchasin g power

parity per ca pita incom e of $2,946 i n 1991
prices fo r China; and th e " rev ised," based
o n $ 1,800 in 1994 which is co nsiste nt with
the rece nt World Bank rev ision s. Both
m eas ures u se the same robust projected
growth rates for China g iven in th e TAC
database.
Project results, ge nerated with the
or ig in al TAC projected poverty m eas u re,
are stro ng ly associated w it h those from th e
ini t ial headcount index (Tab le 1 ). Si111il ar
res ults a re not surprising because the
es tim ate of the TAC proj ected p ove rty
m easu re for China as a whole is 0.106
whi c h is only slightly lower than the initial
nation a l h ea dcount index of 0.112 (expressed as a proportion) for China. In the
same ve in , the rev ise d TAC poverty weig hts
g ive results quite si 111ila r to the revised
headcount ind ex. China's per cap ita
in come of $ 1,800 in purc h asi ng power
poverty in 1994 combined w ith t he proj ec ted growth rates resu Its in a poverty
weight of 0.432, which is about four times
hi g her than the earlier esti111ate of 0.1 06.

Assessing the Impact of Poverty on the

Project Rankings
In conve nti o n a l scor in g models for priority
sett in g, we ig h ts wou ld be app li ed to cr iteria
to arriv e at a project score. In o ur app lication, we esse nti a lly have tw o c ri te ri a:
effi c ien c y in th e form of n et prese nt value
and eq uity in the form of pove rty intensity.
For a co mparative treatm e nt, b ot h of th ese
va ri ab les wo uld be resc a led (o r nor 111 a lizecl )
between 0 and 1, repr ese ntin g th e lowest
a nd highest proj ect va lu es for each crite ri o n. A rb itrary we ig hts wo uld then be
ass ig ned to eac h criterion to rank th e
projects. Cha ngi ng the weights a ll ows the
analy st to re iate the sensitivity of the project
rankin gs to judg ments about wh at is
des i rab Ie.
This procedure for assessing th e effects
of different objectives on proj ect rank in gs in
a scor in g model of prior ity setting is tr icky

b l ind ad herence to co nventio na l practice

can lead to di stoned results b eca u se iss u es
re la ted to the brea dth of coverage of th e
proj ect are not co n sid e recl. Th e size of
proj ec t net be nefits in th e fon11 of Net
Prese nt Value (NPV) should mirror the size
of th e op p o rtunity to be g rasp ed or th e
probl e m to be so lved . Th erefore, the cost
of t he proj ect refl ects co n sid erat ion s of
critical mass to get the job clon e. Because
of a f ixed size of the proble111 to be so lved
or opportunity to be ex ploited , doubling t h e
siz e of th e project is unlikely to lead to a
twofold in crease in NPV. Hen ce, multiplyin g net ben efit s by th e poverty 111 easu1es in
Tabl e 1 and estimatin g a poverty-wei g hted
NPV is perh ap s the 111ost infor111ative wa y to
evalu ate how the projects rate on the equity
di111 en sio n wh il e preserving th e rea lity of
proj ec t size .
Th e effects of i n corpmati ng a po ver ty
criterion are g rap h ed in Figure 2, w h ich
contain s three ve rsions of the composition
of NPV in C IP 's resea rc h portfolio. Th e
first , on the le ft- h and sid e, is bas ed on
avera ge project NPV. Th e second and third
entries show t he co mposition of pove rtyweigh ted NPV with th e two ve rsi o ns of th e
TAC 111 eas u1e give n i n Tab le 1.
In go in g from efficie nc y to the two
po ve rty -m od if ied ra nkin gs in Fi g ure 2,
several findin gs e 111 e rge. As expected from
Tabl e 1, the proj ec t th at loses the most
ground is sweetpotato v iruses. Its contribution to total NPV drop s fro111 5.5% to 2% to
3% d ep endin g o n th e pove rty m eas ure we
use to weigh benefits by. In other wo rds,
the project loses abou t 50% of its imp o rtan ce. In tegrated pest 111anagement (IPM) of
sweetpo tato is t he 111 a in beneficiary of
mod ify in g the results fo r poverty. Its va lu e
share ri ses by 40% from 5% to 7%. Some
other proj ects, in c ludin g late blight, trn e
pota to seed, n atura l reso urce manage111e nt
in tropi ca l m o unta ins , and several of th e
smaller projects, co ntribute more to th e
value of Cl P's portfolio w hen benefits are
adjusted to reflec t the poverty inten sity of
the reco111111 e ndation domain.

(A lsto n et a l. , 1996) . In our app li cat io n,

CIP Progrom Report 1991-98

25

Efficiency

Poverty
TAC
original

Poverty
TAC
revised

SDM

20

------------ --- PLB

SDM

PIPM ------- ...... .............................. .

10

TPS

.. "' ---- .
PV ..........

TPs9'' ... ,

., ,

' '
...........
"-;;>(,...
........ .,

..................

SPo.,.

.... ....... ,

,....,,.......

'

',

PV

...":.

PBW

SV

SIPM

----...::---.
--

sv

NRM ........... .

,,,...,..e.fl

PSS

f "'

~iE:::::::::::::
O

ARTC

NRM

-."::,:::o~=ir.,.,,

... .-.............................. ~~~

PS IC
PU
ARTC

Figure 2. Composition of project benefits (in % of

NPV) by one efficiency and two poverty

scenarios.
Almost all projects for w hich China does
not figure as a target or spillover country
gain in importance w hen benefits are
weighted by the original TAC poverty
measure. Potato integ rated pest man agement is th e exce ption because of high
projected growth rates in Colombia, w hich
contributes abo ut 30 % to project benefits.
Strong projected growth results in an
estim ated poverty weight of on ly 0.180 for
Colombia.
In general , the use of the rev ised pove rty
estimates in the third column of Fi gu re 2
leads to the reestablishment of the efficienc y ranking. In other words, poverty
matte rs as a modifier whe n the ear li er lowe r

26

FeolureArlicle

Concluding Comments
In our application , modifyi ng prio rity
setti ng for the geographic incid ence of
poverty did not appreciably change the
effic iency-based results in terms of project
rankings. The extent to w hich poverty
matters to project outcomes depends on th e
incidence and seve rity of poverty in China.

SPD
PIPM

PBW

estimates on its in cide nce in China are

used . We prefer to use our revised TAC
po ve rty we ight that in corporates the recent
rethi nkin g on a hi gher incidence of poverty
in Ch in a toget her w ith the projected high
rates of eco nomi c growth .

Estimates of the geographic in cide nce of

po ve rty provide only a rough guid eline of
the impact of technical change on poor
peop le . For example, poorer farmers grow
proportionally more sweetpotato than
richer farmers in Shandong Provin ce, and
the former have benefited di sproport ionate ly more than the latter from a new v irusfree propagation program related to research and trainin g in the sweetpotato virus
project (Fuglie et al., 1999).
Targeti ng for the rural poor has been
di fficult to implement at a more disaggregate spat ial level w ithin a country beca use
increasing geographic specificity is usually
accompanied by large r sampling errors in
national surveys measuring standards of
liv in g. Some inn ovat ive approaches to
pove rty mapping, such as combining
cens us data with information from national
li vin g standards surveys, show promise in
enhancing the resolution of geographic
pove rty tra ps (Minot, 1998). Th ese developments should lea d to increased precision
in facto ring the geograph y of poverty into
priority setting in agricultural research.
Computationally, modifying the res ults of
priority setting for poverty is an easy, albeit
uncertain, exercise . Disentanglin g the
effec ts of tec hnolo gica l change on the rural
and urban poor is considerably more
difficult. We plan to draw as man y lesso ns

as poss ib le from our case stud y resea rch

(Wa lke r an d Crissm an , 1996) to improve
the pove rty foc us of o ur resea rch and
trai nin g.
References

Als ton , J.M ., G .W. Norton , and P.G.

Pard ey. 1996. Scien ce un der sca rc ity :
Prin c ipl es and practi ce fo r ag ri cultural
resea rc h eva luat ion and prio rity sett ing.
Co rn e I I U niversity Press, Ith aca, NY,
USA. 585 p.
Chen, S. and M . Rava lli o n. 1996. Data in
transiti o n: A sses sin g rur al li vin g standard s in Southern Chin a. Chin a Eco n.
Rev. 7(1 ):23-56.
CIP (Intern ational Potato Center). 199 7.
M edium -Term Plan 19 98-2000. CIP,
Lim a, Peru .
Dein in ge r, K. and L. Squire. 1996 . A new
data set meas ur ing in co me ineq uali ty.
The Wo r ld Ban k Eco n. Rev. 10(3) :565 59 1.
Fu gli e, K. O ., L. Zhan g, L. Sa lazar, and T.
W alker. 19 99 . Econ o mi c im pac t of
viru s-free sw eetpotato seed in Sh andong
Provin ce, China. CIP, Lim a, Peru .
G rysee ls, G., J.P. Groenewo ld , and A.
Kassa m. 1996 . TAC database fo r
qu anti tat ive analysis of CGI A R pri orit ies
and stra teg ies. Food and Agr icultu re

O rga ni za ti o n of th e U nited Na ti o ns,

Rome, Ita ly. 8p . (Unpu bl ished Ma nuscr ip t.)
M in ot, N . 1998. Genera tin g d isagg regated
pove 1ty maps : An applicati o n to Vietnam. MS SD Di scussion Pap er No. 25.
IFPRI, W as hin gton, D.C., USA.
Ravallion , M. 1994 a. Poverty comp ari son s. H arwoo d Ac ademic Publish ers,
Geneva, Sw itze rl and. 145p.
Ravalli o n, M. 199 4b . Measu rin g soc ial
we lfare wi th and w ithout pove rty l in es.
Am. Eco n. Rev. 84( 2): 35 9-3 64.
TAC (Tec hni ca l Adviso ry Co mmittee to the
CG IAR). 1996 . CGIAR pri oriti es and
strateg ies . Foo d and Agriculture O rga niza tion of th e United Nations, Rom e,
Italy. 77p .+Ann exes.
Th e Eco no mi st. 1996 . How poor is Chin a?
O ctobe r 12, 1996. p. 35-3 7.
W al ke r, T. and M. Co llion. 199 7. Prio rit y
sett in g at CIP fo r th e 199 8-2 000 Medium-Term Pla n. U nproce ssed. C IP,
Lim a, Peru.
Wa lker, T. and C. Cri ssman. 1996. Case
studi es of the eco nomic impact of CIPrelated tec hnolo gy. CIP, Lima, Peru. 157

p.
W o rld Ba nk. 19 93. China: Strateg ies fo r
redu c in g poverty in the 1990s. Th e
Wor ld Ba nk, Was hin gton D .C. , U SA.
14 0 p. + A nn exes .

CIP P1ogrom Repo1t 199798

27

Research on Potato
Impact on a Changing World

Characterization of Phytophthora infestans Populations

in Peru
W. Perez1, S. Gamboa1, M. Coca 2, R. Raymundo\ R. Hijm ans1, and R. Nelson 1

Th e late blight (LB) pathoge n (Phytophth ora

in festans) is wide ly believed to have
originated in the Tolu ca Va ll ey of Mex ico
and to have sp rea d arnu nd the wor ld
through two major mi grat ion s. The first of
th ese occurred in the nin eteenth ce ntury
and led to the Iri sh Potato Famin e. DN A
fingerprinting of global co llecti ons of th e
pathoge n has indic ated that thi s pathogen
mi grat ion led to th e wo rld w ide di sse min ati on a singl e lin eage of the path oge n,
des ignated US-1 (the "o ld" lin eage) . More
recent wo rld wide mi gration s of diverse,
aggressive and fu ngic id e- re sistant strain s of
P. infestan s ("new" pop ul ations) have made
management of the disease in cre asi ngl y
difficult since the 1970 ' s (Fry et al., 1993).
Relatively littl e is known about co ntemporary popul atio ns of P infestans in Peru ;
systemati c stu di es ha ve not been co ndu cted
since the 1980s. Mu ch of CIP 's wo rk on
breeding potato (So lanum tuberosum) for
improved resistance to LB is co ndu cted in
Peru, and nat ional prog rams around the
world use the products of this breeding
program. It is therefore important to know
w hether th e pathogen populations in Peru
are rep resentat ive of the new pathoge n
pop ulation s that are cu rrently predo min ant
in many countries.
Available data on P infestans populati ons in Ecuado r and Bolivia suggest th at
both the A 1 and A2 mating types might be
present in Peru. Rece nt pathogen population analyses demonstrated th at Ecua dori an
P infestans population s belonged to th e

1 CIP, Lima, Peru.

2 Present address: Uni vers id ad Ma yor de Sa n And res, La
Paz, Bol ivia.

new migration, but showed the A 1 matin g

type (Fo rb es et al., 1997), w hil e Bolivian P.
infestans popul ations belonged to the A2
mat in g type (Fu ndaci 6 n PR OI NPA
[Prom oc i6n e ln ves ti gaci6 n de los
Productos And i nos] Bo li v ia, pers. co mm. ).
Sexual reco mbination between A 1 and A2
strain s co uld lead to in creases in pathogen
diversity and aggressiveness. It is impo rtant,
th erefo re, to know if reco mbin ati o n occ urs,
and w here, and to do cume nt th e consequ ences.
The present stud y was undertaken to
ch aracte riz e Peruvian populations of P
infesta ns. Samples were taken from Pasco,
Junin, Cusco, and Pun o departments, w hi ch
cover th e central to southern hi ghl ands. Th e
samplin g sites included Cl P's screening
sites at Comas and Oxapampa.

Materials and Methods

Sampling
Th e sa mpl ing strategy was designed to
aI low assessme nt of the dominant pathogen
popul ati ons at two sites in the central
hi ghl ands of Peru (Com as and Oxapampa),
and to determin e the population structure
in th e two southernmost departments of
Peru (P uno and Cusco) . Th e specific
sampli ng route was based on maps constru cted usin g geographi c systems, indicating zones of hi gh estim ated LB seve rity.
Approx im ately 30 sa mpl es each were
tak en from Oxapampa and Com as. For th e
depa rtm ents of Puno and Cusco, a hi erarchical sampling strategy was used to allow
an alysis of divers ity within and betwee n
fields , sites, and regions. For eac h department, several loca liti es (neighbor in g

CIP Progrom Repo111997-98

31

villa ges) we re sam pl ed. For eac h lo cal ity, 3

or 4 field s were sa mpl ed, w ith 10 sa mples
take n per- field (unl ess fewe r th an 10 we re
found because the leve l of LB was low).
Each infec ted leaflet was maintained
after co ll ec tion in a sea led Petri dish
containin g a la ye r of 1 .5% wate r aga r.
Usin g this method, in fec ted ti ss ue co u ld
be m ai ntai ned for 7-10 days b etwee n
colle ct io n in the field and isola tio n in the
labo rato ry
Co ntro l DNA sa mp les were used to
al low compa riso n w ith popul at io ns of P.
infes tans elsew here. G . Forb es (CIP-Quito)
and W. Fry (Co rnell University, USA)
prov id ed these samples.

Pathoge n isolation, culture, and storage.

Pl ates w ith infected t issue we re in c ub ated
fo r 7 d at 15- 1 SC, and sporangia we re
coll ected and rin se d o n a 10- m filter. The
filt er sys tem al lowe d recovery of isolates,
even w hen the infected tissue was seve ral
days ol d and con taminated w ith ba cteria
and sa proph ytes.
Th e spora ngia l su spensio n was refr ige rated at 5-SC to prom ote the lib eratio n of
zoospores, w hich we re then u sed to
in oculate potato slices (va r. Hua yro) . Tuber
slic es we re incubated at 1SC for 5-7 din
mo ist chamb ers. M yce li al frag ments we re
tran sfe rred ase pticall y to Rye B agar and VS agar pl ates. After 1 to 2 wk, grow in g
colonie s were transferr ed to Rye A aga r and
maintain ed at 15C.

Determination of mating type. The

matin g type of 2S7 different isolates was
determin ed by pa irin g eac h unkn ow n w ith
two iso lates o f a kno w n A 1 mating type
(Peruvian iso lates 22 S and 1696) o n 10%
clarifi ed V-S agar. Pl ates w ere p laced in an
incubator at 15C in th e dark an d exa m i ned
fo r th e prese nce of oospores afte r 4 wk. A
subset (n = 7S ) of the isolates was tes ted for
a poi y m erase chain rea ction (PCR )-based
marker linked to th e matin g type loc us
(Jud elso n, 1 996; see below fo r D NA
metho ds).
32

Potato

Virulence assays
Th e spec ifi c v irul ence of 114 iso lates
w as dete rmined b y inoculation of deta ched
leaflets of a d ifferent ial set of potato
culti vars ca rry in g the 11 kno w n major (R)
ge nes for r es istance . Differenti als we re
obta in ed fro m th e Resea rch In stitute for
Plant Protection , Wage nin ge n, the Netherlan ds. Leaflets col lected fr om th e midd le
part of eac h di ffe rentia l culti va r at 45-60
da ys o f age were inoculated i n in ve rted
Pet ri dishes lined w ith 1.5 % aga r wa ter,
suc h that lea flets lay in th e lid s below the
agar layer. ln ocu la for v iru le nce tests we re
obtained fro m tu b er sl ices inc ub ated for 6
to 7 cl at 1 SC in a mo ist chamber. A 20 -L
drop of a sporan g ia (a ppro x. 5 x 10 3
spo rangia/ml) wa s placed on th e abax ial
surface of eac h leaflet. Each test inc luded
th e susce ptib le cu lti v ar Chata Blanca
(co ntai n ing no know n R genes) as a co ntrol.
Th e virulence assa ys we re repeated at least
tw ice .
Metalaxyl resi stance. A su bset (n = 276 )
of iso lates was p lated on 10% V-S aga r
co ntai nin g metalaxy l at concentrations of 0,
5, 50, and 100 ppm . Isolates w ere co nsidered re sista nt to th e systemi c fungicide if
growth was:::: 40 % of the 0- ppm co ntrol at
any metalaxy l conce ntrat io n. Iso lates we re
consi dered mod erate ly resistant if growth
was redu ced to < 40% of th e co ntro l va lue
at 100 ppm, but not at 5 ppm . Iso lates were
co nsid ered suscept ible if growth was
retarded at 5 ppm to <40% of th e 0- ppm
co ntro l.
Isolation of DNA . Cultures of P. infestans
we re grown in pea broth (filtrate from 120- g
autoc lave d froze n peas per liter) in 5-6
di spos able Petri di shes at 1 SC in a dark
incubator w ithou t shaki ng. Each Petri di sh
was inoc u Iated wi th mycel i um from an
acti ve ly growing co lony in Rye B aga r. After
10 cl ays, th e fungal tissue was harvested b y
vacuum fil tration , froze n at - 70C fo r
seve ral hours, and then lyoph ili zed for 4-5
cl. Lyo philizecl ti ss ue was th en gro und in
liquid N , . DN A was ex tra cted fr om the
powde red myce lium (De Paul o and Powell ,
1995).

Polymerase chain reaction (PCR). A

subset of the iso lates from Pun o and Cusco
(n = 78) was analyzed by PCR usin g th e
prim e1s Sl A and Sl B (Jud elso n, 1996).
Th ese prim ers amplify a fragment of DNA
of approximately 1,250 bp cones ponding
to loc us Sl, which is linked to th e A 1determinin g allele of th e matin g type locus.
It represe nts a tandemly rep eated arr ay of
DNA elements that is typi ca ll y prese nt in a
hemi zygo us state in A 1 iso lates, but is
absent in A2 iso la tes (Juclelson , 1996). Th e
reaction was performed as descri bed by
Jucl elso n. DNA from A 1 and A2 mating
type iso lates (positive) and negat ive (water)
controls w ere used .
Restriction fragment length polymorphism (RFLP). DNA hybridi za tion ana lys is
was perform ed u sin g th e probe RC 57
(Goodw in et al. , 1992). Three g of DNA
from eac h iso late w ere di gested w ith
Eco RI. Hybridi za tion and dete ction we re
clon e fo llowin g the protocol of th e nonradio acti ve kit ECL 3
Amplified fragment length polymorphism (AFLP). The AFLP assay wa s ca rri ed
out usin g th e protocol of J. Dun ca n,
Scottish Crop Resea 1ch ln st itute4 , with
minor modification. The primary tem plate
was prepared in a one-step restri ctionli gat io n reaction, w ith th e restri ct io n
enzy mes Eco RI and Ms e I (Va n de r Lee et
al., 1997). The sequence of the primers
used in th e prea mplification step was E +A
and M +A (core sequences: E = 5' ACACTCCCTACCAATTC; M = 5'CATCACTCCTCACTAA). Fm se lect ive
amplification, th e seco nd ary template DNA
was amp li fied w ith prim ers contai nin g two
se lec ti ve 3' nucleotides. Res ults we re
v isu ali ze d by silver staining accord in g to
th e m anufacturer's manual (Prom ega) .
Three p rim er combinations wer e tested: E +
AA and M + AC, E + AC and E + CA, and E
+ AA and M + CA.

3 ECL 1 "'

of Arne rsh;:im , Inc.

4 SCR I, Dundee, Scotland .

Diversity and cluster analysis. Di ve rs ity

estimates were ca lcul ated usin g th e fo rmul a
of Ne i (1987) . RFLP and A FLP f in gerpr ints
we 1e sco 1ecl v isually for the presenc e (1) or
abse nce (0) o f polymorphic DNA fr ag ments
in bin ary characters. Cluster analysis w as
conducted usin g the unweighted pai1 gro up
method with arithm etic mean algorithm in
the softwa re pro g1a m NTSYS-pc (Rohl f,
199 2). Sim il ari ty was ca lc ulated usin g Dice
coeffici ent.
Results
A tota l of 409 iso lates were collected in
1997 and 1998 from Pasco, Ju11i11, Cusco,
and Puno, co nti guous departm ents spa11 11i11 g the central to southern Peruvi an
And es .
A set of 287 iso lates collected in 1997
and 1998 we re tes ted fo r matin g type by
pairin g w ith two known A 1 teste rs of
Peru v ian ori gin. None of them form ed
abund ant, typical oospores in th ese pairings. Th e iso lates were thus tentatively
considered to belong to the A 1 matin g typ e.
Prime1s amplifying the Sl locus, which is
linked to th e P infestans mating typ e loc us,
we re us ed to co nf irm this assignm ent fo1 a
sub set of the iso lates. No am p I ificat io n was
detected fo r any of th e 70 isol ates tested by
the PCR assay (as ex pected for A 1 st1a i11 s),
althou gh co ntrol amplifications usin g DNA
from A2 strain s showed amp I ification.
A total of 277 of the P infestans isolates
were analyz ed by DNA fingerprintin g usin g
the RFLP probe RC 57 (Figure 1). Ten
genotypes we re detected; these formed 6
groups bas ed 011 clus ter anal ys is of the
RFLP band data (Fi gu re 2) . Three genoty pes
accounted for 95% of the collection. Th e
calculated diversity estimates were thus
rath er low (H = 0.33). The collection s from
Pasco and Junin eac h had a diversity of
zero (only on e ge notype detected); the set
from Cusco had a div ers ity of 0.17; and th e
set from Pun o h ad a diversity of 0.0 5. 111
som e cases multi pie pathogen Ii neages
we 1e dete cte d in sin g le fields. For example , three li neages were detected in eac h
CI PProgram Reporl 1997-98

33

;;

.0

;:

"'
N

"'"'~ ;:';;;

"'
"'
M

r--

.0

;:: ;:

- A

__ ....- ___ ----

---- -- ---~---

- -= ----- ---

-= ~

Figure 1. DNA fingerprints obtained by (A) AFLP; and by (B) RFLP analysis using the hybridization probe RG57.

The molecular weight standard is al kb DNA ladder.

of two fields sampled in Paucartambo,
Cusco. One field had lin eages EC-1 , PE-3,
and PE-6, w hereas th e other had US-1 ,
EC-1 , and PE-3.
Lin eages US-1 and PE- 3 were domin ant
in Pun o. US-1 is th e 'o ld ' lineage, fo rm erl y
found wo rld w ide (Fry et al., 1993) and
prev iou sly found to be predominant in Peru
(Tooley et al., 1989). Th e ' new' lineage,
EC-1, was dominant in Pasco, Junin, and
Cusco. Of the 287 iso lates anal yzed in this
stud y, 174 (60.6 %) belonged to EC-1 .
Three va ri ants of EC-1 we re also detected
in the Peru v ian collection (Figure 2) .
Th e all eles present in the Peru v ian
collection we re all prese nt in th e CornellCIP database established by Forbe s et al.
(199 8) . Some of these alleles had been
34

Potato

detected in isol ates from Japan and the

Ne th erlands.
Two hundred and seve nty-six iso lates
we re tested for sensitivity to th e systemic
fun gic ide metal axy l. The majority of the
iso lates tested (230/ 276 = 82.2 %) we re
res istant. A ll isol ates from Pa sco and Junin
we re fu ll y resi stant to metal axy l. A few
iso lates from Cusco we re moderately
res ista nt (2/ 111 = 1 .8 %) or susc eptible
(5 / 111 = 4.5 %) to th e fungi c id e. A greater
proportion of th e iso lates from Pu no we re
mod eratel y resi stant (14/7 8 = 17.9% ) or
susceptible (28/78 = 35 .9 %). Th ese trends
co rrespond to th e d istribution of pathogen
lin eages in the different departm ents .
A set of 114 iso lates was tested fo r
v irul enc e on a set of 11 potato lin es

Iso late
CZ
BTLM11
BTLMS
BTLM4
BTLM3
TSP16
1473

684
P02

4
TSP20
TSP10
PCZ101
PPU036
1394
1385
PE96005
PC0018
POX00 1
1021
PCZ033
PCZ098
PCZ024
PPU005
1452
1393
1696
PCZ036
PPU013
US-1
PE84006
PE850019
PE840028
PCZ111

228
PCZ00 1
PCZ110
US-7
US-8
2939

I I
0.72

I I
0.79

I I
0.86

I I
0.93

Origin
Cusco, Peru
Lima, Peru
Lima , Peru
Lima, Peru
Lima, Peru
Lima, Peru
Argentina
Cajamarca, PerU
Ancash , Peru
Junin , Peru
Lima, Peru
Lima, Peru
Cusco, PerU
Puno, Peru
Junin , Peru
Junin, Peru
Junin, PerU
Junin, Peru
Pasco, PerU
Ecuador
Cusco, Peru
Cusco, Peru
Cusco, Peru
Puno, Peru
Junin, PerU
Junln, Peru
Arequipa
Cusco, PerU
Puno, Peru
USA
Junln, Peru
Junin, Peru
Junin, Peru
Cusco, Peru
Jun in, PerU
Cusco, Peru
Cusco, Peru
USA
USA
Ecuador

Mr
A1
A1
A1
A1
A1
A1
A2
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1
A1

A1

Year
198?
1997
1997
1997
1997
1997
1996
1996
199?
1997
1997
1997
1997
1997
1994
1994
1994
1997
1998
1996
1997
1997
1997
1997
1994
1994
1995
1997
1997

A1

A1
A1
A1
A1
A1
A1
A1
A2
A2
A2

1984
1985

1984
1997
1985
1997
1997

1996

Number
n=1
n=1
n=1
n=1
n=1
n=1
n=1
n=1
n=1
n=1
n=1
n=1
n=93
n=2
n=1
n=1
n=1
n=46
n=29
n=1
n=1
n=2
n=8
n=51
n=1
n=1
n=1
n=2
n=31
n=1

n1
n1
n=1
n=1
n=1
n=1
n=2
n=1
n=1
n=1

Genotype
des ignation
(RG57)
PE-7
PE-7
PE7
PE-7
PE-7
PE-7
EC1.4(t)
EC-1.4 (t)

~g:~:: m
EC-1.4{t)
EC-1
EC1
EC-1
EC-1
EC-1
EC-1
EC-1
EC-1
EC-1.2 (t)
EC-1.3 (t)
PE-3
PE-3
PE-3
PE-3
PE-3
US-1.xa
US-1
US1
US-1
US-1
US-1
US-1.xb
PE-2
PE-5
PE-6
US-7
US-8

I
1.00

Coefficient
Figure 2. Annotated dendrogram derived from RFLP data obtained using the hybridization probe RG57, using
the unweighted pair group method with arithmetic mean algorithm . *MT = mating type.
carrying known resistan ce genes (differential genotypes). Lineages EC-1, PE-3, and
PE-5 showed broad-spectrum virulence
(Table 1). Most, but not all, isolates of US-1
showed narrow-spectrum virulence. The
isolates resistant to metalaxyl all carried
virulence to between 5 and 10 R genes.
Isolates moderately resistant or susceptible
to metalaxyl were virulent to between two
and seven R genes.
To determin e wh ether the AFLP technique would allow greater differentiation of
diversity, DNA from 43 isolates was
analyzed by AFLP fingerprinting. Among
these, 23 were Peruvian isolates; the others
were control DNA samples . A comparison
of the dendrograms yielded by RFLP and
AFLP fingerprinting reveal ed that, as
expected, the AFLP technique allowed
greater differentiation (Fi gure 3).

Discussion
By analyzin g subsets of the collection by
various methods, it could be concluded that
the P. infestans population in central and
southern Peru is of the new type. Although
all of the isolates tested showed the A 1
mating type, the pathogen population is
now relatively diverse, broadly virulent,
and resistant to metalaxyl.
The P. infestans population has shifted
dramatically since the 1980s. Twenty
isolates collected in the department of Junin
in 1984 and 1985 showed virulence to
between 0 and 2 of the 10 R genes tested,
with the greatest number (40% ) showing
avirulence to all of the genes (Tooley et al.,
1989). In contrast, none of the 25 isolates
collected from Jun in in 1997 showed
virulence to fewer than 4 of 11 R gen es
tested in th is study. The majority of the

CI PProgram Report 1997-98

35

Table 1. Genotypes of Phytophthora infestans defined by DNA fingerprinting using the hybridization probe RG57,
and their associated characteristics (virulence, reaction to metalaxyl, site of coll ection, and corresponding
number of isolates).

Genotype

EC-1

Virulence

Metalaxyl

1,2,3,4,5,6,7,8 ,10, 11
1,2,3,4,6,7,8 , 10, 11
1,2,3,4,7,8,10, 11
1,2,3,4,7
1,3,4,7,8,10,11
1.3,4 ,7,8, 11
1,3,4,7,11
1,3,7,11

Comas
Comas
Cusco y Comas
Comas
Cusco y Comas
Cusco
Comas
Cusco

Cusco

1
8
13

2
9
3
1
1

1,3,4,7,8 ,1 0,11

PE-5

1,3,4,7,10,11

Cusco

PE-3

1,2,3 ,4 ,7,8 , 10, 11

1,2,3,4,8 , 10, 11
1,2,4 ,7,8,10,11
1,4,8, 10, 11
1,4,8,10,11
1,4,8
2,8
8,10, 11
8, 11

R
MR
R
MR

Pu no
Pu no
Pu no
Pu no
Cusco
Pu no
Pu no
Pu no
Pu no

4
1

Pu no
Pu no
Pu no
Pu no
Pu no

1
1

1,2,3,4, 7,8, 10, 11

2,3, 11
2,8
4,8
8, 11

isol ates had viru lence to eith er eight or nine

R genes.
Among 212 isolates col lected from
potato in Ecuador between 1990 and 1993,
Forbes et al. (1997) detected on ly three
genotypes (two lineages): US-1 , EC-1, and a
var iant of EC-1. The Peruvi an co ll ect io n
described here was rich er in pathogen
diversity, with 10 genotypes detected in the
199 7 co ll ection s taken from Pasco, Junin ,
Cusco, and Puno. Th e pathogen population is also appa rent ly distinct from that
prese nt in Boli v ia, wh ich is dominated by
th e A2 mating type. Thi s study was undertake n in part to determin e where the
Peruvian A 1 population might encounter
th e Bo li via n A2 population. Such a site
was not identified in this study.
Pototo

R
R
R
R
R
R
R
R

N=

EC-1.2

US-1

36

Site

s
s
s
s
s
R

s
s
s
s

1
1
2

2
1
2

Differenti ation was seen even w ithin

Peru . Co ll ections from Puno consi sted of
US- 1 and PE-3; those in Cusco consisted of
US-1, PE- 3, and EC- 1; and those from Pasco
and Jun in cons isted of EC-1. Such differentiation shou Id be taken into account when
pathogen isol ates are moved for experimental or other purposes.
Iso lates from Comas and O xapampa,
Cl P's key screenin g sites for LB in Peru,
belon g to th e new lineage EC-1 , found also
in Ecuador and Colomb ia (Forbes et al. ,
1998). The iso lates take n from th ese sites
show broad-spectrum vi rule nee and
resistanc e to meta laxy l. Thus, the pathogen
population being used by CIP's LB breedin g
program is rep rese ntative of the current

Isolate
PC001 B
PPU 036
PCZ116

if~:c5

PCZ024
1452
1393
1696
PPU005

1473 ----~

PCZ116
PCZ10 1
1021
PCZ09 8
PC Z033
PCZ098
PPU005
PCZ024
1452
1393
1021
1696
PPU 036
PPU0 48
PCZ11 1
PC ZOO I
PPU013
PCZ036
PC001 8
PEB4006
PPU 048
PE8 5001 9
PPU013
PE8 40028
PCZ036
US-1
PE850019
PCZ111
PE840028
228
~---- PCZ001
~----- 14 73
PE84006
US-1
, - -- - - - - U S ?
US- ?------~
- - - 228
- - - 2939
293 9 ;=;::::;=;::::;=:;::::;:::::;:::::;:::::;::::;;::::;:::::;:::::;:::::;::::;::::;:::;:::::;:::::;::::;"

0.69

0.76

0.84

0.92

1.00

Coefficient of similarity

:ECJm~

1.00

0.92

0.84

0.76

0 .69

Coefficient of similarity

RFLP (RG57)

AFLP

Figure 3. Comparison of dendrogram s derived from RFLP analysis using the probe RG57 and from AFLP
Ii ngerpri nli ng.
domin ant popu Iati on in th e And es as
determin ed by this study.
Th e predorn i nant EC-1 l ineage showed
res istance to th e systemic fun gici de
meta laxy l. Farrne rs sti 11 use fun gici des
co nt ai nin g rn etala xy l. Beca use th e syste rnic
fun gic id e is mi xed w ith contact fun gic ide(s)
in th e produ cts, the in effec ti ve ness of the
more ex pensive systemi c produ ct is not
necessa rily appare nt. Farm ers should be
made awa re that metala xy l is no longe r
effect ive.
In thi s stud y, as in oth ers, A FLP fi ngerprintin g revealed rnore di ve rsity th an did
RFLP/ RG57 (K arnoun et al. , 1998). Th is
tec hniqu e merit s further appli catio n in
po pul ation ge netics studi es of P. infestans in
cases w here discrimination is des irabl e.
Acknowledgments
We thank Dr. Ladis lao Palo min o of IN IACusco, and Emi li o Barah o na of IN IA -Pun o

for ge nerou s support and collaborati o n

durin g th e co ll ec ti o n of iso late s from Cusco
and Pun o, res pect ive ly. We thank William
Fry of Corne l l U ni versi ty, USA, fo r providing the clo ne RG5 7. We thank Grego ry
For bes o f CIP-Q uito fo r prov idin g D NA of
isola tes from Ec uado r. Th is wo rk was
parti all y suppo rted by th e Global Initi ati ve
on Late BI ight.
References Cited
De Paul o, J.J. and CA Powe ll. 1995 .
Ext raction of doub le-strand ed RNA from
p la nt tissues w ithout the use of o rga ni c
so lve nts. Plan t D is. 79 :246- 248 .
Forbes, GA, X.C. Escobar, CC. Ayala, J.
Reve lo, M.E. O rd o i'i ez, B.A . Fry, K.
Dou ce tt, and W.E. Fry. 199 7. Popul ati o n
ge neti c stru cture of Phytophthora
infes tans in Ecuador. Phytopath o logy
87:375 -380 .
Forbes, G .A ., S.B. Goodw in, A. Drenth , P.
Oyarzun , M .E. Ordoi'iez, and W.E. Fry.
1998. A globa l marker datab ase fo r

CIPProgram Repo111 997-96

37

Phytophthora in fes tans. Plant Dis .

82:811 -8 18.
Fry, W E., S.B. Good w in, A .T. D ye r, J.M.
Ma tu sza k, A. Drenth , P. W. Tooley, L.S.
Suj kowski , Y. J. Koh , B.A. Cohen, L.J.
Spi elm an, K.L. D ea hl , and D .A. In gli s.
1993. Hi sto rical and rece nt mi grati o ns of
Ph ytop hthora in festans: Chronology,
path ways, and impli ca tion s. Pla nt Di s.
77:653-66 1.
Goodwin, S., A . Drenth , and W .E. Fry.
1992. Clo ning and ge netic analyses of
two hi ghl y pol ymorphi c, mod erately
repetitive nuclear DNAs from
Ph ytop hthora infestans. Cur. Gen et.
22(2 ):107-1 15.
Jud elson, H .S. 1996. Chromosomal
heteromorph ism lin ked to th e matin g
type loc us of the oomycete Ph y tophth ora
infesta ns. Mo l. Gen. Genet. 252: 1551 61.
Kamoun , S., T. va n der Lee, G. va n den
Berg-Ve lthuis, K. de Groot, an d F.

38 Potato

Cove rs. 19 98. Loss of produ cti o n of th e

eli c itor protein INF1 in th e c lona l lin eage
US- 1 of Phytophthora infestans. Ph ytopatho logy 88 :1315 -1 323 .
N ei, M. 1987. Mo lec ul ar evo lu tionary
geneti cs . Col um bia University Press . NY,
USA.
Rohlf, F.J. 1992 . N TSYS-pc: N um eri ca l
taxo no my and mul tivar iate analysis
system. Ve rsio n 1.70. Exe ter Softwa re,
Seta nket, NY, USA.
Van der Lee, T., I. De Witte, A . Drenth , C.
A lfo nso , and F.Govers. 1997. A FLP
l inkage map of the oomycete
Phytophthora infestans. Fun ga l Gen.
Bio l . 21 :278-291.
Too ley, P.W., C.D. Therrie n, and D .L . Ritch.
1989. Ma tin g type , race composition ,
nu c lear DNA conte nt, and isozy me
analysis of Peruv ian isolates of
Phytophthora in festans. Ph ytopathology
79: 478-481

Genetic Diversity among Isolates of Phytophthora

infestans from Various Hosts in Ecuador
L.J. Erselius1, H.R. Hohl 2, M.E. Ord6fiez3, P.J. Oyarzun 1, F. Jarrin1, A. Velasco1, M.P.
Ramon\ and G.A. Forbes 1

The late blight (LB) pathogen (Phytophthora

infestans Mont. de Bary) is globa ll y distributed on potato (Solanum tuberosum) and
tomato (5. lycopersicon ). It is believed to
have originated in Central M ex ico
(Goodw in et al., 1992) . Elsewhere, populations conta in only a subset of the genetic
diversity fou nd in Mexico and, until
recently, cons isted of on ly the A 1 mating
type. In the ea1ly 1980s, th e A2 mating
type was detected in Europe (Hohl and
lse lin, 1984). New genotypes ha ve sin ce
appeared in many countr ies (Drenth, 1994;
Fry, 1996; Goodwin, 1997). Th e development of molecular markers has helped to
track the wor ldwid e movem ents of the
pathogen and two major migrations from
Mexico have been docum ented (Goodw in,
1997). The possibility of other migrations
from Mexico to South America has also
been suggested (Andr ivon, 1996; Tooley et
al., 1989).
Much of the discussion about the orig in s
and migrations of P infestans, recently
rev iewed by severa l authors (Anclr ivon,
1996; Fry et al., 1993; Goodwin et al.,
199h is based primarily on what is known
about isolates from potato. That is large ly
because LB is an economically important
disease 011 potato in the temperate zone,
where most of the research on the pathogen
has bee n carried ouc and where there
appear to be few altern ative hosts. P
infestans on tomato has also rece ived
considerab le attent ion, but the dynamics of
the pathogen population on this host are

1 CIP, Quito, Ecuador.

2 Zurnikon, Swilzerlond.
3 Universidad Tecnol6gica de f Equinoxial, Quilo, Ecuador.

not c lear. Some studi es have id enti fied

genotypes of the pathogen th at infect potato
and tomato equal ly (Fry et al. , 1991;
Goodwin et al., 1995; Legard et al., 1995;
Spielman et al., 1989) suggesting 110 host
specialization . Other studies, however,
indi cate c lear genetic differentiation
between potato and tomato populations
(B rommonschenkel, 1988; Goodwin and
Fry, 1992; l<oh et al., 1994; Lebreton and
Anclr ivon, 1998; Oyarzun et al., 1998).
Uncertainty exists about relati ons
between pathogen populations infecting
potato and tomato. But even less is known
about the po ss ible role that other less
important crops and wi ld species play in
the dynamics of the pathogen population.
In Central and South America, alternative
hosts exist in the same geograph ical
environment as potato and tomato.
Noncultivated Solanum spp. from this
region has been assessed for resistance to P.
infestans (Co lon et al., 1995; Glendinning,
19 83), but relative ly few studi es have been
done on pathogen popu lations attacking
these plants in the wild. Disease severity
was assessed on seven wild hosts in
Mexico, but genetic analyses of the pathogen were not carr ied out (Rivera-Pena,
1990). Matuszak et al. (1994) included
some isolates from wild spec ies in a study
on metalaxyl resistance in Mex ico. They
showed a similar frequency of res istance
and sens iti vity to isolates from cultivated
species, suggest in g that the isol ates from
wi Id spec ies formed part of the same
popul at ion.
Genetic character izatio n of the pathogen
populations attack in g alternative hosts of P

CIP Progrom Reporl 1997-98

39

in fes tans cou ld gi ve new insigh t in to

b io logy and li fe hi sto ry of the pathoge n. For
exa mp le, recen t iso lati ons from 5.
brevifo lium and 5. tetrapetalum in Ecuado r
have revea led th e prese nce of th e A2
matin g type (O yarz un et al. , 1998 ) in
additio n to the A 1 matin g type popul at ions
attac kin g potato and to mato in th at country
(Forbes et al. , 1997; O ya rz un et al. , 199 8).
Th e A2 isolates fr o m 5. b revifolium and 5.
tetrapetalum di ffe r fro m al l P in fes tans
ge noty pes represe nted in a globa l marker
database co mpi led fo r thi s path oge n
(Forbes et al. , 199 8).
In thi s stud y, we prese nt in fo rm ati on
from recent studi es in w hi c h we have
beg un to c haracteri ze po pul ation s of P
infestans attac k in g So lanaceo us hos ts in
Ecuado r. Some prelimin ary data we re
prese nted (Ersel ius et al ., 1998 ). We give
here a more detai led desc riptio n of th ese
pop ul atio ns, and di sc uss their re lated ness
to eac h other and to oth er clo nal I ineages
wo rl dw id e.

Materials and Methods

Hosts of P. infestans
A ll hosts of P infestans are in th e ge nu s
5o lanum and includ e fi ve cul tivated
spec ies: tomato (5. lycop ersico n), to mato
tree (5. betacea), pota to (5. tub erosum),
dipl oi d potato (5. phureja ), and pepin o
du lc e (5. muricatum); and seve n wi ld
spec ies : 5. brevifolium, 5. ca ripense, 5.
co lomb ianum, 5. ochrantum, 5.
tetrap etalum, 5. andrea num, and 5.
tuqu errense. The identi fica tion of th e wild
spec ies is based on co mparison w ith
pu bli shed descr iption s (Co rrell , 196 2), and
sho uld be consid ered te ntati ve. A ll the
w i ld species we studi ed grow in the same
A ndea n habitat as potato.
Collection and isolation
No predetermin ed sam plin g p lan w as
used beca use di stributi o n of th e hosts and
di sease on host col oni es was patc hy.

Iso lates we re trap ped w ith potato tu ber

sli ces (Fo rbes et al. , 1997), o r isol ated o n
sel ective medium (Oya rzun et al. , 1998).
So me iso lates, notab ly those from 5.
och ra ntum, grew poo rl y and we re very
diffi cult to iso late. A small er sampl e of
is o lates fro m 5. te trapetalum and 5.
bre vifo lium we re d esc rib ed for matin g type
(O ya rsun et al. , 19 97). Some of th e iso lates
fro m potato and to mato w ere also described (Forbes et al ., 1997; O yarzun et al. ,
1998 ). Iso lates were ma intai ned for short
petiods on Rye A or Rye B mediu m (Cate n
and Jinks, 1968 ) at 18C in the dark, and
stored fo r lo nge r peri ods (more th an 1 mo)
on Rye A aga r slant s at 15C w ith a 12-h
photo peri od.

Characterization of isolates
A t least 10 iso lates fr om each hos t
exce pt 5. tuqu errense and 5. andrea num
we re c haracte ri zed w ith three or mo re
genetic and phenotyp ic markers (Table 1).
Restr iction fragme nt le ngth po lymo rp hi sm
(RFLP ) fin gerprints we re obtained for
iso lates usin g th e moderately repetiti ve
prob e RG 57 (Good w in et al. , 1992). Two
g of DN A from eac h isolate w ere di gested
w ith EcoRI fo r 24 h, th en underwe nt
electroph ores is o n 0.7 o r 0.8% aga ro se ge ls
(56 V, 20 mA) fo r 24 -4 5 h in 1 X TBE .
H yb ridi za tion and detect ion w ere do ne
usin g th e no nradi oacti ve kit ECL '" 4 acco rdin g to th e manu fac turer' s in stru ction s.
M itoc ho nd ria l D NA hap lotypes we re
dete rmin ed by ampli fica ti o n of DN A of
eac h iso late usin g prim ers design ed fo r
sp ec ifi c reg io ns of th e mitochondri al
genom e of P in fes lans (Griffith and Sh aw,
1998). Di gesti on o f th e amplified reg io ns
w ith Cfo l, Mspl , an d EcoRI restrict io n
enzy mes y ield s ba nd patte rn s by w hi ch th e
isolates ca n be c lass ifi ed into four
hapl otypes: la, lb, Il a, and llb (Carter et al. ,
1 990; Griffith and Shaw, 1 998) .
Th e intern al transcr ibed spacer regio n 2
(ITS2) of th e ribo so mal DNA (rDN A) o f P.

4 A m ersh;im , In c.

40

Potato

Table 1. Host species and number of isolates of P. infestans assessed with neutral markers.
Host species

5. tuberosum (pototo)
5. lycopersicon (tomato)
5. phure;a
5. tuquerrense
5. colombianum
5. muricatum (pepino dulce)
5. caripense
5. betaceo (tree tomato)
5. brevifolium
5. tetrapetalum
S. ochrantum

5. andreanum

mtDNA

Mating type

Gpi 0

Pep'

185
125
8
6
17

225
140

38
77

107
92

40
23

8
6
13
16
26
13
22
13
5
3

8
2
10

8
6
16

11
15
9
22

19
22
13
22
22
11

6
0
5
l

11

15
28
13

22
22
23
11

22

Fingerprint

7
4
16
11
4
4

a. Gpi = glucose-6-phosphate isomerase, Pep = peptidase, mtD NA = mitochondria l DNA.

infestans ca n be amplifi ed using po lymeras e c hain reaction (PCR) and th e

specifi c primers ITS3 and PINF2 (Tool ey et
al. , 1997). DNA from iso lates was amp li fied using th ese prim ers and a templ ate of
about 10 ng DNA as desc rib ed previously
(Tooley et al. , 1997).
Iso lates were tested for res istance to 5
and 100 m g/ml m etalaxy l in 10%
uncl arifi ed VS m edium and c lass if ied as
res istant, intermedi ate, o r se nsitive. Cond itions of the test and criteri a foi- classifi cat io n
were described previou sly (Forbes et al.,
199 7).
lsozyme e lectrophores is for the enzymes
glucose-6-p hosph ate isome rase (Gp i) and
peptid ase (Pep) was don e o n starch ge ls
(Sp ielm an, 199 1) and also polyacryl amide
ge ls. Po lyac ry lamide ge l electrophores is
(PAGE) was don e usin g 1 mm thick 7.5%
ge ls with 25 mM Tri s-0. 19 M glyc in e, pH
8.8 as sepa rating ge l and electrod e buffer.
Bands were cl ea rer w hen a 1 c m stackin g
ge l (2.5% ac ryl amide 0.06 M Tri s- HCI, pH
6.7) was used (Dav is, 1964). PAGE ge ls
were run with a constant current of 5 mA
for 1 h, then in c reas ed to 10 mA. Voltage
rose continu ous ly throu gho ut, from about

50 to 280 V. Electrop ho resis was terminated w hen the bromop henol blu e dye
re ac hed th e bottom of t he ge l, about 16 cm.
All ozyme ge notypes were scored as
described by Spielman et al. (199 1), w hi c h
represent th e relati ve mo bi I iti es of th e
enzym e all eles to an all ele designated as
100. Iso lates w ith kn own allel es from the
coll ect io n of W. E. Fry, Corn ell Uni ve rsity,
were used for compari so n.
Gpi, Pep, and m atin g type data w ere
combin ed w ith RFLP f in ge1prints as described prev io usly (Fo rbes et al. , 1998), to
c rea te multilocus genotypes. M ultil ocus
ge noty pes of iso lates from 5. brevifolium
and 5. tetrapetalum were co mpared with
publi shed ge notypes o f P infestans taken
from a globa l m arker database (Forbes et
al. , 1998) using clu ster analys is. D ata
anal yses and presentation of res ults we re as
desc ribed previou sly (Fo rbes et al. , 1998).

Pathogenicity on potato and tomato

Two separate tests of pathogeni c aggressive ness were done. In th e first, o ne iso late
from tom ato was co mpared w ith three
isol ates from 5. muricatum for pathoge nicity o n three tom ato culti va rs (F lora D ade,
Caribe, and FMX-1 93) and on e unknow n

CIP Prog ram Report 1997-98

41

va riety of 5. murica tum . In th e second test,

one isol ate from potato and six from 5.
co lombianum we re co mpared for path ogen ic ity o n three potato c ultivars (C hata
Blanca, Cruza 148 , and Yun gay) an d a
sin g le pl ant of 5. co lombianum. A ll potato
and tom ato c ulti va rs used in th ese te sts a1e
co nsidered free of m aj o r res istance genes (R
ge nes for potato and Ph genes for tom ato),
w hich interact w ith pathogen races
(Oyarzun et al. , 1998). lnocu lum was
prod uced o n leaves of the ori g in al hosts
before in oculation. After 7 days, lesion
d iameter was measu 1ed and th e presen ce of
nec ro sis noted.
Results
A l I leaf les ions from w hich isolates were
take n resembled tho se produ ced b y
infection w ith P infestans, although w i Id
spec ies showed differing symptoms in th e
fie ld. Leaves of 5. brevifolium, for in stance,
w hich are small (less than 3 c m lo ng) and
thin , bla c kened 1apidl y and spo rulation was
usuall y v isible onl y at the edges of lesions.
5. ochrantum leafl ets are large (up to 20 c m )
and flesh y. The y showed extensive chl o rosis, w ith sporangi a format ion ove r most of
th e ch lorot ic area. Al I iso lates reacted as P
infestan s and clos ely- re lated spec ies in that
th ey all gave the 4 59 kb band after PCR
w ith th e ITS2 region pr imers (Too ley et a l. ,
1997). A lso, they grew poorly or not at all
o n nitl'ate m edium (Ga lind o and Hohl ,
1985) . Th e morphol ogy of sporang ia fro m

all iso lates was ty pi ca l o f P infestanslimon iform w ith a short pedice l (Erw in and
Rib eiro, 1996) . Sporangial dim ensions
we re cons istent w ith th ose pub I ished for P
infestans (Erw in and Ribe iro, 1996) for al l
iso lates except som e from tom ato , 5.
muricatum, and 5. betacea, w hi c h w ere a ll
large r than those reported prev iou sly.
Althoug h thi s is not co nc lu sive ev iden ce
th at these iso lates we re P infestans, no
other described spec ies accommodates
th em better.

Neutral markers
Ana lyses o f the populations w ith neutra l
m arkers res ulted in th e id ent ificat ion of
seve ra l multilocus ge notypes, w hi c h belon g
to four clon al lineages (Table 2). Ve ry
limited amounts of polymorphism for RFLP
fin gerprints we re found w ithin th e clo nal
I in eages attac king potato, tomato, and th e
w ild spec ies 5. brevifo lium and 5.
tetrapetalum. The limited polymorphism
w ith the popul ations attac kin g potato and
tomato confirms prev ious reports (Forbes et
a l. , 1997; Oyarz un et al. , 1998).
O nl y two plant spec ies we re associated
w ith more th an one clonal lin eage of th e
pathogen ; in both cases the lin eages were
US- 1 and EC-1. On e plant spec ies w as 5.
ochrantum, for w hi c h th ere was a geog raphical separation of the two lin eages
(EC- 1 is found in more northerly sites) . Th e
other was 5. andreanum, for w hi c h both
lin eages we re fo und at th e same site. Ea c h

Table 2. Clonal lineages of Phytophthora infestans found in Ecuador to date and host species from which they
were isolated.
Clonal lineage

Gpi'

Pep'

MtDNA'

US-1

86/l 00

92/ 100

IB

EC-1

90/ l 00

96/ 100

llA

EC-2

l 00/ l 00

76/l 00

New

EC-3

86/100

76/l 00

IA

Hosts

S. muricatum, S. caripense, S. ochrontum,

S. andreanum
Potato, 5. colombianum, S. tuquerrense,
5. andreanum, S. ochrontum, 5. phureja S. spp.
Potatob, S. phureja b, 5. brevifolium, S. tetropetalum
5. betacea

Tomato,

a. Gpi = glucose-6-phosphote isomerose, Pep = peptidase, mtDNA = mito chondrial DNA.

b. Only one isolate of this genotype found on th is host.

42

Pototo

of th e other host plant species was attacked

by on ly one c lona l lineage of P infestan s.

The patterns described here w ere

co nsi stent for repea ted iso lat ions. However, two exceptions occurred. We iso lated
th e EC-2 lin eage once from potato in
Pi chin cha Prov ince (ce ntral Ecuado r) and
once on 5. phureja in Loja Provin ce
(so uth ern Ecuador) . Th e potato EC-2 iso late
was on ly weak ly pathogeni c when reinoculated on potato. Th e isol ate from 5. phureja
has not yet bee n reinocul ated on th at host.

4
-25
-24

In co ntrast, three of th e four clonal

lin eages of th e pathoge n we re assoc iated
w ith more than one host spec ies. For
exa mpl e, iso lates from potato, 5. phureja, 5.
tuquerrense, and 5. colombianum al I
belon ged to th e EC- 1 clon al Ii neage
prev iou sly desc ribed on potato in Ecuador
(Forbes et al. , 1997). Isol ates from tom ato,
5. ca ripense, and 5. murica tum belo nged to
th e US-1 clon al lin eage, previo usly desc ribed as th e primary pathogen popul ation
attac kin g to m ato in Ecuador (Oya rzun et
al., 1998). Iso lates from 5. brevifolium and
5. tetrapetalum fo rm ed a distin ct group
belon ging to an unreported c lonal lin eage
that we here name EC-2 . A va ri ant of EC-2
w ith three RFLP band differences ha s bee n
nam ed EC-2. 1. We beli eve th at th e EC-2 .1
genotype has evo lved w ithin th e EC-2
clon al lin eage.
Iso lates from 5. betacea were al I A 1
matin g type and had the Gpi genotype 86/
100, whi c h is generally assoc iated w ith th e
US-1 clonal lin eage (Goodwin et al. , 1994).
Th ey rese mbl ed the EC-2 I in eage for so me
characteri sti cs, in c ludin g th e unu sual Pep
genotype 76/1 00 and three nove l RFLP
band s (Fi gure 1 ). Before this stud y, th ese
RFLP band s had been found on ly fo r EC-2
(u npubl.). Th e RFLP fing erprint of iso lates
from 5. be ta cea also lacked band 1. Th e
absence of band 1 had not been 1eported
prev iously for clonal lin eages of P infestans
(Forbes et al., 1998). Th e genotype of th ese
isolates from 5. be tacea has bee n des ignated EC- 3.

-21 -20a
-20

-16
-14a
-14
-13
-12
-10

-Ba

-7

-5

-3

-2

-1a

-1

Figure 1. RFLP banding patterns from probe RG57

representing principal genotypes of three

clonal lineages of Phytophthora infestans
found in Ecuador. Lanes 1 and 2 = EC-1 ,
lane 3 = EC-3, and lane 4 = EC-2. Bands
1a, Ba, and 20a are unique to EC-2 and
EC-3.
Multilocus genotypes, consi st in g o f RFLP,
allozym e, and mating type marker data,
w hi ch are found in Ecuador (US-1 , EC-1,

CI PProgram Report 1997 -98

43

EC-2 , EC- 2. 1, and EC- 3) and th ose fo und in a

global d ata base of P in fes tans m arker data
(Forbes et al. , 1 998) were co mp ared by
c lu ster analys is (Fi gure 2). EC-2, EC-2 .1 ,

. . . . - - - - - - - EC3 (EC)
EC2. 1 (EC ) *

.----------1

EC-2 (EC )*
. . . . - - - - - - - - - - - CA-3 (CA ) *
FR I (FR )
CA2 (CA)

and EC-3 we re dista nt to al I ot her know n

ge notypes repo rted thu s far from cl onal
pop ul ati o ns. EC- 3 cl uste red w ith EC- 2 and
EC- 2.1 beca use all co ntain the three uniqu e
RFLP ba nds and th e unu sual 76/100 Pe p
ge noty pe. Beca use th ese all eles are uniqu e
in th e database of publi shed ge notypes
from c lonal lin eages (Fo rbes et al. , 1998 ), it
is not surpris in g that th e new ge notypes
fro m Ecuador should c lu ster away fro m al l
others.

BR-I (BO ) *
RW-1 (RW)
RW2 (RW)
EC-1 (CO )
US- 4 (US )
US6 (US )
US-5 (U S)
US-2 (US)
. . . . - - - - - - EE-2 (EE)
. . . . - - - - US- 1 (CA)
- - - RU-1 (RU )
. . . . . - - - - - - US-3 (US )
- - - - - US-2 (AU)

.-------f------

JP-1 (J P) *

Metalaxyl resistance
In ge nera l, the p opula t io ns of P. in festa ns
that we studi ed were se nsit ive to th e
systemi c fung ic id e metalaxy l. Some
iso lates fro m all spec ies had intermedi ate
re sistance, but onl y f ive iso lates we re
res istant, and all came fr om potato. Hi gh
fr equ enc ies of metalaxy l resistance in th e
EC- 1 lin eage have bee n re ported prev iously
(Fo rb es et al ., 1997 ).

- - - - - - CA-5 (C A) *

..-------1.--------

CA-7 (CA ) *

- - - - - - CA-6 (CA )*
. - - - - - - - US-8(US)*

.-------1

1..------

P0-57 (RU) *
IL- 1 (IL)*
US-7(US ) *

' - - - - - - CA- 4 (CA ) *

. . . . - - - - - - EE-1 (EE )
. - - - - - - P0-4 (BY)
~--

ES -1 (ES)

. . . . - - - - - - - CR-1 (CR)
' - - - - - - - - AU- 1 (AU)

1.0

0.5

Genetic distance

Figure 2. Cluster analysis of clonal genotypes of

Phytophthora infestans, including isolates
from wild hosts in Ecuador, based on a
distance coefficient (Jaccard) for
multilocus genotypes consisting of RFLP
fingerprint with the RG57 probe, mating
type (asteris ks indicate A2 mating type),
and dilocus allozyme genotype (Gpi and
Pep). Genotype labels are the International Organization for Standardization
(ISO) two-letter country code plus a
unique number (Forbes et al ., 1998).

44

Pololo

Pathogenicity
Th e prima ry purpose of pat hoge ni c ity
tests was to detect host sp ecific ity amo ng
iso lates of th e sa m e c lon al lin eage fro m
di ffe rent hos ts. Two syste m s we re tested in
thi s stud y. Th e iirst in vo lve d on e iso late
fr om to m ato and three fro m 5. m urica tu m
(5. mu r.) (U S-1 ). Th e seco nd invo lve d o ne
isola te fro m potato and six fr om 5 .
colomb ianum (5. co l.) (EC-1 ). In th e f irst
test, eac h isolate wa s mo st agg ress ive
(ba sed o n les io n length) o n its ow n host.
Th e iso lates fro m to m ato did not attac k 5.
muricatum, but iso lates fro m 5. murica tum
attacked toma to, although less agg ress ively
th an th e iso late fro m thi s ho st (Fi gure 3).
Th e interacti o n betwee n iso late sou rce and
in oc ul ated spec ies w as hi ghl y signi fica nt
(P = 0.0002).
Th e seco nd test gave di ffe rent res u lts.
Isola tes fr om potato and 5. co lombianum
attac ked eac h host w ith equ al aggress iveness (Fi gure 3) . Th e intera cti o n betwee n
isol ate so urce and ino cul ated spec ies was
not signifi ca nt (P = .7355) .

which are 1e ported for the fir st tim e here:

EC- 2 from 5. brevifolium and 5.
tetrapetalurn and EC- 3 from 5. b etacea. At
thi s tim e, EC-3 lin eage is represe nted by
one ge noty pe as no pol ym orph ism with in
th e Ii neage was d etected. Ea c h Ii neage,
except EC-3, was assoc iated w ith mo re than
on e host, but rarely were ho sts associated
w ith more th an one lin eage.

A - S. colombianum
4

Isolates
from
S. colombianum

Isolates
from
potato

0
S. col.

Potato

S. col.

Potato

B - S. muricatum
5

Isolates
from S. murica tum

Isolates
from tomato

0
S. mur. Tomato

S. mur.

Tomato

Figure 3 . (A) Histogram of mean lesion length on

leaflets on potato and 5. col. plants

caused by isolates collected from each host
in Ecuador. (BJ Histogram of mean lesion
length on leaflets of tomato and 5. mur.
plants caused by isolates of P. infestans
collected from each host in Ecuador. Error
bars represent 95 % confidence intervals.
Discussion
This study revea led that all isolate s assessed
belong to one of two lin eage s previously
reported fo r Ec uador: US-1 and EC-1
(Forbes et al. , 1997), and two new lineages

Iso lat io n of pathogen genotypes from a

host is not an indi cat ion that th ey are
prim a1y pathogens of that host. Th ere is
ve ry stro ng hos t spec ifici ty in th e population of P. infestans in Ec uador fo r potato and
tomato, ye t o n rare occas ion s we have
found potato genotyp es on tom ato
(Oy arzun et al., 1998).
Pathoge ns ma y at tim es infec t alternat ive
hosts , albeit wea kl y. Whe n we collec ted
isol ates from 5. ochrantum in Ca rchi (a ll
we re EC-1 ), les ions we re diffi cult to find.
Mo st o f th em occurred w ithin th e pl ant
canopy, not o n exposed leaves. Whe n th ese
isolates we re rein ocu lated on leaves of 5.
ochrantum, they were not as aggress ive as
the US-1 iso lates coming from th e sa me
ho st in th e ce ntral part of th e co untry.
Th erefore we ass um e that the US-1 popu lation is th e primary pathogen pop ul ation of
5. ochrantum, and th at th e EC-1 population
is only w ea kly pathogeni c on 5. ochrantum.
Disease on wi ld hosts is generally patc hy,
and th e US-1 popul at ion may not have
bee n prese nt in Carchi w hen we sa mpl ed .
Th e EC- 1 iso lates from 5. ochranturn are
identi ca l for our mark ers to tho se from
pota to; both ho sts may be attacked by the
same pop ul at io n. If thi s we re tru e, 5.
ochranturn wo uld have bee n exposed to
abund ant aeria l in ocu lum beca use th ere is
extensive potato production in th e reg ion .
Th erefo re, EC-1 may be a weak path oge n of
5. ochrantum, causing so me les ion s on
old er leaves gro wi ng in highl y hu m id
conditions w ithin th e ca nop y.
Th e EC-2 isolate from potato was also
found to be weak ly patho geni c w hen
rein oc ul ated o n potato, indicatin g th at it is
CIP Program Report 1997-98

45

probably not a p1im ary p at hoge n of th at

host. We are analyzing patho ge nic aggressive ness w ith EC-1 and US-1 isolated from
5. anc/rean um to dete rmine if one lin eage
represents the primar y pathogen population
of th at host.
Our analyses w ith n eutral m ark e1s gave
littl e in sight into ho st spec ificity w ithin
c lonal linea ges in Ecuador. To get m o re
information on host sp ec ific it y, we did tw o
pathogenicity tri a ls, eac h in the fo rm of a
quadratic che ck, similar to what was done
to co nfirm host sp ec ificity for po tato and
tomato i n Ecuad o r (Oyarz un et al., 1998) .
O ne test cl ea 1ly d emonstrated that iso lates
of US-1 attackin g tomato are di ffere nt from
iso lates of US-1 attack in g 5. muricatum.
Th e results of this tes t are consistent w ith
fi e ld observations. We have seen the two
plant species grow in g in close pro x imity
and onl y one in fec ted w ith P. infestans. We
d o not know if thi s differential re ac tion is
d ue to v irulen ce (different Ph ge nes) or
1ep rese nts differe n ces in quan t itati ve ho st
spec ifi c ity, as in th e case of pota to and
tomato (Oyarzun et al., 1998).
Th e seco nd agg 1ess iveness te st failed t o
demonstrate a differe nce be twee n iso lates
attac king potato and those attacking 5.
co lombianum . Th erefore, we conclude that
th e sa me patho ge n population attacks both
c ul tivate d potato and 5. colombianum.
Future research w i 11 in volv e more pathoge n icity trials of thi s ty p e to determin e
w het her other cases of host sp ecific ity ca n
be detected w ithin clo nal lineages in
Ecuador. We also plan to use D NA fin ge rp1in t in g, usin g th e amplified fra gment
length polymorphi sm (AFLP) t ec hnique in
this endeavo r. Fo r now, howeve r, phenotyp ic eval uation of pathogenic aggressi ve n ess is the tool th at g ives the greatest
di sc rimination of hos t-specific popul ations
occ urring within the sam e lineage.

Conclusions
Th e ge netic di ve rsity of this pathoge n in
Ec uador is extremely w ide . Four clonal
lineages were found in association with
46 Potato

d iffe rent host species in the genus 5olanum.

Th e di ss im i larity between the novel c lon al
lin eages (EC-2 and EC-3) and other known
clonal l in eages leads us to the hy poth es is
that if the origin of EC- 2 is Mex ico, its
introducti o n into South Ame rica must h ave
in vo lved an unreported m igration. Both
line ages are quite di ffe 1ent from any
d esu ib ed to date. It ap pea 1s highl y unlikely
th at e ither w as introdu ced o n potato tu bers,
w hich is co nsi dered th e primary means of
lo n g distance transp o rt of this pathoge n (Fry
et al., 1993 ). It is extremel y difficul t to
speculate on the tim e when EC-2 and EC-3
we re introduced in to South A merica , or on
th e m ec hani sm of introdu ctio n.
EC-2 and EC-3 are also of interest as
clonal lin eages bec ause of t heir simil arity to
each other. Both share three novel RFLP
b and s and the unu su al 76/ 100 Pep ge notype . We ha v e no goo d hypo theses to
explain w h y EC-2 and EC-3 are genetically
si mil ar. A nother interes ting aspect of EC-3
is that it possesses th e 86 Gpi allele.
Prev iou sly, this allele h ad bee n associated
w ith only the US-1 c lon al lineage, and a
few others w hich are ve ry similar to it
(Goodw in , 1997), although it has also been
fou nd in Mexico (Forbes et al. , 199 8).

References Cited
Andrivon, D. 1996. The origin of
Ph ytophthora in festa ns populations
present in Europe in th e 1840s: A critical
rev iew of histori ca l and sc ientifi c
ev iden ce. Plant Pathol. 45:1 027-10 35 .
Brommonschenk e l, S. H. 1988. Pathoge nicity, co mpatibility, cyto ge netics and
isoe n zy m e pattern s of Brazilian iso lates
of Ph ytophthora infestans (Mont.) de
Bary. M.S. Thesi s, U ni v ersidade Federa l
de Vi<;:os a, Vi<;:osa , Brazi I. 82 p .
Carter, D. A., S.A. Archer, and K.W. Buck.
1990. Restriction fra gment length
polymorphisms of mitochondrial DNA of
Phytophthora in festans . Myc ol. Res.
94:1 123-1128.
Caten, CE. and J.L . Jinks. 1968. Spontaneous variability of single isolates of

Phytophthora infestans. I. Cultural

variation. Can. J. Bot. 46:329-348.
Colon, L.T., D.J. Budding, L.C.P. Keizer, and
M.M.J. Pieters. 1995. Components of
res istance to late blight (Phytophthora
infestans) in eight South American
Solanum spp. European J. Plant Pathol.
101 :441-456.
Correll, D.S. 1962. The potato and its wild
relatives. Texas Research Foundation,
Renner, TX, USA.
Davis, B.J. 1964. Disc electrophoresis - 11.
Method and application to human serum
proteins. Ann. NY Acad. Sci. 121 :404427.
Drenth, A. 1994. Molecular genetic
evidence for a new sexually reproducing
population of Phytophthora infestans in
Europe. Ph.D. Thesis, Wageningen
Agricultural University. 150 p.
Erselius, L.J., M.E. Ordonez, P. Ramon, P.
Oyarzun, and G.A. Forbes. 1 998.
Phytophthora infestans in Ecuador:
Evidence for population segregation on
different hosts. 7th International Congress of Plant Pathology held 9-16 Aug .
1998 in Edinburgh, Scotland. British
Society for Plant Pathology.
www.bspp.org. u k/i cpp9 8/abstracts/6/
166 .htm l
Erwin, D. C. and O.K. Ribeiro. 1996.
Phytophthora Diseases Worldwide. APS
Press, St. Paul, MN, USA.
Forbes, G.A., X.C. Escobar, C.C. Ayala, J.
Revelo, M.E. Ordonez, B.A. Fry, K.
Doucett, and W.E. Fry. 1997. Population
genetic structure of Phytophthora
infestans in Ecuador. Phytopathology
87:375-380.
Forbes, G.A., S.S. Goodwin, A. Drenth, P.
Oyarzun, M.E. Ordonez, and W.E. Fry.
1998. A global marker database for
Phytophthora infestans. Plant Dis.
82:811-818.
Fry, W.E., A. Drenth, L.J. Spielman, B.C.
Mantel, L.C. Davidse, and S.S. Goodwin.
1991 . Popu I ation genetic structure of
Phytophthora infestans in the Netherlands. Phytopathology 81 :1330-1336.
Fry, W.E., S.B. Goodwin, A.T. Dyer, J.M.
Matuszak, A. Drenth, P.W. Tooley, L.S.

Sujkowski, Y.J. Koh, B.A. Cohen, L.J.

Spielman, K.L. Deahl, D.A . Inglis, and
K.P Sandlan. 1993. Historical and recent
migrations of Phytophthora infestans:
Chronology, pathways, and impli cations.
Plant Dis. 77:653-661.
Galindo, J. and H.R. Hohl. 1985.
Phytophthora mirabilis, a new species of
Phytophthora. Sydowia 38:87-96.
Glendinning, D.R. 19 83. Potato introductions and breeding up to the early 20th
century. New Phytologist 94:479-505.
Goodwin, S.S. 1997. The population
genetics of Phytophthora. Phytopathology 87:462-473.
Goodwin, S.S., B.A. Cohen, and W.E. Fry.
1994. Panglobal distribution of a single
clonal line age of the Irish potato famine
fungus. Proc. Natl. Acad. Sci. USA
91:11591-11595.
Goodwin, S.S., A. Drenth, and W.E. Fry.
1992. Cloning and genetic analyses of
two highly polymorphic, moderately
repetitive nuclear DNAs from
Phytophthora infestans. Curr. Genet.
22:107-115.
Goodwin, S.S. and W.E. Fry. 1992. Host
specificity as a pre- and post-reproductive isolating mechanism in
Phytophthora. Phytopathology 82:1139
(abstr.).
Goodwin, S.S., L.J. Spielman, J.M.
Matuszak, S.N. Bergeron, and W.E. Fry.
1992. Clonal diversity and genetic
differentiation of Phytophthora infestans
populations in northern and central
Mexico. Phytopathology 82:955-961.
Goodwin, S.S., L.S. Sujkowski, and W.E.
Fry. 1995. Rapid evolution of pathogenicity within clonal lineages of the potato
late blight disease fungus. Phytopathology 85:669-676.
Griffith, G.W. and D.S. Shaw. 1998.
Polymorphisms in Phytopthora infestans:
Four mitochondrial haplotypes are
detected following PCR amplification of
DNA from pure cultures or from host
lesions. Appl. Environ. Microb. 64:40074014.
Hohl, H.R. and K. lselin. 1984. Strains of
Phytophthora infestans from Switzerland

CIPPrngrom Report1997-98

47

with A2 matin g type behav iour. Tran s. Br.

M yco l. Soc. 83 :529-530.
Koh , Y.J., S.B. Goodwin , A.T. D ye r, B.A.
Cohen, A. O gos hi , N . Sato, and W.E. Fry
1994. Migrations and displaceme nts of
Ph ytophthora in festa ns popu Iations in
eas t As ian countries. Phytopatholo gy
84:922 -92 7.
Lebreton , L. and D. Andrivon. 1998. French
isolates of Ph ytophthora infestans from
potato and tom ato differ in ph enoty pe
and genoty pe. European J. Plant Path.
104:583-594.
Lega rd , D.E. , T.Y. Lee, and W. E. Fry. 1995.
Pathoge nic specia Iization in
Ph ytophthora infestans: Aggress iveness
on tomato. Ph ytopa tholog y 85:1356 1361.
M atuszak, J. M., J. Fernandez-Elqu eza bal ,
W.K . Gu, M. J. Villarreal-Gonzal ez, and
W.E. Fry. 1994. Sensitivity of
Phytophthora infestans popu Iati o ns to
metalaxyl in Mex ico: Distribution and
dynamics . Pl ant Di s. 78 :9 11-916.
Oyarzun, P.J. , M .E. Ordo r'iez , and G.A .
Fo rbes. 19 97. First report of
Phytophthora infestans A2 matin g type in
Ec uador. Pl ant Di s. 81: 311.
Oyarzun, P.J. , A . Po zo, M. E. Ordon ez, K.
Doucett, and G.A. Forbes. 1998. Host
spec ific ity of Phytophthora infestans on

48 Potato

tomato and potato in Ecuador. Ph yto pa thology 88 :265- 271 .

Ri ver-a -Peria , A. 199 0 . Wild tuber-bearing
species of Solanum and incidence of
Ph ytop hthora infestans (Mont. ) de Bary
on the wes tern slopes of the volcano
Nevada de Toluca. 5. Type of resi stance
to P infestans. Potato Re s. 33:479-486.
Spi elman , L.J. 1991 . lsoe nz yrnes and
popu lation genetics of Ph ytophthora
infestans . In: Shatto ck, R.C. , D.S. Sh aw,
L.R. Cook , and J. A . Lu cas (eds. ).
Phytoph thora. Cambridge University
Press, Camb rid ge, UK. p. 231-241.
Spi elm an, L.J., B.J . M cMaste r, and W .E. Fry.
1989. Ev idence aga in st potato and
tom ato host specifi c ity in Phytophthora
infestans. Ph ytopatholo gy 79:114 5
(abstr. ).
Tool ey, P.W., B.A. Bunyard , M .M. Carras,
and E. Hatzi louka s. 1997. Developm en t
of PCR primers from internal tran scribed
space r region 2 for detection of
Ph ytoph thora species infecting potatoes.
A ppl. En v iron. M icrobial. 63:146 7-1 47 5.
Too ley, P. W., C.D. Therrien , and D.L. Ritch .
1989. Mati ng type, rac e composition,
nu c lear DNA content, and isoz y me
analys is of Peru v ian iso lates of
Phytophthora infestans. Ph ytopatholo gy
79 :478-481.

Host Specificity of Phytophthora infestans on Tomato

and Potato in Uganda and Kenya
L.J. Erselius1, M.E. Vega-Sanchez1, A.M. Rodriquez1, 0. Bastidas1, H.R. Hohl2, P.S.
Ojiambo3, J. Mukalazi3 , T. Vermeulen4, W.E. Fry4 and G.A. Forbes 1

Host spec i ficity of Phytophthora infestans,

the potato late blight (LB) pathogen, has
obvious epidemiologica l consequences in
are as where two o r more potential hosts
grow in close proximity. If inoculum can
pass readily from one host to the other, or if
more than one host is cultivated, then
disease management activities must take
both hosts into consideration. In addition to
that practical co nsid era tion, specificity of P.
infestans to potato (Solanum tub eros um)
and tomato (5. !ycopersicon) is an interesting ph enomenon for study beca use it
appears to be ca used by quantitati ve
differences in epidemic components, rather
than an abil ity to cause disease (Oyarzun et
al., 1998) . Elucidation of the mec hanisms
that govern this type of ho st spec ifi c ity
co uld provide new insi ghts into the nature
of host-pathogen relation s.
In sp ite of numerous stud ies on host
specificity of P. infestans to potato and
tom ato , severa l aspects remain unclear.
Many isolates attacking tomato and potato
in some regions of the Netherlands (F ry et
al. , 1991) and in North America (Goodwin
et al., 1995a; Legard et al., 1995) could not
be distinguished by restriction fragment
length polymorphism (RFLP) fingerprint,
dilocus all ozyme genotype, or mating type.
That implies they might be of th e same
pathogen population. Furthermore, many
isolates co ll ected from tomato in North

1 CIP, Quito, Ecuador.

2 Zumikon, Switzerla nd.
3 CIP, Sub-Saharan Africa.
4 Corne ll University, Ithaca, New York, USA.

America were highly aggress ive on potato

in a detached lea f assay (Legard et al. ,
1995), again indicating that one population
attacks both hosts. In contrast, the same
genetic markers have shown that distinct
genotypes are associated with eac h host in
Brazil (Brommonschenkel, 1988), northwest
Mexico (Goodwin et al., 1992a), the
Philippin es (Koh et al. , 1994), one region of
the Netherlands (Fry et al., 1991 ), France
(Lebreto n and Andrivon, 1998), and
Ecu ador (Oyarzun et al. , 1998) .
Studying the population of P. infestans in
North America, Legard et al. (1995) concluded that pathogenic aggress iveness on
tomato evo lved from with in the potatoaggressive popul at ions. All tomatoaggressive genotypes were sti 11 aggressive
on potato, so aggressiveness on tomato was
not associated w ith any measurable loss in
aggressiveness on potato. However, this
mode l of evo luti on toward aggressiveness
on tomato is not universally accepted.
Turkenstee n (1973) argued that if a genotype of P. infestans were able to acquire
equal aggressiveness on both ho sts with no
los s of fitness, it wou Id rapidly repla ce
other genotypes that are aggressive on only
one host. Th e presence of host-specific
populations in many parts of the world
supports the hypothe sis that dually aggressive genotypes do not readily occur, or if
the y occur they are less fit th an those
occurring on one host are. Geographical
separation of hosts could theoretically lea d
to separate pathogen populations. In
Ecuador, however, tomatoes and potatoes
growing in c lose proximity to one another
were found to have sepa rate pathogen
populations (Oyarzun et al., 1998) .

CIPProgrom Rep0111 997-96

49

In cases w here host spec ificity has been

corroborated w ith mo lecu lar mark ers, the
pathogen genotypes were from different
c lonal l ineages. Thi s adds an unquantified
variabl e, beca use it is not known if the
highl y spec ialized forms of the path ogen
evo lved sy mpatricall y, or were in troduced
from sepa rate so urces. A n ideal system to
study host specific ity woul d in vo lve two
host-specific populati ons that evo lved
sympatrica ll y. Potatoes and tomatoes are
grown together or in c lose proximity to one
another o n sma ll farms in Uganda and
Kenya. Our prelimin ary investigation
(unpu bl ished data) in d icated th at iso lates of
P infestans from th at reg ion may provide
such a syste m . Th e obj ect ive of the
researc h repo rted here was to test th e
hypo th es is tha t potato and tomato populations from Ken y a and Uganda differ bu t
belon g to the same clonal linea ge. We also
discu ss ev idence supporting sy mpatric
evolution of host- spec ific population s in
th is region.

Characterization of isolates
RFLP finge rprints we re obtai ned fo r al l
iso lates from both co ll ections usin g the
moderately repetitive probe RG 57
(Goodw in et al., 1992b). Two g of DNA
from each isolate were digested wi th fcoR I
for 24 h, the n und erwe nt electrophoresis on
0.7 or 0.8 % agarose gels (56 V, 20 mA) for
24-45 h in TBE 1 X. H ybr id ization and
detection we re do ne usi ng the no nradioacti ve kit EC LTM 5 according to th e
manufactu1er's instructions.
M itochon drial DNA (mt D NA) haplotypes
were determi ned fo r th e 1995 col lection by
amplificat ion of DNA of each iso late using
primers desig ned for specific regions of the
mitoc hondrial genome of P infestans
(G1iffith and Sha w, 1 998). Di gestion of the
amp li fied regio ns with Cfol, Mspl and fcoRI
restr iction enzy mes y iel ded band patte rn s
by w hi ch the isol ates co uld be c lassified
in to fo ur hap lotypes: la, lb, Ila, and ll b
(Carter et al., 1990; Griffith and Shaw,
1998).

Materials and Methods

lsozy me electropho res is for th e enzy mes
glucose-6-ph osp hate isomerase (Gpi) an d
peptidase (Pep ) was done on ee l lu lose
Data reporte d here co me from two
co llection eve nts. Th e first took pla ce in
aceta te (Goodwi n et al., 1995b) fo r the
March 1995 in Ken ya and Uganda. A
1997-98 co llection and on pol yacryla mide
sample of 14 isolates from tomato and 1 0
ge ls for the 1995 co l lection . Po iyacry lafrom potato were col lected in Ken ya; 7
mid e ge l electrophoresis (PAGE) was done
isolates from tomato and 8 from potato
usin g 1 mm thick 7 .5% ge ls w ith 25 mM
we re co ll ected in Uganda. Th e specific
Tri s-0. 19 M glycine, pH 8.8 as separa tin g
regi ons w ith in countries we re recorded
ge l and electrode buffer. Band s were
(data ava il able upon req uest). Th ey are not
clearer whe n a 1 cm stacki ng gel (2.5%
reported here becaus e we did not find
ac ry lamide 0.0 6 M Tris-HCI , pH 6.7) was
var iab ility assoc iated with co lle ct ion site. A
used (Dav is, 196 4). PAGE ge ls were run
second co llection trip was made in Kenya
wit h a co nstant current of 5 mA fo r 1 h,
and Uganda in late 1997 and early 1998.
then in creased to 1 0 mA. Vo ltage rose
Fifty-eigh t isolates we re co ll ected on th is
continuous ly th ro ughout, from about 50 to
trip, al l from potato.
280 V. Electroph oresis was term inated
, when the bromoph enol b lu e dye reac hed
P in festa ns w as iso lated from potato
th e bottom of th e ge l, abo ut 1 6 cm.
leaves by trapping the isolates in potato
A llozyme genotypes fo r both cel lu lose
tuber sl ices (Fo rbes et al., 199 7) . P
acetate an d PAGE we re scored as described
infestans from to mato grew poorl y on tuber
in Spielman (1991 ), w hi ch represent the
slices and was, therefo re, isolated by
relative mobiliti es of th e enzyme al leles to
pla c in g 1 c m2 pieces of infected tissue on a
se lective medium (Oya rzun et al. , 1998).

Collection of isolates

5 Ame rsham, Inc ..

50

Pototo

an allele designated as 100. Isolates with

known all eles from the co ll ect ion of W. E.
Fry were used for comparison.

P P

P T T T

The iso lates collected in 1997-98 were

tested for resistance to 5 and 100 mg/ml
metal axy l in 10% un c lar ifi ed VS med ium
and c lass ifi ed as resistant, intermedi ate, or
sens itive. Cond ition s of the test and criter ia
for classification we re described previou sly
(Forbes et al. , 1997).
Two pathogenicity tests were done to
compare aggressiveness of a subset of
iso lates from 1995 on potato and tomato.
In the first, five potato and five tomato
isolates were inocul ated on three potato
cultivars (Yungay, Cruza-148, and Chata
Blanca) and three tomato cultivars (Caribe,
Flora Dade, and FMX-193) . Th e second
test in vo lved six isol ates from the 1995
collection. Four were rep ea ted from the first
test (two from each host) an d two (o ne from
potato and o ne from tomato) were assessed
for th e first time. The same potato and
tomato cu ltivars were used in th e seco nd
test. Experim ental design, sta tistical
mod els, and evaluation procedures were as
reported previously (Oyarzun et al., 1998).

~~~

-21
-20
-16
_ -14a
14

-13

-10
-9

... -7

-5

' -3

Results
-1

All isol ates collected in 1995 and 1997 -98

belonged to the US-1 c lona l lineage, based
on RFLP f ingerp rint (Fi gure 1), Pep phenotype of 92/ 100, and, for the 1995 collection, an mtDNA haplotype of IB (Goodw in
et al., 1994; Griffith and Shaw, 1998). All
isolates from potato had the US-1 Gpi
phenotype 86/l 00 (Goodwin et al., 1994),
which is typical of US-1 , but all tomato
isolates were 1 00/ 100 at this locus (Figure

Figure 1 . RFLP fingerprints of three potato (P) and

three tomato (T) isolates collected in

Kenya and Uganda in 1995. These are the
same isolates shown in Figure 2. All
represent a typical US-1 fingerprint.
Numbers on the right represent conventional numbering of bonds (Goodwin et
al., 1992).

2).

A hi gh level of metalaxyl resistance was

found among the iso lates collected from
potato in 1997-98 (those co llected in 1995
were not assessed). The overall percentage
of resistant isolates was 73%, with 86% in
Kenya and 59% in Uganda. Metalaxyl
resistance was found in all districts

sampled . There were no pronounced

differences between th ese areas. We have
no quantitative inform at ion on th e use of
meta laxy l on potato in the two countries,
but worke rs in the region believe that
meta laxyl is seldom used on either crop (P.
Ewe ll , CIP, Nairobi, pers. comm.).

CIP Program Report 1997-98

51

Isolates
from tomato

86-

100-

Isolates
from potato

~ 3
c:

c:

Figure 2. Glucose-6-phosphate isomerase bonding

pattern for three potato and three tomato
isolates collected in Kenyo and Ugando in
1995 . These are the same isolates shown
in Figure 1. The three potato isolates
have an 86/l 00 phenotype ; the tomato
isolates are 100/100. Electrophoresis was
done with cellulose acetate.
Th e 1995 sampl e co nsisted of A 1
iso lates onl y. Th e iso lates from the 199798 sa mpl e we re tested for matin g type at
Co rn ell University and also fo und to be A l .
Ear li er th ey we re paired w ith an A 1 tester in
Kenya, and th e numbe r of oospores was
low (1-1 O/ plate) in 20% (13/ 66) of th e
pairin gs. Variou s oospore counts we re also
fo und in 31% (18/5 8) of the 1997-98
iso lates of sin gle cultures. A few iso lates
produced hundreds o r even th ousa nds of
oos pores in sin gle culture. H oweve r,
se lii ng and matin g resul ts we re not strictl y
repea tabl e.
Th e in te raction betwee n the host and th e
so urce of th e pathogen (potato o r tomato)
was ev id ent by v isual ex amination of
p lotted mea ns of les io n length (Fi gure 3).
Thi s interaction was hi ghl y sign ifica nt for
both tests at P <0.0001 (Tabl e 1 ). Les ions
o n both hosts produ ced by iso lates from
potato we re always acco mpani ed by
no ti cea bl e nec ros is (Fi gure 3) . Iso lates from
tomato produ ced no necrosis on tom ato
leave s durin g th e 7 days after in fect ion ,
eve n thou gh ab und ant sporul at io n could be
seen. On th e oth er hand , w hen inocul ated
onto potato, these sa me isol ates indu ced
necros is. W e co nc lu de that host spec ific ity
is quantitat ive rath er th an qu alitat ive,
because iso lates we re more aggress ive o n
th eir prim ary hosts. Thi s is similar to a
situ ation rece nt ly desc ribed in Ecuador

52

Potato

~ 2

Q)

c:

"'
::!:
Q)

potato

tomato

potato

tomato

B
4
Isolates
from potato

Isolates
from tomato

.i:::

Ci 3
c:

c:
0

'iii 2
~

c:

"'
::!:
Q)

potato

tomato

potato

tomato

Figure 3 . Histograms of mean lesion length on

potato and tomato plants caused by
isolates of P infestans collected from
potato in Kenya and Uganda in 1995. (A)
Data from seven isolates each from
tomato and potato. (B) For the second
test, four isolates were repeated (three
from each host), and two new isolates
were assessed (one from each host) . Error
bars represent 95 % confidence intervals.
(O ya rzun et al., 1998 ) for iso lates of
different c lonal lineages attac king potato
and tomato.

Discussion
O ur data c lea rl y show that th e pathogen
popu lati ons attac kin g potato and tomato in
Kenya and Uga nda are different from each

Table 1. Analyses of variance from two experiments that test effects of origin (potato or tomato) of isolates of

Phytophthora infestons and inoculated host species (potato or tomato) on diameter of lesions (cm) in a
detached-leaf inoculation assay involving isolates collected in Kenya and Uganda in 1995.
Source

Degrees of freedom

Mean square

F Value

NT'
NT

l
8
4
43
60

3.71
17.23
21.19
3.37
14.43
0.36
0.25

4
4
24
108

l.52
42.9
30.6
5.43
3.27
0.48
0.38

Pr > F

First experiment
Isolate origin (O)
Host species (H)
0* H
Isolate embedded in Origin (1 0)
Plant embedded in host (PH)
lo* PH
Residual error b

58.78
9.36
40.00

0.0001
0.0001
0.0001

NT

Second experiment
Isolate origin (0)
Host species (H)
0*H
Isolate embedded in Origin (1 0)
Plant embedded in host (PH)
10 * pH

Residual error b

NT
NT
63.52
l l.27
6.79

0 0001
0 0001
0 0001

NT

NT = not tested. Mo in effectsof isolote origin ond host species were not tested because of their highly significant interaction,
0 * H. Thisinteraction was tested using the mea n square for theinteraction between individual isolates and plants, Isolate
* Plant (O ' H) This intera ction was also used to test Isolate (0), isolate embedded in origin, and Plant (H), plant genotype
embedded in plant species.
b Based on variance among petridishes, which are pseudo replicationsof the experiment.
0

other. Althou gh al I iso lates had the US-1

RFLP fingerprint, iso lates from tomato and
potato could be separated based on Gpi
genotype, agg ressiven ess on th e two hosts,
symptoms o n to mato, and even iso lation
c haracteri st ics. Isol ates from potato we re
eas ily trapped o n potato tuber sli ces;
iso lates from tomato were most eas il y
iso lated w ith se lecti ve med ium.
Spec iali za tion of P. infestans on potato
and to mato has been detected at other sites
(B ro mmon sc henkel, 1998; Fry et al., 1991;
Goodwin et al. , l 992 a; Koh et al. , 1994;
Oya rzun et al., 1998). In those cases the
two popul ations belonged to different
c lona l lin eages, or were different genotypes
from areas whe re sex ual recombi nation
occurs. Sub-Saharan Africa seems to be the
o nl y area studi ed to date w here two

coex istin g, host-spec ific popul ations belon g

to th e sa me clonal lin eage.
A lthou gh o ur resea rc h demonstrates a
common origin for th ese populations, it
does not prove that th ey developed sympatri ca l ly. Th e population attackin g tomato
cou ld have been an introduced population
th at evo lved independentl y of potato LB .
W e beli eve, however, that it is mo re likely
th at the tomato population in Kenya and
U ganda evo lved from the potato population
and was not introdu ced. The principal
re aso n for thi s is the apparent sca rc ity of
transport mec hanism s by w hi ch it wo uld
have been introd uced. We are unaware
th at to mato pl antl ets o r fruits are imported
in to thi s regio n, nor is th ere evid ence th at P.
in festa ns ca n be transported w ith to mato
seed.
CIPProg ram Report 1997-98

53

Potato tubers are generally considered to

be the primary long-distance transport
mechanism for the LB pathogen (F1y et al. ,
1 993). Reduced aggressi veness of the
tomato population on potato foliage and
tubers argues against this type oi transport.
It is more likel y that P. infestans w as
introduced into the region on potato and
then evolved agg1essiveness on tomato.
This vie w is consistent w ith that of the
authors of a recent stud y in the USA (Legard
et al., 1995), who proposed that tomato
aggressiveness develops within potato
populations. Our data, however, indicate
that aggressiveness on tomato in Kenya and
Uganda is associated with a loss oi aggressiveness on potato. Our isolates from
tomato were less aggressi ve on potato than
those isolated from the potato host. Our
tomato isolates also did not grow well in
potato tubers, as evidenced by the difficulty
in isolating them with tuber slices.
The time at w hich the mutation of the
Gpi locus occurred remains unclear, as w ell
as factors which may have led to the
apparent universality of 100/100 Gpi
phenotype within the tomato population .
Mutation from 86/ 100 to 100/100 within
the US-1 clonal lineage has been reported
previously (Forbes et al. , 1998; Goodwin et
al., 1994). There is no ev idence, nor logic
we kno w of, that would suggest a genetic
Ii nkage between tomato aggressi veness and
the Gpi genotype 100/ 1 00. In South
America , the 86/ 100 Gpi genotype occurs
in the US-1 clonal lineage attacking tomato
(Brommonschenkel, 1988; Oyarzun et al. ,
1998).
Our data also demonstrate that the
populations of P. infestans in this region are
not substructured geographicall y. In none of
the characters w e assessed was there an y
detectable pol ymorphism that could be
associated with a geographical region. The
area that was sampled in Uganda is in the
south w estern part of the countr y, the
principal potato-growing region. It is
contiguous with the potato-growing region
of Rwanda, so our resu Its can probably be
extrapolated to Rwanda.
54

Potato

The high level of resistance to metalaxyl

is noteworth y because it demonstrates the
degree to w hich this characteristic can
develop within the US-1 lineage. Metalaxyl
resistance has frequently been associated
with other lineages or sexual populations
that have occurred as a 1esu It of rec ent
migrations oi the pathogen (Fry et al. , 1993;
Goodwin , 1 997 ).
References Cited

Brommonschenkel , S.H. 1988. Pathogenicity, compatibi I ity, cytogenetics and

isoenzyme patterns of Brazilian isolates
of Phytophthora infestans (Mont.) de
Bary. M.S. Thesis, Universidade Federal
de Vir;:osa , Vir;:osa , Brazil. 82 p.
Ca1ter, D.A., S.A. Archer, and K.W. Buck.
1 990. Restriction fragment length
polymorphisms of mitochondrial DNA of
Phytophthora infestans. M ycol. Res.
94:1123-1128.
Dav is, B.J. 1964. Disc electrophoresis - 11.
Method and application to human serum
proteins. Ann. NY Acad. Sci. 121 :404427.
Forbes, G. A ., S.B. Goodwin , A. Drenth , P.
O y arzun , M.E. Ordonez, and W.E. Fry.
1998. A global marker database for
Phytophthora infestans. Plant Dis .
82 :811-818.
Forbes, G.A. , X.C. Escobar, C.C. Ayala , J.
Revelo , M.E. Ordoi'\ez, B.A. Fry, K.
Doucett, and W.E. Fry.1997. Population
genetic structure of Phytophthora
infestans in Ecuador. Ph ytopathology
87:375-380.
Fry, W.E., S.B. Good w in, A.T. Dyer, J.M.
Matuszak, A. Drenth, P.W. Toole y, L.S.
Sujkowski, Y.J. Koh , B.A. Cohen, L.J .
Spielman , K.L. Deahl, D.A. Inglis, and
K.P. Sandlan. 1993. Historical and recent
migrations of Ph y tophthora infestans:
Chronology, pathways, and implications.
Plant Dis. 77:653-661.
Fry, W.E., A. Drenth , L.J. Spielman , B.C.
Mantel , L.C. Davi dse, and S.B. Goodwin.
1991. Population genetic structure of
Phytophthora infestans in the Netherlands. Ph ytopathology 81 :1330-1336.

Goodwin, S.B. 1997. The population

genetics of Phytophthora. Phytopathology 87:462-473 .
Good wi n, S.B., L.S. Sujkowsk i, A.T. Dyer,
B.A. Fry, and W.E. Fry. l 995a. Direc t
detect ion of gene flow and probable
sexual reproduction of Phytophthora
infestans in northern North America.
Phytopathology 85:473-479 .
Goodwin, S.B., R.E. Schneider, and WE .
Fry. l 995b . Use of ce llul ose-acetate
electrophores is for rapid identification of
al lo zyme genotypes of Ph ytophthora
infestans. Plant Dis. 79:1181-1185.
Goodwin , S.B., B.A. Cohen , and WE. Fry,
1994. Panglobal distribution of a single
clonal lineage of the Iri sh potato fam in e
fungus. Proc. Natl. Acad. Sci. USA
9 1:1 1591-11595.
Goodwin , S.B., L.J . Spielman, J.M.
Matuszak, S.N. Bergeron , and W.E. Fry.
1992a. C lonal diversity and genetic
differe ntiation of Phytophthora infestans
populations in north ern and central
Mexico. Phytopathology 82 :955-961 .
Goodwin, S.B., A. Drenth , and WE. Fry.
1992b. C lonin g and genetic ana lyses of
two highly polymorphic, moderately
repetitive nuclear DNAs from
Phytophthora in festans . Curr. Genet.
22:107- 115.
Griffith, G.W and D.S. Shaw. 1998.
Polymorphisms in Phytopthora infestans:
Four mitochondrial haplotypes are
detected fo llow in g PCR amp li fication of

DNA from pure cu ltures or from host

les ions. Appl . Environ. Microb. 64:4007 4014 .
Koh , Y.J. , S.B . Goodwin, A.T. Dyer, B.A.
Cohen , A . Ogos hi, N . Sato, and W.E. Fry.
1994. Migrations and displacements of
Phytophthora infestans popul at ions in
East Asian co untri es. Phytopathology
84:922-927.
Lebreton, L. and D. Andrivon. 1998. French
iso lates of Phytophthora infestans from
potato and tomato differ in phenotype
and genotype. European J. Plant Path.
104:583-594.
Legard, D.E., TY. Lee, and WE. Fry. 1995.
Pathogenic specia li zation in
Phytophthora infestans: Aggressiveness
on tomato . Phytopathology 85:13561361.
Oyarzun , P.J., A. Pozo, M.E. Ordonez, K.
Doucett, and G.A. Forbes. 1998. Host
specificity of Phytophthora infes tans on
tomato and potato in Ecuador. Phytopathology 88 :2 65 -271.
Spielman, L.J. 199 1. lsoenzymes and
popu lat ion genetics of Phytophthora
infestans. In: Shattock, R.C., D.S. Shaw,
L.R. Cook, and J.A. Lu cas (eds.).
Phytophthora. Cambridge University
Press, Cambr id ge, UK. p. 231 -241 .
Tur kensteen, L.J. 1973. Partial resistance of
tomatoes aga in st Phytophthora infestans
the late blight fungus. Th esis.
Wa ge ningen Agricultural University.
88 p.

CIP Program Reporl 1997-98

55

Genotype by Environment Reaction of Potato to the

Late Blight Pathogen
G.A. Forbes 1

Quantita ti ve or hori zo nta I resistance of

potato (So lanum tuberosum) to the
oomycete Phytophthora infestans (Mont.)
de Bary, th e ca usa l age nt of late blight (LB ),
does not co nfer abso lute protection. It is
genera ll y considered effect ive, howeve r,
aga inst all known rac es of th e pathogen
(Forbes and Jarvis, 1994). In co ntrast,
qualitative or vertical res istance to the
pathogen is hi ghly rac e-spec ific an d is
rapidly rend ered in effective because of
changes in the pathogen population
(Goodwi n et al., 1995; Wastie, 199 1). Th e
effecti ve ness of quantitative res istance
agai nst different populati ons of the pathogen was demonstrated rece ntl y both in
Europe (Turkensteen, 1993) and North
America (Inglis et al., 1996). In both cases,
th e authors demonstrated that res ista nce
rank in gs of ve ry old cultivars have remain ed co nstant over tim e, alth oug h
pathogen population s have changed (F ry
and Goodwin, 1997).

In response to the lack of inform at ion on

the stability of quantitative resistance to LB ,
CIP and a group of co ll aborators deve loped
a genotype by env iro nm ent (GXE) proj ect to
study resistance to P. infestans in a stand ard
set of culti vars at several sites worldwide.
Th ese culti vars differ in their leve ls of
resistanc e. Th eir res istance has not broken
dow n over time and is therefore considered
to be quantitative, althou gh that has not
been co nfirm ed by genetic studies. Th e
cu Iti vars possess few or no R genes that
co nfe r qua I itative resista nce.

Materials and Methods

Participants

O ne aspect of quantitative resista nce that

is still un c lea r is th e deg ree to wh ich a
culti va r se lected as res istant at one site w ill
also be res ista nt at anoth er. Haynes et al.
(1998) rece ntly demonstrated that res istance was stable at eight sites in the USA.
Another ex periment showed that cv. Alpha
was more res istant in New Yo rk, USA, th an
in Tol uca, Mexico, relati ve to three other
culti vars that behaved simil arl y in both
states (Pa rk er et al., 199 2). It appea rs,
howeve r, that no studi es of this type have
been done for a bro ader geograp hi c range.

El eve n sites were in cluded in the study,

but not all were used in ea ch of th e 3 yr of
th e ex perim ent (Tab le 1). Fourteen potato
cultivars were used in th e study: Alpha,
Bintj e, Robijn, Ei genh eim er, Record an d
Pimpern el from Netherl ands; Stirling,
Teena , an d Torr idon from Scotland; Chata
Blan ca and Yu ngay from Peru ; Monserrate
from Colombia; Seseni from Democ rati c
Republi c of th e Congo, and Cru za 148,
bred in M ex ico an d now grown extensive ly
in sub-Saharan Africa. The Scottish Crop
Res ea rch In st itute (SCRI), National Agri cultural Research In st itute (INRA) in France,
and CIP provided tubers or in v itro materia l.
In vestigators in the north ern latitudes found
it was diffi cult to get so me of th e shortdayle ngth mate ri als to produce tubers. In
tho se cases plant cutti ngs were used instead
of tub ers.

1 C IP, Quito, Ecuador.

A simil ar expe rim ental design was used

at eac h site, but there were som e sitespecifi c mod ificat ions. Cultivars were
sow n in pl ots of fou r or more plants in a

CIP Prngrnm Repoit 1997 98

57

Table 1. Sites and trial (years) of participation in the assessment of phenotypic stability of resistance in potato to

Phytophthora infestans.
Altitude
Participating institution'

Country

INTA, Balcarce
Agriculture Canada
Landbrugets Kartoffelfond

Argentina

International Potato Center

INRA, Ploudoniel
IN RA, LeRheu
Pl CTI PAPA

Ecuador
Fronce
Fronce
Mexico
Scotland

SCRI
CPRO, Wogeningen
Cornell University
Washington State University

Canada
Denmark

Latitude

Longitude

(m)

4600' N
5542' N
00 22' s
4800' N
48 01' N
1912' N
5523' N

63 00' w
0912' E
78 33' w
0400' w
01 43' w
99 35' w
432' w

15
79
3,058
36
100
2,640
110

Netherlands
USA
USA

Cultivar set used in each trialb

Trial l

Trial 2

Trial 3

1,2
1,2

1,2
1,2
1,2
1,2
1,2
1,2

1,2

1,2
1,2

1,2
1,2
1,2

1,2
1,2
1,2
1,2
1,2
1,2

a. INTA = Institute Nocional de Technologio Agropecuario, Argentina; INRA = lnstitut National de lo Reserche Ag ronomique,
Fronce; PICTIPAPA = Programo International Cooperation del Tizon Tardio de lo Popa, Mexico; SCRI = Scottish Crop Research
Institute; CPRO = Centre for Plant Breeding and Reproduction Research, Netherlands.
b. Cultivor set l = Bintie, Choto Blanco, Cruza 148, Monserrate, Sesini, Alpha, and Yungay; cultivar set 2 = Stirling, Teena,
Torridan, Reco rd, Rabiin, Pimpernel, and Eigenheimer.

randomized complete block design with 3

or 4 rep Iications. In some cases spreader
rows planted between the plots we re
inoculated. In other cases the plants being
studied were inoculated. In a few cases
plants we re left to natural infection.
In the 1st yr, distribution to some sites
was incomplete. Therefore not al I cultivars
were assessed at all sites. During the 1st yr
seven cultiva1s (set 1), Bintje, Chata Blanca,
Cruza 148, Monserrate, Sesini, Robijn, and
Yu ngay were assessed at seven sites
(A rgentina, Canada, Ecuador, France
[Ploudaniel]), Netherlands, Scotland, and
USA (Trial 1, Table 1); and the remaining
seven cu ltivars (set 2) w ere tested only in
Argentina, Canada, Netherlands, and
Scotland. In year 2, all 14 culti va rs were
assessed at nine sites (Trial 2, Table 1) and
in year 3 all 14 cultivars were assessed at
seven sites (Trial 3, Table 1). Onl y data from
the 1st seven sites are presented here.

58

Potato

At all sites in all years, percentage

infection at the plot level was estimated
visually every 4-8 d. Values were converted to the relati ve area under the disease
progress curve (AU DPC) as described
previously (Fry, 1978; Shan er and Finney,
1977).

Statistical analyses
Variances for each site-by-year com bi nation were calculated for culti va r, block, and
error effects (Table 2), and examined for
homogeneity between sites. Nonparametric analyses were conducted separately for
each year. Th is test does not require
homogeneity of variances and was developed for determining stability of culti v ars in
multilocation studies (Huehn, 1990a). This
nonparametric approach was used recently
in a similar study done in the USA (Haynes
et al., 1998).
In several cases in this study, values
were missing for certain cultivars at certain
sites. Although use of stabi I ity parameters

Table 2. Mean square errors (type Ill) for cultivar, block, and error effects of stability experiments carried out at sites in nine countries over 3 yr in the genotype x
environment (GXE) study.

Source

Argentina

Canada

INTA, Balcarce Agriculture

Trial

Canada

Denmark

Ecuador

France

Mexico

France

Netherlands

Scotland

USA

USA

Landbrugets

CIP

INRA,

PICTIPAPA

INRA,

(CPRO)

(SCRI)

(Cornell)

(WSU)

LeRheu

Kartoffelfond

Ploudaniel

Year l
Cul ti var

0.065

0127

0.357

0.032

0.114

0.086

0.164

Block

0.007

< 0.001

0.001

0.001

0.008

0.003

0.005

Error

0.003

0.002

0.004

0 001

0.041

0.001

0.003

Cultivar

0.020

0152

0.004

0.273

0.048

0.050

0.180

0.116

0.023

Block

0.017

0.005

<0.001

0.007

0.01 l

< 0.001

0.023

0.002

0.006

Error

0.004

0.006

0.001

0.010

0.001

0.001

0.008

0 001

0.001

Year 2

Year 3

3
~

U1
'()

Cultivar

0.003

0.034

0.008

0.018

0.042

0.002

0.031

Block

< 0.001

0.001

< 0.001

0.001

0.001

0.006

0.004

Error

< 0.001

0.003

< 0.001

0.001

0.001

0.049

0 001

is theoreticall y rob ust fo r mi ss in g d ata

(Hu ehn , 1990a), the co m p uter code u sed to
ca lc ul ate th ese param eters (Lu , 1995) did
not wo rk w ell w ith mi ss in g d ata. Fo r thi s
reaso n, missin g data we re repl aced w ith the
average rank of the cultivar fo r the year in
qu es tion. For the 1st y r, average 1anks we re
u sed for C hata Blanca and Robijn in
Argentin a and fo r Bintj e, Chata Blanca, and
Robijn in Ecuador. For th e 2 nd y r, average
ranks were used for Bintj e, Tee n a, Reco rd ,
and Pimpernel in th e USA (W as hington
State Unive rsity); Robi j n and C h ata Blanca
in Argentina; Stirling in Ec u ador; and
Torrid on in France (LeR heu). For the 3 rd y r,
ave rage ranks we re used fo r C hata Blanca,
Cruza 148, and Pimpern e l in M ex ico;
Robijn and Record in Arge ntin a; and
Stirlin g in the USA (Co rn e ll Uni ve rsity).
Th e or ig inal disease se verity d at a are also
bein g an alyzed in an effo rt to co nfirm th e
accu racy o f data co ll ect ion and co mpilation , and to poss ibl y red u ce th e number of
cases w here av eraged ranks mu st be u sed.
Cu lti va r stability was also exa min ed
graphi ca ll y by plottin g re lative AUD PC
(ve rti c al axis) for each c ultivar at eac h site
(hor izo ntal axis). For graphing purposes,
rel ative A UDPC was standardized by
dividin g it by the site m ea n . Th at was done
because some sites had very sma ll relati v e

AUDPC va lu es w hil e o th ers had ve ry large

va lu es.

Results
Mean square errors in ANOVA vari ed by
site for cu Iti var, b loc k, and error t erm s
(Tabl e 2). For th at reason , nonparam etri c
analyses we re used. Th e two n onparametri c param eters used in th e stud y are
denoted Sl and S2. Sl is th e average ra nk
dev iatio n of a cu lti v ar and for th at reaso n
has rath e r cl ear b io log ical inte rpretatio n.
S2 is th e va rian ce of the ranks and is hi ghl y
co rre lated w ith Sl. Both S parameters are
based o n corrected A UDPCs. The co rrection factor e limin ates the o ve rall effect of
resist ance (Hu ehn , 1990a) and results in
new AUDPC va lu es, w hich may not b e
co mpl ete ly corre lated w ith th e original
va lu es. T herefore, Sl and S2 are paramete rs of stab ility after correcting for res istan ce . Both param eters h ave been eva luated theoreti ca ll y (Hu ehn , 1990a) and in
practi ca l app li cations (Huehn , 1 990b).
Th e Sl and S2 parameters were quite
low in a ll alm ost a ll ca ses (Tabl es 3, 4 , 5) .
None of th e indi v idu al Z statistics we re
c lose to the va lu es for significance at P =
0 .05. N e ither were total Z statisti cs (a
m eas ure of overa ll c ulti va r stab ility )

Table 3 . Mean AUDPC, rank of mean AUDPC, mean of absolute rank differences of a clone (51), approximate test
of significance of 51 (Zl), common variance of ranks (52), and approximate test of significance of 52
(Z2) ; analysis of seven cultivars at seven sites in the lst yr of the GXE project.
Cultivar
Cruzo 148
Monserrate
Seseni
Yungay
Robi jn
Chata Bla nco
Bintje
Total

Mean AUDPC

Rank

51

Zl'

52

Z2

0.08
0.1 l
0.20
0.22
0.30
0.43
0.53

2
3
4
5
6
7

2.48
2.29
2.38
1.24
2.19
2. 10
2.76

0.14
0.00
0.04
4.32
0.04
0.14
0.89
5.57

4.67
3.67
3.95
l.1 4
3.81
3.14
5.29

0.18
0.04
0.00
3.30
0.01
0.30
0.67
4.50

a. Zstatistics ore measures of stability. The tests for the significance of the sum of Zl or Z2 are compared to a )(1 va lue of 14.07.
Individua l Zl or Z2 values are compared to a )(1 va lue of 7.24.

60 Pototo

Table 4. Mean AUDPC, rank of mean AUDPC, mean of absolute rank differences of a clone (Sl), approximate test
of significance of Sl (Zl), common variance of ranks (S2), and approximate test of significance of S2
(Z2); analysis of 14 cultivars at 9 sites in the 2nd yr of the GXE project.
Variety
Stirling
Torrid on
Cruza 148
Monserrate
Teena
Robijn
Seseni
Yun gay
Record
Eigenheimer
Chata Blanca
Pimpernel
Alpha
Bintje
Total

Mean AUDPC
0.04
0.05
0.09
0.15
0.17
0.17
0.21
0 22
0.30
0.35
0.36
0.36
0.38
0.44

Rank
2
3
4
5
6
8
9
10
11
12
13
14

Sl

Zl 0

S2

3.67
4.39
3.94
3.94
4.56
3.78
4.50
4.67
367
4.33
6.50
5.06
4.33
3.94

l.32
0.09
0.68
0.68
0 01
104
0.03
0.00
1.32
0.13
4.79
0.24
0.13
0.68
11.14

1000
14.53
l 0.75
12.28
15.44
10.94
14.94
14.94
10.19
13.19
30.36
21.44
13.44
11.75

Z2
1.28
0.10
0.99
0.52
0.02
0 92
0.06
0.06
1.20
0.31
6.54
0.89
0.26
0.67
13.81

Zstatistics are measures of stability. The tests for the significance of the sum of Zl or Z2 are compared to a X2 value of 23.68.
Individual Zl or Z2 va lues are compared to a X2 va lue of 8.49.

significant. But this statistic was much

high er in the 3rd yr than in the 1st an d 2nd
yr, indicatin g that c ultivar ranks were more
variable the 3rd yr. Overall, the nonparametric tests did not provide ev id ence for
instability of the res istan ce reactions of the
cu ltivars.
Visual examination generally corroborated the nonparametric statistics. Highly
res istant and hi ghly susceptible cultivars
were generally consistent at al I sites. The
cultivars ap pea red to vary more in their
reactions in the 3rd yr than in the 1st and
2nd yr (compare Figures 1 - 5), as indicated
by the nonparametric analysis.
There were some unusual results for
particular site-by-c ultivar combinations, but
most of these were not rep eated. For
exa mple, Chata Blanca was relatively
resistant at Cornell University (USA) in the
2nd yr (Figure 2), but was suscept ibl e in th e
3rd yr (Figure 5). Chata Blanca was also

more resistant than normal at Washington

State University (USA) in the 2nd yr (Figure
2). Data from the 3rd yr at that site are not
yet available for comparison.
The most surprising result that occurred
more than once was that Stirling was more
suscepti ble in Argentina than at other sites
(see Figures 2 and 5). Otherwise, Stirling
was one of th e most res istant cultivars at all
sites.

Discussion
This research indi cated that horizontal
resistance to LB is relatively stable. Tropically-adapted resistant materials (e.g.,
Cruza 148) were resistant in the temperate
areas of this study, and Torridon and Stirling
from Scotland were re sistant in Quito,
Ecuador. The primary cases of instability
appear to be iso lated eve nts, except for the
repeated increase in susceptibility of
Stirling in Argentina.

CIPProgram Report 1997-98

61

Table 5. Mean AUDPC, rank of mean AUDPC, mean of absolute rank differences of a clone {Sl), approximate test
of significance of Sl {Zl), common variance of ranks {S2), and approximate test of significance of S2
{Z2) ; analysis of 14 cultivars at 7 sites in the 3rd yr of the GXE project.

Variety

Mean AUDPC

Rank

Sl

Zl '

52

Z2

0.03
0.03
0.09
0.09
0.10
0.12
0.13
0.14
0.17
0.18
0.20
0.21
0.24
0. 28

2
3
4
5
6
7
8
9
l0
11
12
13
14

6.00
5.52
4.95
2.86
2.38
4.10
2.76
6.38
3.62
4.10
4.76
4.67
5.90
4.95

1.79
0.75
0.09
3 09
4.96
0.29
3.43
2.93
l.02
0.29
0.01
0.00
l.54
0.09
20.28

26.57
21.81
16.81
5.57
3.95
11.24
5. 29
28.33
9.33
14.81
16.29
15.24
23.81
17.1 4

2.52
0.73
0.01
2.70
3.58
0.59
2.84
3.45
l.1 3
0.05
0 00
0.02
l.35
0.02
18.99

Torrid on
Stirling
Teena
Cruza 148
Seseni
Monserrate
Robiin
Yun gay
Record
Pim pernel
Eigenheimer
Alpha
Chata Blanca
Bintie
Total

o. Zstatistics ore measures of stabil ity. The tests for the significance of the sum of Zl or Z2 ore compared too X2 value of 23.68.
Individual Zl or Z2 values ore compared too X2 value of 8.49.

Standardized, relative AUDPC

4
- - Bintje
Chata blanca
............ Robijn

Yun gay
- - - Monserrate
Cruza 148

- - Seseni

......

... .............. ........ . .. -... ,

.............. ...............

......... .......

1-

.............. -

..... ............................. ............ .

-

....................

..

0
Argentina

Canada

Ecuador

France

Netherlands

USA-Cornell

Figure 1. Seven cultivars tested at seven sites in the lst yr of the international genotype x environment (GXE)
experiment.

62

Potato

Standardized, relative AUDPC

. Cha ta blanca
Record
Pimpernel
Sesen1

... ..............

..

"" /
-...
/
"\
-----.:"'-~---..,.
~..,...,. ........ -~m ............... .

..... .

\ \

, ............. ......

'

........ . . . . ...

. . . . . ... . ..

Denmark

Argentina

France
LeRheu

Ecuador

Canada

....

Scotland
France
Ploudaniel

USA
Wash.

USA
Cornell

Figure 2. Seven cultivars tested at nine sites in the 2nd yr of the international GXE experiment.

Standardized, relative AUDPC

Bin tj e
.. Alpha
Eigenheimer
..................... Yungay
Monserrate
Cruza 148
.... . ..... Torridon

.....
..

.. .... -""

/ /
1 /

1 ~/
All

..........

~,..-@ ~-""*"<>;~;-

.......................... ,'

~<

.................................................................................................. ...................,...:.

............

... ...

----

..I

IO

- - -- - -- ........ - -

- - - I

I>

I.

,' ' ,,,

0
Argentina

Denmark
Canada

Ecuador

France
LeRheu

Scotland
France
Ploudaniel

USA
Cornell

USA
Wash.

Figure 3. Seven additional cultivars tested at nine sites in the 2nd yr of the international GXE experiment.

CIP Prngrnm Report 1997-98

63

Standardized, relative AUDPC

.............. Monserrate
Torridon
Cruza 148

---Bintje
- - - - - Eigenhe ime r
- - -- Alpha
- - - Yungay

...... ......

,,,, ,,,,

~.............

-:

""

---

- - -:,,,,,.,.,.::..:....----~

.,,,,. .,,,,. .,,,,.

-1

- -;1.>..- - - - - - - - ;
.....
""
'

: :. '." '. '":~. . . :....,

, .... . . . . . . . . .~1~.- . . . .~.:">'.~
.

Denmark

.,;

.,,.."

'

........................ .....

. .....

........

,.

0
Argentina

-----

France
LeRheu

France
Ploudaniel

Mexico

" ......~ '

Scotland

~ ...

USA
Cornell

Figure 4. Seven cultivars tested at seven sites in the 3rd yr of the international GXE experiment.

Standaridized, relative AUDPC

........
--- - ---

2.0

1.5

.,,, ..."'
1.0

'

Cha ta Blanca
Record
Pimpernel
Teena

,' ,'" ................. . ..... ,,-

Robijn
Seseni

.....

-"

. -.......... .......

--

,-..' ,'

,'

Stiriling

......... ... ...

-- -

- \- - - """'"' ""' '"'"' "" "'"" ""

-- ~ ... ...... ~ "'""'''....... ~_ ... ,"":::_:-... -.....
-: . .: ........ "' \
/
.._ ~
\ \,/ /

'

0.0
Argentina

Denmark

Fran ce
LeRheu

France
Ploudaniel

Mexico

Scotland

USA
Cornell

Figure 5. Seven additional cultivars tested at seven sites in the 3rd yr of the international GXE experiment.

64

Potato

The hi gh leve l of res istance of Stirling in

Ecuador and Scotland would appear to
argue against a relation between daylength
se nsitivity and resistan ce . Thi s relation
co uld have been tested with greate r
ce rtainty had more short-d ayl ength sites
participated in the study. Non ethel ess, other
factors may explain th e differenti al reaction
of Stirling in Argentina. Arge ntin a is know n
to have an unusual population of the
pathogen com prisin g both A 1 and A2
matin g types (Forbes et al., 1998). It
ap pears important to reeva lu ate Stirling in
Argentina to determin e whether it is truly
susce ptible there. Subs equ ently, res ea rch
could be designed to dete rmin e whether
,
the putative effect is related to environmental factors or pathogen genotype.
In ge nera l, this study co rrob orates a
rece nt and similar study in th e USA
(H ay nes et al. , 1998). S parameters in th e
USA study were also in signifi ca nt, althou gh
somew hat higher than those reported here.
Thus, two independent studies have arrived
at the same general conclusion: there is
littl e ev idence for phenotypi c in sta bility of
qu antitative resistance to potato LB.
Acknowledgment
I w ish to thank the followin g peopl e w ho
have wo rk ed ve 1y hard in the GXE st ud y:
A rgentina (S. Capezio, S. Distel, M. Huarte,
M . Van D amme), Canad a (G. MacKenzie,
H. Platt, R. Tarn), Denmark (H. G. Kirk, K.
Tolstrup), Ecuador (0. Bast id as, G. Chacon),
France (D. Andrivon, D. Ellisseche, R.
Pelle), Mexico (A. Hernand ez-Vil chis, H.
Lozoya -Saldana), Netherlands (L . Colon),
Scotland (J. Duncan, R. Lowe, G. Mackay, J.
McNicol, H. Stewa rt), USA (C. Brown, WE.
Fry, D. In glis, H. Mayton).
References Cited
Forbes G.A., S.B. Goodwin, A. Drenth, P.
Oyarzun, M.E. Ordof'i ez, and W.E. Fry.
1998. A global marker d at abase for
Phytophthora infestans. Plant Dis.
82:811-818.

Forbes G.A. and M.C. Jarvis. 19 94 . H ost

res istance fo r mana geme nt of pot ato late
blight. In : Ze hnd er, G., R. Jansson, and
K.V. Raman, (eds. ). Advances in potato
pest biology and management. Am eri ca n
Phytop ath o logic al Society, St. Paul , MN,
USA . p. 439-457.
Fry, WE. 1978. Quantification of ge neral
res istan ce of potato cultivars and fungicide effects for integ rated control of
potato late bli ght. Ph ytopatholo gy
68:1650-165 5.
Fry, W.E . and S.B. Goodwin. 1997. Res urgence of the Irish potato famine fungus.
BioS c ience 47:363-371.
Goodwin, S.B., L.S. Sujkowski, and W E.
Fry. 19 95 . Rapid evolution of pathogenicity within cl o nal lineages of the potato
late bli ght dis ease fungus. Phytop athology 85:669-676.
Ha ynes, K.G. , D.H. Lambert, B.J. Christ,
D.P. We in ga rtn er, D.S. Dou ches, J.E.
Ba cklund , G . Secor, W Fry, and W
Stevenson. 1998 . Phenotypic stability of
res istan ce to late blight in potato c lones
evaluated at eight sites in the United
States . Am. J. Potato Res. 75:211- 217.
Hu ehn, M. 1 990a . Nonparam etric measures of phenotypic stability. Part 1:
Theory. Euphytica 47:189-194.
Hu ehn , M. 1 990b. Nonparametric measures of phenotypic stability. Part 2:
App li cations. Euph yti ca 4 7:195-2 01 .
In gli s, D.A., D.A. John so n, D. E. Lega rd ,
W.E. Fry, and P.B . Hamm . 1996. Relative
resistances of potato clones in res ponse
to new and old populations of
Phytophthora infestans. Plant Dis.
80:575-578.
Lu, H.Y. 1995 . PC-SAS program for est im atin g Hu ehn 's nonparametric stability
statistics. Agron. J. 87:888-891.
Parke r, J.M ., H .D . Thurston , M.J. Villarea lGonzal ez, and W . E. Fry. 1992 . Stability
of di sease expression in the potato late
bli ght pathosystem: A preliminary field
study. Am. Potato J. 69:635-644.
Shan er, G. and R.E. Finney. 1977. Th e effect
of nitro gen fertilization on the expression
of slow-mildewing resistance in Knox
wheat. Phytopathology 67:1051-10 56.

CIP Program Report 199/.98

65

Turkensteen, L.J . 1993. Durable resistance

of potatoes against Phytophthora
infestans. In: Jacobs, T. and Parlevliet,
J.E . (eds.). Durability of disease resistance. Klu w er Academic Pub lishers,
Dordrecht, The Netherlands. p. 115-124.

66

Pototo

Wastie, R.L. 1991. Breeding for resistance.

In: Ingram, D.S. and Williams, P.H .
(eds.). Phytophthora infestans: The cause
of late blight of potato. Academic Press,
San Diego, CA, USA. p. 193-224.

Identification of QTLs for Late Blight Resistance in a

Cross Between S. phureja and a Dihaploid S. tuberosum
and Association with a Plant Defense Gene
M. Ghislain, B. Trognitz, R. Nelson, Ma. del R. Herrera, L. Portal, M. Orillo, and
F. Trognitz 1

Race nonspecifi c res istance to late blight

(LB) disease in potato, caused by
Phytophthora infestans, is thou ght to be
quantitatively inh erited and has been
c haracterized in seve ra I cultivated South
American potato spec ies (Co lon et al.,
1995; Trognitz et al., 1997). A cross between a LB resistant accession of th e
diploid cultivated potato, 5olanum phureja
(phu), and a susceptible dihaploid c lon e
from 5. tuberosum (d ih-tbr) produc ed a
seg regat ing (PD ) population of 246 progenies. The resulting hybrids were assayed
at two field sites under natural infection by
P. infestans . Evaluations of field resistance
revea led large variation-from hi gh partial
resistance to susceptibility (Trognitz et al. ,
1997). Randoml y ampl ified pol ymorph ic
DNA (RAPD) and amplified fragm ent len gth
polymorphic (AFLP) markers were used to
genotype 94 hybrids selected randomly
from the PD popul at ion (Ghislain et al.,
1997). To better understand the inh eritance
of horizontal 1esis tanc e to LB , we developed genetic map s and quantitative trait
loci (QTL) analysis using the field eva luation data. Assoc iation between QTL and
known genes was also tested by searching
fm linkage diseq uilibrium in selective
genotypes of the phenotypic extremes.

Materials and Methods

The PD population is the progeny of the
cross of 5. phurej a (phu) clone CHS-625 by
dihaploid 5. tuberosum (dih-tbr) clone PS-3
(Trog nitz et al., 1997). Field evaluations

1 CIP, Lima, Peru.

with naturally occurring P infestans

populations were done in both Quito,
Ecuador, and Huancayo, Peru. Horizontal
resi sta nce was quantified by the average of
the diseased foliage area (AD FA) taken six
tim es at weekly interval s and area under the
disease progress curve (AUDPC).
Molecular techniques used in the study
have been described elsew here (Ghislain et
al., 19 99) . Template D NA of five parents of
reference potato mappin g populations was
provid ed by H. van Eck and J. R. van der
Voort, W age ningen University, Netherlands.
Sixty-eight simple sequ ence repeats (SSR) or
micro sa tellite primer pairs corresponding to
loci on the potato genetic map were
provid ed by R. Waugh, Scottish Crop
Re sea rch Institute, Dund ee, Scotland. Two
restri ct ion fragment len gth polymorphism
(RFLP) markers per chromosome were
selected from the database of potato-tomato
RFLP probes. They were used to anchor the
phu and dih - tbr linkage gro ups to the
potato genetic map by linkage analysis.
RFLP probes were provided by C.
Gebh ardt, Max Planck ln stitut, Germany,
and S. Tanksley, Cornell University, USA .
Single-marker analysis was used to
detect assoc iations betwee n marker
(genotypic) classes (presence or absence of
the band) and their resp ective phenotypic
valu es. Significance of association was
determined with a two-sampl e Student's ttest. Interval mapping method using LOD
score (Lan de1 and Botstein, 1989) was
appli ed through the use of MAPMAKER/
QTL vl .1 b software.

CIP P<ogram Repo<t 1997-98

67

Results
Phenotypic data of the PD population
Fie ld eva lu atio ns of the 94 hybr id s used
to develop the genetic maps were compared between the two environments.
Good co rrelation, r = 0.75 and 0.77, were
observed between data sets using ADFA
and AUDPC, respecti vely. Such corre latio n
for data co ll ected in two distinct fie ld
environments under natural i nfestation, is
very high and may ind icate that sorlie Q TL
of resis ta nce to LB are common in both
env ironm ents.
Genetic linkage maps
DNA markers for the two parents are
expected to have a 1 :1 seg regat ion ratio
(pre sence of the band-absence of the band )
in th e F1 c ross between heterozygo us
parents under a pseudo-test cross model.
Several markers displa yed a skewed
segregation tested by a X2 test at P<0.01.
The number of markers w ith skewed
segregation observed from the phu parent is
similar to that reported for other genetic
analyses in plants. In contrast, 24% of
markers from the dih-tbr parent have a
skewed segregat ion at a P<0.01 , which is
unusuall y high. Both female- and malederived genet ic maps have been constru cted using the linkage analysis program
MAPMAKER/EXP v3.0b 2 The phu link age
map consists of 13 link age groups totaling
993.6 cM, of w hich 11 have been anchored
to known potato chromoso mes (Figure 1A3) .
The d ih-tbr lin kage map co nsists of 12
linkage groups totaling 889.9 cM, eac h
with markers anchored to the potato
genetic map (Fi gure 1 B).

Th e anc hor ing was performed with

inform ative RFLP, SSR, and AF LP markers
with known positions o n the potato genetic
map (Figure 1).
AFLP ancho rs : Ei ght out of 12 primer
combinations wi th known mapped AF LP

2 Wh itehead Institute, web: http://www-gen ome.wi.m it. edu/

genome_software; ftp:<ftp://ftp-genome. wi.mit.edu/
distribution/software/ mapmaker3>
3 Figures follow references c ited section.

68 Pototo

markers (Rouppe van der Voort et al.,

1998) turned out to be informative w ith
PD material and we re used to identify
the AF LP markers w ith known map
positions.
SSR anchors: 26 out of 68 prim er
combina tions proved to be informative
for the PD maps (polymorphic and
heteroz ygous lo ci) and amplified loci
w ith kno w n locatio ns in the potato
ge netic map.
RFLP anchors: 1 3 RFLP markers were
informative for the PD genetic maps.
Genetic analysis of field resistance
Notably, markers associated with fie ld
resistance were primarily from the dih-tbr,
su sce ptible parent, and in particular on
chromosomes Ill and XII (Tab le 1).

Seve ral QTL were detected using

interval-mapping analyses (Figure 1A and
1 B). Th e QTL analysis (significance of QTL
detection w ith a LOD score 2'. 2) re vea led
on ly minor QTL positioned on chro mosomes V II , XI, and XII of the phu map, and
3 QTL located on the dih-tbr map: 2 minor
on chromoso mes Ill and V, and one large
(exp lain in g up to 36% of the trait variation)
on chromosome XII , using the field data.
Selective genotyping of the
phenotypic extremes
A se ri es of anonymous RFLP markers, R
gene primers amplifying R gene-l ik e
seque nces (Rls) , and candi date genes we re
tested fo r linkage disequilibrium with the
most resistant and susceptib le extremes of
the PD population to test the hypothes is of
associat ion between known QTL for LB
res istance, defense genes in other material
or plant-pathogen systems and the PD trait
variation.

A set of 21 RFLP markers, previously

found to be associated w ith effects on
quantitati ve resistance to LB (LeonardsSchippers et al. , 1994), was used to probe
DNA extracted from the most resistant an d
susceptible plants of the PD population (1720 plants fo r each group). Among these,

Table 1. Number of markers associated with LB resistance (!-test at P< 0.05) and chromosome assignment of

S. phureja CHS-625 and the di haploid 5. tuberosum PS3.

Potato chromosomes

II

phu

Markers

AD FA-QUI

associated with

AUDPC-Q UI

PS-3

IV

v VI
0
0
0
0

0
0
0
0

19

0
0
11

0
0

VII

VIII

IX

AUD PC-HYO

0
0
0
0

Total analyzed

168

30

Markers

ADFA-QU I

41

18

AUDPC-QUI

39

22

LB resista nee

ADFA-HYO

0
0
0
0

associated with

36

38

0
0

0
0
0

48

lO

46

18

56

19

4
11

CHS-625 LB resista nee

dih-tbr

Ill

AD FA-HYO

AUDPC-HYO
Total analyzed

279

2
15

29

x XI

XII

0
0

11

12

0
0
0

29
29
29

20

29

0
0

10

9
2

0
0
11

29

ADFA = average of the diseased foliage area, QUI = Quito,

AUDPC = area under the disease progress curve, HYO = Huancayo.

four markers (CP116, GP313 , Cl21 , and

GP1 86) showed band s associated w ith
res ista nce in th e PD popul ation (Tabl e 2).
O ne of th ese mapped at a QTL pos iti on on
chromosome 111.
Of 47 segregating band s detected usin g
5 primer pairs correspo ndin g to co nserved
R gene domains, on e has been found to be
associated with th e quantitat ive res istance
observed in th e field w ith the PD population (see Rl s in Tab le 2) .
Several ca ndidate genes were se lected to
be tested for assoc iation w ith LB res istan ce
based on their role in plant defen se. The
intron of the PR 7Oa gene (formerl y STH-2)
amplifi ed and digested w ith Taql (c leaved
amplifi ed po lymorphi c sequ ence, CA PS)
displ ayed weak assoc iat ion w ith res istan ce
(Tab le 2). Three genes of th e phenylpropanoid pathwa y were tested by DNA hybrid ization analysis: ph enylal anine ammo ni um
lyase (PAL), 4-co um arate-CoA li gase (4CL),
and chalcone fl ava none isomerase (CHI).
Band s hybridi z ing with all three of these
genes showed signifi ca nt associat ions w ith
res ista nce . Seve ral PA L bands segregat ing

from both pare nts mapped on chromosome

Ill to a QTL position for the dih-tbr pare nt.
Thi s resul t mi ght be of function al signi ficance for this QTL. Th e osmotin clon e
pA 13, cl o ned from 5. commerson ii, hybridized w ith several ba nd s of w hi ch two from
the dih - tbr parent showed significa nt
assoc iati on w ith res istance. Th ese band s
cosegregated with th e res ista nce-associated
band hybrid izin g w ith GP313 (te ntati ve ly
on chrom oso me V). Oth er candida te genes,
probes encodin g lipoxyge nase, glucanase,
sol anidin e glu cosy l tran sferas e (SGT),
ca talase, and genes in vo lved in the
Pseudomonas-tom ato in co mpatible reaction Pto, Pti-4, Pti-5, and Pti-6, were tested
but no cl ea r association w ith res istan ce was
found.
Conclusion

Th e Q TL analys is reported here for the PD

population revea led an important QTL on
chromosome XII th at seems to be the most
interestin g. It was identified w ith both
parental maps and w ith a large effect on the
va ri ation of field res istance to LB in
Hu ancayo and Quito. Other Q TL were

CI PProgramReport 1997-98

69

Table 2. Candidate genes and markers associated with one of the contrasting phenotypic classes tested by xi
statistics, the PD population, and chromosome assignment to the phu and the dih-tbr maps {-not
mapped).
Marker type

Band/kb

RFLP probe

Cl2 l/fcoR VI 6.5

GPl 86/ EcoR 1/8
GP313/Xbol /6 5
CPl 16/ Drol / 3
ASl-52/0 23

Rls
CAPS'
Candidate genes
0

STH!Toql/4
PAUEcoRl/3
PAUEcoRl/2.8
PAUEcoRl/2
PAUEcoRl!l .9
PAUNdel/6 9
PAUNdel/6
PAUNdel/5 9
PAU Ndel! 4.4
PAUNdel/4
PAUNdel/3.6
PAUHindlll/5.5
PAU Hindi 11/5
PAU Hindi 11/2 8
PAU Hindi 11/2.7
4CUfcoRl/9

4CUHindi11/8
4CUNdel!4 5
osmotin/fcoRV/13
osmoti n/ Xbol!l 2
osmotin/Xbal/6.8

Total tested'

30
25
29
25
17
33
34
34
34
34
34
34
34
34
34
34
34
34
34
34
33
34
34
33
33
33

31
21
29
23
16
31
33
33
33
33
33
33
33
33
33
33
27
27
27
27
32
33
33
33
33
33

18
3
9
19
12

5
11
20
l
3
16

31
25

10

70

Potato

5
23
24
5

10

12
26
26
22
9
9
26
8
15
9
9
24
25
9
10
24
9

a. r = resistant, s = susceptible, * = P< 0.05, *' = P< 0.01, " '

sequence, CAPS = cleaved amplified polymorphic sequence.

detected in both parental genetic maps, but

with lowe r contributions to the phenotypic
variation. Some of the QTL detected in the
PO genetic maps locate at different map
positions homologous to those found by
Leonards-Schippers et al. (1994). Hence,
the genetic base of LB field resistance in ou1

xi

Observed

6
27
26
8
21
4
22
22
11
12
22
23
l0
24
=

po

Chromosome'

(D) (D) (D) V

(D) (P) (D) IX
(P) Ill
(D) Ill
(D) (P) Ill
(D) Ill
(D) (P) Ill
(D) Ill
(D) Ill
(D) Ill
(D) Ill
(P) (D) Ill
(D) Ill
(D) (D) Ill
(D) Ill
(D) (D) (D)-

7.78
6.11
4.17
14.8
4.87
6.48
10. l
12.7
5.52
4.37
13.6
22.4
8.67
9.55
8.77
9
9.32
4.15
8.96
8.96
4.19
4.19
5.85
5.12
5.76
6.82

P< 0.001, - = not mapped, Rls

Rgene-like

PO population seems to differ from the

published map locations. Therefore,
breeding for LB resistance can broaden the
genetic complex of field resistance. Future
deve lopments will be on the application of
QTL-associated markers to selection of
genotypes with field resistance to LB.

Acknowledgments

The authors are grateful to the dedi ca ted

students wh o contributed to thi s wo rk : J.
Soli s, G. Casa ll o, C. Vas qu es, 0. Hurtado,
D. N in o, P. M anosa lva, and R. Castill o.
References Cited

Colon , L.T. , R.C. Janse n, and D .J. Buddin g.

199 5. Parti al resistance to late bli ght
(Phytophth o ra infestans) in hybrid
progeni es of four South Am eri ca n
Solanum spec ies crossed wi th dipl oid 5.
tuberosum. Theor. Appl. Genet. 90 :691 698 .
Ghisl ain , M. , Z . Dapeng, and M a. del R.
H errera. (ed s.) 1999. M o lec ul ar bi o logy
laboratory protoco ls: Pl ant genotypin g,
2nd Editi o n. Crop Improvement and
Geneti c Resources Departm ent trainin g
manu al, Interna tional Potato Ce nter
(CIP ), Lim a, Peru. 38 p.
Ghisl ain , M ., B.R. Trogn itz, M a. de l R.
H errera, A. Hurtado, and L. Portal . 1997 .

D NA mark ers fo r th e introgressio n of late

bli ght resistance in potato. Intern ational
Potato Ce nte r Progra m Rep ort 1995 1996 . p. 84-89.
Lander, E.S. and D . Botstein . 1989. Mappin g Mend eli an factors und erl y in g
quantitati ve traits usin g RFLP linkage
maps. Ge net ics 121 :185 -1 99.
Leo nard s- Sc hippers, C. , W. Gi effers, R.
Sc hafer-Preg l, E. Ritter, S.J. Kn app, F.
Salamini , and C. Gebh ardt. 1994.
Qu antitative resi stance to Phytophthora
infestans in potato: A case stud y fo r QTL
mapp ing in an all ogamo us pl ant speci es.
Geneti cs 137:67 -77.
Rouppe va n der Voort, J.N .A.M. , H.J. Va n
Eck, J. D raa istra, P.M. Va n Za nd voo rt, E.
Jacobse n, and J. Bakker. 1998. An online
cata logue of AFLP mark ers cove rin g the
potato genome. M o l. Breed . 4:73 -77 .
Trognitz, B.R. , M. Esla va, L. Portal, and P.
Ramon. 1997 . Resi stance to late bl ight
fro m d ive rse w ild so urce s. In te rn ati onal
Potato Ce nter Progra m Report 1 995 1996. Li ma, Peru , 127- 137.

CIPProgrom Report 1997-98

71

II

I
e31r54.e

III

4. 4 -

e4m42.a
N3.7ffl

6.8 17.7 -

5.6 -

20.4 -

6.5 -

AM1242r

K8.78r
- e36r~
- GP2)6
e45r . '

e45m59.c
e34m34.a

7.8 e46r42.d

23.9 -

33.7 -

17.1 e39rn50.a

e31r58. f

e:Ja61. t
e32-n51.e

4.5 10.8 -

14.5 -

10.? -

- - -;>-STM0038
112.7!'f

e45r59.b
N1 . 1750
09. 97r

--~ STM1029

13.5- 1
e.34m34.c

e32rn61. w

e32r61 .d

e36r49 d
13.5 -

19.4 AQJ.1700

e46r42.f

e35r50.c
e22m48.d
e33m33 b

4.40.0
10.2 ------

10.3 e3.2(,JJ.f

e34m33. o

VII

VIII

r
~

7
Sl3 it>P219

20.3 -

==~~TM3009
3>"

Z4 .600

00 .4 5 0

12 6

11.45.2

Wl6.700

l .1

G3.950

OO

--,r-- e32

~
Jr

I0 .6 _/ -

.C

E6 .14 00

--~ STMI021
225

45

14 8

e31m58.d
Zl.800

e46m42t
e36m36.d

23.5 -

e3.8~9
. eSTM0024
--0- - -

2.6 -

6. 2 -

IX

e3 1m58 .e
e32T6 1.a

e3 1m48.j

6. 1 10.4 -

e31m58. b

88. 700

302

6 .8

5.8

e34m33. n

_;-+-e4569d I

Tl 18 10
H1 0 2400
J3-1 15 0

12%

Figure 1 A. Genetic linkage map of the phu parent including markers anchored to the potato genetic map:
= SSR, = AFLP, . = RFLP. Genetic di stance in cM are indicated left; markers at right.

Location of QTL associated with field resistance to LB , position (' ), and the respective percentage
of phenotypic variation explained (bottom of bars) . Fi eld dota are ADFA in Huancayo (
and in
Quito (IT] ).

I )

72

Potolo

IV
84. 78r

VI
e32r61.c

e31m36.a

7.1 -

-M~ooSTF\JAc58

11. 2 -

15.4e46m42.e

BB. 1190
BG3.13'r

19.o -

_ ".:'._2~STMOCO I

13.5-

e39m50.c

e32m61 .b
11.8 4.0 5.1 3.34.7 -

814. 790
e45m42.b
e38m49.b
e31m37.b
e31 m48.g

21.2 -

e46m42.b
R6.2100
e31r58.c
AJB.1120
e32m48.b

Z1 1.400

BG l. 11 rr

29_3 -

e32m33.h

- - -'>-GP 186
--~STM00 1 3

32.0 -

e36r36.a

XII

XI

PLG13
e45m42.c

e33m'.34 .c

e31 r58.g

12.4 -

18.2 B21 ffl0

10. 4 e31 r54.b
12.1 -

35.9

--"'"''M"'"
e32rr61 .c

6.9
KBBSO

e36r50 .d

6.1
45

22

PLG12

I~

11 %12%

e46m42 j

e36m50.e
e35m48 .a

e32r48 .g

13.5

_ -3>STMCID3
G13 .880

14.7 -

9 .4 -

""""I
e34r33j

e3Bm32 .c

5 o/o

192
e46 r42.c

e34r33.k

10.4 -

292

AMl2100

GP l2)
- - 3>sTM I009
--~
93.950

27. 8 -

16.9 -

--~ 1'58
e36r36.c

23 -

e32r48.e
e34r33.h

10.s e32m34.b

CIPPrngrnm Repo1t I 997-98

73

II

I
Bl 5.1500

III

A----~ GP313

---STMI~
AM3.1600

21.2 -

BG17.800
14.8 e38r49.b

17.6 e31m36.h
8 .4 -

e32m51 .b
e32m48 .c

B13.68r

10.1 -

9.4 -

1.2
10.6 -

AR1.12rr
__e~2.'.:_~TM1029
e32r49.gSTM2J20
AMl.159r

15.8 -

e39m50 .e

15.7 e39r50.g

17.78.8 Z1.1400

B2.65r
- - - - > STMl020
e38m49.I STMJ025

8.8

e33m33.f

G15.4rr
e31m37.e

20.9 22.a -

e35r61 .e
e32r34.h

6.9 2.2 7.6 -

e36m50.f
e31 m37 .g
__J.;J.1_~ STMCXJ38
e32m33.I

VII

&45r42.b
&45m60.h
e31 r58.a

~-

VIII
e38m32.f
--~ GP l 2 1

e33m3 2.b

---->

e32m48.d

STM 105 7
8.S -

R13.7 40
C9.l600 STMl02 I
---->

BG4.$40
G9.2700
__ - >STM20I 3

16%
11 %

IX

_an_~ STMCXJ24

23.6 -

M5.550

e::!im48.I

e3Jm61 .c
e:!lm50.k

e39m50.m

23.4 e46m42.b
e38m32.e

e34m34.e
15.1

16.2 e31 m54.c

e39m50. n GP 127

3.3
13. 3

---->
.A._

e32r3:3. g STM002 8
- - -_:::,..;.STM3009
___
GPl29

e38m32.a
7.3 K3 .175r
12.a -

>

AM12.53r
16.4-

STM1102
- - - - > STMl05i
110.850

Figure 1 B. Genetic linkage map of the dih -tbr parent includ ing mar ke rs anchored to the potato genetic map:
e = SSR, = AFLP, .A. = RFLP. Genetic distance in cM are indicated left; markers at right.
Location of QTL associated with field resistance to LB , position (~ ), and the respective percentage
of phenotypic variation explained (bottom of bars) . Field data are ADFA in Huancayo (
and in
Quito (

D).

74

Pololo

I)

IV

__ -)llo-STRJ Ac58
02.1300

e38r49.e
10.6-

VI

7.1Y18.1250

e39150.I

e31154.t
e31 r58 .tJTMl0 20

4.022.4-

AD3.2400
17.4-

e32m61 .e

14. 1

2.6::;
3.3 / 4.5 / 4.5 /_
4.7
8.5-

e3l"~.I

7.3

ALB. 85r
8. 7

G9.700

::_3~1cJP188

--~TM\100

B1. IT
e36m36.c
e33m33 .e
e33m34.k

4.7-

e36r37.b

__ 82~J::b21

15.7-

11.9-

e46m42.c
e45r59.e

7.0e46m42.a
8.2e35r48.e
R11.8rr

4.5-

9.3N15.850

.. ~
17.2-

_ Q1.~1JsTMOO I

-w~GP\86

11%

x
5.8

XII

XI

e36m36 .e
- - - -~ STM005
l15.85r
I
e32r61.h
e35m48.c
e36r50.j
e32r48.e
e31158.g
e32m48 .I
R13.5rr
e31m54.e

e35n48.k

BG4.1rrr
8.94.9-

----~STM0003

M4.970

AM171!20

e31m36.g
M15.820
e36150.b
e32m48.b
R15.970
B5.198r

' 36%

H9340

e31 m36 .j
20.6--

eJlr32.h
8.3 e39r50.j

e32 r49.f

A---

520.1100

e35m48.b

- -ai:m~ STM0030

-)llo-CP58

8451159.a

21%

9.oe33n51 .k

A____>

Cl'l'iO
GP34

CIPProgram Report 1997-98

75

Parasitic Fitness and Temperature Response of New

Lineages of Phytophthora infestans from Peru
J.L. Andrade Piedra\ G.A. Forbes 2,

W.E. Fry', and R. Nelson1

The main meth ods for co ntrollin g late blight

(LB) ca used by Phytophth ora in festa ns
include th e use of fungicides and resistant
cu lti vars (U maerus et al. , 1983; Sc hw inn,
1983) . Computer simul ati on tec hniqu es are
being used to improve the efficiency w ith
w hi ch these components ca n be used
(Bruhn and Fry, 1981 ). The ep id em iology
of the disease is complex, and mathematical mode ls using simplifyin g eq uati ons help
in its analys is. Simulators are used to
generate hypotheses, id enti fy topics for
resea rch, and deve lop forecasts (Fry, 1982).
Simul atio n should also be useful for
adapting LB management strategies to a
range of agroecosystems, fo r ex ante imp act
assess ment, and for evaluating ri sk ac ross
time an d space.
This study is part of the development,
ca li brati on, and validation of the simul ato r
SIMPOL (Hi jm an, Forbes, Grunwald, and
Andrade, personal communication), w hi ch
is based on th e simulator LATEBLIGHT
(B ruhn and Fry, 1981 ). Th e co mponents of
res ista nce o n which the simul ator is based
were quantified in American cu lti vars
infected by isolates of the" o ld " P infestans
lin eage US - 1, which wa s once dominant in
man y potato production areas of the world,
but w hi ch is now relati ve ly rare. The
equations that defin e th e effect of temp erature on pathogen deve lopme nt we re
calc ul ated based on data of Crosier (1934),
w ho very probabl y wo rk ed w ith US-1 . This
c lonal lin eage w as prev iously predominant
in Pe ru (Too ley et al., 19 89; Goodwin et al.,
1994) and probably in Ecu ado r (Forbes et
al. , 1997), but recent studies show that it
has been replaced by "new" populations

(Forb es et al. , 1997) dominated by lineage

EC-1 . The greater agg ressive ness of so me
new c lonal lin eages has been documented
(Kato et al. , 1997; Day and Sha.ttock, 1997;
Mizubuti and Fry, 1998; Miller et al., 1998),
but the compo nents of parasitic fitness of
the Andean I ineages have not bee n studi ed.
Mi zubu ti an d Fry (1998), studying th e
effect of temperature on the developm ent of
lin eages US-1 , US-7, and US-8, found that
th e interact ion of temperature by lin eage
was signifi ca nt fo r spora ngial germin ation,
but not fo r compone nts of parasiti c fitness .
The Andean agroecosyste m, in co ntrast to
the potato production areas of the USA, is
characterized by its extreme env ironm ental
heterogeneity. Temperature adaptatio n of
And ea n iso lates of th e same clonal lin eage
comin g from different altitudes is possible.
Such adaptation wou ld have important
implications fo r simul at ion.
Th e objectives of thi s study we re to (1)
determin e the agg ress iveness of new
popul ati ons of P infestans of Peru , (2) stud y
the effect of temperature on th e de ve lopment of the predominant lineage of Peru,
(3) determin e if adaptation occurs in
isolates of the sa me clonal lineage co ll ected
from different altitud es, and (4) obtain data
to va lid ate the simul ato r SIMPOL und er
highl an d trop ica l co nditions.

Materials and Methods

Two exper im ents were ca rried out. Exper iment 1 was done in September and October of 1998 at CIP-Lima. Experim ent 2 was
conducted in November and December of
1998 at Corn ell University, Ithaca, NY,
USA .

1 CIP, Lirna, Peru.

2 Cl P, Quito, Ecuador.
3 Cornel l University, Ithaca, New York, USA.

CIPProgram Rep0il l99798

77

For experiment 1, w e used 11 isolates

representing the clonal lineages US-1 (n =
3), EC-1 (n = 4), and PE-3 (n = 4). For
experiment 2, w e used eight EC-1 isolates
collected in Comas (2,400 masl ) (n = 4 ) and
Oxapampa (1 ,800 m) (n = 4). Both sites are
in the highlands of Peru. All isolates
collected were maintained on Rye A agar.
Before the experiments , isolates were
transferred to tuber slices and then to Rye B
agar. lnoculum for studies was collected
from leaves of cv. Tomasa Tito Condemayta
(experiment 1) and Superior (experiment 2).
Peruvian isolates were carried from Peru
under a permit from the lnstituto Nacional
de Recursos Naturales and into the United
States under a permit from the Animal and
Plant Health Inspection Service of the
United States Department of Agriculture.
At Cornell, Peruvian isolates were transferred in a laminar-flow biosafety cabinet
(class II , type A) in a level P2 laboratory. At
the end of the experiment, all materials
were autoclaved before disposal.
The potato cu ltivars Tomasa Tito
Condemayta (very susceptible), Yungay
(moderately susceptible), and Amarilis-INIA
(resistant) were used in experiment 1. The
plants were grown in a greenhouse at
temperatures between 14 and 24C. In
experiment 2, Superior (a susceptible
cultivar) was used. In this case, plants were
grown in a greenhouse with temperatures
between 18 and 25C and a 14-h day
photoperiod.
Inoculation basicall y followed the
method described in Mizubuti and Fry
(1998 ). For experiment 1, inoculum
concentration was 5 x 1 0 3 sporangia/ml.
Leaflets were incubated at 1 8C with a 14-h
day of weak light (0.5 mEm2 s1 ) . Incubation
period (IP) and latent period (LP) were
evaluated at 36 , 42, 48 , 54, 60 , 66, 72, 78,
84, 90 , 96, 102, 114, 120, 126, and 138
hours after inoculation (h.a.i.) using a
dissecting microscope. For experiment 2,
inoculum concentration was 15 x 10 3
sporangia/ml and leaflets were incubated at
8, 13, 18, 23, and 28C with a 14-h day of

78 Potato

intense light (150 mEm 2s 1) . IP and LP were

evaluated at 24, 30 , 36, 42 , 48, 54 , 60, 66,
72 , 78, 84, 90, 96, 102, 114, 120, 126,
138, 144, 150, 162, 168, 174, 186, 192,
198, and 210 h.a.i . In both experiments,
lesion area (LA) and sporulation capacity
(SC ) were measured as described by
Mizubuti and Fry (1998). IP w as defined as
the time necrotic spots appeared; LP was
defined as the time when sporangia appeared.
Analysis of variance was done using a
model that made it possible to compare
clonal lineages in experiment 1 or sites in
experiment 2. In both cases, the error term
w as based on the variability between
isolates within clonal lineages or within
sites. We used con trasts to compare
treatments or groups of treatments. All
analyses w ere done using SA S software
v6.11 4 .
Results
Experiment 1

The interaction lineage by culti var was

not significant for IP (P = 0.1596), LP (P =
0.0975), and LA (P = 0.2839 ), which means
that the effects of lineage and culti var could
be studied separately. Amarilis had an IP
significantly shorter than Yungay (P =
0.0001) and Tomasa (P = 0.0012), while
Yungay and Tomasa had similar values
(P = 0.4560 ).
In contrast, Amarilis had a significantly
longer LP than Yun gay or Tomasa (P < 0.01 )
and, again , differences between these last
two cu ltivars were not detected (P =
0.0753) . The LA on Tomasa was significantly larger than on Amari I is or Yungay (P
< 0.01), w hich were statistically equal (P =
0.9087) for that component. We did not
find differences between clonal Ii neages for
IP (P = 0.0682), LP (P = 0.4678), or LA
(P = 0.3484 ).

4 SAS Institu te, Inc. , Cary, NC, USA.

For SC, the interaction between lin eage

and cu lti var was highl y sign ifi cant (P =
0.0009). Through a simple effects ana lys is,
we found highly significant differences
(P <0 .01) for all the clonal lin eages within
every cult ivar and vice versa (Figure 1). The
lowest SC values occurred with Amarilis
and the highest with Tomasa. Nonorthogonal compar isons showed that the new
population of P infestans (EC-1 and PE-3)
sporu lated significantl y less than the old
population (US-1) on all the cultivars (P
<0 .01) . The SC for lin eage EC-1 was
significantly le ss than that of PE-3 in the
three cultivars (P <0.01).
The interaction cultivar by isolate within
eve ry c lon al lineage was not significant for
any component: IP (P = 0.5248), LP (P =
0.9796), LA (P = 0.4406), and SC (P =
0.9689). However, differences between
iso lates within clonal lineages were found
for IP (P = 0.0120), LP (P = 0.0001 ), and SC
(P = 0 .0001 ), but not for LA (P = 0.3650).

Experiment 2
The interaction between site and temperature was not significant for IP (P =
0.8608), LP (P = 0.2275), LA (P = 0.7171 ),
or SC (P = 0.7946). The isolates collected in
Comas were similar to those co ll ected in
Oxapampa for IP (P = 0.1321), and SC (P =
0.7371 ). Differences were significant for LP
(P = 0.051 ): isolates from Comas sporulated
at 99.8 h.a.i . and those from Oxapampa
at 93.5 h. a.i. Isolates from Oxapampa
ca used les ions that were 0.72 cm 2 larger
than those caused by isolates from Comas
(P = 0.0139).
The effects of temperature were highly
significa nt for all the components (P =
0.0001) (Figure 2). The models that
expla in ed the variab ility of each component as a function of temperature were
highly significant (P = 0.0001 ). The
determination coefficients (R 2) were 0.7318
for IP, 0.9173 for LP, 0.7377 for LA, and
0.5019 for SC. The equations for IP and LP
had a quadratic shape with the minimum
value at 27.28C for IP and 22.99C for LP.

Sporangia per lesion area (cm,)

8.0
-----------,

6.0

D
D

US-1
PE-3

EC-1

4.0

Ama rilis

Yungay

Tomasa

Cultivar

Figure 1. Sporulation capacity of three clonal

lineages of Phytophthora infesfans on

three potato cultivars. Each data point

represents the average of three or four
isolates inoculated on each cultivar.
Vertical bars indicate the standard error
of the mean.
Models for LA and SC also had a quadratic
shape w ith the maximum at 22.22C for LA
and 17.15( for SC.

Discussion
The displacement of US-1 by a new
population (EC-1 , PE-3) in Peru could not
be exp lain ed by the compo nents of parasitic fitness studied. The three clonal
lineages had simi lar va lu es for IP, LP, and
LA. Unexpectedly, the SC of the old
population was significantly greater than
that of the new population.
Previous reports (Mizubuti and Fry,
1998 Miller et al., 1998; Kato et al., 1997;
Chyc~sky and Punja, 1996) attributed the
displacement of US-1 to a greater aggressiveness of new populations, partly due to
lower va lu es of IP and LP, and high er va lu es
of LA. Day and Shattock (1997) found that
the new populations presented a greater
fitness ind ex (infectio n frequency by
number of sporangia/lesion).
We speculate that the greater aggress iveness of the new Andean populations of P

CIP P1ogrom Reporl 1997-98

79

Incubation and latent period (h)

200.0

~~~~~~~~~~~~~~~~~~~~~~~~~-.

LP = 0.436 Temp, + 20 .025 Temp + 294.190

160.0
120.0
80.0

I~

4o.o

0.0

=- - -

--

- - =--

IP= 0.061 Temp, - 3.280 Temp + 77 .01T

1
~~~~~~~~~~~~~~~~~~~~~~~~~~

13

18

23

28

Lesion area (cm')

12.0
10.0

LA= 0.044 Temp,+ 1.946 Temp - 13.173

8.0
6.0
4.0
2.0
0.0
8

13

Sporangia per lesion area (cm

10.0

18
)

(x10

23

28

~~~~~~~~~~~~~~~~~~~~~~~~~-.

SC

= 66.4 Temp 2 + 2278 .7 Temp - 12916.0

18

13
- - - Latent period -

23

28

- Incubation period

Temperature (C)

Figure 2. Effect of temperature on (A) incubation and latent period, (B) lesion area, and (C) sporulation capacity

in the EC-1 clonal lineage of Phytophthora infestans. Each data point represents the average of eight
isolates at each temperature. Vertical bars indicate the standard error of the mean.
80

Potato

infestans could be due to other components

of parasitic fitness such as infection frequency, infectious period, or improved
adaptation to en vi ronm enta I factors Ii ke
temperature, humidity, or so lar rad iation.
Further data would need to be gathered to
test these hypotheses.

The low SC of the new population was

also observed in experiment 2. Under
similar conditions, Mizubuti and Fry (1998)
found SC values of between 3.2 and 5.3 x
10 4 sporangia/cm 2 of lesion for lin eages US7 and US-8. In our experiment we did not
find values higher than 0.8 x 10 4 sporangia/
cm 2 A lower parasitic fitness has been
reported for metalaxyl-insensitive isolates
(Dow ley and O'Sullivan, 1985; Davidse et
al., 1989). Dowley (1987) found that this
reduction was due to a lower sporu lation.
Th e majority of EC-1 iso lates from Peru are
in sensitive to metalaxyl (Raymundo, 1998);
the lower SC could thus be assoc iated with
res istance to the fungicide.
In contrast to observations on the
pathogen, differences between cu ltivars
were easily detected. The previously
reported reaction (Franco, 19 94) of cultivars
in th e field was confirmed in this study.
Res ista nce was associated with lower
values of IP, LA , and SC, and hi ghe r values
of LP. The seemingly illogica l negative
cor relation betwee n IP and LP in our study
occurred because the res ista nt cu ltivar had
a rap id visible response to inoculation.
Necrosis occurred earlier in the res istant
cultivar Amarilis-INIA than in the other
cultivars.
The response to temperature of EC-1
isolates was similar to that previously
reported by Mizubuti and Fry (1998). They
found similar values for IP, LA, and SC in
both old and new populations. We found
that the interaction of site by temperature
was not significant for the components that
were evaluated. This suggests isolates of
the same clonal lineage collected from
different altitudes have not adapted to
temperature in relation to IP, LP, LA, or SC.

The movem en t of the pathogen th rou gh

infected seed could prevent the isolation
necessary for the adaptation of P. infestans
to temperature. Seed movement between
potato produ ct ion areas is common in the
Peruvian Andes.
The data obtained will make it possible
to continue with the va lidation of the
SIMPOL simulation model. The regional
impact of new disease management
technologi es will be est imated by linking
this simulation model with geographic
information syste m s.
Literature Cited
Bruhn , J.A. and W.E. Fry. 1981. Analysis of
potato LB epidem iology by simulation
modeling. Phytopathology 71 :612-616.
Chycosky, Cl. and Z.K. Punja. 1996.
Characteristics of populations of
Phytophthora infesta ns from potato in
British Columbia and other regions of
Canada during 1993 to 1995. Plant Dis.
80:579-589.
Crosier, W. 1934. Studies in the Biology of
Ph ytophthora infestans (Mont.) de Bary.
Memoir 155. Cornell University Agricul tural Experiment Station, Ithaca, NY,
USA. 40 p.
Da vidse, L.C., J. H enk en, A. Van Dal en,
A.B. Jespers, and B.C. Mantel. 1989.
Nine years of practical experience w ith
phenylamide resistance in Phytophthora
infestans in the Netherlands. Netherlands J. Plant Path. 95:197-213.
Day, J.P. and R.C. Shattock. 1997. Aggressiveness and other factors relating to
displac ement of populations of
Phytophthora infestans in England and
Wales. European J. Plant Path . 103:379391.
Do w ley, L.J. 1987 . Factors affecting the
survival of metalaxyl-resistant strains of
Phytophthora infestans (Mo nt. ) de Bary
in Irel and. Potato Res. 30:473-475.
Dowley, L.J. and E. O'Sullivan. 1985.
Monitorin g meta laxyl resistance in
populations of Phytophthora infestans.
Potato Res . 28:531-534.

CIP Progrom Report 1997-98

81

Forbes, G.A., X.C. Escobar, C.C. Ayala, J.

Reve lo, M. E. Ordonez, B.A. Fry, K.
Doucett, and W.E. Fry. 1997. Population
. genet ic structure of Phytophthora
infestans in Ecuador. Phytopatho logy
87:375 -3 80.
Franco, E. (ed .). 1994. Catalogo de sem illa
basica de papa en el Peru. Proyecto de
Apo yo a la Produ cc i6n de Semi I la e
ln vest igac i6n para Mejorar la
Productividad de Papa en el Peru
(SE INPA). Li brary, In ternationa l Potato
Cente r, Li ma, Peru. 49 p. (photocopy)
Fry, WE. 1982. Prin c ipl es of plant disease
management. Academic Press, New
Yo rk , USA. 378 p.
Goodwin , S. B., B.A. Cohen, K.L. D ea hl,
and W E. Fry. 1994. Panglobal d istr ibution of a sin gle clona l lin eage of the Irish
potato famine fun gus. Proceed. Nat.
Acad. Sci. 91:11591-11595.
Kato, M., E.S.G . M izubut i, S.B. Goodw in,
and W. E. Fry. 1997. Se nsiti v ity to
protecta nt fun gic id es and pathogenic
fitness of clonal lineages of Ph ytophth ora
in festa ns in th e United States. Phytopathology 87:973-978.
Miller, J.S., D.A. John so n, and P.B. H amm .
1998 . Aggressiveness of iso lates of
Phytophthora in festa ns from the Co lu mbia Basin of Washington and Oregon.
Phytopathology 88: 190-19 7.

82

Potolo

Mizubuti, E.S.G. and WE. Fry. 1998.

Temperature effects on developmental
stages of iso lates of three clonal lineages
of Phytophthora infestans. Ph ytopatho logy 88:837 -843.
Raymundo , R.M . 1998. Estructura racial y
reacci6n a metalaxyl en pobla ciones de
Phytophthora infestans en culti vos de
papa en Comas, Cuzco y Pun o, Peru.
Thesis for ln geni ero Ag r6no mo.
Un iversidad Naciona l del Centro del
Peru , Hu ancayo, Peru. 79 p.
Schwinn, F.J. 1983. New developments in
chemical co ntro l of Phytophthora. In :
Erwin, D .C., S. Bartnicky-Garcia, and
P.H . Tsao (eds. ). Phytophthora. Its
biology, taxonomy, eco logy and pathology. The Amer ican Phytopathologi ca l
Society, St. Paul , Mn, USA. p. 327-334.
Tooley, P.W., C.D. Therrien , and D.L. Ritch.
1989 . Matin g type, race composition ,
nuclear DNA co ntent, and isoz im e
ana lys is of Peruv ian isolates of
Phytophthora in festans. Phytopathology
79:478-48 1.
Umaerus, V., M. Umaerus, L. Erj efa lt, and
B.A. Nilsson. 1983. Control of
Phytophthora by host resistance: Problems and progress. In: Erw in , D.C. , S.
Bartnicky-Garcia, and P.H . Tsao (eds .).
Phytophthora . Its biology, ta xonomy,
ecology and pathology. p. 315-326.

Estimating the Global Severity of Potato Late Blight

with a GIS-Linked Disease Forecaster
R.J. Hijmans1, G.A. Forbes 2 and T.S. Walker1

Late blight (LB), caused by Phytophthora

infestans, is the most important disease of
potato worldwide (Hardy et al., 1995). In
many parts of the world, potato production
depends on the control of LB with the
routine use of fungicides. Poor farmers in
deve lopin g countries, howeve r, often
cannot control the disease, losses can be
heavy, and crop abandonment may occur
(Ortiz et al., 1999). Although the extent of
these losses is diffi cult to estimate, LB is
generally considered the prim ary biotic
constraint to potato production in developing countries (Forbes and Jarvis, 1994;
Haverkort, 1990).
CIP has a research program on integrated
management of LB that provides resistant
varieties and disease management principles to farmers in developing countri es .
The extent of LB poses a probl em for
allocating the limited resources available
for this work. Ideally, allocations should be
made to activities that contribute most to
poverty alleviation, food security, and
environmental quality. Thus, decisions
about resource allocation should be
supported by systematic estimates of the
loss due to LB in different parts of the
world. Unfortunately, that kind of inform ation currently available is incomplete and
highly uncertain.
As an alternative, agroecological zoning
has been used to guide resource allocation
for agricultural research and development
(Greyseels et al., 1992). Howeve r, generic
agroecological zones may not be very
usefu I to deal with specific produ ction

1 CIP, Lima , Peru.

2 CIP, Qu ito, Ecuador.

constraints such as potato LB. Instead, an

agroecological zoning approach that is
spec ific for an agricultural technology or
production co nstraint can be used (Wood
and Pardey, 1998; Corbett, 1998).
The strong technological progress in
computer hardware and geographical
information systems (GIS) software has
made the manipulation of large
georefe renced databases relatively straightforward. Th e use of th ese databases for
crop or constraint-specific zoning requires
th e development or identification of zoning
criteria. Criteria that link climatic data with
crop growth and production constraints are
encapsulated in simulation and forecasting
models. Crop growth simulators have been
linked to GIS to examine potential crop
growth (Han et al., 1995; Van Keulen and
Stol, 1995), effects of ab iotic constraints
such as drought (Van Keulen and Stol,
1995), and effects of biotic constraints
(Seem, 1995). Effects of global climate
change on plant disease severity were
evaluated for rice leaf blast caused by
Pyricularia oryzae using disease simulation
mod els linked to GIS (Luo et al., 1998).
In this paper we describe the use of a
GIS-linked LB forecaster to estimate global
disease seve rity. Running a potato LB
forecaster for a complete growing season
allows estimating the number of fungicid e
sprays needed for effective disease management. By linking the forecaster to interpolated climate data on a grid, the number of
sprays need ed to manage the disease can
be ca lculated for the whole world. The
number of sprays needed is an adequate
and easy-to-understand proxy for disease
severity. Differences in fungicide use have

CJP Program Report 1997-98

83

clear impacts on farmer safety, crop

productivity, environmental contamination,
and th e potential for successful disease
management.

Dail y leaf wetness and precipitation we re

estimated with stochastic weather generators (See Supit et al., 1994, for precipitation).

Materials and Methods

In temperate regions there may be one

clearly distinguishable season in which
potato is grown. However, in Mediterranean and (sub)tropical areas, potato may be
grown at different times of the year, or even
year-round (Forbes et al., 1998; Haverkort,
1990). There is no comprehensive database
on planting dates and growing seasons of
potato worldwide. Therefore, we determined the most likely growing season(s) for
each 10-minute grid cell following rules
adapted from previous studies (Stol et al.,
1991; Van Keulen and Stol, 1995). In each
grid eel I there can be between 0 and 12
growing seasons a year, with emergence on
the 15th of each month. Potato can be
grown in a month where minimum temperature is above 3C and average temperature is below 22C. At temperatures below
3C, frost risk is very high; at temperatures
above 22 C, tub erization is strongly
reduced (Stol et al., 1991 ). A growing
season can start in any month that perm its
an accumulation of at least 1,000 degree
days (Cd) in consecutive months without
reaching either of the temperature limits .
Degree days were calculated as the sum of
daily average temperatures minus a base
temperature of 2C. The growing season
ends when 1,500 Cd are accumulated, at
140 days after planting, or when either of
the temperature limits is reached .

The LB forecaster used in this study was

deve loped at Cornell University, USA (Fry
et al., 1983). It uses daily rainfall , average
temperature, and hours of leaf-wetness
(estimated to occur when relative humidity
;:::90%), to predict the need for protectant
fungicide sprays on a susceptible, moderately susceptible, or moderately resistant
cu ltivar.
The necessity for the first spray is
forecast on the day that accumulated blight
units (BU) reach a threshold. Either a BU or
accumulated fungicide unit (FU) threshold,
whichever is reached first, predicts subsequent sprays. BU accumulation is based on
daily average temperature, hours of leaf
wetness, and level of host resistance. FU
acc umulates with time (unit= day) and
rainfall, and represent fungicide obsolescence. BU accumulation and thresholds are
dependent on cultivar resistance, but FU
acc umulation and thresholds are not. There
is a minimum interval of 5 days between
sprays.
The calculation of BU, FU, and the
threshold values were established by
iterations of a well-validated LB simulator,
also developed at Cornell (Bruhn and Fry,
1981 a,b). In the development of the
forecaster, cv. Hudson was used as the
susceptible, Rosa as moderately susceptible, and Sebago as moderately resistant
(Fry et al., 1983).
Climate data were taken from GCLIMl 0,
a database with global monthly climate
data on a 10-minute grid (Hijmans, 1999).
Minimum and average temperatures,
precipitation, number of days with precipitation, and minimum and maximum relative
humidity were used in this study. Daily
temperature data were derived by linear
interpolation between monthly averages.
84

Pololo

These rules reflect the fact that growing

seasons are generally longer in cool er
climates due to slower crop development.
Because potato is an irrigated crop in many
areas, precipitation is not taken into
account to determin e length of growing
seasons, and irrigation is assumed to be
available where needed.
The LB model was run for each grid cell
and for each growing season, from emergence unti I 1 week before the end of the
growing season. The number of sprays
needed to fully protect a potato crop from

yield loss du e to LB was reco rd ed for

suscep tibl e and res ista nt culti va rs. Because
of th e stoc hasti c character of the leaf
wetness and rainfall ge nerato rs, eac h run
was repeated 30 times, and th e number o f
sprays/ run was averaged. For grid ce ll s w ith
more than o ne grow in g season, th e model
was run 30 times for eac h season and th e
avera ge of all run s was used .

res istan t va rieti es (F igure 3). In most

cou ntri es w here LB is a signifi ca nt problem,
the actual number of sp ra ys is below
optimum (Fi gure 4). Large differences
between the optimum and actu al number of
spr ays indi ca te th e potential for in c reas in g
p roduct ivity t hrough im proved m anagement of LB.

Discussion
Co un try ave rages we re ca lcul ated usin g
a geo refere nced d atabase of world potato
produ ct io n th at d elin eates th e potato
produ ction zo nes in eac h co untry and
indi cates t he area planted w ith potato in
eac h zone (Hu acc ho and Hijm ans, 1999).
For eac h potato produ ct io n zo ne, th e
fracti on of th e nation al potato area was
multiplied by the ave rage number of sp rays/
ha and then agg rega ted by co untry. The
resul ts we re co mpared w ith ear l ie r estimates of actual fun gic ide use in d eve lop in g
countries. Th ose est imates we re made
ind epende ntl y of th is stud y fo r CIP research
priority settin g (Walker and Col lion , 1997) .

Results
Th e es tim ated number of fun gicid e app li cations req uired for co mpl ete co ntrol of LB in
a susceptible potato variety is shown in
Fi gu re 1. Th ere are stri k in g differences in
LB seve rity between potato p ro du cti on
zones. Th e optimum numb er of sprays is
espec ially hi gh in th e tropical highl and s of
Latin America, Africa, and Asia; in weste rn
Europe, the east coast of Ca nad a, and the
north ern USA; southeast Brazil, centralsouth China, and in m any coas tal areas.
Major potato producing areas with low LB
press ure in c lude north ern China; th e plains
in Indi a and Bangladesh, w here irri gated
potato is produced in the coo l dry season;
and no rth weste rn USA.
By co mparin g Fi gures 1 and 2, one ca n
qualitative ly appreciate th e benefits of a
shift from a susceptibl e to a res istant
cultivar. When com p ared o n a country-bycountry basis, there is an ave rage reduction
of 1 5% in the optimum number of sprays if
suscept ibl e varieties are repl aced by

Th e GIS-1 in ke d fo recaste r ap pears to be a

va li d tool for distinguishing areas w ith
different LB severity on a globa l scale. The
too l is proba bl y most useful fo r compa riso ns of relative effects, e.g., the relative
benefi ts of host res istance in two co untri es.
On the w ho le the predicted optimum
num ber of sprays see m s to be low . For a
number of co untries the est imated curre nt
fun gic ide use is even hi gher than what is
pred icted to be optimal (Fi gure 4). Thi s
could be du e to erro rs in the c limate and
po tato distribution data, in th e d elimitation
of the grow in g seaso ns, in the forecaster,
and in the estimates of current fungicide
use. Errors in fungicide use est im ates m ay
be due to a bias toward th ose areas w here
LB is a problem when the estimates are
made. Nonetheless, our res u lts seem
reasonab le. The actu al nu mbe r of sp 1ays
approaches or su rp asses the optimum
number of sprays in co untri es w ith a kn ow n
hi gh leve l of control, such as Costa Rica,
and is below optimum in countries w ith a
known fung icide use deficit, suc h as
Rwa nd a (Figure 4).
Th e acc ura cy of o ur predicti o ns in
abso lute te1ms is dependent on th e quality
of the c lim ate and potato distribution data,
but espec iall y on the d ec ision criter ia.
Th ese c riteria are enca psul ated in a forecasting system that has been va lid ated
prev iously at th e field leve l in th e USA (Fry
et al., 1983), and th e Toluca Valley of
Mexico (N . Grun wa ld, Cornell Uni ve rsity,
pers. co mm ., 1999). Thi s fo recaster was
deve loped from iterati o ns of a we l 1va I id ated LB simu lato r (Bruhn and Fry,
19 81 a,b), which is driven by weather
CIP ProgramReport 1997-98

85

Q)

..0

::>

Do
86

Pototo

c:
E

::>

c..

...

'-.J

co

~
~

co

1-4

9-12
> 12

Not suited for potato

D s -s

Optimum Number of Sprays

(Resistant Variety)

Figure 2. Optimum number of sprays to control late blight in a resistant potato cultivar.

..,.

'

~ !

Predicted number of sprays (resistant)

20

15

0
0

10

15

20

Predicted number of sprays (susceptible)

Figure 3. Optimum number of sprays for a susceptible cultivar versus a resistant cultivar. Aggregate values for all
potato producing countries. Line: y = O.BSx; R2 = 0.97.

va riabl es. A fo reca ster developed by

empirical ex perim entatio n would be
l imited to the conditions in which it was
deve loped.
Fewer actual than opt im al sprays in
many deve loping countr ies, es pec iall y in
sub-Saharan Africa, were to be expected. In
th e African hi ghlands, fung icides are
genera l ly co nsidered to be und erused, one
of the reaso ns behind the w id esp read
cultivat ion ofCruz a- 148, a hi gh ly resistant
cultivar (Forbes and Jarvis, 1994). Th e
oppos ite case is Ind onesia . Th ere the
curren t number of sprays is above optimum
and the dominant suscept ibl e culti va r,
Granola, is hard to rep lace because of
mark et prefe rence. In those cases th ere may
be a pote nti al ga in in profitability throu gh
optimizing fun gicide use.
The potenti al imp ac t of th e introduction
of LB resist ance is huge, espec iall y cons id-

88

Potato

ering the level of resistance th at was used

in the forecaster is lowe r th an th at used in
current breedin g pro grams . Although ho st
resista nce is wide ly used in Sub-S aharan
Afr ica, our results illu strate th e potentia l
benefit of a further in crease in the deployment of res ista nce. In addition, in creas ing
access to fungicides and ensurin g th ei r safe
and effect ive use could ma ke a grea t di rec t
impact in this regio n. A different strategy
might be emp loyed fo r co untri es such as
Mex ico . Here fungi c id e usa ge is close to
th at predicted (Figure 4), so that the introdu ction of host resistance and opt imi zing
fun gicide use (Thiele et al., 1998) may have
a relati ve ly greater imp act. Countries nea r
th e ori gin in Figu 1e 4 should be ass igned a
low priority in the al location of fu nd s for
globa l LB researc h.
Our zoning is bas ed on a number of
assumptions. We think they are transparent
and simpl e, and helped us to present a

Number of sprays (actual)

20
y=x
a Costa Rica

15

a Indonesia

10

a Colombia
a

a a

a Ecuado r

5Cl

mm

Diii

Cl

1111

DD

ma

111

ca

China
Rwanda
Bolivia

a Bhutan

o~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

10

15

20

Number of sprays (predicted)

Figure 4. Optimum number of sprays for a susceptible cultivar versus the actual number of fungicide sprays to

control late blight. Aggregate values for developing countries.

c lose approx imation to rea lity. They are,
however, subject to discussion and improvement. For examp le, o ur met hod of
estimatin g p lanting dates should be more
fo rm al ly tested. How farmers adapt to LB
pressure wou ld also be of interest. In the
African highla nd s, for examp le, farme rs
tend to p lant ea rl y o r late to avo id the
wettest seaso n, that most co ndu c ive to LB ,
thereby accept in g the in c reased risk of
drought stress (Devaux and Haverko rt,
1987). In these areas, the introdu ct io n of
res istance ma y actually lead to higher
fun gicide use if the high er level s of res istan ce all ow piant in g in the rainy seaso n.
References Cited

Bruhn, J.A. and W.E. Fry. 1981 a. Ana lysis of

potato late b li ght ep id emiol ogy by
simulation modeling. Phytopatho logy
71:612-616.
Bruhn, J.A . and W .E. Fry. 1981 b. Deve lopment and eva lu ation of a comp uter

sim ul ation model-generated potato late

bl ight fo recast. Presented at the
Ph ytop hthora Symposi um , UCR, Ri verside, CA.
Corbett, J.D. 1998. C lassify in g maize
produ ct io n zo nes in Ke nya th rough
multivariate cluster analysis. In: Ha ssan,
R.M. (ed.). Ma ize technology development and t1ansfer. A G IS appl icat io n for
resea rc h and planning in Kenya. CAB I,
Wal lin gfo rd , UK . p . 15-25.
Deva u x, A. and A.J. Haverk ort. 1987. Th e
effects of shifting plant in g dates and
mul chin g on late b light (Ph ytophthora
intestans) and drought stress of potato
crops grown und er trop ica l hi ghl and
co nditi ons. Exp. Agric. 23:325-333.
Forbes, G.A., R.J. Hijmans, and R.J. Nelson.
199 8. Potato blight: A wor ld probl em.
7th Internat iona l Congress of Plant
Pathology, 9-16 August 1998, Edinburgh,
Scotland . Brit ish Society for Plant
Pathology. Ava il able on:

CIP Program Reporl 199798

89

www.bspp.org.uk/icpp98/abstracts/5 .1 I
9S.html.
Forbes, G.A. and M.C. Jarv is. 1994. Host
resistance for management of potato late
blight. In: Zehnder, G. , R. Jansson, and
K.V. Raman (eds.). Advances in potato
pest biology and management. American
Phytopathological Society, St. Paul, MN,
USA. p. 439-457.
Fry, WE., A.E. Apple, and J.A. Bruhn. 19 83.
Eva luation of potato late blight forecasts
modified to incorporate host res istance
and fungicide weathering. Ph ytopathology 73:1054-1059.
Greyseels, G., C.T. De Wit, A. Mccalla, J.
Monyo, A. Kassam, E. Crasswell, and M.
Collinson. 1992. Setting agricultural
research priorities for the CGIAR. Agric.
Sys. 40:59-103.
Han , S., R.G. Evans, T. Hodges, an d S.
Rawlins . 1995 . Linking a geographic
information system with a potato simulatio n model for site-specific crop management. J. En v. Qua! . 24:772-777.
Hard y, B., B. Trognitz, and G. Forbes . 1995.
Late blight breeding at CIP. CIP Circular
21 :2-5 .
Haverkort, A.J . 1990. Ecology of potato
cropping systems in relation to latitude
and altitude. Agric. Sys. 32:251-272.
Hijmans, R.J. 1999. Global Climate Surfaces at a 1 0-m i nute Resolution ,
GCLIMl 0, Vers ion 1.0. Production
Systems and Natural Resource Management Department Working Paper 2.
International Potato Center, Lima.
Huaccho, L. and R.J. Hijmans. 1999.
GPOT97, A global geo-referenced
database of potato production for 19951997 . Production Systems and Natural
Resource Management Departm ent
Working Paper 1. Internationa l Potato
Center, Lima .
Luo, Y. , P.S. Teng, N.G. Fabellar, and D.O .
Tebeest. 1998. The effects of global
temperature change on ric e leaf blast
epidemics: A simulation study in three

90 Potato

agroecological zones. Agric. Ecosystems

En v. 68:187-196.
Ortiz , 0. , P. Winters, H. Fano, G. Thiel e, S.
Guaman, R. Torres, V. Barrera, J. Unda,
and J. Hakiza. 1999. Understanding
farm ers' responses to late blight. Evidence from Peru , Bolivia, Ecuador, and
Uganda. Impact on a Changing World.
CIP Program Report 1997-98. p.101-109.
Seem , R.C. 1995. Geographic in formation
systems for localized pest predictions. JT
Bulletin OEPP 23:639-646.
Sto l, W., G.H.J. De Koning, A.J. Haverkort,
P.L . Kooman, H. Va n Keu len, and F.W. T.
Pennin g de Vries. 1991. Agro-ecological
characterization for potato production . A
simulation study at the request of the
Intern at ional Potato Center (C IP), Lima,
Peru. Report 155 , CABO, Wageningen.
Supit, I. , A.A. Hooijer, and C. Van Diepen.
1994. System description of the
WOFOST 6.0 crop simulation mode l
implemented in CGMS. Vol . I: Theory
and algori thms . Joint Resea rch Centre,
Lu xem bourg.
Thi ele, G., 0. Navia, and E.N . Fernandez Northcote. 1998. Analisis econ6mico de
la estrategia del contro l quimico de!
ti z6n tradio (Phytophthora infestans) para
culti vares de papa suscept ibles en
Cochabamba, Boli v ia. Fitopatoiogia
33(3):176-181.
Van Keulen, H. and W. Stol. 1995. Agroeco logical zonation for potato production. In : H ave rkort, A.J . and D.K.L.
MacKerron (eds.). Potato ecology and
mod eling of crops under conditions
limitin g growth. Kluwer, Dordrecht,
Nether lands. p. 357-379.
Walker, T.S. and M. Co lli on. 1997. Priority
setting at CIP for the 1998-2000 medium-term plan . International Potato
Center (CIP), Lima , Peru. 48 p.
Wood , S. an d P.G . Pardey. 1998.
Agroecological aspects of eva lu atin g
agricultura l R& D . Agric. Sys. 57:13 -41.

Implementing IPM for Late Blight in the Andes

R. Torrez\ J. Tenorio 2, C. Valencia 2, R. Orrego3, 0. Ortiz3, R. Nelson3, and G. Thiele 4

To manage late blight (LB) Andean farmers

use resistant potato cultivars and fungicides, as well as other control techniques.
Since the spores of the LB pathogen
(Phytophthora infestans) are too small to
see, farmers have neither a clear understanding of the disease or of methods for
control Ii ng it. As a resu It they suffer heavy
losses (Ortiz et al., 1999). CIP has been
developing potato genotypes with strong
levels of partial resistance which slow
rather than prevent the epidemic, resulting
in less disease and higher yields with fewer
fungicide sprays. To get the most out of
these new genotypes, however, farmers
need to understand the disease better and
integrate resistant cultivars with improved
fungicide use and other practices.
The farmer field school (FFS) approach is
a relatively new extension method based on
hands-on learning that has been implemented widely in Asia (van de Flier! et al.,
1995). It has been used extensively for
managing insect pests in rice, but has also
been applied for disease management. The
FFS approach should have a number of
advantages for implementing integrated
management of LB. The season-long
training and research format creates an
opportunity for farmers, extensionists, and
researchers to exchange information and
experiences. For example, by learning
farmers' reactions to new genotypes,
researchers have a better understanding of
which characteristics to include in advanced breeding materials. The FFS also
gives farmers an opportunity to enhance
their knowledge of diseases and disease
management through experiments and

1 Fundaci6n PROINPA, Bolivia.

2 CARE, Peru.
3 CIP, Lima, Peru.
4 CIP, Bolivia.

other hands-on learning activities, as well

as through discussions and games. The field
experiments provide farmers with earlier
access to new partial Iy resistant cu Iti vars
and give them an opportunity to systematically evaluate this material. Follow-up
experiments al low farmers and scientists to
develop integrated management practices
appropriate to local conditions and needs.
In 1 997-98, pi lot testing of the FFS approach for managing LB was carried out in
Peru and Bolivia. This paper provides a
critical assessment of early progress.

Methods
We selected seven pilot sites to test the new
approach on the grounds that potatoes
grown by small farmers were the main
crop, LB was the principal production
constraint they faced, and local institutions
were interested in collaborating. In
Cajamarca, Peru, the non-governmental
organization (NGO) CARE was responsible
for four, whereas !NIA (lnstituto Nacional
de lnvestigaci6n Agraria) led a fifth. In
Morochata, Bolivia, the National Potato
Research Program (PROINPA), led two FFS,
and began to involve an NGO, ASAR
(Asociaci6n de Servicios Artesanales y
Rurales). Fixed groups of farmers interested
in learning more about LB and its control
participated at each pilot site.
We developed and refined field guides
for use by FFS facilitators. These contained
different training sessions sequenced to fit a
whole growing season and to progressively
facilitate farmers' acquisition of knowledge
and skills in managing LB. In Bolivia, a
strategy for controlling LB in susceptible
cultivars had already been extensively fieldtested, and th is formed the focal point for
the FFS (Thiele et al., 1998). The strategy
CIPProgramReportl997-98

91

inc lu des ea rl y p reve nti ve app li cat io n of a

systemic fungicide, alternat in g th e use of
systemi c and contact fungic ides, and
c lim ate dependent i nterva Is betwee n sp ra ys
(N av ia et al., 1995) . In Peru , the FFS
cu rri cu lum was b roader in c lu din g aspects
such as resea rch princ ip les, bio log ica l
aspec ts of LB , and fungicide act io n (Tab le

1). At the co re of the FFS in Peru was a tria l

m anaged by the fa rm ers to evaluate new LB
res ista nt cu lti va rs and understand th e
interacti on betwee n 1es istance and fu ngicid e use. In both Peru an d Bo li v ia emph as is
was placed o n learn in g underl y in g biologica l p 1in c ip les, alth ough teac hin g sty les
di ffered between th e two sites (Table 2) .

Table 1. Activiti es carri ed out during th e FFS (Pe ru).

Session no.

Topic

Type of activity

Introduction to FFS

Group discussion

Review of activities planned

Evaluation of farmer knowledge and practices

Experimenta l design: concepts of experimentation and

randomization

Field exercise
Hands-on lea rning activity

Field experiments: varieties vs. strategies for chemical

control

Seed quality

Group discussion

Tuber blight

Honds-on learning activity

Supplementary small experiments

Sources of inoculum

Hands-on learning activity

Dia gnosis and disease cu lture experiment

Life cycle of P. infestons

Results of disease cultureexperi ment

Group discussion and hands-on

learning activity

Evaluation of tuber bl ight

Disease and environment

Group discussion

Results of life cycle experiment

Hands-on learning activity

What is integrated disease management?

Group discussion

Chemical control

Hands-on learning activity

Exercise with fungicides

Fungicide application
Nozzle selection

Field exercise
Hands-on learning activity

Ba ckpack calibration
Good spraying practice
Results of fungicid e exercise

Simulation of disease spread

Ha nds-on learning activity

Genetic resistance

10

Precautions in the use and proper storage of pesticides

Group discussion and practice

Chemica l control strategies

11

Positive selection

Group discussion and field

12

Harvest and economic analysis

Group discussion

exerci se
Discussion and future plans

92

Pololo

Table 2. Training methods compared.

Pilot areas

Bolivia

Peru

Morochata

San Miguel

Collaborating institutions

Research Center, some NGO

NGO, some Research Center

Number of field schools

Participants in each FFS

10 to 20

5
7 to 25

Sessions in FFS

10

12

Potato varieties used

One moderately susceptible

12 with different levels of resistance

Emphasis

Technology adaptation and training

Knowledge acquisition by farmers

and access to new materials with LB

in fungicide use

resistance
Focus of technological development

Validating existing strategy for

chemical control

Selecting resistant varieties and

developing strategy for combining
resistance with levels of fungicide
use

Experimental design

Each farmer had single plot where

strategy was tested and compared to
farmers without training

Experiments were conducted at the

schools and consisted of 12 varieties
and three intensities of fungicide
use

Principal learning activities

Games, observation, hands on

learning

Impact indicators

Group dynamics, hands on learning,

observation

Farmer knowledge, practi ces,

Farmers opinions about benefits

AUDPC, yield, and economic

(impact evaluation will begin in

benefits

1999)

Integrated pest m anagement (IPM)

technology was tested at both sites. In
Bolivia, this principally involved farmers
validating th e existing strategy for managing
LB based on improved fungicide use with
the moderately susceptible cultivar
Waych'a (grown by nearly all farmers in the
pilot region). There was no experimental
design as each far mer tried out what he had
learned on a whole plot. Farmers purchased
all th eir own inputs. Facilitators collected
data on the fun gicides applied, doses and
dates of application, and measured LB
damage periodically on the plots of 20 of
the participating farmers. Yields were
measured by taking two 10 m crop cuts on
eac h plot. The LB damage data were used
to calculate the area under the di sease
progress curve (AUDPC). The same data
were collected from a control group of 15

farmers who did not participate in th e

training. Farmers in the FFS evaluated all of
the plots at flowering and harvest to
appreciate the contribution of different
control regim es.
In Peru, no fungicide spray strategy
existed. The overall aim was to work with
farmers to deve lop efficient site specific
management strategies integratin g fun gicide
use with different levels of partial resistance. Trials of cultivars and breedin g lines
were intended to allow farmers to evaluate
new materials with partial resistance at
different levels of fungicid e use. An expe rimental design was used with 12 cu ltivars
showing different types of resistance and
three levels of fungicide use: low (2 sprays),
medium (4 sprays), and high (6 sprays).
Each treatment was identified with a

CIPProgram Report 1997-98

93

colored plastic flag and a group of fa rm ers

was in charge of spraying and eva lu at in g.
Farm ers helped collect data on LB damage
and y ields, and drew co nc lus ions through
th ei r observations and discussions ove r the
course of th e season. Th e CARE fac ilitators
and CIP staff record ed and analyzed th e
data and ca lculated AUDPCs. Apa rt from
th e fo rm al eva luation s of disease damage,
farmers eva luated th e different culti va rs and
trea tm ents inform all y as th e crop was
growing, by studying foliage and check ing
y ield s.
Results

In Bolivia, participation in th e FFS led to an

in crease in farmer kn ow ledge (Ta bl e 3).
Farm ers w ho rece ived training began
sprayin g on ave rage five days before
farm ers without training and co ntinu ed for
five days longe r. Th ey app lied fungicides
more frequently, w ith doses closer to
recomme nd ed leve ls and better cove rage of
foliage. Thi s was transl ated into lowe r
leve ls of AUDPC among the FFS fa rmers
even w hen different numbers of app li cation s of fun gicide are co ntroll ed for (Figure 1).

Costs we re slightly higher fo r fa rm ers

w ho participated in training as they made
more fungic id e app li cat ions. But the plots
of fa rm ers w ith training y ielded over 8 tons
more on ave rage (an increase of 86 %),
giv in g an add iti o nal net benefit to FFS
fa rm ers of more than US$2 ,000 (Torrez et
al. , 1999).
The 1997-9 8 c rop pin g seaso n in Peru
was unu suall y humid beca use of El Nino,
and LB pressure was particularly hi gh. The
effects of thi s ca n be see n in the AUD PCs
for the different treatments, w ith six genotype cul tivars hav in g high AUDPCs even
w ith hi gh leve ls of fungicide use (Figure 2).
Since it was such an exception al yea r
fa rm ers focused th eir attention more on
va rietal response to th e di sease and less on
eva lu atin g strateg ies for fungicid e use. Onl y
seven cu lti va rs had a yield of at least 1 kg
per plot (40-plant plots from w hi ch tubers
from surv iv in g plants in th e two center rows
we re harves ted [a max imum of 20 p lants)) .
Farm ers elimin ated two of th e seven
culti vars from th ei r final se lecti on because
of poo r culinary quality (Fi gu re 3). Th ey
selected three culti va rs and one clo ne

Disease culture experiment: An example of hands-on learning

In the fourth session of the farmer field school in Peru , farmers prepared a spo rangial suspension by was hin g infected foliage. Using a mini-mi croscope and hand
lenses, the farmers observed the sporangia in the wash-water and on th e lesio ns that
were the sources of th e inoculum. They enhanced th eir observations by sketchin g
the spec imens.
H ea lthy potato leaves were dipped in th e spo rangial suspension and th en placed
in plastic boxes (d isposa bl e containers obtained at a loca l supermarket) w hi ch
served as humid growth chambers. Farmers also dipped healthy leaves in clea n
wa ter and placed th em in pl ast ic boxes. Farmers observed changes in the leaves
over the subsequ ent week, and analyzed the res ults in the FFS session a week later.
By ask in g question s about what happen ed and w hy, farmers discove red for
themse lves mechani sms of infection. This exe rc ise provided farmers the opportunity
to know w here late blight comes from and to understand that this disease is ca used
by a livin g entity and not by humid climatic co nditi ons that onl y contribute to its
deve lopment. In thi s way, scientific informati o n helps farmers build new und erstandin g.

94

Poloto

Table 3. Knowledge of LB and control measures in Morochata, Bolivia.

Farmers responding correctly(%)
Learning component (question)

FFS group (20 persons)

Control group (15 persons)

Sporulation on the back of the leaf is symptom of LB

65

Concentric rings or yellowing is symptom of Alternaria

Asystemic fungicide moves within the plant

65
90

27
13
33

Acontact fungicide only protects where it reaches

80

26

Afungicid e is for plant diseases

45

An insecticide is for insects

60

27
27

Name lwo or more contact fungicides

65

40

Name lwo or more systemic fungicides

70

26

Sou rce: Veizaga, 1999.

Mean AUDPC

1600
1400
1200
without training

1000
800
600
400

638
263

261

115

200
0
0

with training

45
5

Number of applications

Figure 1. AUD PC and number of applications of fungicide, Morochoto, Bolivia.

(Amar ilis, Chagllina, Atahualpa, and
377740.1) for further evaluation and use
because these were perce ived to have good
market acceptability (si mil ar to loca l
traditional cultivars) and because they
yielded well eve n with low leve ls of
fungicide. Farmers identifi ed access to new
cu ltivars that are not yet avai Iab le on the
loca l market as an important benefit of
participating in the FFS (Groeneweg and
Schouten, 1999), and th ey have begun

multiplying seed of the materials they

selected on their ow n.
A baseline study of LB management in
Peru showed some differences in farmers'
knowledge about th e biology of LB amo ngst
FFS farmers (Tab le 4). So far, evaluation of
impact on fa rm ers' production systems has
been large ly qualitative. Two focus groups
were conducted to identify farmers' opinions about the benefits of the FFS. Farmers

CI PProgramReport 1997-98

95

4500
3500
2500
1500
500
High fungicide

L:;;:~ium fungicide
ung1c1de

Genotype

Figure 2. AUDPC and levels of fungicide use, Caiamarca, Peru.

7.00

~1~~----------~---1

6.00
5.00

0
c..

4.00

Cl

3.00
2.00
1.00
0.00

Genotype

Figure 3. Yield and levels of fungicide use, Caiamarca, Peru.

96

Potolo

Table 4. Knowledge of LB and control measures in Cajamarca, Peru.

Farmers responding correctly(%)
Learning component (question)

FFS (10 persons)

Non-FFS (30 persons)

70
70
70
70

4
4
0
0

LB is caused by rain, fog, or lightening (false)

LB is caused by a fungus {true)
Any dark spot is LB (false)
Sporulation as symptom of LB {true)

Source: Ortiz and Winters, 1998. Part of a baseline study of LB conducted in Cajamarca, Peru, in 1998. Unpublished.

indi cated that the most impo rtant benefit

was improved understandin g of the princ ipl es involved in disease progression and
management. This helped them understand
that LB was not something "magical" but
so m ethin g that could b e controlled.
Knowing this also enhanced their co nfidence and self-esteem. A three-yea r impact
eva lu ation project will begin in 1999 to
assess the extent to which new knowledge
helps farmers to reduce losses and in c rease
profits.

Some Difficulties to Overcome

This was very much a lea rnin g expe rience for those developin g the FFS approac h
for integrated pest manageme nt of late
bli ght, and not everything went perfectly. In
both Peru and Bolivia, FFS facilitators
rece ived littl e training in th e m ethod and as
a con seq u ence there was a tendency to use
traditional exte nsion m ethod s and not fully
exp loit hands-on learnin g. Intensive FFS
facilitator training is bein g plann ed to
overcome this critical diffi c ulty. Th ese pilot
FFS effo rts focused on LB control but
farm ers and faci I itators noted that other
pests were also important. N ew ve rsions of
the FFS cu rriculum unde r deve lopm ent
should h elp address this weakness.
In Pe ru , the FFS attempted to develop
understanding of more complex concepts,
including ex perimentation , randomization,
and ge net ic resistance. Examples were
taken from the farmers' own co ntext but the
link between the example and rea lity was
not always clear to the farmers. In the

cultivar by fun g ic ide trial, some farmers

failed to see that th e objective of th e
experiment was to identify best fungicide
treatments for eac h cultivar. One group
declined to continu e the studies for a
second season because they did not like to
see potato plants dying (G roeneweg and
Schouten, 1999). This suggests that more
time and effort needs to be taken to d eve lop
farmers ' understanding of the role of
experimentation in lea rning. Finally, th e
FFS presents a new ap proach to which
farm ers are not yet accustomed. Th e FFS is
designed basically to provide information
and knowl edge but farmers have come to
expect m ater ia l inputs (seed, fertilizer,
pesticides), as th ey h ave received th ese
materials from other organizations in the
past. Care needs to be taken in clarifying
the new philosophy w ith farmers at the
outset.
In Bolivi a, one weakness of impl em entation was th e almost exc lusive concentration
on chemical control methods. The curriculum is being expa nded to give more
importance to genetic resistance and FFS
farmers will also eva luate new resistant
material s. Th ere were also problems w ith
the sequenc ing of some of the sessions. A
key session on th e life cycle of the pathogen took pl ace after a session on th e
strategy of c hemi ca l co ntrol. As a res ult,
farmers were less able to understand the
principles und er lying the control strategy.
As the curriculum is reworked more ca re is
being taken with appropriate sequencing.
Finally, women took no part in the FFS.

CIPProgram Report 1997-98

97

Eve n though men typi ca ll y make most

dec isio ns rela ting to pesti c ide use, more
attention need s to be paid to un derstand ing
how wome n ca n be approp ri ately in vo lved.
Desp ite th ese di fficulti es, in both Peru
and Bo li v ia the FFS parti c ip ato ry approac h
was c lea rl y a success as jud ged by th e
eno rm o us enthu sias m o n th e pa rt of th e
fa rm ers in lea rnin g mo re about LB.
Conclusions

Wh i le th ere we re some eleme nts in com mon betwee n the ex peri ences at the two
sites, the1e we re also substant iai di ffe rences. To a ce rtain extent th e expe ri ences
w ere co mpl ementary. In Bo li v ia, th e focus
wa s o n adaptation of an ex istin g tec hn o logy
as fa rm ers tested a strategy fo r c hem ica l
co ntro l w ith a suscepti b le cul tivar. In Peru
the emphas is w as on learn ing to wo rk w ith
res istance, and o n improv in g fa rme rs
know ledge abo ut LB and its manageme nt
w ith o ut p rom ot ing a spec ific strategy. Thi s
co mpl ementa ri ty mean s th at much ca n be
lea rn ed by sharin g inform ati o n betwee n
co untri es and partners.
A ca refu I stud y of th e Ionge r-te rm
impac ts of the approaches take n in Peru
and Bo li via should help us des ign FF S
bette r. Impact studi es are pl ann ed for Peru
to dete rmin e w heth er in th e case of LB ,
enh anc in g kno wl ed ge abo ut bi o log ica l
prin c ip les and increasin g ava il ab ility of
c ulti vars w ith res istan ce res ults in in creased
y ieid s and profits. It is also imp o rtant to
determin e how inform ati o n imparted
th ro ugh FFS spread s, as it may be th at
co mpl ex knowle dge abo ut bi o ph ys ica l
prin c ip les does not sprea d eas il y fro m
farm er to fa rm er. A careful co mpar iso n of
the lo nge r term costs and benefits o f th e
tec hn ology adaptation approac h take n in
Bo li v ia co mpa re d w ith the kn owledge
enh anci ng approa ch ta ke n in Peru sho uld
help us to und erstand th e appro pri ate
we ight to be given to eac h in th e FFS.
It is c lea r fro m the first yea r's ex peri ence
th at an FFS requires co nsiderabl e tim e and

98

Pororo

reso urces. Nevertheless, as the data fr o m

Bo livia show, th e returns th at ca n be
ac hi eved by hel p ing fa rm ers to co ntro l LB
are ve ry hi gh, ind ee d m uc h greate r th an
th ose repo rted fo r in sect pests in ri ce,
(stu d ies in Java indi ca ted yield ga in s o f 12
to 24%) w here FFS has bee n w id ely use d
(va n de Fli erl et al., 199 5) . A nd altho ugh
FFS are intensive and re lati ve ly cos tl y, th e
fin anc ial in vestm ent in FFS fo r LB is hi ghl y
attracti ve. A proj ect supported by th e
Intern at io nal Fund for Ag ri cu ltural Deve lo pme nt has rece ntly bee n ini tiated to sca le up
th e ini ti ative repo rted here, involv in g
Ch ina, Ba nglades h, Uga nd a, and Ethio pia,
as wel l as Per u and Bo l iv ia. in each co untr y, pa1tn ers fro m national ag ri cul tura l
resea rch ce nters and NGO s wo rk togeth er
w ith fa rm ers . Criti cal el ements to success
are ensurin g that partner in stituti o ns
und ersta nd and are com m itted to FFS and
th at th ose fac ili tat in g FFS are themse lves
pro perly train ed.
References Cited

Gro eneweg, K. and M . Sch outen. 1999 .

Farm er field sc hool appro ac h: A di ffe rent
settin g, a di ffe rent transl ati o n . A pil ot
farmer field sc hoo l ex peri ence in San
M iguel, Peru. A nalyti cal Report Departme nt o f Co mmunicat ion s and Inn ovati o n
Studi es, W A U, Th e Netherl and s. 87 p.
Navia, 0., H. Equi ze, and E. Fern and ezNorth co te. 1995. Estrategias para el
co ntro l quimi co del tiz 6n. Fi to pato logia
Fi cha Tecn ica 2. PROI N PA,
Coc haba mba, Bo li v ia.
Orti z, 0., P. W inters, H . Fano, G . Thi ele, S.
Gu aman, R. To rrez, G. V. Barr era, J.
Un da, J. H aki za 1999 . Und ersta nd in g
fa rme rs' res po ns es to LB (LB ) in Peru ,
Bo li v ia, Ec uador and U ga nd a. Impact o n
a Ch angin g W o rld . CIP Pro gram Repo rt
1997-98 . p. 101-109 .
Th ie le, G. , 0. Nav ia, and E. Fern and ez Northcote. 1998 . A nali sis eco n6mi co de
la estrateg ia del co ntrol qufmi co del
ti z6 n tradi o (Ph y tophth o ra infestans) para
cu!ti va res de papa susc eptibl es en
Coc haba mb a, Bolivia. Fitopato log fa
33 (3) :1 76- 181.

Torrez, R. A., Ve izaga, E., Macias, M.,

Sa laza r, A., Gandaril las, 0. Nav ia, and
G. Th iele. 1999 . Capacitaci6n a
ag ri cultores en el manejo integrado de!
tiz6n de la papa en Cochabamba.
Working Document, PRO INPA Foundation , Cochabamba, Bolivi a.
Van de Fli ert, E., J. Pontiu s, and N. Roling.
1995. Searching for strategies to replicate
a successful extens ion ap proac h: tra ining

of IPM trainers in Indones ia. European J.

Agr ic. Edu. Ext. 1 (4):41- 63 .
Ve izaga, A. 1999. Evaluac i6n parti cipat iva
de! proceso de adopc i6n de! co ntrol
quimi co de! ti z6 n, Phytophthora
infestans, en I a zo na de Morochata,
Th es is In g. Agronomo. Universid ad
Mayor de Sa n Simon, Cochabamba,
Boliv ia.

CI P Progra mReport 1997-98

99

Understanding Farmers' Responses to Late Blight:

Evidence from Peru, Bolivia, Ecuador, and Uganda
0. Ortiz1, P. Winters 2, H. Fano1, G. Thiele3, S. Guaman3 , R. Torrez3, V. Barrera4,
J. Unda4, J. Hakiza 5

It is ultimate ly a farmer's decision whether

or not to adopt a technological solution to a
problem. In developing co untries, however,
inform ation for farmers' decision-making is
scarce, and the infrastructure for generating
and maintaining deci sion-support systems
to assist fa rmers is defi cient. Therefore, in
order to id entify potent ial points for intervention in disease management, it is
necessary to understand how farmers
respond to diseases such as late blight in
potatoes. As a first step toward this understanding, a baseline study was conducted in
Peru, Bolivia, Ecuador, and Uganda
between November 1997 and August 1998.
The results of this study will help the
participatory development of integrated late
blight (LB) management. Specific objectives
of the study were (1) to estimate losses in
yield and income due to this disease; (2) to
document farmers' knowledge and practices (particularly of fungicide use) in
relation to LB control; and (3) to document
potato-related activities of extension
organizations at the pilot si tes.

Methods
For each of the four countries, a semistructured questionnaire was the main data
collection tool. Additionally, informal
interviews, focus groups, non-parti cipant
observation, and field eva luations were
used to enhance the validity of the research. Th e research was ca rried out in
collaboration with personn el from the
corresponding national programs.
1 CIP, Lima, Peru.

Pilot sites were selected where potato is

an important crop, and where LB is en demic. In Peru, th e work was focused in
the Department of Cajamarca with the
participation of 13 1 fa rmers. In Bolivia,
data were collected from 45 farmers in
Ayop aya Province. Two hundred and seven
farmers in the Ecuadorian provinces of El
Carchi, Cotopaxi, Chimborazo, and Bolivar
were included, and in Uganda, Kabale,
Kisoro, and Mbarara districts were included
in th e research with data collected from 118
respondents. This paper summarizes the
ma in findings of this study in the four
countries . Information is based on the data
collected by teams wo rking in eac h of the
countries and the reports written by those
teams.

Results and Discussion

Farmers and the potato systems
Th e potato crop was important for
incom e and food sec urity at each pilot site,
although it was one of several crops
man aged by farmers. Most informants in the
study were small farmers with less than 5
ha of land. The majority owned their land.
In Uganda and Ecuador, a sizabl e minority
(28% and 14%, respectively) were tenants.
Most of the parti cipa ting households
derived the bulk of their income from
agri culture and Iivestock managem ent.
Som e farmers (less than 30% of informants
in Peru and Uganda) also depended on offfarm acti v ities to generate supplementary
incom e, such as small businesses, temporary labo r, and artisanal activities.

2 University of New Engla nd , Armadale, Australia.

3 PROINPA, Bolivia.
4 INIAP, Ecuador.
5 NARO, Uganda.

CIP Program Report 1997-98

l0l

On average, Bolivian and Peru v ian

farmers planted less than 1 ha of potato,
w hi ch is typical for subs istence farmers in
the Andean Region. In th e Ecuadorian case,
the figure was hi gher (1.8 ha) because the
samp le included comme rcial areas such as
El Carchi Province. In Uga nda, howeve r,
fa rm ers grow about a fifth of a hectare (0 .1 8
ha ) in potato because they ha ve less land
available for cultivat ion. Yields varie d
across cou ntries. In Bolivia 12. 7 t/ ha, and in
Ecuador 1 2.6 t/ ha were reported. Th ese
numbers represent relative ly good yie lds
compared with Peruvi an farmers w ho
harveste d an average of 4.9 t/ha. In Peru ,
y iel d was greatly influenced by the adve rse
weather conditions caused by " El Ni11o"
du1in g ear ly 1998. In Boli via, Ecu ado r, and
Peru , y ields were eval uated by sampling
fie ld s; in Uga nd a, yie lds we re estimated
from information pro vided by fa rme rs
where an average of 7.0 t/ ha was reported.

Importance of late blight for farmers

From th e farmers' point of view, LB is
considered e ithe r as th e primary or secondary potato pest in th e pilot sites in th e four
co un tries. Fi gure 1 presents the indexes
obta in ed from th e farmers' rankin g of pest

Other fungus

Other insects

D Uganda

PTM

13 Peru

Rot

Epitrix

APW ~illill'!~mllll!!m::l~!i!!ifilll
BW !=::=~::=::=:=::=~::=~:==
LB

~~~~~~~~E::::J
0

20

40

60

80

100

Figure 1. Index of pest problems perceived by

farmers in Peru (n = 131) and Uganda

(n = 118). Note that farmers in Peru
consider frost a pest. Key: LB: Late blight,
BW: Bacterial wilt, APW: Andean potato
weevil, PTM: Potato tuber moth.

l 02

Potato

probl ems in Peru and Uganda. Th e Peruv ian case also appli es to Ecuador and
Boli via. In all countries, however, LB was
not the only problem. Th ere were other
pests that caused conce rn and were
perceived as reducing potato yie lds and
quality. In Uganda , for exam ple, bacter ial
wi lt was ranked as more important than LB.
In Ecua dor, Bo li v ia, and Peru , the Andean
pota to weevil ranked ju st below LB as a ke y
pest (Ortiz et al. , 1996). H ence, in designin g prog rams for the improved management
of potato pests, it is important to recog niz e
that farmers face multipl e problems.
Late bl ight dam age was evide nt during
field eval uations in the Andean countries,
but damage va rie d acco rdin g to a number
of facto rs (Table 1 ). Th e d isease in creased
dur ing th e potato seaso n, and tho se fi e ld s
that were planted late (co inciding with th e
rainy pe1iod) tend ed to ha ve more damage.
Late blight damage was in ve rs ely correlated
to fungicide applications. Plot locat ion and
the cu lli var o f potato that farmers used also
influenced disease damage.
Potato fie lds were eva luated durin g the
grow in g seaso n and also at harvest time in
Peru, Ecuado r, and Boliv ia. In Uganda, th is
part of the study will be repeated in 1999.
Cons idering potato yie ld as a depen dent
va1iab le, and di sease damage as an independ ent va ri abl e and co ntrol Ii ng for ot her
factors, it was clear in Per u, Bolivia , and
Ecuador that LB reduces potato y ields.
Figure 2 illustra tes the negative re lationship
between LB d amage and y ield in four
Bolivian co mmuniti es, w here an R2 of 0.8
was obtained. In addition, data from Peru
and Ecuador also prov id ed evidence of the
devastating effects of this disease at the
field leve l.
A conserva ti ve 20% difference in
severity between approp ri ate and in appropr iate LB contro l can be used as the lowe r
bound to esti m ate losses. Accord in g to
estimates, eac h 20 % in crease in LB seve1ity
reduces y ie ld from between 1 t/ha (i n Pe1u )
to 4 t/ha (in Ecuador). This conservative

Table 1. Regres sion of late blight severity in the Peruvian sample.

Variable
Days after planting
Days ofter planting squared
Plantin g dote
Fungicide applications

I-statistic'

Coefficient

0.3 1
0.002
0.55
- l.2 1

2.06 **
3.10 ***
10.18 ***
- 2.88 ***

- 14.40
l 5.20
- 9.45
- 8.86
l.01
4.35

- 3.74 ***
4.1 0 ***
- 1.84 *
- 2.70

10.08
8.58
- 7568.41

2.4 5 **
2.59 ***
- 10.18
355

Potato varieties:
Amarilis
Concho n
Chou cha
Liberteiia
Perricholi
Yun gay

0.18
l.18

Places
Contumazo
San Miguel
Constant
Number of observations
R2

0.54

' I-sta tistic reference loca tion is Cajomarco and varieties are those that are either native or not used by a substantia l number of
form ers. *, **, ond **' indicate significa nce with 99%, 95%, and 90% confidence, respectively.

estimate is lowe r th an th e figures reported

by Bay lon (1987) who indi cated losses of
about 6 t/ ha in Peru , and those reported by
Thi ele et al. (1998) who indicate th at an
avera ge of 6.5 t/ ha co uld be lost to LB in
Bo li v ia.
With 1998 p ri ces, U S$140/t, a reduct ion
in potato y ield would mea n a loss of
between US $140/ ha to US$560/ ha in Peru ,
and betwee n US$250/h a and US$ 1,000/h a
in Bo li v ia where po tatoes are va lued at
US$250/t. For small farm ers, y ield reducti o n caused by LB ca uses a significant loss
of income from th eir lan d, and 1edu ces
food ava il ab ility for the ho usehold.
It is difficult to extrapol ate th e resu lts of a
pilot study to th e situat ion in an entire
country, but the extent of the losses ca n be
estimated. In Peru, fa rm ers grow an ave rage
of 250,000 ha of potatoes per year. Using
conse rvative estim ates, 20% of th e total

area pl anted to potatoes cou Id be severe ly

affected by LB each year. Thi s would mean
a tota l mon eta ry loss of between US $7
million and US$25 million per year. In
Boli via, Thi ele et al. (1998) repo rt th at LB
affects 20,000 ha. U sin g 1998 potato
pr ices, th is rep res ents a loss of betwee n
US $5 million and US$20 million per year. It
is c lear that this range of losses justifi es
in vestment in developing control strategies.
And any strategy th at can contribute to th e
red uction of LB damage w ill have a direct
effect on increasin g y iel ds, and in turn ,
farmers ' income and food secur ity (Torrez et
al. , 1999).

Farmers' knowledge of late blight and

practices to control the disease
Unlike in sects, diseases are difficu lt to
see; farme rs do not usually kn ow th e ca usa l
agent of plant diseases (Be ntl ey, 1990,
199 1). Thi s was confirmed by the fin d in gs
of th is study that showed that fa rmers lack

CIP Program Report 1997-98

l 03

Piusilla

20
18
16
14
ro 12
~ 10
8

"

Qi

;;::
4
2
0

R2 = 0.81

10

15

20

AUD PC

Pata Morochata

kno w ledge of late blight biology. Most of

them did not kno w that the real cause of
late blight is a microorganism. Th ey
associated the disease with the weather
conditions that favo1 its occurrence such as
heavy rainfall, fog , cold weather, and even
li ghtning. Add ition a ll y, farmers were unable
to correctly diagnose the disease. More
than 88 % of informants confused th e
sym ptoms of late blight with oth er funga l
diseases o r w ith " le af burning " ca used by
excess pestic id e use.

25
20

ro

15

"'~
";;::Qi 10
5

Y =20.31 - 1.69 AUD PC

Fl'= 0.84

0
0

AUD PC

San Isidro

16
14
12
-10

"'~"

";;::Qi

8
6
4

Y-=15.51 - 0.64 AUDPC

R 2 = 0 .89

0
0

10

15

AUD PC

Compai\ia Pampa

20

~
Y=21. 18 - 2 .59 AUDPC
A~ = 0 .77

o ~~-'--~~~~~~-'-~~~~

AUD PC

Figure 2. Relationship between late blight damage

and yields in four Bolivian communities
during 1997-1998 cropping season. Key:
AUDPCr (area under the cur ve - an index
for disease progression) .

l 04

Potolo

When asked about different control

practices, farmers in all the participating
countries we re most familiar w ith fun g ic id e
use, and LB control is largel y based o n this
chemical contro l method. In Ugand a about
69 % of farmers used fungicidal produ cts
compared to more than 95 % in th e Andean
Countri es. Yet it w as c lear that most farmers
did not know how to differentiate contact
fungicid es from system ic fungicid es, and
the y tend ed to spray any av ailable product
according to their expe rience .
During the study period, farmers sprayed
an ave rage of 3 times per cropping seas on
in Uganda and Boli v ia, 5 times in Ec u ador
and 7 tim es in Peru (3 to 4 sprays are
common in Peru in a normal y ear). Th e
number of sprays w as influenced b y local
weather co ndition s. According to a multivar iabl e regre ssio n analysis run in Peru,
fungicide use was influenced by a number
of factors (Table 2). Younger, educated, and
wealthier farm ers tended to spray more.
The plantin g season also influen ced
fun g icid e use; those farmers w ho planted in
the rainy season spraye d more often
becaus e this period is most conduci ve to
the disea se . Plot location also in fluenced
the pattern of fungicide use, so farmers
spray more frequently in those pla ces that
have hi gher relative humidity (Sa n Mi guel
and Contumaza shown in Table 2). It is
importan t to note that there is no evidence
that th e use of resistant culti vars influ ences
fungicide use in Peru. The reason m ay be
that most farmers plant res istant and
susceptibl e cu lti vars together (in strips) and

Table 2. Regression on fungicide use in the Peru vian sample.

Variable

Coefficient

Age

-004

I-statistic'

- l.95'

Educa tion level

0.26

2.33* *

Family labor

0.01

0.06

Land owned

0. 53

2. 16**

Potato vo rieties:
Amaril is

-3.66

-0.66
0.84

Ca nchan

0.47

Chaucha

l.38

1.63

Liberteiia

-0.35

-0.62

Perricholi

0.93

l.22

Yun gay

-0.18

-0.37

Planting dote

0.21

3.27***

3.34

4.24***

Places
Contumaza
San Miguel
Constant

3.38

4.29** '

- 286.03

- 3.22 *'*
131

Number of observations

R1

0.56

' I-stati sti creference location isCajamarca and varieti es are those that are either nati ve or not used by a substantial number
of form ers;*,**, *** indicate signifi ca nce with 99%, 95%, and 90% confid ence, respectively.

tend to spray both at th e sa me ti me.

Ev idence fro m th e co un tri es in cl uded in this
study shows th at eac h fa rm er may have hi s
ow n co ntro l strategy, as o pposed to the
uni fo rm co ntrol strateg ies used in temperate
countri es (M ac kenz ie, 198 1).
Th e Peru vian and Bo li vian data in d icated th at th e number o f sprays was
in ve rse ly and si gnifi ca ntl y assoc iated w ith
th e exte nt of LB damage. Pota to fields in
w hi c h fa rm ers sprayed mo re tended to have
less da mage.
Farm ers used a number of co mmerc ial
prod ucts w ith a range of acti ve ingredi ents
(Tabl e 3). In U ga nd a ju st two acti ve in gred ients were fo und on th e mark et. Between six
and eightee n we re prese nt in th e And ea n
co un tri es. M ancozeb was by fa r th e most
co mm o nl y used acti ve in gredi ent in
Uganda beca use it was relati ve ly inexpen-

si ve . At th e Bo li vian p il ot site, farm ers used

up to six acti ve in gred ients; propa noca rb
was th e most co mm o n. The Peru v ian p il ot
area prese nted mo re va ri abi I ity of active
in gredi en ts. A total of 1 3 diffe rent in gred ients was fo un d in the sa mple, and
metalaxy l was th e most co mmonl y used.
Th e Ec uado ri an case had an eve n hi gher
vari ab i I ity o f chemi ca l produ cts to co ntro l
LB (up to 18), and the most comm o n acti ve
in gred ient, mancozeb, was used in a to tal
of 47% of app li ca ti o ns. Thi s va ri abili ty of
c hemi ca l products used to co ntro l LB,
mi xed freq uen tl y w ith in secti c ides, rep rese nts a threat to oth er li v in g o rga ni sm s
(Pilling and Jepson, 1993; Schu ster and
Scho reder, 1990) . The hi gher vari ability of
acti ve in gred ients in Latin Am eri ca also
refl ec ts the grow in g market for pest ic ides in
th e reg io n, in compa ri so n to th e less
deve loped market fo r th ese produ cts in
Afri ca (Repetto and Ba li ga, 1996) .

CI P Program Reporl 1997-98

l 05

Table 3. Main active ingredients used as fung icides by fa rmers in the pilot sites.
Country

Active ingredients

Uganda

Bolivia

Ecuador

18

Peru

13

Most common

Proportion of

active ingredients'

sprays(%)

mancazeb
metalaxyl
prapanocarb
afurace + ma ncazebh
chlorothaloni l
mancazeb
cymoxanil + mancazeb
cymoxnil + propineb
sulfur
melalaxyl + mancazeb
mancozeb
propanocarb
metiram

90
9

44
25
19

25
16
9
6

42
24
19
7

Only the most common active ingredients are included in this column.
) Mea ns both active ingredients were combined in the same commercial product.

b(+

Th ere was also a grea t dea l of va 1iab ility

of fungic ide closes used by farmers to
contro l LB. In Peru, fm exa mpl e, farmers
tended to use less th an the reco mm ended
close. In co mmercial potato grow in g areas
in Ecua do r, howeve r, fam1ers tend ed to use
a hi gher close than that recomm end ed,
generall y by mi x in g different c hemi ca l
produ cts (Crissman et al. 1998). In
Uganda, 30% of farmers also used mme
than th e reco mmend ed close. But, acco rcli ng to fa m1 e1s and exte nsio n officers, th e
1eason is the ex isten ce of ad ulterated
products. Fam1ers in creased th e close in
mcle1 to make su1e that they co uld co ntrol
the d isease.
Th e cos t of fun gic ides was diffi c ult to
eli cit from fa rmers beca use in most cases
they d id not kee p reco rd s of th e products
th ey use. Howeve r, cos ts in Peru and
Ecuadm ave rage US$140/ ha an d US$1 50/
ha, 1a ngin g from zero to US$500/ ha. Th ese
est im ates are equiva lent to 10-1 5% of total
product io n costs. Thi is consistent w ith
data reported by Bay lo n and Otazu (1987)

l 06

Potato

in Peru , and Crissman et al. (1998 ) in

Ec uado r. An impmtant issue is th at pest ic ide use represe nts an out-of-pocket cost.
Th e c hall enge is to deve lop less expe nsive
met hods of pest co ntrol usin g eith er
nonc hemi ca l or mo re appropr iate c hemi ca l
met hods, so th at far mers can 1eclu ce the
amount of fun gic ide th ey use.
Cultivar use and LB damage
Between 50% an d 100% of farmers in
th e four co unt1i es we re aware of the
differences in LB res istance of potato
c ulti va rs. Howeve r, fa rmers p 1efe tTed
particu la1 culti va rs fo r more reaso ns th an
ju st LB res ista nce. Farmers in Uganda
preferred culti va 1s that matured ea 1ly. They
wan ted to harvest potatoes in approx imate ly four months in orde1 to have the
land free fo r anothe r crop . In Peru, ea rlin ess
was less important because most fa rm ers
have o nl y one grow in g season per yea r.
Among th e Peru v ian farme1s, 85% preferred
cultiva 1s w ith 1es istance to LB , wh e1eas in
Uganda only 12% of farme 1s fe lt that

resistance was an important attribute. LB

was not their main source of concern.
Evidence from Peru suggests that LB
severity is statistically correlated with
cultivar use (see Table 1 ). There were
cultivars (Amarilis, Chaucha, and Libertefia)
that showed resistance at the field level.
There were also susceptible cultivars such
as Canchan. This suggests that genetic
resistance can piay a key role in reducing
LB damage at the field level, and that
resistant cultivars can be a crucial component of an integrated management strategy
against this disease. The continuous
deployment of resistant cultivars would
mitigate the risk associated with resistant
cultivars currently in the field becoming
susceptible to LB.

Extension activities and LB control

In Peru, Bolivia, and Ecuador, nearly all
institutions working with agriculture in the
pilot areas focused to some degree on
potato production. In Uganda, however,
this was not the case because potatoes are
not as important a crop there as they are in
the Andes. Government extension services
in all four countries were inappropriate in
terms of number of extension personnel,
coverage and facilities. However, nongovernment organizations (NGOs) have
been increasingly participating in providing
these kinds of services to farmers
(Bebbington and Thiele, 1993).
The main potato-related activity carried
out by extension organizations is the
provision of technical assistance, particularly related to the use of agro-chemicals,
and the provision of credit though revolving
funds for seed, fertilizer, and pesticide. The
participation of NGOs as pesticide providers to farmers has increased in recent years.
However, evidence suggests that extension
institutions are also important sources of
new potato seed for farmers, so that new
cultivars could be distributed through this
channel.
Farmer training on LB biology and how
to control this disease is limited. Most

organizations sti 11 use a top-down technology transfer approach based on providing

recommendations. This could have
negative implications when trying to
promote integrated pest management. This
technology requires horizontal and participatory methods of exchanging information
and knowledge as suggested by Roi ing and
Van de Fliert (1994), and Torrez et al.
(1999).

Conclusion
According to farmers' perceptions, LB ranks
as the most important pest problem in the
Andean countries included in the study,
and as the second most important pest in
Uganda. Farmers' concern with LB is
understandable considering that, as this
study shows, this disease indeed reduces
potato yields, and is a real threat to potato
production , food security, and farmers '
profits. This suggests that investment in
developing and disseminating control
strategies in developing countries would be
profitable. However, LB is also part of a
complex pest system that depends on
location, type of farmer, and management
strategies. Development of control strategies must be flexible enough to integrate
other local pests of economic importance.
In addition, pest control is only one of
several endeavors for farmers, therefore any
control strategy should be adaptable to
existing production activities and goals.
Farmers' lack of knowledge about
biophysical principles related to pest
control was a common feature in the
participating countries. Any effort to
improve pest control at the field level
shou Id start by providing farmers with
missing information, so that they can
acquire new knowledge. However, to what
extent new knowledg e about biophysical
principles would be reflected in better
decision-making about the control of LB is
a question that remains unanswered.
In spite of the negative effects of fungicide use on peoples' health and the environment, these products are still the
CIP Prngro mReport 1997-98

l 07

primary control measure at the field level.

Fungicide use does reduce LB damage and
increases potato yields. Evidence suggests
that there is room for including fungicide
use in LB control strategies, at least until
other control options are availabl e. Farmers
in the developing countries studi ed,
however, lack knowledge about the
appropriate use of fungicides.
The study demonstrated a clear relationship between potato cultivars and LB
damage at the field level. Hence, genetic
resistance can play an important role as
part of an integrated management strategy
against this disease, and may be a way to
reduce fungi c ide use per hectare. However,
so far there is no evidence that fungicide
use has been reduced.
Inter-institutional cooperation becomes a
key issue in overcoming d ifficu Ities encou ntered by extension services and to facilitate
farmers ' access to new information, knowledge, and resistant cultivars. Participation
of all persons involved, particularly farmers,
is required to develop and disseminate
control alternatives in order to reduce the
devastating effects of LB in developing
countries .
Acknowledgement
Sandra Guaman, who died in 1998, led the
work on our Bolivia baseline study under
difficult circumstances. This work would
not have been possible without her dedicated research.
References Cited
Baylon , Y. and V. Otazu. 1987. Aspectos
econ6micos en el control de rancha P
infestans en dos zonas con diferentes
ni veles de incidencia. In: Resumenes de
IV Congreso Latinoamericano , X
Congreso Peruano de Fitopatologia.
Lima, Peru.
Baylon, Y. 1987. Estudio econ6mico del
control de la rancha (Phytophthora
infestans) (Mont De Bary) en papa en

l 08

Pololo

lugares de di ferentes ni ve les de

incidencia . Agronomy Thesis.
Universidad Nacional del Centro.
Huancayo, Peru.
Bebbington , H. and G.H . Thi ele . 1993.
Non -gove rnmental organizations and the
state in Latin America. Rethinking roles
in sustainable agriculture development.
Routl edge ., London , UK . 290 p.
Bentley, J. 1990. Conocimientos y
experimentos expontaneos de
campesinos hondurei'ios sobre el maiz
muerto . Manejo lngegrado de Plagas
(Costa Rica ) 17:16-26.
Bentley, J. 1991._zQue es el hielo?
Percepciones de los campesinos
hon du rei'\os sob re I as enfermedades del
frijol y otros cultivos. lnterciencia

16(3):131-137.
Crissman, C. , P. Espinosa, C.E.H. Ducrot,
D .C. Col e, and F. Carpio. 1 998. The case
study site: Physical , health , and potato
farming systems in Carchi Province. In:
C. Crissman, J.M. Antl e, and S.M.
Capalbo (eds.). Economic, environmental, and health trade offs in agriculture:
Pesticides and the sustainability of
Andean potato production. Kluwer
Acad emic Publishers, Norw ell , MA,
USA. p. 85-119.
Mackenzie, D.R. 1981. Sch eduling fungicide applications for potato late blight
with Blitecast. Plant Dis. 65:394-399.
Ortiz, 0. , J. Alcazar, W. Catalan, W.
Villano, V. Cerna, H. Fano, and T.
Walker. 1996. Economic impact of !PM
practices on the Andean potato w eevil in
Peru . In: Walker, T. and C. Crissman
(e ds. ). Case studies of th e economic
impact of CIP-related technolog y. CIP,
Lima, Peru . p. 95-110.
Pilling, E.D. and D.C. Jepson 1993 . Synergism between EB! fungicid es and
pyrethroid insecticide in the honeybee
(Apis mellifern). Pesticide Sci. 39:293-

297.
Repetto , R. and S. Baliga. 1996. Los
plaguicidas y el sistema inmunitario:
Riesgos para la salud publica. World
Resources Institute, Washington, D .C.

Roling, N. and E. Van de Fliert. 1994.

Transforming extension for sustainable
agriculture. The case of integrated pest
management in rice in Indonesia . Agric.
Human Values 11(2-3):96-108.
Schuster, E. and D. Schoreder. 1990. Side
effects of sequentially and simultaneo usly applied pesticides on non-target
soil microorganisms in laboratory
ex periments . Soil Biol . Bioch em. 22:375384.
Thiele, G., 0. Navia, and E. FernandezNorthcote. 1998. Analisis econ6mico de

la estrategia de control quimico de! tiz6n

tardio (Phytophthora infestans) para
cultivares de papa susceptibles en
Cochabamba, Bolivia. Fitopatologia
33(3):176-181.
Torrez, R., J. Tenorio, C. Valencia, R.
Orrego, 0. Ortiz, R. Nelson, and G.
Thiel e. 1999. Implementing IPM for late
blight in the Andes. Impact on a Changing World . CIP Program Report 19971998. p. 91-99.

CIP Program Reporl 1997-98

109

Sensitive Detection of Ralstonia solanacearum in

Latently Infected Potato Tubers and Soil by
Postenrichment ELISA
S. Priou, L. Gutarra, H. Fernandez, and P. Aley 1

Ralstonia solanacearum is the causal agent

of th e disease known as bacterial wilt (BW)
or brown rot, the second major constraint
to potato (So lan um tuberosum) p rodu ction
in tropi ca l and subtropical region s wor ldwide (Haywa rd , 1991 ). Brown rot has
recently become a serious threat to potato
seed produ ction in cool, temperate countries of North ern Europe (Janse, 1996).
Since th e pathogen is mainly transm itted
throu gh tub er see d, the use of hea lth y
plantin g mater ial is the mo st effective
means to co ntrol th e disease (H aywa rd ,
1991 ). Under cool conditions, th e plant can
be infected without exhibiting visible
symptoms, resulting in latent infection in
vascular ti ss ues of progeny tubers (Ciampi
et al., 1980; H aywa rd , 1991 ). Efficient
detection tec hniqu es for routin e use in
quarantin e proc ed ures and seed ce rtification schemes are crucial.

In the European Union, stand ard methods to monitor the occurrence of R.

solanacea rum in potato seed tubers invol ve
the use of indirect immunofluoresce nce
antibody staining (IFAS) and isol at io n on a
selective med ium (OE PP/ EPPO, 1990) . The
most commonly used technique s to co nfirm
positive samples in I FAS are the bioassay on
tomato pl ants and the detec tion of R.
solanacearum- specific DNA sequences by
polymerase c hain reaction (PCR ) from pure
bacterial cultures that have bee n isolated
on a selective medium (OEPP/ EPPO, 1990).
Other tec hniqu es have become ava il able
and have bee n rev iewed by Sea l and
Elphinston e (1994). There is still no method
to detect latent infection in tubers th at
combines th e advantages of high se nsitivity

and specificity, low cost, and suitability for

routin e use in deve lopin g countries where
laboratory fac iliti es and sk ill ed pe1sonnel
are often limited (Bla c k and Elphin sto ne,
199 8) .
Another major compon ent of BW
mana gement is the planting of po ta to in R.
so/ana cea rum-free soils. Th e pathogen can
survive in pl ant debris and in the rhi zosphere of potato, many weeds, and other
crop s (Haywa rd , 1991 ). Studies o n th e
ecology, ep id emiolog y, and con tro l of R.
solan acea rum have been limi ted by th e
inability to detect the pathogen in low
popul ation s in soil. The most widely used
method s to detect R. solana cea rum in soil
con sist of strea king soil solution onto a
spec ifi c medium (G ranada and Sequeira,
1983; Englebrecht, 1994). Th e plating
techniqu e req uires skilled perso nn el to
distin gui sh co lonies of th e path oge n from
other bacterial saprophytes, limiting its
extensive us e. Sensitivity of th e method also
varies by soil sa mple becau se so il antagonisti c mi cro flora can impair the gro w th of
R. solanacearum. The us e of indi ca tor
plants such as tomato has been repo rted
(Graham and Lloyd , 1978), but th e sensitivity of the bioassay is low (10 4 -10 5 ce lls/g
soil). Indirect enzyme-linked immuno sorbent assay (ELISA) has also been developed (Robinson-Smith et al., 1995), but it
lacks se nsiti v ity (104 cell s/g so il ) and
spec ific ity.
Detectio n se nsiti vity can be in c reased by
enrichment, i. e. , incubating the ex tracts in a
selective broth to multiply the bacte ria . In a
preliminary study (Garris et al., 1997), the
existin g methods of enrichm ent o f the

1 CIP, Lima, Peru.

CIPPwgwrn Repo1t 1997 98

ll l

bacter ia in tuber extracts we re compared

and refin ed to in crease th e se nsiti vity of
sero log ica l techniques . Maxim um grow th
ra te of R. solanacearum durin g enrichment
was obtai ned usin g citrate buffer fo r sample
extract ion from tubers and th e modified
se lective broth (M-SMSA) deve loped by
El ph instone et al. (1996). This paper reports
th e improvement of the im mun oassay on
ni trocel lul ose membran e (NCM -ELISA) and
of the antiserum spec ifi c ity. For the detection of R. so lanacea rum in soi l, double
antibody sa ndw ich immunoassay (DASELISA) in microtitration plates has been set
up using the sa me antiserum used fo r
detection in potato. The detection effecti veness of postenrichment ELISA is com pared
with microbiological techniques and a
nu cleic acid spot hybridization proced ure
(NASH ).
Materials and Methods
Tuber extracts
Preparation of tuber extracts and
enrichment. Tubers we re washed and
disinfected. A thin slice of th e tuber was
removed from around th e sto lon end w ith a
sca lpe l, and 3 x 3 mm stri ps along the
vasc ul ar rin g we re removed. Th e tuber
fra gments we re put into a pl astic bag on ice
and weig hed. Two ml of ster il e 0.1 M citrate
buffer (pH 5.6) per g of tuber tiss ue were
added , and tub er fragments were sq uas hed
w ith a rubbe r ma llet. Th e enri ched tuber
extracts we re prepared by mi x in g 500 L of
the supernatant tuber extra ct w ith 500 L of
M-SMSA in a 1.5 ml Eppendorf tube, and
in cubat in g for 48 hat 30( with co nstant
ag itat ion (1 70 rpm) or manu al ag itation
tw ice a day.
Preparation of inoculated tuber extracts. A health y tuber extract (HTE) was
prepared w ith R. so/ana cea rum-free tubers
(cv. Revol uc i6n) produ ced by CIP Seed
Unit. A suspe nsion of R. so lanacea rum in
citrate buffer was prepared by cu ltu ring
stra in CIP 204 (biovar 2A) for 48 h at 30(
on mod ifi ed Kelman's medium (M KM;
French et al., 1995), but witho ut tetrazolium chloride. Th e HTE was mixed wi th the
11 2 Potato

suspe nsion of R. sola nacearum to obtai n a

final bacteria l co nce ntration of 2 x 10 8 ce ll s/
ml. Ni ne, tenfold d i luti on ser ie s in HTE
were prepared to reac h the theoretical
co ncentration of 0.2 ce lls/ ml . Eac h di lu tion
w as enr iched in M-SMSA . The HTE and th e
M-SMSA broth were incubated for 48 h at
30 ( as contro ls. Th e population of R.
so lanacearum in the initial inoculated tuber
extract was assessed by spreading 5 0 L of
a 1o-s dilu tion in sterile, distilled wate r on
MKM (3 plates per dilution) and by cou ntin g co lonies of th e pathogen after 48 h
incubation at 30 C. The experi ment was
repeated three tim es.
Preparation of naturally infected tuber
extracts. Ei ght sin gle-tuber extracts were
prepared from infected tubers (cv. Yungay)
harves ted in a rac e 3-i nfested field
(Carhuaz, Peru ). Ten, tenfold dilution series
of th e 8 infected extracts in HTE we re
prepared, and each dilution was enri ched
in M -SMSA. Popul ations of R.
so lanacearum in th e orig in al infected
extracts were assessed by sprea din g 50 L
of eac h of th e dilutions on MKM (3 pl ates
eac h).

A total of 255 extracts we re prepared

fr om tubers of 122 different potato clo nes
and culti va rs: Revo luci6n, Yun gay, and,
Canchan (all susceptible), Mo linera (to lerant), and Cruza 148 (resistant), w hich we re
harvested from symptomless plants in th e
sa me germpl asm evaluation trial. The tuber
extracts were prepared by mixing th e tuber
fragments removed fro m 20 tu bers/ge notype take n at random from th e harvest of 5
pl an ts . Iso lations of R. solanacearum were
perfo rm ed on MKM from the non enri ched
extracts .
Soil extracts
Preparation of inoculated soil extracts.
A nonsteri le R. so lanacearum-free so i l from
Hu ancayo, Peru , was in oculated by mi xin g
10 ml of a water suspension of R.
solanacea rum strain 204 (b iova r 2A) with
90 g of so il to obtai n final concentrations of
10 7, 104, 10 2, 20, 2, and 0.2 cells/g so il.
Three replications of each concentration of

soil inoculum were prepa red. Inocul ated

soi Is were kept at room temperature for 24
h before use. Soil suspensio ns we re prepared by mixing 10 g of so il w ith 90 ml of
PBS buffer. The so i I suspension was ag itated
for 30 min and allowed to sett le for 40 sec.
The supernatant was eith er used direct ly for
detecti on or enrich ed. The enrich ed so il
extra cts were prepared by mi xi ng in an
erlenm eye r fl ask 2 ml of soi l solution w ith
38 ml of M-SMSA supplemented w ith
potato broth (1 :1 ), and in cubating th e
mixture for 48 h at 30( w ith constant
agitation (170 rpm ). Fifty L of each so il
so luti on we re plated on M-SMSA med ium .
Three plates per concentration were
incubated for 48 h at 30C and colonies of
R. solanacea rum were co unted to est im ate
the original so il popul atio n.
Preparation of naturally infested soil
extracts. Ten soi I samples were taken from
a rac e-1 infested field (San Ramon , Peru)
and 15 from a rac e-3 infested field
(Ca rhu az, Peru) in th e rh izosph ere of
infected and symptoml ess potato plants (at
approximately 20-30 cm depth ). Soil
extracts were prepared 1 or 2 d after
sa mplin g as previously described for
inoculated so il s, and 8, te nfold serial
diluti o ns were done in ste ril e, di still ed
water. No ndiluted and diluted soil so lu tions
were enri ched in M-SMSA-potato broth and
plated in M -SMSA medium to estimate th e
original so il populati on of R. solanacearum.
Polyclonal antisera production
Antigen preparation. Strains of CIP 204
(biovar 2A) and CIP 104 (biovar 2A) of R.
solanacearum were used for the rabbit
immuni zat io ns. Th ey were cultured on
MKM without tetra zo lium chloride fo r 48 h
at 30C. Th e ce lls were harvested in 0.01 M
phosph ate- buffered saline, pH 7.4 (0.5X
PBS), ce ntrifu ged for 10 min at 10,000 x g
and was hed 3 times in 0.5X PBS. Th ey we re
resusp end ed in 0.5X PBS, and their number
was estim ated by absorbance at 600 nm
and the concentration adjusted to 2 x 10 9
cells/ml . The cells w ere then fixed with 2%
glutarald ehyde following the method of
A llan and Kelm an (1977).

Rabbit immunization. Femal e Rex x

New Zea land er rabb its were immuni zed by
intradermic inj ect ion of 2 x 10 9 glutara ldehyde-fixed w hol e ce ll s of an eq ual mixture
of strai ns CIP 204 and CIP 104 in 1 ml of
0.5X PBS, w hich was emu lsifi ed in an eq ual
volum e of Freund 's in co mplete adjuvant
(Di feo). The rabbits we re immuni zed
intram usc ul ar ly 32 d later in eac h hindl eg
w ith a total of 106 glutaraldehyde-fixed
whole ce ll s (sa me strain mi xture) in 1 ml of
0.5 X PBS, which again was emulsified in an
equal volume of Freund's incompl ete
adjuvant. This la st inj ect ion was repeated
once a week for 9 wk. Blood was co ll ected
from the latera l ea r vein 52 d after the first
immuni zat ion, and then weekly. Th e blood
was all owed to clot at 4(, sep arated by
centrifugation (15 min at 10,000 g) and the
serum fraction collected. Antibody leve ls in
th e serum we re determined by NCM-ELISA.
Antiserum adsorption. On e ml of eac h
antiserum was mi xed with 9 ml of a
bacterial suspe nsion in 0.5X PBS conta inin g
a total of 1. 5 x 10 10 ce ll s of an equ al
mixture of 5 spec ies: Erwinia carotovora
subsp. atroseptica (CIP 421 ), E. carotovora
subsp. carotovo ra (CIP 400), E.
chrysanthem i (CIP 367), Pseudomonas
syzygii (CN PPB 3792), and R. pickettii
(NCPPB 3899) . Thi s mixture was incubated
1 h at 32( with co nsta nt agitation (50 rpm )
and ce ntrifu ged for 10 min at 10,000 g.
Supernata nt was collected and the antibody
leve l determined by N CM- ELISA .

NCM-ELISA. Dot-blotting and NCMEllSA were perform ed as described in a CIP

trainin g manual (CIP, 1997) with th e
following modifi cat ions. The blocking and
antibody so lution s we re co mposed of 2%
nonfat powdered milk in TBS buffer; R.
solanacea rum-spec ifi c antiserum and th e
goat-antirabbit antibodies conjugated to
alkaline phosphatase (B iorad) were diluted
1:1,000 and 1:4,000, res pecti ve ly. Th e
concentrations of NBT and BCIP solutions
in N-dimethylformamide used to prepare
the subst rate solution were doubl ed (40 mg/
ml and 20 mg/m l, respectively).

CIP Progrnm Report 1997-98

113

DAS-ELISA. DAS-ELISA was performed

as described by Clark and Adams (1977)
using the immunoglobulins (lgG, diluted
1 :1,500) purified from the polyclonal rabbit
antiserum used for NCM-ELISA. The lgG
were conjugated to alkaline phosphatase
and diluted 1 :1,500 for detection. Bioreba
microtitre plates were coated with the lgG
for 4 hat 37C, incubated with the samples
overnight at 4C, and then for an additional
4 h with the conjugated lgG at 37C.
Absorbance at 405 nm was determined
using a Biorad Model 2550 plate reader
following 1 h incubation at room temperature. ELISA readings were considered
positive when they exceeded two times the
mean of the negative controls.
NASH. NASH was used to check the
specificity of NCM-ELISA. For dot-blotting
of the samples, NCM was immersed for 15
min in 5X SSC. Bacterial DNA from tuber
extracts on the NCM was released with
alkali , neutralized, and fixed by UV
crosslinking . A 2 Kb DNA fragment specific
for R. solanacearum race-3, developed by
Cook and Sequeira (1991 ), was used for
hybridization of R. solanacearum in tuber
extracts. Probe labeling and hybridization
at 55C were performed as described in a
CIP training manual (CIP, 1997). Only the
first three washing steps were applied.
Genomic DNAs of R. solanacearum (strains
CIP 204 and CIP 311, biovar 2A) and f.
carotovora . subsp. carotovora (CIP 400)
were used as positive and negative controls
and were prepared according to Cook and
Sequeira (1991 ).
Bacterial isolates and cross-reaction
tests
A total of 259 potato isolates of R.
solanacearum from Cl P's international
collection including different biovars (170
isolates of biovar 2A, 20 of biovar 2T, 48 of
biovar 1, 15 of biovar 3, and 6 of biovar 4)
originating from several countries, were
used to check the antiserum. Other bacteria
were also used and are listed in Table 1. All
bacterial strains were cultured for 48 hat
30C on MKM without tetrazolium chloride, and the cells were harvested in citrate
114

Potato

buffer. The suspensions were adjusted to 2

x 10 9 cellsi mlfor NASH, and 2 x 10 8 cells/
ml for NCM-ELISA. Suspensions of the 11
identified non-R. solanacearum bacterial
strains listed in Table 1, and of 11 unknown
saprophytes isolated from enriched tuber
extracts, were also diluted in HTE to a final
concentration of 2 x 10 6 cells/ ml and
enriched to check the cross-reactivity of
NCM-ELISA after enrichment. Since the
tuber extracts were prepared from the tuber
vascular ring, very few saprophytic bacteria
could be isolated from enriched extracts
(unlike with soil extracts). Those 11 isolates
were a representative sample of the variabi I ity of colonies isolated on MKM plates.
To verify the specificity of DAS-ELISA after
enrichment, 66 unknown saprophytes were
isolated from soil extracts of 3 different soils
(Huancayo, Carhuaz, and San Ramon,
Peru). These isolates were diluted in soil
extract to a final concentration of 2 x 1 0 6
cells/ml soil solution and enriched.

Results
Improvement of antiserum specificity
Several antisera had previously been
obtained by immunizing rabbits following
various protocols. The immunization
schedule reported in this paper produced
more specific antibodies and a higher titre.
When immunization was done only with
strain CIP 204, the antiserum reacted with
all R. solanacearum potato strains of
biovars 1, 2T, 3, and 4 tested. The antiserum did not, however, recognize a group of
strains of biovar 2A (data not shown ).
Therefore, two strains, CIP 204 and CIP 104
(biovar 2A), the latter belonging to the
nonrecognized group, were used for
immunization. From the two antisera
produced, antiserum P-359 was chosen
because it recognized all strains of R.
solanacearum tested at 10 8 cells/ml of
citrate buffer. After adsorption of the
antiserum , all 259 R. solanacearum strains
tested were still recognized by the adsorbed
antiserum. The cross-reactions of
saprophytes were significantly reduced
(Table 1). After enrichment, none of the 11
saprophytic bacteria isolated from enriched

potato ex tracts cross-reacted in NCMELISA. Only Pseudomonas celebensis, the

banan a blood disease bacte rium geneti ca ll y
closel y re lated to R. solanacearum, still
reacted stro ngly in NCM -ELISA, but thi s
pathogen is unlike ly to occ ur in tube rs.
In NCM-ELISA, ma ny cross-reacti ons
were obtained after e nri chme nt of so il
extracts inoc ulated with sa proph ytes (Table
1 ). Since no c ross-react io n was obse rved in
DAS-ELISA with the 66 sap rophytes iso lated
from so il extracts, DAS-E LISA was c hosen

for the detection of so iI popul ati ons of R.

solanacea rum. P celebensis was sli ghtly
detected in DAS-ELISA but not P syzygii,
th e age nt of the Sumatra disea se of clove,
also present in Asi a n soi ls.
All R. solanacea rum biovar 2A stra ins
tested we re hybridi zed by th e probe used,
but none of the non-R. solanacearum
bacte ri al strains tested. NASH thu s provided
a tool to evaluate th e spec ificity of th e
serol og ica l method s in assessin g th e risk of

Table 1. Cross-reactions obtained with non-Rolstonio solonoceorum bacterial strains in NCM-ELISA using the
crude or the adsorbed antibodies {Ab), and in DAS-ELISA using purified immunoglobulins {lgG), when
in pure culture in citrate buffer at 2 x l0 8 cells/ml without enrichment {- E) and after 48 h enrichment
( + E) from an original concentration of 2 x l 06 cells/ml tuber extract or 2 x 106 cells/ml soil solution,
CIP, 1998.

NCM-ELISA'
Nonodsorbed Ab

Adsorbed Ab

-E

+ E

- E

+ E

-E

+ /+ /+/-

++

+/-

++

Bacteria

Strain (no.)

Erwinia carotovora
subsp. carotovora
Erwinia carotovora
subsp. atroseptica
Erwinia chrysanthemi
Pseudomonas syzygii
Ralstonia pickettii
Burkholderia cepacia
Pseudomonas aeruginosa
Pseudomonas putida

CIP 400

CIP 421

+/-

CIP 367
NCPPB 3792
NCPPB 3899
NCPPB 2993
NCPPB 1965
NCPPB 1806
NCPPB 1808
UW443
uw 446
11 isolates

+/++
+
+ /+/+/-

Pseudomonas ce/ebensis
Unknown bacteria isolated
from enriched tuber extracts
Unknown bacteria isolated
from three different soil
extracts
Unknown bacteria isolated
from three different enriched
soil extracts

29 isolates

37 isolates

DAS-ELISA'
Purified lgG

+ E

++

++
++
+
(6/ 11)
Not
done

++
++
+/(7/ 11}
Not
done

Not
done

Not
done

++
++
+/-

++
++

++
++
Not
done

(6/l l)

+ /(2/29}

(11/ 11)
+ /(6/29)

(29/29)

(29/29)

+/(2/37}

+/(10/37)

(37/37)

(37/37)

+/+/Not
done

a. +/-, +, and++ = coloration intensities; +/-equivalent to those obtained with R. solanacearum in citrate buffer at
concentration of l 06 cells/ml, + at concentration of l 07 cells/ml, and + + at concentration of l 08 cells/ml. - = not detected.

(IPProgram Report 1997-98

115

cross-reactions in tuber extracts naturall y

infected with R. so/anacearum biovar 2A.
Comparison of the detection sensitivity
of NCM-ELISA and NASH
Inoculated tuber extracts. The positi ve
samples appeared as bluish purple spots,
the coloration rang in g in intensity from li ght
pink to dark purple depending on the
concentration of R. solanacearum. In our
work, o nly the purple coloration obtained
with R. solanacearum at a concentration of
at least 2 x 10 7 cells/ml was rated positi ve.
Unacceptable coloration included a light,
pinkish purple (equal to or lighter than that
for the concentration of 10 6 cells/ ml),
similar to that caused by some saprophytic
bacteria, and brown (co lor of the tuber
extract). The detect ion thresholds obtained
in NCM-ELISA, NASH, and after plating on
MKM are shown in Table 2.
After enrichment, the lowest bacterial
concentration that could be detected in
NCM-ELISA and NASH va ri ed between 2
and 20 cells/ml among th e replications. Th e
efficiency of isolation on MKM after

enrichm ent va ried by sample becau se of

the presence of bacterial saprophytes
overgrowing R. solanacearum colonies.
Thus, the plating on MKM w as further used
to correlate with the resu Its obtained in
ELISA before the extracts were enriched.
NASH was used as th e tool to eva lu ate the
cross-reactivity of the se rolo gica l R.
solanacearum method after enrichment,
since its se nsitivity was similar to that of
NCM-ELISA.
Naturally infected tuber extracts. Log 10
v iable counts in the infected tuber and soil
extracts we re estimated from three MKM
plates at the dilution enablin g the counting,
and are presented in Table 3. The sensitivity
in NCM- ELISA after 48 h enrichment
averaged 6.45 cells/ ml, which is close to
the sensiti v ity obtained w ith inoculated
extracts (Table 2) . The same sensitivity was
observed in NASH after enrichment (Table
3). NASH was tenfold less se nsitive than
NCM-ELISA w ithout enri c hm ent, w hich
co rresponds to the results obtained w ith
inoculated extracts.

Table 2. Minimum populations of Ralstonia solanacearum detected in inoculated tubers by plating on modified
Kelman 's medium (MKM), NCM-ELISA, and NASH without enrichment (-E) and with enrichment ( +E)
after 48 h incubation of the inoculated tuber extracts in modified SMSA broth at 30(, CIP, 1998.
- E'

Cells/ml

Plating

NCM-

tuber extract

on MKM

ELISA

2X10
2x10 7
2x10 6
2x10 4
2x10 2
20
2
0.2
TE b
a.

+
+
+ /-

NASH
+

Plating

NCM-

on MKM

ELISA

+
+
+

+
+
+

+
+
+
+
+
+
+/-

NASH
+
+
+
+
+
+
+ /-

+ = detected, - = not detected, +/- = variable among the three rep lications of the whole experiment.

b. TE

116 Potato

+
+
+
+
+

+ E'

R. solanacearum-free tuber extract used for the dilutions of the inoculated tuber extract.

Table 3. Sensitivity of NCM-ELISA and NASH for the detection of Ralstonia solanacearum (Rs) in naturally-infected
tuber extracts without enrichment (- E) and with enrichment ( + E) in modified SMSA broth for 48 hat
30C, Cl P, 1998.

Original

last dilution in

population in the

which Rs could be

population detected

tuber extract

detected in:

after enrichment

NCM-ELISA
Cells/ml

Log 10

SD b

-E

+E

Estimated minimum

NASH

-E

+E

Cells/ml

]QS
lQI
lQ-4
lQ-9
]QI
lQI

8.6
6.7
6.3
2.3
8.4
10.0

]Q9
lQB

7.6

{cells/ml)

8.66x10 5
6.73 x 10 5
6.33 x 10 4
2.30xl0 9
8.40x10 7
J.QQxlQ 8
1.66 x 109
7.60xl0 8

5.938
5.827
4.801
9.360
7.924
8.003
9.220
8.880

0.015
0.046
0.021
0.055
0.021
0.005
0.015
0.023

lQ-5
10-s
1Q4

10 2

ND'
ND
]Q2
10 1

1Q9
lQI

10-1

1Ql

]Q9
]QB

]Q I

ND

l.7

a. Estimated from colony counts after plating on three plates containing modified Kelman's medium incubated for 48 hat 30C.
b. SD

standard deviation.

c. ND

bacteria detected only in the nondiluted extract.

Efficiency of NCM-ELISA for the

detection of R. solanacearum in field
samples
Of 255 tuber samples tested from the
germpl asm eva luati o n tri al, 56.5% were
found negative in N CM-ELISA after 48 h of
enri c hm ent. The sa me res ult was obtained
with NASH and isol ation o n MKM, demonstrating no la ck of sensiti vity of NCM-E LISA.
In all th e sa mpl es found positive in N CMELISA (42.4%), the presence of R.
solanacearum has been co nfirm ed eith er by
iso lation o n MKM o r by DNA-hybridization , or both, revea ling no c ro ss-reac ti v ity
of ELISA.

Detection of R. solanacearum in soil

Sensitivities of DAS-ELISA and plating on
M -SMSA were great ly in creased by the
enri chment proced ure; as few as 20 and
1 00 eel ls/g soi I cou Id be detected from

inocu lated extracts, respectively (Tabl e 4).

Th e enrichment efficiency from soi I extra cts
has bee n improved by adding potato broth
to the M-SMSA broth (1: 1) beca use the
bacteria do not grow we ll in th e M-SMSA
broth in the absence of potato extract. In
prelimin ary assays to detect R.
so lanacearum in two natu rall y infested
soils, minimum populati ons detected
avera ged 17 ce lls/g so il for San Ram o n and
141 cells/g so il for Ca rhu az soils (Table 5).
The lower sensitivity obtained with Carhuaz
so il could be due to slower growth in MSMSA broth of race-3 strains in Ca rhu az
soil as co mpared with th e rac e- 1 strains in
San Ramon so il. Another possibility is the
occurrence of different antago nistic so il
mi crofl ora, which impaired optimal growth
of R. solanacearum in the enri chment
broth.

CIP Prog rnm Report 1997-98

l l7

Table 4. Minimum populations of Ralstonia solanacearum detected by plating on modified SMSA medium (MSMSA), DAS-ELISA without enrichment (- E) and with enrichment ( + E) alter 48 h incubation of the
inoculated soil extracts in modified SMSA-potato broth at 30C. Same results were obtained with the
three replications of inoculated soils, CIP, 1999.

+ E'

- E'

Cells/g soil

Plating an
M-SMSA

+
+

10 7
10 4
102

DAS-ELISA

Plating on

DAS-ELISA

M-SMSA

20
2
0.2

+
+
+

+
+
+
+

SP
a.

+ = detected, - = not detected.

b. SE

noninoculated soil extract.

Discussion
R. solanacearum was successful ly detected
by postenrichment NCM-ELISA (tuber
sa mpl es) and DAS-ELISA (so il samp les)
eve n at low population leve ls in both
in oc ul ated and naturall y- infected extracts.
Sensitivity and non-cross- react iv ity of
postenrichment ELISA was demonstrated in
255 field tuber sa mpl es and 25 field so il
sa mpl es. Th e spec ifi c ity of ELISA using the
adso rbed antiserum was satisfacto ry, in th at
no cross-reaction was detected after
enr ichment (except w ith P ce /ebensis),
w ith out res ulting in a failure to detect all
st rain s of R. solanacearum as reported by
Robinson-Smith et al. (1995).
Postenrichment NCM-ELISA comb in es
the adva ntages of hi gh sens itivity, low -cost
(about $0.30/sa mpl e fo r suppli es), ease, and
speed (6 h after enrichment of the extracts),
and does not require exte nsive labo ratory
equipment. The sensitivity of NCM-ELISA
after a 48 h enrichment was sign ifi ca ntl y
higher than the one obtained w ith IFAS,
indirect ELISA, and PCR (Janse, 1988;
Elphinstone et al., 1996, Ca ru so et al. ,
1998; Seal, 1998). Th e lowe r sens iti vity of

118 Pototo

poste nri chment ELISA obtained by

Elphin sto ne et al. (1996) ma y be due to the
use of PBS buffer fo r sa mple extraction,
because the sa me enrichment broth was
used in the present study. Indeed, the
growth of R. solanacea rum was mu ch lower
in this buffer than in citrate (Gorris et al. ,
1997). Sensitivity of postenrichment NCMELISA co rresponded to postenrichment
nested PCR (Elphinstone et al., 1996).
However, this latte r tec hniqu e is not
suitable fo r seed testing in reso urce-poo r
countries. It is expensive ($ 0.70/sample
[Seal, 1998)) and requires extensive
laboratory fac iliti es, reso urces, and hi gh ly
skilled personnel. Moreover,
postenrichment ELISA detects only v iab le
ce ll s of R. solanacearum, w hereas PCR may
detect dead ce ll s as we ll.
An NCM -ELISA kit for th e detectio n of R.

solanacearum in tuber extracts, including

the M-SMSA enr ich ment broth , has been
developed at CIP. It is currently being
distributed to national agricultural research
in stitutes and seed programs in deve lop in g
countries for seed qua I ity testing and for
assessing susceptibility of breeding lin es to
BW. Adequate samp lin g strategies are being
investigated.

Table 5. Sensitivity of DAS-ELISA for the detection of Ralstonia solanacearum (Rs) in tenfold dilutions of naturally
infested soil extracts from two different fields without enrichment (- E) and with enrichment ( + E) in
modified SMSA broth for 48 hat 30(, CIP, 1999.
Original

Last dilution in which

Estimated minimum

population

Rs could be detected in

population detected

in soil

DAS-ELISA (cells/g)

after enrichment

Cells/g 0

Log 10

SD b

-E

+ E

Cells/g

2.5
2.6
5.7
29.0
1.0
1.4
1.6
26.0
54.0
47.0

Son Ramon soil

2.5 x 106
2.6 x 10 7
5.7 x 10 5
2 9 x 10 6
1.0x10 7
1.4X10 5
1.6 x 106
2.6 x 106
5.4X 10 5
4.7x 10 6

6.398
7.4 15
5.756
6.462
7.000
5.146
6.204
6.4 15
5.732
6.672

0.1244
0 0473
0.0106
0.0212
0.1244
0.0883
0.0770
0.0473
0.0685
0.0657

lQ6
lQI
lQS
lQ4
lQI
lQS
lQ6
lQS

6.5 18
5.4 14
6.079
6.447
4.724
6.653
6.017
5.4 62
6.653
6.4 62
6.000
5.792
6.146
6.447
6.000

0.124
0.269
0.061
0.017
0.174
0.013
0.023
0.064
0.097
0.151
0.124
0.100
0.088
0.045
0.000

lQ4
lQ4
lQ4
lQS
lQ2
104
104
lQ-4
lQS
lQS

10 4
lQS

Carhuaz soil

3.3 x l 06
2.6X10 5
1.2x10 6
2.8 x 10 6
5.3 x l 04
4.5 x 106
1.0 x 106
2.9X10 5
4.5 x 106
2.9 x 106
1.0 x 106
6.2X10 5
1.4x10 6
2.8 x10 6
1.0 x 106

10 4
l Q4
lQ4
l05
l Q4 .

330.0
26.0
120.0
28.0
530.0
450.0
100.0
29.0
45.0
29.0
100.0
62.0
140.0
28.0
100 0

a. Estimated from colony counts after plating on three platescontaining modified SMSA medium incubated for 48 h at 30C.
b. SD = standard deviation.

For th e detection of so i I pop ul ati ons of R.

so lanacea rum , the se nsiti vity obtain ed w ith
postenri chm ent DAS-ELISA was simil ar to
th at reported usin g platin g on spec ifi c
med ium (G ranada and Sequ eira, 1983;
Engl ebrecht, 1994) w ith in ocul ated as we ll

as naturall y-in fested so il s. However, th e

se ro log ica l tec hniqu e is ea sier to use
becau se it does not require sk ill ed perso nnel to d ifferenti ate betwee n saprophytic and
path ogeni c co lo ni es. The se nsiti v ity of
detect ion was consid erably in creased by
CIP Program Report 1997-98

11 9

the enrichment procedure compared with

that repo rted by Robinson-Smith et al.
(1995) for indirect ELISA.
A DAS-ELISA kit is being developed at
CIP for the detection of soil populations of
R. solanacearum. Quantification of soil
populations is important in research for the
development of control compo nents
effective in reducing soil inocu lum. Relative
semiquantification of soil inoculum potential could be achieved by diluting the soil
extract befo re enrichment and recording the
last dilution in whic h R. so/anacearum was
detected.
Acknowledgments

Th e authors thank Ida Bartolini , Vidal

La za rte, and Christian Delgado from the
CIP Virology Laboratory for their technical
assistance and advice.
References Cited

Allan, E. and A. Kelman. 1977. lmmunofluorescent stain procedures for detection

and identification of Erwinia carotovora
var. atroseptica. Phytopath. 67:13051312.
Black, R. and J.G . Elphinstone. 1998.
Developing appropriate detection
methods for deve loping countries. In :
Prior, Ph., C. Allen, and J. Elphinstone
(eds. ). Bacterial wilt disease: Molecular
and eco logical aspects. Springer Verlag,
Berlin, Germany. p. 123-12 7.
Caruso, P., P. Llop, J.L. Palomo, P. Garcia,
C. Morente, and M.M. Lopez. 1998.
Eva I uation of methods for detection of
potato seed contamination by Ralstonia
solanacearum. In: Prior, Ph., C. Allen,
and J. Elphinstone (eds. ). Bacterial w ilt
disease: molecular and ecological
aspects. Springer Verlag, Berl in , German y. p. 128-132.
Ciampi , L. , L. Sequeira, and E.R. French.
1980. Latent infection of potato tubers by
Pseudomonas so/anacea rum. Am. Potato
J. 57:377-386.

120 Potato

CIP (International Potato Center). 199 7.

Techniqu es in plant virology. L.F. Salazar
and U. Jayasinghe (eds .). Trainin g
Manual. CIP, Lima, Peru.
Clark, M .F. and A.N. Adams. 1977. Characteristics of the microplate method
enzyme-linked immunosorbent assay for
the detection of plant viruses. J. Gen.
Virol. 34:475-483.
Coo k, D. and L. Sequeira. 1991. The use of
substractive hybridization to obtain a
DNA probe specific for Pseudomonas
solanacearum race 3. Mol. Gen. Genet.
227:401-410.
Elphinston e, J.G ., J. Hennessy, J.K. Wilson,
and D .E. Stead. 1996. Sensitivity of
different methods for the detection of
Pseudomonas solanacearum (Smith) in
potato tuber extracts. EPPO/OEPP
Bulletin 26:663-678.
Engl eb re cht, M.C. 1994. Modification of a
semi-selective medium for the isolation
and quantification of Pseudomonas
solanacearum. Bacterial Wilt Newsletter
10:3-5 . Australian Center for Intern ation al Agriculture Research (AC IAR),
Canberra, Australia.
French, E.R. , L. Gutarra, P. Aley, and J.
Elphinston e. 1995. Culture media for
Pseudomonas solanacearum: isolation,
identification and maintenance.
Fitopatologia 30: 126-130.
Garris, M.T. , S. Priou , L. Gutarra, E.R.
French , M. Cambra, and M.M. Lopez,
1997. Mejora de la sensibilidad de las
tecnicas serologicas y moleculares de
deteccio n de Ralstonia solanacearum
mediante enriquecimiento selectivo.
Fitopatologia 33:32 (Abstr.).
Graham, J. and A.B. Lloyd. 1978. An
improved indicator plant method for the
detection of Pseudomonas so/anacearum
race 3 in soil. Plant Dis. Rep. 62:35-37.
Granada, G.A. and L. Sequeira. 1983. A
new selective medium for Pseudomonas
solanacearum. Plant Dis. 67:1084-1088.
Haywa rd , A.C. 1991. Biology and epidemiology of bacterial wilt caused by
Pseudomonas solanacearum. Annu. Rev.
Phytopathol. 29:65-87.

Janse, J.D . . 1988. A detection method for

Pseudomonas solanacearum in symptomless potato tubers and some data on
its sensit ivity and specificity. OEPP/EPPO
Bul letin 18:343 -351.
. 1996. Potato brown rot in western
Europe - history, present occurrence and
some remarks on possible origin,
ep id emiology and contro l strategies.
OEPP/ EPPO Bulletin 26:679 -695.
OEPP/ EPPO (Organisation Europeenne pour
la Protection des Plantes/ European Plant
Protection Organization) . 1990. Quarantine procedures no 26. Pseudom.o nas
so /anacearum. OEPP/ EPPO Bulletin
20:255-262.
Robinson-Smith, A., P. Jones, J.G.
Elphinstone, and S.M.D. Forde. 1995.
Production of antibodies to Pseudomo-

nas so lanacearum, the causative agent of

bacterial wi lt. Food and Agr icultural
Immunology 7:67-79.
Seal, S.E. and J.G. Elphinstone. 1994 .
Advances in identification and detection
of Pseudomonas solanacearum. In :
Hayward, A. C. and G. L. Hartm an (eds.) .
Bacterial wilt: the disease and its causat ive agent Pseudomonas solanacearum.
CAB Internationa l, Wallingford, UK. p.
35-58.
Seal , S.E . . 1998. molecular methods for
detection and discrimination of Ralstonia
solanacearum. In : Hayward, A . C. and G.
L. Hartman (eds .). Bacterial w ilt: the
disease and its causative agent
Pseudomonas so /ana cearum. CAB
Internationa l, Wallingford, UK. p. 103109.

CIP Progrnm Report 1997-98

l 21

An Improved Method for Fighting Bacterial

Wilt: NCM-ELISA Detection Kit and
Training Materials
Detecting bacterial wilt disease has become easier for our partners and clients
thanks to development of the NCM-ELISA Kit, which can be used to monitor
Ralstonia solanacearum-the agent that causes bacterial wilt or "b rown rot" in
potato tubers. Detecting the disease-causing agent is essential for seed production
systems and varietal evaluation of resistance to bacterial wilt.
Post-enrichment NCM-ELISA (enzyme-linked immunosorbent assay on nitrocellulose membrane using enriched samples) is as sensitive as the Double-Antibody
Sandwich (DAS-ELISA), but is much easier and quicker. Samples and reagents are
supported by nitrocellulose membrane instead of micrometer plates and they can be
stored for several weeks before testing or sending to outside labs. After NCM-ELISA,
the membrane is stable and can be stored for extended periods. R. so/anacearum in
latently infected tubers (those without any visible symptoms) is detected in ELISA
after incubating tuber extracts in a semi-selective broth prior to testing. This enrichment procedure increases the sensitivity of the assay by nearly one million. As few
as 10 bacteria per milliliter of extract can be detected with this method.
The kit can also be used with roots and stems of potato and other plants for
research on disease epidemio logy. The package is available complete with an
instruction manual and video in Chinese, English, and Spanish. Learning more
about the kit or training manual is as easy as contacting Dr. Sylvie Priou
(s.priou@cgiar.org) in CIP's Crop Protection Department or accessing OP-publications on our web page at: www.cipotato.org.

122

Potato

Survey of the Durability of Extreme Resistance to PVY

Derived from Solanum tuberosum subsp. andigena
E. Mihovilovich, L.F. Salazar, F. Saguma, and M.W. Bonierbale 1

Potato virus Y (PVY), is one of the most

important viruses attacking potato (5olanum
tuberosum subsp. tuberosum). It is distributed worldwide and causes significant crop
losses. The most effective strategy for
controlling PVY is the use of durable
resistance genes. Durable resistance is
unlikely to be overcome by diverse pathogen strains or under variable production
conditions.
The most complete types of resistance to
PVY have been grouped into two main
classes: hypersensitivity (expressed as a
necrotic response to infection), which is
conferred by dominant N genes; and
extreme resistance (with very little or no
virus accumulation in the plant) conferred
by dominant R genes. PVY can usually be
recovered following inoculation from plants
carrying N genes, but not from those
carrying R genes (Barker 1996). Ry genes
are considered durable, since they have
proven effective against a range of PVY
strains known at present. Ny genes are not
considered durable because they are strainspecific. 5. tuberosum subsp. andigena and
5. stoloniferum are sources of Ry genes.
Most breeding for resistance to PVY is
based on the introduction into 5. tuberosum
subsp. tuberosum cu ltivars of Ry genes
derived from these two species. The
extreme resistance of European varieties
comes from 5. stoloniferum (Ry,,). Conversely, resistance in some varieties and
parental clones developed by CIP comes
mainly from 5. tuberosum subsp. andigena

(Ry,d)
New strains, multiple virus infections,
and interactions with subviral pathogens

are the main factors leading to the breakdown of virus resistance. During the past
10 yr, a new strain of PVY, tuber necrotic
ringspot disease strain (PVYn"), has become
established in Europe. It causes severe
necrotic symptoms in tubers of PVYsusceptible varieties. It has been reported
that the independently replicating potato
spindle tuber viroid (PSTVd) has a synergistic effect on PVY multiplication (Querci et
al., 1997). They also showed that PSTVd
has the ability to decrease resistance to
potato leafroll virus (PLRV), and to be
encapsidated by this virus and transmitted
by aphids. These factors led us to test the
durability of the Ry 1g gene with respect to
its ability to provide protection against
pyynn, and to provide extreme resistance
against combined infection by PVY and
PSTVd.
<l(

Materials and Methods

Four Ry ac1g-bearing clones, all developed by
CIP, were used in this study. They included
cv. Costanera and Tacna, which are
adapted to the arid Peruvian coast and
carry extreme resistance to PVY, and the
two parental clones YY.5 and TY.4. Both
Peruvian cultivars carry the Ryadg gene in a
simplex (Ryryryry) condition. The parental
clone YY.5 is duplex (RyRyryry) and TY.4 is
triplex (RyRyRyry). These and other multiplex progenitors (possessing more than one
copy of the dominant resistance allele) have
the ability to transmit resistance to their
progeny with a higher frequency than
would be the case with simplex clones
(Mendoza et al., 1996). A PVY extreme
resistance European variety, Pirola, carrying
the Ry, 10 gene, and three PVY susceptible
varieties-Bintje (Europe), Desiree (Europe),

1 CIP, Lima, Peru.

CIP Program Reporl 1997-98

123

and Atlantic (No rth America)-were used

as controls. Plants of all clones we re grown
from vi rus-tested in vitro plantl ets by first
transplanting them to jiffy pots an d then to
individual plastic pots containing a 1 :1 :1
mixture of soil, sand, and peat. Three PVY
strains, the ordinary (PVY 0 ), necrotic (PVYN),
and new tuber necrotic ringspot (PVY"")
strain s we re used. All were from CIP stock
isolates, and we re provided in plants of
Nicotiana occidentalis. PSTVd (severe
strain ) was provided in tomato plants,
w here it is maintained for resea rc h purposes. Experiments were carried out in the
greenhouse at CIP's La Molina Experiment
Station from September to December 1997
(primary infection trial ), and from March to
Jun e 1998 (secondary infection trial).
Inocul at ion in the primary infection trial
involved mechanical , aphid, and graft
transmission using five plants of eac h clone
for eac h inoculation m ethod and strain.
Mechanical inoculation was performed by
rubbing freshly extracted sa p from infected
N. occidentalis plants onto carborundumdusted potato plants. For graft inoculation,
scions (s hoot apices) from infected N.
occidental is plants were cleft-grafted onto
the stems of test plants . Aphid inoculation
was ca rried out by setting 15 infective
aptero us aphids, prev iousl y fed o n PVYinfec ted plants of cv. Atlantic, onto eac h
test piant.
For the PVY-PSTVd combined infection
trial, 45 plants of each clone were first
mechanically inoculated at the seedling
stage using a suspension of sap from
PSTVd-infected tomato plants. Infection
was ver ified b y nucleic acid spot hybridization (NAS H ). Al I of the va ri et ies described
above, but not the multiplex parental
c lon es, were used in this cross-infectivity
test. Two weeks after PSTVd ino culation ,
five pre-infected plants (PSTVd ) of each
genoty pe were mechan ica I ly inoculated
with each of the three PVY strains. The 30
remaining PSTVd-infected pl ants of each
c lon e were graft inoculated (15 plants) or
ap hid inoculated (15 plants) with PVY 4 wk

124

Potato

after PSTVd inoculation , using 5 plants per

strain and inoculation method.
Tuber s recovered from the primary
infection trial of both experiments were
used to obtain plants for the evaluation of
secondary infection . Visual foliar evaluations and enzy111e-linked immunosorbent
assay (E LIS A) tests for PVY were perform ed
in both exper imen ts , at 30 and 45 days after
inocul at io n during the pri111ary infect ion
cycle, and at th e sa me interva ls after tuber
plantin g during the secondary in fect ion
cycle. Tub ers obtained from the prim ary
and secondary infection cycles were
evaluated v isually at harvest and after 1 1110
of storage at 24C.
Results
Reaction to PVY inoculation
Plants of the susceptible controls showed
mosaic sympto111s 7 d after inoculation with
all strains (PVY 0 , PVY N and PVY" 11 ). During
the evaluation period of 45 cl , Bintje and
Atlantic showed severe mosaic and
crinkled sy111ptoms, and a strong hypers ensitive react ion, i.e ., severe systemic necrosis, and some dea d plants. On the other
hand , Desiree developed mild mosaic and
strong hype rse nsiti v ity.

All plants of the susceptible varieties

reacted po sitivel y to ELISA, except those
that showed seve re systemic necrosis.
Plants infected with PVY 0 showed much
more severe symptoms than those in fec ted
with the other strains, which suggests that
the PVY 0 ino culum used in this study
represe nts a highly v irulent isolate. More
severe symptoms deve loped after mechanical and graft inoculation than aphid
inocul at ion. Var ieties and parental c lon es
carrying Ryodg did not show foliar symptoms
against any of th e strains, and reacted
negatively in ELISA. Conversely, Pirola
(understood to carry Ry,,) develop ed a 111ild
necrosis in some leaves and stems, but
v irus was not detected in these plants by

ELISA .

Table 1 . Reactions of selected potato varieties and parental clones against PVY 0 , PVYN, and PVY'".

Varieties

PVY

and parental

resistance

clones

genotype

Costonera
Tacno
YY.5
lY.4
Pirolo
Desiree
Atlantic
Bintje

Ryr4i
Ryr4i
Ryal!
Ryal!
Ry""
Ny""
IY"*1.'Ys1o,ny""
'Y"*1.%o, nyl.ii

Foliage reactions

Tuber
reactions

PVY
I

(-)
(-)

(-)
(-)
H:MN; (-)
MM; H:SN; (+ )
SM,C; H:SN; (+)
SM,C; H:SN; (+)
I

PVYN
(-)
(-)
(-)
(-)

H:MN; (-)
MM; H:SN; (+ )
SM,C; H:SN; (+)
SM,C; H:SN; (+)

PVY""
(-)
(-)
(-)
(-)
H:MN; (-)
MM; H:SN; (+)
M,C; (+)
M,C; (+)

pyynn

Necrotic crocking b

a. Reaction to PVY by ELISA is shown in parenthesis. Key reactions: MM = mild mosaic, M= mosaic, SM = severe mosaic;
C= Crinkle; H:MN = hypersensitivity: mild necrosis, H:SN = hypersensitivity: severe necrosis.
b. Also observed on plants infected with PVYN.

In th e secondary infection tri al, the

susceptibl e co ntrol s reacted with mosaics
aga in st all strain s. PVY infect ion was
co nfirmed by ELISA in all cases. The PVYres ista nt va ri ety Pirola (Ry51 ) , howeve r,
conti nu ed to show a mild hyperse nsitivity
react ion (foliar and stem mild necrosis) in
some plants with all strains. Virus was not
detected in th ese pl ants by ELISA. Th e
resistant va ri eti es and parental clones w ith
Ry 0 c1gdid not develop symptoms w ith any
stra in, and continued to test negat ive in
ELISA. Tab le 1 summ ari zes the reactions
observed.
Tubers recove red from primary infection
did not show any symptom of virus infection, rega rdl ess of vari ety o r strain . Co nversely, tubers from the seco nd ary infect io n
trial of the susceptibl e variety Atlantic
(inocu lated w ith PVYN and PVY 1111 ) showed
severe necrotic cracking les ion s typical of
virus infection. Thi s symptom was not
observed o n tubers of any other va ri ety or
parental clone harvested from th e seco ndary infection cycl e. Rin gspot necros is,
reportedly indu ced by th e PVY 1111 strain, was
not found on tubers of any variety or
parental clone, either at harvest or after
1 mo of sto rage at 24C.

Reaction to combined PVY and PSTVd

inoculation
When inocu lated w ith PSTVd and PVY
strain s, PVY-susceptibl e va rieti es deve loped
much more severe mosa ic, crinkled, and
necrot ic symptoms during both infect ion
cycles than those observed under PVY
infection alone. Again, more severe symptom s we re obse rved o n plants infected w ith
PVY 0 than the oth er two strain s. PVY
infect ion was confi rm ed in th ese va ri eties
by ELISA. In contrast, the doubl y inoculated Ry 0 c1 ,-bearing va ri eties, Costanera and
Tacn a, di~ not show any symptom assoc iated w ith PVY, and reacted negati ve ly in
ELISA. The Ry,10 -bea rin g variety Piro la
showed the same hypersensitive reaction
(mild nec ros is of some leaves and stems) as
that shown with PVY infection alone and
continued to test nega ti ve for PVY in ELISA.
Severe stunting was observed in pl ants of
thi s va ri ety as well as in the susceptible
controls Atlantic and Bintje. Tabl e 2
summarizes the reaction s of var ieti es
against comb ined in ocul ation of PVY and
PSTVd .

No sympto ms of virus infection were

observed on tubers of either PVY-resista nt
or PVY-suscept ibl e varieties recovered from

CIPProgramReporl 1997-98

125

Table 2. Reactions of selected potato varieties against combined inoculation with PSTVd and PVY.

Varieties

Tuber reactions

Foliage reactions 0

PVY resistance

pyynn

genotype

PVY

PVYN

Costanera
Tacna
Piro la

Ryodg
Ryodg
Rv,,,

(-)
- ; (-)
St; H:MN; (-)

(-)
(-)
St; H:MN; (-)

(-)
(-)
St; H:SN; (-)

Desiree
Atlantic
Bintje

Nytub
ryodg, ryslo, nytub
rv,dg, rv,,. nytub

MM; H:SN; (+)

SM,C; H:SN; (+)
SM,C; H:S N, St; (+)

MM; H:S N; (+)

SM,C; H:SN, St; (+)
SM,C; HSN, St; (+ )

MM; H:SN; (+)

SM,C, St; (+)
SM,C; St (+)

-;

Spindle shape
Spindle shape
Spindle sha pe,
and severe
size reduction
Necrotic crocking b

a. Reaction to PVY by ELISA is shown in parenthesis. Key reactions: MM = mild mosaic, M = mosaic,
SM = severe mosaic; C = crinkle; St = stunting; H:MN = hypersensitivity: mild necrosis,
H:SN = hypersensitivity: severe necrosis.
b. Also observed on plants infected with PVYH

primary or seco ndary infection, except for

the susceptible va ri ety Atlantic. Tubers of
Atlantic recove red fro m plants infected w ith
PVYN and PVY"" aga in showed necrotic
crack in g. In add ition , tubers from Piro la,
Costa nera , and Tacna showed severe size
red uct ion as we ll as defo rm atio n (sp indl e
tubers) due to v iroid in fectio n. No oth er
var iety showed this effect. Ringspot
necrosis-the reported sympto m of PVY""was not fo und on any of the tubers.

hypersensitive respo nse towa rd symptom

deve lopme nt. Piro la was not expected to
deve lop symptoms to known strains since it
carri es th e Ry 510 gene for extreme res ista nce.
Although v iru s was not detected in Piro la
plants, nec rotic spots deve loped on its
leaves and ste ms in reaction to all of the
strains tested in thi s trial. Barker (1996)
descr ibed simil ar necroti c reacti o ns in this
cu lti va r and suggested that they may be
ca used by difference between iso lates and
env ironm ental cond iti ons.

Discussion
Reaction against PVY 0 , PVYN, and PVY""
Reacti o ns to PVY"" did not differ from
reactio ns to the other strain s. Furthermore,
it was not possible to detect PVY"", PVY 0 , or
PVYN in inoculated pl ants of Ry ac1g -co ntaining va ri et ies and parental c lo nes. Some
unexpected reaction s were obse rved in
plants not ca rry in g thi s gene. Des iree,
understood to ca rry a gene for hypersensitivity (Ny,"6 ), deve loped mosaic symptom s
w ith all strain s in additi o n to the ex pected
hype rsensiti ve rea ction. Ny genes are
kn own to be temperature-se nsiti ve. Temperatures above 22( during the primary
and seco ndary infecti o n cyc les may have
accou nted for Desiree's ove rcom in g the

126

Potato

Temperatures, above 27( we re encountered during primary infection beca use of El

Nino. Th at, and flu ctuatin g temperatu res
betwee n 20 and 25( during th e seco ndary infect io n trial , m ay account for th e
reactio n in Pirola as well as for th e un com mo n hypersensitive reaction observed in
susceptibl e control s Bintje and Atlantic.
Observation by Le Rom ance r and Kerlan
(1992) of hyperse nsiti vity against the PVY 00
strain in an Ryadg-carryin g c ulti va r, V-2 , and
th e Ry510 -ca rrying va ri ety Corine led th em to
suggest that this stra in behaved as a
res istance-brea kin g strain. Howeve r, they
did not find the ringspot necrosi s symptom
th at character izes the pathogeni c effect of

pyynn on the tubers of the cultivars they

studied. As mentioned by Barker (1996),
that can be caused by the presence of other
potyviruses in the isolate.

this experiment might be tolerant of this

strain, since even though they developed
severe foliar symptoms, tubers did not
develop necrosis.

As the necrotic reaction observed is a

nonspecific response of plants to virus
infection, and not a damage effect produced in the plant by the virus itself, this
reaction cannot be taken as a mechanism of
resistance breaking. The necrotic reaction
seems to be induced by the replication of
virus in a few cells before hosts that possess
resistance genes trigger the resistance
response. In the case where susceptible
cultivars develop this reaction, as in our
experiment, the virus succeeds in reaching
high levels of replication and the pathogenic effect finally develops . No symptoms
associated with PVY multiplication develtherefore it
oped in plants containing Ry ad;
g
can be asserted that resistance to PVY
replication of varieties and parental clones
carrying this gene is extremely strong and
stable.

No case has yet been reported in which

potato plants show tuber reactions against
pyynn without developing foliar symptoms.
Hence, there is no evidence to support a
resistance-breaking effect. Since no foliar or
tuber reaction to pyynn was found in the
Ry ar1g-bearing cultivars tested, we conclude
that Ryadg-mediated resistance is also
effective against this strain.

Tubers recovered from plants of the

resistant and susceptible cultivars tested did
not develop the tuber necrotic ringspot
symptoms reportedly induced by the PVYo 11
strain. Kus (1992), after testing 40 European and North American cultivars against
this strain, concluded that tuber reaction
induced by this strain is character-specific
for each cultivar. He observed that within
susceptible cultivars, the occurrence of
ringspot necrosis differed in its incidence,
ti me of appearance, and severity. For
instance, Igor, a common European variety,
reached 100% tuber symptoms in the first
cycle of infection, while Russet Burbank,
among other susceptible varieties, never
developed symptoms. Conversely, Desiree,
which was also tested in our experiment,
developed interveinal mosaics, and only
8% of secondarily infected plants exhibited
ringspot necrosis on tubers. Although we
also observed mosaics on plants of Desiree,
we did not find this tuber reaction even on
secondarily infected tubers. Our results
suggest that susceptible cultivars used in

Reaction against combined inoculation

of PVY and PSTVd
Combined infection by PVY and PSTVd
resulted in much more severe symptoms of
PVY in susceptible varieties than those
observed under PVY infection alone. This
effect and severe stunting were also
described by Singh and Somerville (1987).
Production of smaller and deformed tubers
with a characteristic spindle shape were
attributed to the viroid infection, since this
symptom is characteristic of PSTVd alone.

There were no differences between the

resistance response of either Ry ac1g- or Ry sto bearing cultivars under double infection
and those observed in the same cultivars
infected with PVY alone. No PVY symptoms developed and PVY was not detected
by ELISA in the presence of these genes.
Hence, there was no evidence to suggest
that combined infection by PVY and PSTVd
can break Ry d -mediated extreme resis' g
tance.
Conclusions

The durability of the Ry ad gene, with

respect to a new strain o?PVY (PVY""), and
the presence of PSTVd was demonstrated in
this study. The qualitative effect of this
gene was confirmed, since one, two, or
three dominant alleles provide the same
level of protection. Additionally, it was
shown that the Ryadg-mediated extreme
resistance gene was more effective than

CIP Program Report 1997-98

127

Ry510 Both resistance genes appear to be

stab le even in presence of PSTVd, whic h is
known to contribute to very high leve ls of
PVY conce ntration through its synergistic
effect. This desirable feature of Ryadg may
also contribute to stability of resistance in
the event that PVY might interact with
subviral pathogens in nature, whereby
vector relations or infectivity may be
altered .
Ry genes are frequently used in breeding, and it is expected that the y soon wi ll
be isolated and used in plant-derived
transgenic re sistance. The continued
testing of the durability of this apparent
"frozen ve rtical resistan ce" (Robinson,
1976), is important if resistant varieties
developed from these programs are expected to replace conventional susceptible
va ri eties. That is particularly true if deployment on a lon g-term, globa l scale is
envisio ned .
References Cited

Barker, H. 1996. Inheritance of resistance to

potato virus Y and A in progeny obtained
from potato cu ltivars co ntainin g ge ne Ry:
Evidence for a new gene for extreme
resistance to PVA. Theor. Appl. Genet.
93 :710-716.
Kus, M. 1992. Potato tuber necrotic
ringspot disease. Var ieta l differences in

128

Potato

appearance of ri ngspot necrotic symptoms on tubers. In : Proceedings of the

E/\PR Meet in g Viro logy Section. Held 29
Jun e-3 July 1992, Vitoria-Gasteiz , Spa in .
p. 81-83.
Le Romancer, M. and C. Kerlan. 1992.
Potato tuber necrotic ringspot disease: A
genetical app roach of the phenomenon
and studi es about hypersensitive or
extre me susceptibl e behaviour of several
cultivars . In : Proceedings of the EAPR
Meeting Viro logy Section. He ld 29 June3 Jul y 1992, Vitoria-Gasteiz , Spain. p.
9 1-95 .
Mendoza, H . A., E.J . Mihov ilovich, and F.
Saguma. 1996. Ident ificatio n of Triplex
(YYYy) Potato Virus Y (PVY) immun e
progenitors derived from Solanum
tuberosum ssp. andigena. Am. Potato J.
73: 13-19.
Querci M , R.A. Owens, I. Bartolini , V.
Lazarte, and L.F. Sa lazar, 1997. Evidence
for heterologous encaps id at ion of potato
spind le tuber viroid in potato leafroll
v irus. J. Gen. Virol. 78:1207- 1211.
Robinson , R.A. 1976. Plant pathosystems.
Springer-Verlag, Berlin, Germany. 184 p.
Singh, R.P. and T.H . Somerville. 1987. New
disease symptoms observed on fieldgrown potato plants with potato spindle
tuber viroid and potato virus Y in fections.
Potato Res. 30:127-1 32.

The Leafminer Fly in Potato: Plant Reaction and

Natural Enemies as Natural Mortality Factors
F. Cisneros and N. Mujica1

The pea leafm in er fl y (LMF) species,

Liriomyza huidobrensis (Blanchard) is
highly polyphagous, attacking many
vegetables and ornamental plants (Spencer,
1990) . LMF ha s become a key pest in all
the countries w here it has been introdu ced
and is resistant to many in secticides. In
many places, the pest has shown significant
levels of resista nce to most carbamate,
organophosphate, and pyrethroid in sect icides commonly used to kill larvae and
ad ult fli es (Parrella et al., 1984, Macdonald,
1991 ). LMF is capable of co mpletely
destroying potato crops in the absence of
adequate co ntrol measures.
Originating in the neotropics, it was
restricted to Central and South America
until the 1980s. Since then, L. huidobrensis
has rapidl y sp rea d to other areas, and is
reported in severa l African, European, and
Asian countri es. Specifically, L.
huidobrensis is reported to be a serious
potato pest in Argentina, Brazil, Chile,
Indonesia, Israe l, Kenya, Malaysia, Mexico,
and Peru, and in Central America and
North Africa (llE, 1996).
LMF occurred at low populations on the
central coast of Peru and had no negative
eco nomi c effect on the potato crop until the
1950s (Wille, 1952). In the following
decade, potato fields were sprayed with
organic in sectic ides (DDT, parathion, BHC)
against the tomato pinworm (Tuta absoluta)
(Campos, 1976). Although in most cases,
these sprays were unnecessary, as there was
no evidence of pinworm infestations
reachin g eco nomic thres holds in potato.
During that decade, howeve r, LMF popula1 CIP, Lima, Peru .

tions had begun to in crease (Herrera, 1963;

Campos, 1978). At this point, the in secticides used were primarily to fight LMF
infestat ions rather than pinworms. Since the
1980s, 10-13 sprays/season have been used
against LMF in potato. The cost of spray ing
has become the most expe nsi ve component
of potato production (Ewell et al., 1990;
Mujica and Cisneros, 1997). Most research
over the last 20 yr has been oriented toward
finding new compounds to repl ace in secticides that have lost effective ness in controllin g LMF larvae and ad ult flies.
The resurgence of LMF as a pest has
been attr ibuted to the destruction of its
natu ra l enemies by ins ecticides (Cisneros,
1986; Yabar, 1988; Ochoa and Carballo,
1993). For farmers, the presence of ad ult
fli es ear ly in the cropping season eli cits
early sprays that create a treadmill effect.
Secondary effects of such frequ ent spraying
are the occurrence of heavy infestati ons of
th e white mite, Polyphagotarsonemus latus
(Banks), and the budmid ge, Prodiplosis
longifila (Gag ne).
This paper reports results of the study of
the reaction of the potato plant to LMF
infestat ion and th e occurrence and role of
natural enemies of LMF with a goal of using
thi s knowledge to determine how to break
the cyc le of high LMF incidence.

Material and Methods

Reaction of the potato plant to
LMF infestation
In the study of LMF potato pl ant interaction s, five aspects were investigated : (1)
adult feeding and oviposition behavior, (2)
pattern of LMF damage in the potato plant,

CIP ProgramReportl997-98

129

(3) hypersensitivity reaction of potato leaves

(egg extrusion), (4) LMF population density
and potato plant phenology, and (5)
seasonal variation of LMF population.

Adult feeding and oviposition behavior.

Adult feeding and oviposition was studied
in the greenhouse using caged (potted)
plants of cv. Canchan. Plastic tubing
(cellulose membrane, 49 mm av diam) was
used to isolate small groups of flies on
leaflets. Newly emerged male and female
flies were introduced into the tubing. Daily
observations were made on the occurrence
of oviposition and feeding punctures to
describe adult behavior.
Pattern of LMF damage in the potato
plant. Evaluation of larv al damage was
carried out with cv. Revoluci6n in an
experimental field in the Canete Valley from
August to October 1995. No insecticide
was applied. Leaf samples were taken from
the bottom, middle, and top parts of the
potato plants. Occurrence of mining in
potato foliage by LMF larvae was recorded
by observing fresh and dry tunnels in the
sampled leaves.
Hypersensitivity reaction of the potato
plant. The occurrence of egg extrusion was
recorded in greenhouse trials in La Molina
in 1996 using Canchan. Two pairs of male
and female flies per leaflet were isolated in
plastic tubing for oviposition. The flies were
removed after 24 hr. Leaflets were th<:n
observed daily in the laboratory. The
number of eggs extruded was recorded as
well as the number of alive and dead
neonate larvae.
LMF population density and potato
plant phenology. Adult and larval populations were evaluated in a potato field (cv.
Canchan) using no insecticide applications
(Canete Valley, June to November 1996).
The adult population was determined by
counting the number of flies/plant on 25
plants. Damage was determined by recording the percentage of infested foliage area

130

Pototo

on each plant. Larval infestation in leaves

was determined by counting the number of
larvae per leaflet on 25 infested leaflets.
Samples were taken weekly throughout the
growing season.

Seasonal variation of LMF population.

Popu la ti on studies were conducted in three
sites (Canete, Callao, and Carabayllo on the
central coast of Peru ) from January to
December over 3 yr, 1996-1998, to look at
the seasonal variation. Potato fields and
other host crops (celery, chinese onion,
lettuce, tomato, bean, pea, spinach, and
cucumber) were evaluated. Adult populations were determined by using a yellow
sticky trap (20 x 20 cm) in each field.
Infested foliage was taken to the laboratory
for further observations (emergence of flies
and parasitoids). Samples were taken
weekly over the 3 yr.
Occurrence and role of natural
enemies of LMF
Parasitoids. Yellow sticky traps were
used to determine the occurrence of adult
parasitoids. The percentage of parasitism
was evaluated weekly by collecting leaves
infested with LMF larvae. The leaves were
stored in petri dishes in the laboratory and
kept for 30 d. Adult flies that developed
from healthy larvae and the number of
parasitoids (ectoparasitoids and
endoparasitoids) recovered from fly larvae
and puparia were recorded.
Predators. Soil-inhabiting predators. Soil
inhabiting predators were captured in pitfall
traps (1-L plastic cups with 500 cc formaldehyde (5% )) placed in potato and other
host crop fields. Traps were collected
weekly and the number of predators
recorded.

Foliage-inhabiting predators. For recording the presence and number of predatory

flies (Condylostylus similis and Orapetis
sp. ), 25 plants/field were sampled weekly.

Results and Discussion

Reaction of the potato plant to
LMF infestation
Numerous field observations (u npublished data) co nfirm that short-cycl e (ea rl y)
potato cu ltivars suffer more severe damage
by LMF than long-cycle (late) cultivars. Th e
most likely exp lanation is that late culti vars
produ ce new leaves th at compen sate for
damaged leaves . Glandul ar trichom es have
been mention ed as a potential protect ion
mechanism aga in st LMF (Mujica et al. ,
1993), but co mm ercial culti va rs w ith hi gh
densities of such trichomes are not ava ilab le. Prev ious studies (Go nzales, 1994) and
our own obse rvations, howeve r, suggest
that a hyperse nsitivity mechanism could be
invol ved.
Adult feeding and oviposition behavior.
The fema le produces two types of les ions in
the potato leaf, a feedin g les ion and an
oviposition les ion. As soo n as th ey emerge,
adult females make superfic ial feedin g
les ion s in th e upper surface of th e lea f w ith
their ovipositor. Lesion s are round (1 mm
diam). Both femal es and males feed o n th e
exudate of the lesion. M at in g occurs 6 to 24
h after flies eme rge from puparia. O viposi tion les ion s are about th e sa me size as
feeding les io ns, but are deeper because th e
femal e lays the egg in sid e the leaf ti ss ue .
Eggs new ly laid on th e lowe r surface of th e
leaf are not readily visible to the naked eye.
Pattern of LMF damage in the potato
plant. New ly hatched larvae burrow insid e
the leaf surface, makin g a se rpentin e min e.
Mine di ameter increases as the larva grows.
A large proportion of grown larvae remain s
close to the midrib. Leaf ti ssue affected by
larval minin g becom es nec rotic and
brownish co lored. Th e length of time before
the damaged area becomes necroti c
depend s on th e cultivar, physiological
condition, and age of the plant. As nec roti c
areas coa lesce in hi ghl y-infested leaves, the
whol e leaf dri es out and di es. Highlyinfested potato fields appear burned. Fly

infes tat ions may start at plant emergence

and co ntinue to pl ant se nescen ce.
Larva l damage in a gro w in g plant
follows a fai rl y we ll establish ed patte rn .
Le af les ions, for both fee ding and ov ipos ition, are scattered over the plant. Lower
lea ves are the first to show dam age. Middle
and upper leaves are progressively damaged as the plant grows and infestation
continues. Dam age in th e upp er leaves
usu ally occ urs wh en the canopy stops
growing. Nec rosis of the infested leaves
follows the sa me patte rn , until th e w hole
plant dries out.
Hypersensitivity reaction of potato
plant. Females oviposit in potato foli age as
soon as th e plant emerges, but larval mines
are sca rce. At thi s stage of plant development, the lac k of correlation betwee n the
number of adult fli es and the numb er of
larval min es on leaves is evident. In fully
grown plants-up to preflowering- there
are more mined leaves in the lowe r part of
the pl ant than in th e upper part.

During growth pha ses, flowering to

sen esce nce, larval min es develop in leaves
anywhere on the pl ant. This is th e peri od
that infestation flares up in potato fields.
Differences in larva l in festation depend on
wh ere on the plant eggs were deposited.
Wh en an egg is laid in a mature leaf and
hatches, almost all the neo nate larvae
successfully start burrowing into th e
pali sade tissue of th e leaf surface. Th at
expl ains why larval infestations start in the
mature, lower leaves of a growin g plant,
and cove r th e whole plant after all foliage is
mature.
When an egg is laid in a growin g lea f, a
hyp erse nsitivity re action makes many
neo nate larvae fail to burrow into th e lea f
tissu e. Ti ss ue surrounding the encrusted egg
starts hypertroph ic grow th (abnorm al
multipli cat ion) of ce ll s that pu sh the egg to
th e leaf surface. Extrud ed eggs are exposed
to advers e weather co nditions th at res ult in
deh yd rat ion and ex pos ure to predato rs.

CIP Program Repo rt 1997-98

131

Neonate larvae of surviving eggs exper ience the same fate, thereby reducing their
chances to burrow into the leaf tissue.
In the laboratory and greenhouse, more
than 90% of eggs were extrud ed; 60% of
those died from dehydrat ion (Sotomayo r
and Cisneros, 1996) . That explains why
potato plants with growing foliage show
sma ll numbers of mined leaves despite high
adult fly populations. Egg extrusion is
hi gher in vigorously gro wi ng potato plants
that are we ll fertilized and irrigated.
LMF population density and potato
plant phenology. It is important to differentiate between the popul at ion densities of
adult and larval LMF. Immigrant ad ults
initiate infestations in small potato plants.

Despite relativel y large numbers of adult

flies, larva l populations initiall y are low
because of egg extrusion. As more leaves
mature (begi nning at the lower part of the
plant), more larvae are able to develop and
the larval population in creases grad uall y.
When the plant stops grow in g, abo ut 1 wk
before flowering, larval population reac hes
its peak and remains at a plateau until
initiation of plant senescence. That is the
time whe n larva l outbreaks may be extremely injurious and the whole plant dries
out (Fi gure 1).
Differentiating adult fly from larva l
populations helps to monitor LMF populations for management purposes . Monitoring
adult populations usually leads to ea rly
app lications of insecticides that could be
avo id ed considering the low larva l populations during vegetative growth of the plant.
Ear ly app lications adversely affect a
comp rehensi ve pest management strategy.
Seasonal variation of LMF population.
LMF host plants are present in the Canete
Va ll ey of Peru yea r-round because of its
favorable climate. In norm al years (those
not affected by El Nino), monthl y average
summer temperature is 23.6( (18.4( 28 .7C); av winter temp erature is 15.6(
(13.2 ( - l 8.0C). Still, there are clea r
132

Potato

seasonal trends in LMF population levels as

data from 1994-96 observations suggest.
Hi ghest populations are reco rded in June
(wi nter) and September (s pring). Lowest
populations occu r in the warme r months
(December to April) (F igure 2). An increase
in larval parasitism see ms to be the most
important facto r in reducing the LMF
population during the warm month s of the
year. In November, parasit ism rates as high
as 92% are commo n.
Another factor that may affect population
fluctuation is a reduction in the rate of
oviposition during the warmer months
(Li za rr aga, 1989). However, popul ation s
remain hi gh und er intensive insecticid e
treatments, co nfirming the stronger effect
the presence of natural enemies has on
reduction of the fly population.
Atypical weather in 1997 and 1998
caused by El Nino resulted in drastic
changes in LMF population densitie s (F igure
2). Temperatures hi gher than norm al were
record ed up to Jun e 1998, followed by a
return to normal beginning in July. Minimum daily temperatures took longer to
rea ch normal leve ls. That explains the low
LMF population during the period of
wa rmer temperatures. As temperature
decreased, the first LMF peak was recorded
in September. But in two other sites on th e
central coast (Sa n Agustin and Carabayllo)
w here vegetable crops are produced under
heavy use of insecticides, LMF populati ons
remain ed fair ly high despite higher temperatures. Population density peaks were
recorded in May and June and in September and October. Records in 1997-98
suggest that factors such as the occurrence
of other LMF plant hosts may also play a
role in the dynamics of the LMF population
(Fi gure 3).
Occurrence and role of natural enemies
of the LMF
Th e occurrence of low population s of
LMF up to the 1950s cannot be explained
by unfavorable climate or lack of suitable
hosts, as it was in the case of El Nino in
1997-98. Th e temperature range on Peru's

Adult flies or plant larvae/leaflet

30
25
20

Damage%

100

adult/plant

--+-- damage

80

---.-

larvae/leaflet

70

90

60

15

50

flowering

40
10

30
20

10
0

0
29

37

44

51

58

79

72

65

86

93

100

107

114

Days after planting

Figure 1. Variation of leafminer fly adult and larvae populations and damage in one cropping season, Canete,

Peru , 1996.

Adults/trap/month

Temperature QC

6000

25

15
10

1000

Flies-96

PC-1 996

1 - - --

Jan

Feb

Flies-97

tsl

PC-1997

Mar

- -- --

Apr

May

---j

Jun

Jul

Aug

Sep

Oct

Nov

Dec

Figure 2. Population densities of th e leafminer fl y in potato fiel ds, and monthly av min daily temperature for

three consecutive yr (1996-97-98) in the Canete Valley, Peru.

ce ntra l coast is favo rab le fo r th e deve lopment of LM F as we ll as its host plants . Th e
most likely sce nari o is that until the 1950 s
natural enemi es kep t LM F pop ul atio ns
be low eco nomi c thres ho ld s. Fo r seve ral

years, LM F pa ras ites and predato rs have

been recorded in potato fie ld s and, to a
limi ted extent, in fie lds of oth er host crops
on the centra l coas t (Redo lfi et al., 1985;

CIPProgiam Repo11l 997-98

l 33

Larvae/1 00 leaflets
70,ir-~~~~~~~~~~~~~~

600

&'.i

Parasitoids

Fly larvae

Celery

Chinese Lettuce

Tomato

Bean

Potato

on ion

Figure 3. Seasonal av leafminer fly larval

infestation and parasitoids in several
vegetable crops , including potato,
Carabayllo, Li ma, Peru , 1998.
Sanch ez and Redolfi , 1 988: Fonseca and
Sanchez , 1 996) .
Parasitoids. Identification of species.
There is a ri ch co111p lex of paras ito id
speci es that attack LMF larvae o n Peru 's
ce ntral coast (Tabl e 1 ). So 111 e have bee n
id ent if ied; oth ers are in th e prncess of being
id entifi ed or c haracteri zed. Most paras ito id
spec ies (17) belon g to th e Eulophida e
fa111i ly. Th e 111ost ab un da nt specie s are the
ecto p aras itoid, Diglyp hus sp. and th e
endoparasitoids , Halticoptera arduin e, and
Chrysocharis sp p . Ot her paras itoid s are
presen t but are less abu nd ant und er th e
co nditio ns found on the ce ntral coast
(Figure 4).

Relative effectiveness of parasitoids. The

two 111 ost i111porta nt p aras ito id s see 111 to be
large ly res po nsible fo r th e seaso nal va ri ation of the LM F popu lati on. Females of H .
arduin e lay their eggs insid e th e bod y of the
LMF larva by introdu c in g th eir ov ip os itor
th rough the epid er111a l layer of the lea f
111ine. Th e parasitoid larva deve lops inside
the LM F pre-pupal larva . Th e p ara sit ized fly
larva surv ives until pupation, allowin g th e
parasito id larva to co 111pl ete its deve lopment. E111 erge nce of p aras ito id adult wasps
occ urs 1 wk afte r th e emerge nce of fli es
134

Polalo

from nonparasitized puparia. Leve l of

paras it ism (meas ured as percentage of fl y
pu pa ri a) by H. arduine commonly reac hes
55-78% from Jun e to A ugu st.
Fe111 ales of Diglyph us sp. p aralyze the
LM F larva in sid e th e leaf 111i11e and lay th eir
eggs next to the larva. The paras itoid larva
feed s on th e juices of th e fl y larva fro111
outside, res u lting in the death of fly larva
before pupat io n. Th e parasitoid pupae are
formed in sid e th e leaf mine. Ad ult wa sps
eme1ge fro111 th e dry m ined lea ves. Th e
leve l of para siti sm by Diglyhus sp. was
est imated to be arou nd 60 -82% of fly larvae
from Septembe r-No ve mber.

Seasonal occurrence of parasitoids. In

Canete, th e potato p lanti ng sea so n starts in
Marc h and ex ten d s fo1 severa l m onths
depe nd in g on weathe r. Late plan tin gs are
harveste d in Decemb er. Whe n littl e or no
in sec tic id e is used , paras it ism of LMF larva e
from Apri l to A ugust us ually va ri es from
35 % to 55 % . But, fro111 Se ptember o nward,
total parasi ti sm increases stead i ly unti I
Novem ber w hen p aras itis111 nears 100% .
H ig h temp erat ures reco rded in 1997 -98
changed thi s pattern (Fi gure 5). Th e occ urrence of new pests, w h itefl ies, budm idges,
and cutwo rm s, led far mers to use insect icides more frequent ly than norm al . Parasi toids in Can ete we re drastica i ly affected;
tota l paras iti s111 was eve n lowe r than in
other va lleys on the cen tral coa st.
Predators. Leafmi ner flies in th e stud y
area are attac ked by 111 any preda tors
belo ngi ng to five in sect orders (Co leo ptera,
H e111 iptera, Diptera, D er maptera , and
H yme noptera) and by sp ide rs . Most
predators are ge neral feeders and p rey
indi scri min ately o n seve ral insect pests.
Some predatory fli es of the fa111ili es
D o li c hop od id ae and Ernpidida e te nd to
prefer LMF.
A ll develo p111 enta l stages of LMF we re
subj ect to predation . The most exposed
stages, howeve r, were mos t v uln erable-th e
ex trud ed egg in th e fo li age and th e pupa in
th e so il. D irect eva lu at ion of th e ro le p layed

Table 1. Species of parasitoids of the order Hymenoptera that attack leafminer fly larvae and pupae in the
central coast of Peru, 1998.
Hymenoptera

Subfamily

Species

Parasitism

lchneumonnoidea
Braconidae

Opiinae

Opius scobriventris
Opius sp.

Endoparasitoid
Endoparasitoid

Gonospidium sp.

Endoparasitoid

Eulophinae

Diglyphus websteri
Diglyphus begini
Dig/yphus sp.

Ectoparasitoid
Ectoparasitoid
Ectoparasitoid

Tetratischinae

Dioulinopsis sp.

Endoparasitoid

Entedontinae

Chrysochoris phytomyzoe
Chrysochoris oinsliei
Chrysochoris sp. A
Chrysochoris sp. B
Chrysochoris sp. C
Chrysochoris sp. D
Chrysochoris sp. E

Endoparasitoid
Endoparasitoid
Endoporasitoid
Endoparasitoid
Endoparasitoid
Endoparasitoid
Endopo rasitoid

Chrysonotomyio sp. A
Chrysonotomyio sp. B
Chrysonotomyio sp. C

Ectoparasitoid
Ectoparasitoid
Ectopa rasitoid

Closterocerus cinctipennis
Closterocerus sp.

Ectoparasitoid
Ectopa rasitoid

Elachertinoe

logrommosomo multilineotum

Ectoporasitoid

Miscogasterinae

Holticoptero orduine
Holticoptero sp. A
Holticoptero sp. B

Endoparasitoid
Endoporasitoid
Endoparasitoid

Cynipoidea
Eucoildae
Chalcideoidea
Eulophidae

Pteromalidae

by specifi c preda tors in th e dynami cs of th e

fl y is extremely diffi cult. For th e time bein g,
efforts are centered on identifyin g these
spec ies and desc ribing th eir predatory
acti vity. Tabl e 2 li sts the most common
predators of LMF found in coastal Peru.

Soil-inhabiting predators. The importance of so il -inh abitin g predators is usually

underestim ated . That is parti cul arly true in
th e case o f LMF, beca use its larv al stage is
always protec ted in side lea f mines. It is
onl y wh en larvae abandon th e min e to

CIP P1og1om Repo1t 1997-98

135

Diglyphus begini
Diglyphus
websteri
32.7%

Chrysocharis phytomyzae

8.9%
Chrysocharis sp.
0.7%

~8D~~~""""1- 3.2%

~~~~~~

Ganaspidium sp.
0.5%
C/osterocerus
cinctipennis
1.2%

41.5%
Halticoptera arduine

Others
1.5%

Figure 4. Distribution of species of parasitoids of leafminer fly larvae in potato and other crops in the Canete

Valley, Peru, 1998.

Parasitism (%)
90.0
80 .0
70 .0

D Calla o

Caraba yllo

,,

50 .0

r.

''

:-

(;_

40.0

r:".

A-

I
30.0
20.0

0.0

D Canete

60.0

10.0

---,

Aug

Sep

Jan

I
Feb

Mar

Apr

May

Jun

Jul

nOct

ii
Nov

Figure 5. Occurrence of leafminer fly parasitism in potato and other host plants at three sites on the central coast

of Peru, 1998.
pu pa te th at th ey beco m e vu lnerabl e to
pr dato rs in the soil. In so me fi e ld s, th e
m ost abundant predatms we re Pteros tichus
sp. and Ca/osoma abbreviatum.

136

Pototo

Pterostichus sp . uaw ls rapidl y on th e so il

surface sea 1chin g fo r prey, includin g LM F
pupa1ia. Ca losom a abbre viatum a1e also
prese nt, but in sm all er numbers.

Table 2. Predaceous arthropods reported to feed on leafminer fly eggs, larvae, pupae, and adults in the central
coast of Peru, 1998.
Class

Order

Family/species

Insecta

Coleoptera

Carabidae

Calosoma abbreviatum
Pterostichus sp.

Prey
Pupae
Pupae

Cicindellidae

Megacephala carolinachilensis

Hemiptera

Staphilinidae
Undetermined spp.
Anthocoridae

Orius insidiosus

Pupae
Pupae
Egg

Nabidae

Nobis punctipes

Larvae

Lygaeidae

Geocoris punctipes
Dermaptera
Dipiero

Labiduridae
Undeterm ined spp.
Dolichopodidae

Condylostylus similis

Egg
Pupae
Adult

Empididae

Drapetis sp.
Hymenoptera
Arachnida

Formicidae

Adult
Pupae

Theridiidae
Argiopidae
Lycosidae
Anyphaenidae
Salticidae

Adult, pupae
Adult, pupae
Adult, pupae
Adult, pupae
Adult, pupae

Araneida

Megacephala carolina-chilensis was fairly

commo n from December throu gh March. A
non id entifi ed spec ies of the famil y
Staphilinidae was commo nly trapped
throu ghout th e year.
Th e seaso nal abundance of so il predato rs was assoc iated w ith hi gher temperatures (Fi gure 6). Several species of ants are
act ive predators in th e so il and fo li age.
Finally, seve ral spec ies of spiders (ge neral
predators) occurred o n th e so i I su rfa ce and
in the foliage.

Foliage-inhabiting predators. Eggs,

larvae, and adult LMF are exposed to
preda tors in the plant fo li age. Adult and
nymp hs of sma ll to minute true bugs,
Ceocoris punctipes and Orius insidiosus,
prey upo n LMF eggs, espec iall y extruded
eggs. Another true bu g, Nabis punctipennis,
prey on eggs and larvae, penetrating th e
leaf epidermi s w ith its proboscis to reac h
larva in side the leaf mine. Th eir predatory
effect o n LMF has not bee n qu antifi ed
directly, but the presence of th ese voracious

CIP Pwgram Report 1997-98

137

Total insects/trap/month

1200
~ Pterostichus sp (C

1000

Laba

.
. oleoptera:Carabidae)
' ura nparia (Dermaptera: Labiduridae)

800
600

Nov Dec

Figure 6. Seasonal occurrence of the most abundant predators captured in pitfall traps at three sites on the
central coast of Peru, 1998.
pred ators co incides w ith slowe r pop ul ati o n
growth.
Eve n m o re d iffi cult to eva lu ate is the ro le
of p reda tory fli es th at ca pture and k ill LMF
ad u lts. Species o f Dol ic hopod id ae an d
Emp idi d ae occ ur in large nu m bers near
hu m id areas w ith abun da nt vege tat io n
along th e ce ntral co asta l va ll eys o f Peru.
W hen in secti c id es are no t used, th ey are
abu ncl ant in po tato , bea ns, and oth er c rops
infeste d by LM F. Th ey ac ti ve ly hun t LMF at
rest and in fli ght. Simil ar behav io r is
ex hibited by bl ack fl y (Orapetis sp.) .
Predato ry t rue bu gs are co nstant ly p resen t,
whereas p red ator y fli es occur o nl y occasionall y, bu t usual ly in large num be rs.

fa c to rs co nstitute th e bas is for th e m anageme nt of th e pest. H ypers ens iti v ity of the
grow in g potato leaf that res ul ts in eggextru sio n ca n be furt her sel ecte d fo r as a
mec hani sm of res ista nce agai nst th e fl y. Th e
complex grou p of parasi toid s pl ays an
im po rtant role in reg ul at in g LM F pop ul at io ns o n th e Peru vi an ce ntra l co ast. Th eir
in t rodu ct ion to oth er areas , parti c ul arl y
w here th ere is littl e ra in , mi ght help to
m anage LM F in fes tat io ns .
A lth o ugh it was not pos sibl e to qu ant itati ve ly eva lu ate th e mo rtality produ ce d by
predators of LMF, their pres ence was a sign
of good bi olog ica l co ntrol and co in c id ed
w ith low LM F po pul ations .

Conclusions

Acknowledgements

Th ere are important natur al m o rt a lity

facto rs in the potato pl ant and in po tato
fie ld s th at affect LMF po pul ati o ns. Th ese

Id enti fica ti o n of th e LM F paras ito id s was

cl o ne by Ora. A dri ana Sal vo o f Co rd oba
U ni ve rsity, A rge nt in a.

138

Potato

Literature Cited
Campos, G. 1976. Control quimico d el
minador de hojas y tallos d e la papa,
Scribipalpula absoluta Me yri ck, en el
vall e d e Ca fi ete. Revista Peru ana de
Entomo logia 19 :102 -106.
__ . 1978 . Contro l quimico de la mosca
minadora , Liriomyza huidobrensis, en el
valle de Cafiete . Revista Peru ana d e
Entomologia 21:109-112.
Cisneros Vera, F.H. 1986. Control biologico
de las pl agas co n espec ial refere nc ia al
cultivo de la papa . In: Val en c ia, L. (ed. ).
Memorias d el c ur so sob1e contro l
in tegrado d e pla gas de papa. 1. Curso
Intern ac iona l so bre Control ln tegrado de
Plagas d e Pap a. Bogota, Colombia. 29
Jun - 19 Jul. Centro Internacion al d e la
Papa (CIP) and ln stituto Colombi ano
Agropecuar io, Bogota, Colom b ia. p. 1 01-

108 .
Ewe ll, P.T. , H. Fano, 1<.V. Raman , J. A lcaza r,
M. Pal ac ios, and J. Carhuama ca. 1990.
Farmer man age ment of potato in sect
pests in Peru: Report of an interdi sc iplinary resea rch project in se lected reg ion s
of th e hi ghlands and coast. Food Systems
Resea rc h Ser i es (CIP ), No. 6. CIP, Lima ,
Peru. 77 p.
Fon seca, D. and G. Sanc hez. 1996.
Fluctua c ion pob lacio nal d e Lirio myza
huidobrensis (Bl anchard) y sus
parasitoid es en los cultivos d e papa y
frijol en el v all e de Canete. In: ln fo rme
de lnves tig ac ion. Convenio Universidad
Nacional Ag 1ar ia La Molina y Centro
In ternac io nal d e la Papa , Lim a, Peru .

32 p.
Gonzales, L.E. 199 4. Extru sion d e huevos
de Liriomyza huidobrensis en fo li o los de
papa cultivad a en Huaral (Lima). Revista
Peruana d e Entomo logia 37:111-114 .
Herrera, J. 1963 . Prob lem as ins ectil es del
cul t ivo d e p ap a en el va l le de Ca nete.
Re v ista Peru ana de Entomologfa 6:1-9.
llE (Interna t iona l In st itu te of En to mo logy) .
1996 . Di stribution maps of plan t pests .
Series A: M ap No . 568, December 19 96.
CAB Internat ion al, Wallingford, UI<. 3 p.

Liz arr<:1ga-Travag lini, A. 1989. Biologia de

la 'mosca rninadora' Liriomyza
huidobrens is Blanchard (Di ptera:
Agromyzidae). B.S. Thesis in biolo gy.
Uni ve rsi d ad Ricardo Palm a. Fac ultad de
Ci enc ias Biologicas , Lima, Per u. 38 p.
Macd o nald , O .C. 19 91 . Re spo nses of the
alien lea f min ers Liriomyza trifoli i and
Liriomyza huidobrensis (D iptera:
Agromy z id ae) to some pes ticid es
schedul ed for their control in th e UK.
Crop Protect ion 10(6):509 -5 13.
Muji ca, N. and F. Cisn eros. 1997. D eve lopin g IPM co mponents for leafminer fly in
the Canete Val ley of Per u. In: In ternation al Pota to Center Program Report
1995-96 . CIP, Lima Peru. p. 177-184.
Muji ca, N., V. Canedo , M. Pal ac ios, and
K.V. Ram an. 1993 . lnflu enc ia de lo s
tricom as glandulares sobre la b io lo gfa de
la s prin ci pales pla gas de la papa. In:
Soc iedad Entomologica del Pen'.!.
Res um enes y programa. XXXV
Con ve nci o n Nacional de Entomologfa.
7-11 Nov 1993. Arequipa, Peru. p. 35 -

36.
Ochoa, P. and M . Carballo. 1993 . Efec to
de vario s in sect icidas sobre Liriomyza
huidobrensis (Diptera: A gro my z id ae) y
su pa1asi to id e Dig lyphus isaea Wa lke r
(H y menopte ra: Eulophid ae) . Manejo
lnteg rado d e Plagas (Costa Rica) 26:8 -12.
Parrell a M. , C. Keil, and J. Morse . 19 84.
Ins ect icide res isten ce in Liriomyza
trifolii . Ca l. Agric. 38(1/2):22 -23 .
Redolfi de Hui za, I. , M. Pala c io s, and J.
Alcazar. 1985 . H ym enopte1a parasito
ides de Liriomyza huidobrensis en papa
cu ltivada en Rfmac , Cat'iete e lea. Re v ista
Peruana de Entomologfa 28:19-21.
San chez, G . and I. Redolfi de Hui za. 1988.
Liriomyza huidobrensis y sus
parasitoid es en papa cultivada en Rfmac
y Can ete, 1986. Revista Peru an a de
Entomol og fa 3 1:11 0-112 .
Sotom ayor, M. and F. Ci sneros . 1 996 .
A lgun as prec isio nes en la reaccio n de
'ex tru sion' de huevos de mosca
minadorJ en plantas de papa. In:
Sociedad Entomol6gica del Peru .

CIPP1og1orn Repo111997-98

139

Resumenes y prog1a111a. XXXVlll

Convenci6n Nacional de En tomo logia .
17-21 Nov 1996. Chincha, Peru. p. 22-

23.
Spencer, K.A. 19 90 . Host specialization in
th e wor ld Agromyzidae (Diptera). Kluwer
Academic Publish ers, Dordrecht,
Netherlands. 444 p.
Wille, L. 1952. Entomologia Agricola del
PerC1. Segunda Edici6n. Editado por Est.

140

Potolo

Exp. A gric. La Molina. Director of

Agriculture, Ministry of Agricultur e,
Lima, Peru. 543 p.
Yabar L, E. 1988. La mosca minado1a de la
papa en el Peru. Spec ial Report No. 2.
ln st ituto Naciona l de ln ves tigaci6n
Agwia y Agroindustrial (INIAA). National Potato Program, Lima, Peru. 37 p.

Taxonomy and Bionomics of the Andean Potato Weevil

Complex: Premnotrypes spp. and Related Genera
J. Alcazar and F. Cisneros 1

Andea n pota to weev il s (APW) are th e mos t

se ri ous pests o f po tato (Solanum spp .) at
hi gh altitudes (a bove 2,800 m) in th e
And ea n reg ion, w here w ild and culti va ted
potato spec ies are th eir onl y hosts. They are
d ist rib ute d fro m A rge ntin a to Ve nez uela
cove ri ng a mountai n ter ritory abo ut 5,000
km long. W hen no co ntro l meas ures are
use d, a co mm o n situ ation for subs istence
fa rme rs, more th an 50% of tubers are
commo nl y in fes ted at harves t. Co mm erc ial
produ cti on is su bj ected to th e use of hi ghly
tox ic insecti cid es appl ied to th e fo liage or
incorpo rated into the so il. Even then 1530% of harvested tub ers are in fested.
In a fe w cases wh ere anc ient ag ri cu ltural
sys tems are still pract iced, as in co mmun al
ag ri culture, APW damage is neg li gible. Th at
is beca use of di sta nt ro tat ions in time (5 -7
yr) and space (potato fie ld s spaced seve ral
kil ometers apart), a system th at ca n ra rely
be practi ced tod ay.
M ost wo rk on APW has been devoted to
th e study of chemi ca l co ntrol o f the pest,
fo rced by th e need to co mbat the pest and
redu ce damage. O th er, mo re basic, as pects
such as id enti fica ti on of th e va ri ous weev il
spec ies , th eir geog raphi ca l di stribu tion, and
simil ariti es/differences in th eir li fe cyc les
and seaso nal hi story have bee n onl y
pa rti all y or no nsys temati ca ll y addressed.

The APW Complex

Farmers co ll ecti ve ly des ignate a group of
weev il s w hose larvae bore into potato
tub ers as Andea n potato w eevil ("go rgojo
de los And es ") o r white grub s (" gusa nos
bl ancos "). Taxonomi sts have desc rib ed

seve ral spec ies, but it is un certai n if th ere

still remain und esc rib ed species, parti cularl y in south ern Peru and Boli via.

Taxonomy of APW
A lth ough spec ies of th e genu s
Premnotrypes are th e mo st w id ely di stributed , th e fir st taxo nomi c desc riptions of
APW w ere o f spec ies in two different
genera. Th e first spec ies desc rib ed was
Ph yrclenus murice us Germ ar in 182 4 . Th e
seco nd was Rhigops iclius tucu manus Heller
in 1906 (Pierce, 191 4). Both we re co llected
in Argentina.
Pi erce (191 4) desc rib ed two new APW
species from infested tub ers in Peru:
Premnotrypes so /ani and Tryp opremn on
latith orax. Th ey we re co nsid ere d th e most
im portant potato weev il spec ies in th e
Peruvia n entomo log ica l li te rat ure fo r many
years (Will e, 1952). In th e in te rim , fo ur new
spec ies had bee n desc ribed : Trypopremnon
san fo rcli Pi erce (1 9 18), Solanophagus vorax
Hu stac he (1933), Plastoleptops solanivora
Hell er (1935), and Premnotrypes
fractirostris Ma rshall (1936) .
Ku sc hel (1956) in c lud ed th e genu s

Premn otrypes, w ithin a new trib e

(Premn otr ypini ), and prese nted a key fo r the
id entifi cation of 11 va lid spec ies w ithin th e
genus . An addi tional spec ies was described
by Alcala (1979a), ma kin g a tota l of 12
val id Prem notrypes spec ies. Th ey are:
P. vorax (Hu stac he), co ll ected in Co lombi a,
Ecu ador and Peru ; P. /atith orax (Pi erce),
collected in Pe ru , Bolivia and Chil e; P.
solanipercla Ku sc hel, coll ec ted in Peru and
Boli v ia; P. clivos us Ku sc hel and P. z ischkai
Ku sc hel, co ll ected in Boli via; and

1 CIP, Lima, Peru.

CIPProgrnm Report 1997-98

l4l

P fractirostris Ma rsha l I, P piercei A lca la,

P pusillus Kuschel , P sa nfordi (Pierce), P
solani Pierce, P. so/anivorax (H ell er), and
P suturicallus Kusch e l, all collected in
Peru .
Some autho rs dea lin g w ith potato pest
probl ems in c lud ed other genera w ithin the
APW co mplex, but th ese spec ies are not
abl e to comp lete their life cycles in potato.
Yaya (1 97 1) reported Scotoeborus sp. and a
spec ies near Adioristus sp. in potato in the
centra l mountains of Pe ru . Other repo rt s
referred to Hyperodes sp. (De lgado, 1975);
three species of Adioristus (A.
subsquameus Hu stache, A. ebenus
Hu stac he, and A. leprus Hustach e); and
Listrode res sp. from th e central mountains
of Peru (Va lenc ia and O ' Brien, 1976).
Naupactus sp. has bee n reported in interna l
docum ents of the ln st ituto Nacional de
ln vest igac ion Agraria , Huancayo, Peru .
Identification of APW Species
Genera and higher categories
Th e three genera in vo lved in th e APW
compl ex belong to different subfami li es of

fami ly Curcu lionid ae (W ibmer and O'Br ien,

1986): genus Premnotrypes, tribe
Premnotrypini, subfamily Entimin ae; ge nus
Phyrde nus, tribe Cryptorrhynch ini , subfamil y Chryptor rh ynch in ae; and genus
Rhigopsidius, tribe Rhytirrhinini , subfamil y
Rh ytirrhininae.
Tribe Rhytirrhinini is formed by o nly one
genus (Rhigopsidiu s H ell er), w ith two
species ( R. tucum anus and R. pie rce i). Van
Emel en (1952), howeve r, considers them as
sy non ym s. Tribe Crytorrhy nchini is di v ided
into seve ral subtribes . The subtribe
Crypto rrh ynchina co ntain s severa l genera
of which one is rel ated to th e APW com pl ex, Phyrdenus LeCo nte syn. of
Cryptorrhynchus Ge rm ar.
Ku sc hel (19 56) described tribe
Premnotrypini, which includes three
genera: Rhinotrypes Kuschel , Microtrypes
Ku sc hel, and Pre mnotrypes Pierce. On ly
spec ies of genu s Premnotrypes are related
to th e APW compl ex . Synonyms recog nized
for genus Premnotrypes are Trypopremnon
Pi erce, Solanophag us Hu stac he, and
Pla sto leptops Hell er.

Characterization of genera
The following cha racteri zation of genera of APW species is based on conventional taxonomic grouping of species. But the first premise is that the weevils are
adul ts of larvae that developed in potato tubers.
1.

Beak rather short. Hind wings vestigia l.

Beak long. Hind wings we ll developed.

2.

Prosternum grooved for reception of beak. Mandibles lacking deciduous piece.

Pronotum w ith a deep median fu rro w widened angulary at middle and also
behind.
R. tucumanus
Prosternum not grooved for reception of bea k. Mandibles have deciduous
piece. Large compound eyes with more than 80 facets.
Premnotrypes spp .

2
Phyrdenus muriceus

Characterization of Premnotrypes species

External morphological c haracteristics and the male's aecleag us are valuab le for
identifying weev il species. The principal chara cters used for ident ification are those
of the head (compound eye), rostrum or beak (epistome, groove for reception of
antenna ! scape, mandibles, mandibular cusp o r deciduous piece (Figure 1 A)) ,
anten nae (antenna ! scape and antenna! club), thoracic sclerites, and elytra (grooves

142

Pototo

and intergrooves, posterior upper [bending] part). Complementary characters are

scales, pubescence, setae, grooves, tubercles, and callus. (See Figure 1 B.)
All Premnotrypes species show sexual dimorphism; females are larger than
males. Average lengths recorded for females and males are as fol low: P. vorax
females 6.8 mm, males 5.6 mm; P. /atithorax females 6.7 mm, males 5.6 mm; P.
suturical/us females 8.0 mm, males 7.5 mm. The apex of the 5th sternite is blunt in
mal es, pointed in females. The end portions of elytra are strongly declivous in
females, less declivous in mal es.
Key to Premnotrypes species (modified from Kusch el, 1956)
1.

Antenna! scape with abundant scales. Dorsal scal es and setae of rostrum mostly
2
oriented forward.
Antenna! scape without scales. Dorsal scales and setae of rostrum oriented
backward.
9

2.

Eighth groove of elytron irregular, as with supernumerary points.

Eighth groove normal; points in a definite simple line.

3.

Base of prothorax with six nodules widely separated forming an arch. Pronotum
dense ly pointed. Rostrum wide. Ocular lobes very prominent. Peru .

3
4

P. solani
Base of prothorax apparently with only four nodules; lateral ones are fused in a
prominent body. Pronotum sparsely pointed . Rostrum narrow, with a deep
suprascrobal naked groove. Ocular lobes slightly prominent. Peru.

P. so/anivorax
4.

Rostrum epistome elevated

5
Rostrum epistome almost fl at, only slightly convex. In the high part of the
elytron with six major tubercles, two on the third intergroove and one on th e
fifth. The latter is located eq uidistant between the other two. Peru and Bolivi a.

P. solaniperda
5.

Antenna I scape short and very thick, does not reach the anterior margin of th e
eyes . Tibia scaly in the lower border. Anterior margin of pronotum with a stron g
elevated structure that divides in two backward, imparting a sigmoid form to th e
whole structure. (See Figure 1C.) Peru.
P. fractirostris
Antenna! scape slender, reac hing somewhat beyond the anterior margin of th e
eyes. Tibia without scales in the lower border. Structure of pronotum different
from that described above.
6

6.

Frans without swelling nea r the eye. Rostrum with strong to moderate dorsal
curvature.
7
Frans with a swelling nea r the eye. Rostrum with a dorsa l curvature less pronounced than above. Elytra transversally convex up to the 7th intergroove,
whose tubercles are similar to those of the 5th intergroove. Elytra fold vertically
to the sides at the 7th intergroove. The 5th intergroove generally ends in a larger
8
tubercle, but without forming a true callus.

7.

Rostrum with strong dorsal curvature. Elytra flat up to the 5th intergroove and
then almost vertical to th e side. Tubercles of the 7th intergroove much smaller

CIP Program Report 1997-98

143

than those of th e 5th intergroove. This one general ly end s in thick cal lu s in th e
uppe r part of th e bend ing part. (Figure 1 0 ). Co lombi a, Ecuador, Peru, and
Ve nezuela.
P. vorax
Rostrum with moderate curvature. Elytra transve rsa lly moderately convex . Third
intergroove with a large tubercl e on the upper posterior (bending) part. (See
Fi gure 1 E). Peru.
P. piercei
8.

Prothorax w ith th e lateral swe lling of th e base b lunt, and as prominent as th e

anterior lateral swe llin g. Lower tooth of the mandibl es pointed. (See Fi gure 1 F).
Peru , Boli v ia, and Chil e.
P. latithorax
Prothorax wi th the lateral swe lling of the base po inted and prominent. Lower
tooth of mandibl es blu nt. Peru.
P. sanfordi

9.

Antenna! scape thin and reac hin g beyond the anterior margin of the eyes. Th e
3d intergroove end s in th e upper posterior (bendin g) part w ith an elevated
tubercle. Boli v ia.
P. clivosus
Antenna! scape at most may reac h the anterior margin of the eyes. The 3d
interg roove does not have a tubercle larger th an th e upper posterior (bendin g)
part.
10

10. Suture in th e upper posterior (bend ing) part w ith a ca llu s. Peru.
Suture in the upper posteri or (bending) part w ith out a ca llus.

P. suturica llus
11

11 . Prothorax more or less rough and scaly. Th e 7th i ntergroove forms a stron g disc
rid ge. The 3d and 5th intergrooves each have a large tubercle. Behind it th e
elytra fall vertical ly. Bo! ivia.
P. z ischkai
Prothorax somewhat irregul ar but smooth. Few sca les in the disc. The 7th
intergroove does not form a pronounced border that limits the disc. Tubercles in
P. pusillus
odd intergrooves small and blunt. Peru.

Geographical Distribution of APW Species

A ll spec ies of genera Premnotrypes,
Rhigops idius, and Ph yrdenus that attac k
potato are nati ve to th e Andes. Th ey
co mm o nly occ ur between 2,800 and 4,700
rn w here th e number of pl ant spec ies is
so mew hat restricted. Practi ca ll y th e o nl y
crop th at ca n be grow n at those altitud es is
potato (Solanum andigenum and oth er
spec ies of tuber-fo rmin g Solanum, in c ludin g bitter potatoes at hi gher altitudes). Th e
And ea n mounta in s extend along Ch il e,
northwest Argent in a, Boli via, Peru, Ecua dor, Co lombi a, and part of Venez uela.
Th e APW speci es are not eve nly distributed a lo ng th e A nd ea n rnou nta in s. Most
spec ies are co ncentrated in Peru and
Bo li v ia, w hich cou ld be co nsidered th e
144

Poloto

cen ter of o rigi n fo r th ese weev il s. Ten of th e

12 spec ies of Premnotrypes occ ur in Pe ru .
P. clivosus and P. z ischkai have bee n
repo rted onl y in Bol iv ia. P. vorax occ urs in
Colombi a, Ecuador, Venezue la, and Peru .
Rhigops iclius tu cumanus occurs in no rth wes t Argentin a, Bo li v ia, and th e so uth of
Peru . Phyrdenus muriceus occ urs in Bo li v ia
and north east Arge ntina .
A lca la (1979) made th e first stud y on th e
di stri buti o n of An dea n weev il s in Peru. We
have upda ted th at wo rk and prese nt it in
Fi gure 2 along w ith di stribution of A PW
spec ies throu gho ut South A meri ca. R.
tucumanus has bee n reported in A rgentin a
(Tu cum an, Sa nti ago del Estero, Juju y,
Catamarca, and Sa lta) (Agostini and Vilte,
1982), Bo li v ia (Potosi, Chuquisaca, Tarij a,
Cochaba mba, and La Paz) (PR O IN PA,

Figure 1. Morphological characteristics of Premnotrypes spp. A) mandibular deciduous piece, P. latithorax; B)

elytra: tubercles and scales, P. piercei; C) dorsal view, P. fractirostris; D) dorsal view, P. vorax; E) dorsal
view, P. piercei; F) dorsal view, P. latithorax. White lines = 1 mm.
1994), Chile (Agostini and Vilte, 1982), and
Peru (restricted to Puno at the border w ith
Bolivia).

and Santa Cruz) (PROINPA, 1994). The

original description of the spec ies was
based on material collected in Argentina.

Phyrdenus muriceus occurs at relativ ely

low altitudes, not more than 2,000 m and is
distributed in the mesotermic valleys of
Bolivia (Coc habamba, Tarija, Chuquisaca,

Life Cycle and Seasonal Occurrence

The most wides pread and important pests

of th e high Andes by far are P. latithorax,

CIP Program Report 1997-98

145

P. solaniperda Kuschel
2
3
4

5
6

P. latith orax (Pierce)

P. sanfordi (Pierce)
P. pusillus Kuschel
P. piercei Alcala
P. suturical/us Kuschel

P. fractirostris Marshall

P. solani Pierce

P. vorax (Hustache)

10

P. solanivorax (Heller)

11

P. clivosus Kuschel

12

P.

13

Rhigopsidius tucumanus Heller

14

Phyrdenus muriceus (Ge rmar)

zischkai Kuschel

Figure 2. Geographic distribution of species of the Andean potato weevil complex in the Andean Region .

P. suturica llus, and P. vorax. Th ey are

allopatri c spec ies in Peru . Th eir life cycles
and th e seaso nal histories in Peru we re
studied in th eir respectiv e areas of di stribution ( P. vorax in Cajam arca, P. suturicallus
in Huan cayo, and P. latithorax in Cusco.
Behavior
The th ree main spe cies of Premnotrypes
show simila r behavior. It is expected th at
146

Potolo

the sa me may be true fo r th e other nine

species of th e genu s; however, distin ct
differences ha ve been observed for
Phyrdenus and Rhigopsidius.

Premnotrypes. Adult ma! es and females

remain hidden during the day beneath soil
clods, stones, dry leaves, or any other
she lter including so il cracks near th e potato
plant. In the evening weevils climb to the

fo li age to eat th e bord er of the leaves and to

mate. Females lay th eir eggs in side straw or
other plant deb ri s near th e plants. As eggs
hatch, neo nate larvae make th eir way into
the so il seek in g potato tubers. The larva
bores in to the tub er and remain s th ere
feedin g until it is rea dy to pupate. Then the
larva aba ndon s th e tub er and d igs into th e
so il to make a pup al ce ll and prepare a
tunn el fo r its eme rgence as an ad ul t.
Weev il s remai n in the pup al cel l until th eir
emergence w ith the o nset of rai ns. Th e
weev il in the pupal ce ll is ca ll ed an ove rw interin g weevi l. Overwintering weevi ls
ca n be id entifi ed by th e prese nce of th e
de c idu o us pieces of th e mandibl es . These
p ieces are deta ched whe n weev il s beco111e
active and emerge from the soi l.
Alt ho ugh thi s sy nchro ni za tion is ev id ent
for weevi ls in th e field, rain is not th e on ly
st imu lus to emergence. Adu lts also e111e rge
fro111 sto res w here rai n is not a factor. In th is
case, however, the eme rgence per iod is
ex tend ed.
Because th e period of ini tial larv al
infes tation of tubers ex tend s over seve ral
weeks, larvae co mmonl y reac h maturity
and aba ndon the tuber fo r pupation before
harvest. In that case, pup at ion ce ll s are
sca ttered all ove r th e fi eld . Larvae th at
reac h maturity at harves t, w hen tubers are
p i led, co nce ntrate und ern ea th th e pil es .
Fina lly, th e last gro up of larvae aba ndon s
th e tubers durin g sto rage. Here, during th e
first 30 d, 97% of larvae leave the tubers fo r
pup ation . Pupation areas are th e so urces of
new infestations.

Rhigopsidius tucumanus. Th e be hav ior

of R. tucumanus di ffe rs from th at of
Premnotrypes spp. in tha t pupation occu rs
in sid e th e tuber. A s a res ult, adults emergin g from th e planted seed tu bers initi ate
field in festation s. Fema les lay eggs in th e
so il nea r the potato pl ants.
Phyrdenus muriceus. Phyrdenus
murice us ad ults mi grate to th e potato fi elds,
eith er by wa lkin g o r by flying (this spec ies

is th e on ly o ne in th e gro up that ca n fl y).

Feed in g weev ils make small holes in th e
leaves . Females lay eggs in the soi l near the
ste111 of the p lant, and larvae feed on both
tubers and roots. Pu pat io n occurs in th e
soil.

Life cycles
Life cyc les of APW have been reported
by seve ral autho rs with differe nt levels of
detai l: Ca rrasco (1 961 ), Ti soc (1989) , and
Carba jal (1992) on P /atithora x; Cal vache
(1986) and Munoz and A lcaz ar (1998) o n
P. vorax; A lcaza r (1976) o n P suturica llus;
Carhu amaca and Tova r (1987) on P. piercei,
and Gil (1991) on P. so lanipercla. Th ey
sho w 111any sim i lariti es and few di ffe rences.
Th e li fe cyc les of P. suturica llus, P.
piercei, an d P. vorax are give n in Tabl e 1.
Th e most remarkabl e differences are in th e
nu111b er of larva l in stars (fi ve fo r P vorax
compared with o nl y four fo r the other two
spec ies), and the s111a l ler qu antity of eggs
laid by P. vorax. Differences in time to
co111pl ete larv al i nstars and deve lopm ental
stages may be du e to spec ies- related
differences or different environm ental
cond itio ns, part icul arl y temperature.
Seasonal occurrence
Th ere is one generation a yea r for th e
spec ies of Premnotrypes studi ed and for
R. tucumanus. Th e d eve lo pmental stages o f
th ese weev il s are sync hro ni zed w ith
cli111 ati c co ndition s, the croppin g system,
and th e phenologica l deve lopment of th e
potato p lant. O ur reco rd s in Peru; studi es
by G il (1991) with P. so lan iperda in Pu no,
Peru; and Ca rb aj al (1992) w ith P. latithorax
in Bo li via all indi ca te th e occurrence of o ne
generati o n a yea r. H owever, Ga llegos
(1995) in Ecuador and Ca lvac he (1986) in
Colombi a, wo rk in g w ith P. vorax, cons id er
th at th ere is mo re than one generation a
year w hen potato c rop pin g is co ntinu ous.

Potato produ ct ion in the high mountain s

is essent iall y ra in fe d. Th ere are two we ll defi ned seaso ns: the dry, co ld, winter
sea so n and th e rainy, mod erately warm

CIP Program RepO!t 1997-98

14 7

Table 1. Life cycle of three species of Andean potato weevil in days : Premnotrypes suturicallus, P. piercei, and
P. vorax, Peru.

Stage
Egg (d)
Larva I
Lorva II
Lorva Ill
Lorva IV
Lorva V
Pupa {d)
Overwintering adult {d)
Cycle (Egg-adult) (d)
Longevity male {d)
Longevity female (d)
Totol cycle {d)
Ovipositing period (d)
Eggs/female (no.)

P. suturicallus
33
11
9
12
57

26
8
7
9
56

54

52
135
293
277
233
549
127
521

115

291
159
126
434
106
631

grow in g season (s prin g-summer). The

weev il overwinters successively as
prepupa, pupa, and overwintering adult in
a pupal ee l I in the soi I (o r in th e tuber in
case of R. tu cumanus). Whe n spr in g rains
begin, overwinter in g weev il s become acti ve
and emerge from the so il . Weev ils remain
in th e field if potato plants are avai lab le, or
mi grate to new potato fi elds.
When imm igration occurs, in festat ions
are heav ier near th e bord er than in the
center of th e fi eld. Th e seaso nal hi story of
adults and immature stages of P vorax was
studi ed in Cajamarca in th e field and in
stores . Adu lt emergence starts by th e last
week of October and co ntinues until th e
end of Febru ary. Migration to potato fields
occu rs from December to February. Hi ghest
ad ult popul ations in th e field we re observed
in February w ith popul ation density
reac hing as hi gh as four weevi ls/ plant.
Larva l dam age to tubers sta rts by th e middle
of March and increases until Apri l. Twe ntyfive to 72% of tubers we re in fested at
harvest in th e eva luated fi eld s. Prepupae
occur from April to September; pupae from
M ay to October; and overw in tering ad u Its
from Jun e to November (F igure 3).
14 8

Potato

P. piercei

P. vorax
43
16
14
17
18
58
50
66
282
168
193
463
119
374

Natural Enemies
Contrary to what mi ght be ex pected of a
native pest comp lex, natural enemi es for
APW are scarce. For man y yea rs tens of
thou sa nds of larvae have been co ll ected
and observed for th e occurrence of paras itoid s. Non e have been recorded.
Predators are also restri cted in number,
occurrence, and effecti ve ness. Th e on ly
refere nce to effecti ve predators in th e
literature is an ant of the genus lridomirmex
(Ga rm endi a, 196 1 ). That has not been
confirmed in th e areas where CIP has
developed integ rated pest management
programs, in cluding those where in secticides are not used or are used on a limited
sea le. Among vertebrates, toads and birds
act as weevi l predators. But toad popul ations in potato fie ld s are so srnall th at th eir
effect is mea ningless. Birds have been
observed look in g for weev il larvae in
rece ntly harvested potato field s. Aga in , bird
popul ations are low and th eir effect on
weev il populations is limited. A practi ca l
use of this ex peri ence has been the utilization of chi ckens as effecti ve predators at
harvest or shortl y thereafter.

Weevils/1-M trap
70

(%)

100
90

!\

60

l(ll

50

80

',

70

1.1

"\

40

60

50

30
40
20

30
20

10

10
0

0
Oct

N ov

Dec Jan

Feb

Mar

Ap r

May

J un

Ju l

Aug

Sep

Oct

Nov

-i--

Adult emergence (mv)

-O-

Migration (plant)

-+-

Tuber infestatio n (%)

__...._

Prepupae (%)

---*"-

Pupati on(%)

---7!E--

Overwintering ad ult (%)

Figure 3. Seasonal history of the Andean potato weevil (Premnotrypes vorax) in Cajamarca, Peru. 1997.
Some so il -i nh ab itin g ground beetl es
(Ca rab id ae and Ten ebrionidae) are general
predators (Harpa lu s turmalinus, Hylitus sp.,
and Metius sp.). Ground beet le have a
limited chance to capture neonate larvae
because larvae are exposed for on ly a short
tim e before they penetrate the so il. Although groun d beetl es readily eat weevil
eggs when th ey are offered, in nature th e
eggs are in sid e straw or other plant debris
and safe from predation.
The o nl y pathogen ab le to develop
disease in APW is the fungus Beauveria
brongniartii. Conditions favorable for the
ep izoot ic are heavy rains and co ncentrated
weev il populations . Th e multipli ca tion and
practical use of th e pathogen is under study.
Host Plant Relations of Potato Weevils
Th ere are two, co mmonl y accepted
misinterpr etat io ns related to th e host range
of the two groups of weevils present in
potato fields.

1. It is believed that members of the APW

com p lex can survive an d multiply in
other ho st plants present in potato fi eld s
or rota tion field s.
2. It is also believed th at othe r weevils th at
inhab it potato fie ld s (as adults in the
fo l iage o r larvae in the so il ) ca n damage
potato tubers.
A ser ies of tests have demonstrated that

Premnotrypes spp. cannot surv ive as larvae

on plants other than potato. That has been
demons trated wit h P suturica llus, P vo rax,
and P latithorax o n oca (Oxa lis tuberosa),
ullu co (U llu cus tuberosum), and mas hu a
(Tropaeolum tuberosum) (Vera et al. , 1994).
Adult s ma y eventuall y feed on leaves of
several pl ants, but in no case ca n a l ife
cycle be completed. As fo r other spec ies of
weevils presen t in potato fields, our tests
indi ca te that they cannot damage potato
tu be rs. Th ese weev ils surv ive by feeding on
weeds in potato fields (Tab le 2). Adults feed
on fo li age; larvae li ve in th e soil and feed
on roo ts (Lazaro, 199 9).

CI PProgrom Report 1997-98

149

Table 2. Host plant relations of the Andean potato weevil and other species of weevils commonly found in potato
fields . Huancoyo, Peru, 1998.

Weevil species

Plant species'

Premnotrypes
suturicallus

Adioristus sp.

(-)
(-)
(-)
(-}
(- )

(+ )
(+)
(+)
(+ )
(+ )

(+ )
(+)

(+ )
(+ )

( +)
(+)
(- )

(+ )
(+ )
(-)

(-)
(- )
(+ )

(+ )
(+ )
(-)

(-)
(+ )
(-)

(- )

Ki kuyo (Pennisetum clandestinus)

Lengua de vaca (Rumex crispus)
Puka puncho (Rumex acetocel/a)
Garbo nzo (Astraga/us sp.)
Yuyo (Brassica campestris)
Aguja-aguja (Erodium cicutarum)
Sara-sora (Gamochaeta sp.)
Potato cv. Yungay

Scotoeborus sp.

Naupactus sp.

(-)
(- )

a. ( + ) = host plant, ( - ) = nonhost plant.

Conclusion
Controve rsial aspects in the literature
related to th e ta xonomy and bionomi cs of
th e APW comp lex ha ve been c larified wi th
this work. Rev iewed taxo nomic keys w ill
help to id entify APW species. Th e upd ated
geographi ca l distribution of the species life
cyc les and host range w ill help to interpret
th e evo lution of th e weev il compl ex.

Acknowledgement
The aut hors are gratefu I to Professo r
G iovann i Onore, Departamento de
Biologfa, Pontifi c ia Universid ad Cat6 1i ca
del Ecuador for re view in g th is paper.

Literature Cited
Agost ini E. and H . Vil te. 1982. Estud io
mo rfol6gico y biologfa de Rhigopsidius
tucumanus Hell er (Co leoptera:
Curcu lionidae) plaga de la papa en la
Queb rada de Hu ama hu aca y Pun a de la
prov incia de Juju y. Rev. Ag ron. de! Nor
Oeste A rgentin a 19:1-4.
A lca la, P. 19 79a. N ueva espec ie de!
Genero Premnotrypes piercei 1914
(Co leoptera : Curculionidae). Rev. Per.
Ent. 22( 1):63-64.

150

Pololo

Alcal a, P. 1979b. El genero Premnotrypes

pierce (Co leoptera: Curculi o nid ae) en el
Per(1. Resu menes XX ll Co nve nc i6 n
Nacional de Entomo logfa held 4-9
Noviembre in Hu ancayo, Peru.
Alcaza r J. 1976. Biologfa y
Comportam iento del go rgojo de los
A nd es Premnotrypes suturica llus Ku sc hel
(Co l. :Curcu li onidae). In g. Agr. Th es is.
Uni versid ad Naciona l de! Centro de!
Peru , Hu ancayo, Pe ru. 80 p.
Cal vache, C. H . 1986. Aspectos biol6gicos
y eco l6g icos de! gusa no bl anco,
Memorias de! curso sobre el co ntro l
integrado de plagas de la papa. Internation al Potato Center and lnstitu to
Co lomb iano Agrop ec uario. Bogota,
Co lomb ia. p. 1 8-24.
Carbajal , C.P. 1992. Biologfa, flu ctuac i6n
de pob lac i6n y co ntro l del gorgojo de los
Andes Premnotrypes latithorax en la
localid ad de Aguirre. Ing. Agr. Th es is.
Uni vers id ad Mayor de Sa n Simon ,
Coc habamba, Bolivia. 105 p.
Carhu amaca, J. and A. Tovar. 1987. Estudi o
de! c iclo biol6gico de Premnotrypes
piercei A lca la. Proceedings XXX
Conve nci6 n N ac iona l de Entom o logfa
held in Caj ama rca, Peru. Entom o log ica l
Soc iety of Peru , Lima, Peru.

Car rasco . 1961. Sistem ati ca y biologia del

gorgojo de los Andes Premnotrypes
latith orax Pierce. Rev. Per. Ent. 4(1 ):3042.
Delgado, M. 1975. Cu lti vo de la papa en la
sierra. ln fo rme Especia l No. 39.
Mini ste ri o de Alim entac io n, Direcc ion
General de lnvest igac ion, CR IA I, Lima,
Peru. p. 56-58.
Ga ll egos, P.G. 1995. Rela cio n entre la
poblacion de ad ul tos de Premn otrypes
vora x (Hustache) al ini c io del cultivo de
papa y el da11o de tub erculos a la
cosecha. Revista lat in oamer ica na de la
papa 7/8 (1 ). 122 p.
Garmendi a, A . 1961 . Observac iones so bre
un posible co ntrol bi o log ico de la
gusanera de la pap a depositad a en
alm acen. Re v. Per. Ent. 4(1 ):76-77.
G il, P. 199 1. Biologia y com porta mie nto de
Premnotrypes solaniperc/a Kuschel !
(Co l.: Curc ul ion idae). In g. Agr. Th esis.
Univers id ad Nac ion al del A ltipl ano,
Pun o, Peru. 72 p.
Kusch el, G. 1956. Revision de los
Pemnotripini y ad ic io nes a los Bago ini
(Aporte 17 sob re Col eo ptera
Curculionoidea): Bol. Mus. Nae. Hist.
Nat. Sant iago de Chil e. Chil e 26:187235.
Laza ro , A. 1998. Estudio de hospederos del
comp lej o go rgojo de los A nd es en la
Sierra Ce ntral. In g. Agr. Thesis.
Universid ad Nacional del Centro del
Peru , Hu ancayo, Peru. (unpub li shed).
Muno z, M. and J. Alcazar. 1998. Biologia y
compo rtamiento del Go rgoj o de los
And es Premnotrypes vorax. Res um enes
del XL Conve nc ion Nacional Entomogia
del PerC1 , Nov. 1998. Li ma, Per u.

Pi erce, D.W. 1914. New potato weev ils

from Andean South Ame ri ca. J. Agr ic.
Res . 1(4):347 -352.
PROINPA. 1994. ln for me Anual.
Compend io.1993-1994. Cochab amba,
Bolivia. 40 p.
Tisoc, I. 1989. Cic io biologico de
Premnotrypes /atithorax bajo
cond ic iones de laboratorio en el Cusco.
Rev. Per. Ent. 32 :89-92.
Va lencia, L. and Ch. O'B rien . 1976 . The
Andean potato weev ils (Co leoptera:
Curcu li onidae) in the Choco n area of
Peru. Intern ational Potato Center, Lima ,
Peru . 8 p. Broc hure.
Va n Emden, F.I. 1952. On the taxonomy of
Rhinchophora larvae : Adelognatha and
Alophina e (lnsec ta:Coleoptera). Pro c.
Zoo l. Soc. Lond. 122(3) :651- 795.
Vera, A ., J. Alcazar, F. Cisneros, and W.
Cata lan. 1994. Eva luacion de la
suscept ibil id ad de lo s principa les
tub erc ulo s And in os al gorgoj o de los
Andes. Res um enes XXXV I Convencio n
Nac iona l de Entomo log ia, Iquitos, Peru .
Wibmer, G. J. and C.W. O'B ri en. 1986.
Annotated c keck li st of the weevi ls
(Curculionida e sens u lato) of South
America (Co leo ptera: Curculionoid ea) .
Mem. Amer. Entomol. In st. 39(1- 16): 1563.
Wille, J. 1952. Entomolog ia Agricola del
Peru. Estac ion Agrico la del Peru .
Esta c ion Ag ri co la de la Molina.
Min isterio de Agricu ltura . Lim a, PerC1.
54 3 p.
Yaya, V. 197 1. Mernori a A nual. Sub
Direcc io n de ln vestigacio nes
Agropecua ri as Zona Ag rari a X.
Min isterio de Ag ri cultura , H ua ncayo,
PerC1.

CIP Program Repo rt 1997-98

15 l

Population Dynamics of the Andean Potato Tuber

Moth, Symmetrischema tangolias (Gyen), in Three
Different Agro-ecosystems in Peru
M. Palacios1,

J. Tenorio 2,

M. Vera\ F. Zevallos 3 , and A. Lagnaoui 1

Andean potato tub er mo th ,

5ymme trischema tangolias, is o ne of the

most ser io us pests of potato, 5o lanum
tuberosum, in Peru and Boli v ia. It is
co nsid ered a pest of economic importance
in potato stores in both countri es (Ewe ll et
al. , 1994; Pal ac ios and Cisneros, 1997;
Arena s et al. , 1998) . Dam age is ca used by
larvae, w hi ch min e tub ers making th em
un suitable for hum an con sumptio n or for
seed. To preve nt damage to tubers in
sto ra ge, fa rm ers use hi ghl y to xic insectic id es, th ereby jeopardi zing th eir hea lth and
th e env iro nm ent. 5. tangolias is also a
potato pes t in the fi eld where larva e mine
stems and tubers . In dry years, a high
in c id ence of the pest in th e fi eld ca n redu ce
tuber produ ct ion. Know ledge o f Andea n
potato tuber moth surv iva l and behav ior in
fi eld and sto rage is mini mal . The objecti ve
of th is research was to stud y the population
dynami cs of 5. tango lias, its seaso nal
trend s, li fe cycle in storage, surviva l, and
behav ior.

Materials and Methods

Th e study was ca rri ed out in three Andean
com muniti es in Peru : Urqui ll os (C usco,
so uth ern highland s), Carhuap acc ha
(Huancayo, in th e ce ntral hi ghl and s), and
Santa Clo tild e (Caj amarca, in th e northern
hi ghland s) .
Urquillos is situ ated at 2,800 m. Potato is
grown year- round. Durin g the rainy season
(October-May), potato is grown at alti tudes

I CIP, Limo , Peru.

2 CA RE , CajomarcJ , Peru.
3 Universidocl San Anto nio Abad, Cuzco, Peru.

of from 3,600 -3 ,800 m. Du ring the dry

seaso n, pota to is gro w n under irri gation in
rotation w ith ot her cro ps at lower altitudes
(2,800 m). Potato production in this reg ion
is for both co mm erc ial use and local
co nsumption .
Carhu apa cc ha is situ ated at 3, 175 m.
Cult ivated areas exte nd up to 3,800 m.
Irrigatio n is ava ilabl e at lower eleva ti ons
a I low ing fan11 ers to produce other commerc ial vege tabl es . Potato is produ ced o nce a
year durin g th e rainy season (O ctobe1-M ay)
in the upper lands of the commun ity (3,5 00
- 3,800 m.). Potato produ ction is commercia ll y or iented and insecticid es use is
intensive.
Santa Clot ild e is situ ated at 3,020 rn .
Pota to is grow n fr om Jun e to Dec embe r and
fr om Janua ry to May in rotation w ith w hea t,
bar le y, and peas . Produ ction is basi ca lly fo1
home co nsumption .

Populations dynamics
Seasonal patterns of PTM abundance.
Moth popul ation s were monitored in potato
fi elds and stora ge sites using ex perim ental
sex ph ero mo ne traps in Santa Clot ilde and
Urquil los, and light trap s in Carhuap acc ha.
Five field s, two in Santa Clotild e and three
in Urqui l los, we re selected for the fi eld
tri als. Fi ve stora ge sites we re monitored,
two in Carh uapa cc ha and three in
Urquillos. One trap by each fi eld and store
wa s used . Modifi ed water pan traps (Bacon
et al., 1976, Raman and Booth, 1983) we re
placed at ground leve l. Rubb er septa
impregnated wi th a 2: 1 mixture of (E,Z) 3J
tetradeca d ienyl acetate and (E)-3tetradece ny l acetate (Gri epink et al., 1995)
CIP Program Report 199798

153

susp ended in eac h trap. A dul t moths

trapped we re co ll ected wee kl y throughout
the year. Match catch data was ex p1essed
as the average number of moth s/ trap/n ight.
Life cycle in stores
Pupae co ll ected from the fields were
used to start a moth co lony to study its li fe
cyc le in rustic potato stores. Co uples of
emergi ng adults we re placed in 0.5 L plastic
co ntain ers (12 cm diam, 6 cm high) and
covere d w ith a 50-rnesh clo th held in place
by a rubber band). Eggs lai d by females
were col lected daily. Newly hatched larvae
were released on potato tubers to comp lete
their developme ntal cycle reaching th e
pupa and adult stages. A dults were sexed
based on abdomen size and then iso lated in
egg- laying conta in ers. Observations we re
mad e for one yea r. Data were collec ted
over the pre-ov ip osition and ovipositio n
per iods from adult emergence until laying
of the last egg batc h. Natural morta lity was
evalua ted for eac h developmental stage .
Survival
Fi eld s and stores we re sampled to assess
damage to potato plants an d tubers, and to
determine th e fate of th e surviving popu lation. A t harvest, 13 fi el ds we re sampled: 4
in Urquil los, 5 i n Carhuapacc ha, and 4 in
Santa Clotilde. Each field was di v id ed into
5 represen tat ive sect ions . A 1 m 2 area was
eva lu ated in eac h sect ion fo r number of
pl ants and for moth damage in sterns and
tubers. Four weeks after harvest, five 1 rn 2
areas in the same fields we re sampled for
number of larvae, and pupae in sterns,
tubers, and soil.

At the encl of sto ra ge period , damage in

tubers was eva lu ated in four stores in
Urqu ill os and five stores in Carhuap acc ha.
A sample of 100 potato tubers was eva luated and sco red as hea Ithy o r darn aged in
eac h store . A second eva luatio n, co ndu cted
1 month after potato tubers were taken from
storage, consisted o f three 1 m 2 areas in the
roof an d wa l ls in th e sto re. Numbers of
pupae and adults in cracks and crevices
were record ed.

154

Poloto

Behavior
Egg-laying preference. Th e ovi pos iti on
prefere nce of the moth was i nvestigatecl on
three levels . Th ese in cl ud ed prefe renc e of
soil vs plant, pa rts of th e p lant (lowe r,
midd le, and upper), and the microhabitat
prefe rr ed (ax ill ae, sterns, leaves). We also
evaluated th e effect of potato deve lopment
stages (vegetative growth, tuber initiation,
tuber bulking) on oviposition. The ex periment was carr ied out in the green ho use in
th e C IP station in H uancayo . Twelve potato
pl an ts in pots (cv Yungay) were used per
deve lopment stage. Potato plants were
pl anted seque ntiall y eve ry 30 cl so that
plants of d iffere nt ages could be infested at
the same time. Each group of plants was put
in cages w here fifty mated couples of 5.
tango lias we re relea sed. After 10 cl, plants
were cut at the base of the ste rn and
evaluated to dete rmin e number of eggs laid
in the upper, middl e, and lower t hi rd of the
plant. Once all plant mate rial was removed
from pots , potato sl ices were put on the so il
surface to determine the numb er of eggs
laid in the so il. For this exper im ent, a
factoria l des ign wi th three factors (2x3x3)
was used . D ata on the number of larvae
per plant was analyzed by ANOVA and
means were com pared using Fis her's least
significa nt difference (LSD) procedure.
Plant infestation preference. Th e
obj ective of th e in festat ion prefe rence
exp eriment was to determi ne the locat ion
of initi al plant damage (lowe r, middl e, or
upper leve l) caused by neonate larvae
according to the stage of development of
th e potato plant (vegetat ive growth, tub er
ini tiation, tube r bulkin g). This ex peri ment
was ca rri ed out in th e green house at th e
CIP stat ion in Huan cayo . Twenty potato
pl an ts (cv. Yungay) we re planted in pots fo r
eac h developmental stage; th e plan ts we re
p la nted seque nt ial ly eve ry 30 cl so that
p lants of differe nt stages co uld be in fested
at the sa me ti me. Each 20-plant group was
subdivided in to 4 subgroups of 5 plants;
each plant was a repli cate. In each gro up
50 neona te larvae of 5. tangolias we re
rel eased on so il , and o n th e lowe r, midd le,
and upper parts of the p lant. Th e eva lu ation

was made 15 d later. For thi s experim ent a

compl etely randomi zed des ign was use d
w ith three fac tors: A = deve lopm ent stage,
B = zon e o f release of neo nate larvae, and
C = zo ne w here initi al damage w as fo und.
The num be r of larv ae d amag ing th e pl ant
was analyzed by ANOVA and mean s w ere
comp ared usin g Fish er's least si gnifi ca nt
difference (L SD ) procedure.

Tuber infestation in the soil. Th e depth

that neo nate larvae penetrate the soil to
infest pota to tubers w as d etermin ed by
pla cin g tu be rs at 5 depth s (0, 3, 5, 10, and
15 cm) in boxes containin g soil. Each
tre atm ent was repli ca ted 3 tim es w ith
exposure to art ifici al in festation with 25
neonate larv ae of 5. tango lias. Tub er
damage was assessed after 35 d. D ata we re
subj ected to ANO VA and mea ns we re
se parated by th e LSD test.
Pupation in the soil. To determin e th e
depth at w hi ch pup ati o n occ urs, three
calibrated boxes (5 0 x 50 x 50 cm ) co ntainin g fi eld so il we re used . Fi fty full y-gro w n 5.
tangolias larvae we re released in eac h box .
Penetrat ion was evalu ated at 45 d by
checkin g so il layers in in crements 1.0 cm
thick. D ata we re subj ected to A NO VA and
mean s we re compared usin g Fish er's least
signifi ca nt differenc e (LSD) procedure.
Flight height. Flight height for monitorin g
and manage ment was dete rm in ed by usin g
stick y traps (transparent pl asti c sheets 1 m
w ide x 2 m hi gh smea red w ith Tan gle Trap
(Tan glefoot , Grand Rapids, Ml ) in potato
fields of two communiti es in Cajam arca.
The traps we re marked at 20 cm -interva ls.
Evalu ati on wa s based on eig ht re co rd s/trap,
taken wee kl y. Catch d ata we re subj ected to
A NO VA, mea n se parati on wa s done usin g
Duncan' s new multipl e range test (DMRT).
Results and Discussion
Seasonal trends
Urquillos. Po pul ati on data tak en from
October 199 4 until O ctober 1996 sh ow ed
5. tangolias was prese nt in potato field s,

eve n in th e absence of the crop (Fi gure 1A).

Durin g th e 1st yr, th e number of adult
moth s reac hed 505/trap/wk betwee n
October and Febru ary w hen th e potato
crop was pre sent. Th e number dec reased to
17/trap/w k betwee n M ay and Septemb er in
th e abse nce of th e c rop. During th e 2nd
year, Dece mber 1995 - December 1996,
th e moth populati on flu ctuated in li ke
mann er, but numbers we re lowe r.
M ot hs were prese nt in potato stores
throu ghout the year (Fi gure 1 B). Minimum
ca pture was 23 moth s/trap/w k; max imum ,
186. Moth s w ere prese nt throu ghout th e
yea r beca use tubers we re stored yea r-round
as a res u It of th e two-crop season s/yr in
Urquill os ; th e first at 3, 500 m and th e
seco nd at 2, 800 m alti tud e.

Carhuapaccha. Th e pest was monitored

in farm er's stores usin g li ght traps from
De ce mber 1995 until December 1996. The
moth was present th rough out th e yea r, but
popul ations w ere low between Janu ary and
May w hen no tubers we re stored. Popul ation s started to in crease in June, pea kin g
first in August (30 mo th s/trap/wk) and th en
again in Nove mber (145 moths/ trap/w k).
Th e seco nd peak co rr esponds to th e end of
th e storage period . D amage cau sed by 5.
tango lias in potato tubers increased fr om
0.8 % afte r 1 mo of storage to 90 % 3 mo
later (e nd of the sto rage period ) (Fi gure 2).
Res ul ts suggest that the pest could have
entered th e stores on or w ithin tub ers as
eggs or ea rl y larva l in stars. D eve lopm ent of
a seco nd ge neration in stores is beli eved to
be respo nsibl e for th e damage in tubers.
Santa Clotilde. M onitorin g farm ers' fi elds
from Febru ary 1996 to May 199 7 showed
th at the moth wa s prese nt throu ghout the
ye ar (Fi gure 3). Popul ation peak ed twi ce
durin g th e yea r (d ee p drop durin g M arch in
first peak w as du e to low temperatures and
rain w hich affected adult moth acti vi ty).
Both pea ks correspo nded to th e fl oweri ng
stage of th e crop. M ax imum capture in the
first po pul ation pea k w as 807 moth s/t rap
per week and 596 moths/trap per wee k in

CIP ProgramReportl99798

155

A
Moths (N 9/trap/wk)

600
Field at 2,800 m

500
400
300
200

100

o t__~~~~~:::::~~~~~~~~~...._.,_===~~~

IPotato

II

Vegetable

1~~

Potato

11

Vegetable

1~s

Potato

1~s

8
400
300

Store at 2,800 m

200
100

0 J

>3,600

N D

M A M JJASOND

2,800
1994

I >3 ,600 11

1995

11

M A M J

A S 0

GM~

2,800

1,996

~~~~~~~~~~~

Figure 1. Andean potato tuber moth population in Urquillos, Cusco, Peru , 1994-1996.
the second peak. Adult populati o n drastica ll y decreased when there was no potato
cro p in the field due to lack of food and low
temperatures at night (frost).

Life cycle
Studies show that, in rustic storage, 5.
tango lias completes its life cyc le in 105
da ys at 13C (Santa Clotild e and
Carhuapaccha) and only 70 days at 17.4C
(Urquillos). Egg la y in g capacity was not
significantly different (P=0 .05) betwee n
loca tio ns, averaging 178-185 eggs/fe male.
Accumu lated natural morta lity of immature
stages was high averaging 75 % in Santa
Clot ild e and Urquillos and 57% in
Carhuapaccha (Tabl e 1 ). 5. tangolias was
found to develop three to five ge nerations
in a yea r (Tab le 1 ).
156

Pololo

Pest survival
In Urquil los, potato stems and tub ers
sho we d no dama ge at harvest. Th at mi ght
be the result of the adequate use of c ultural
practices such as irrigation , hilling up , and
timely harvest. In addit ion , stores close to
the fields may be a source of food and
refu ge for the pes t throughout the year.
Th at m ay have had an influence on th e low
!eve l of damaged tubers in the f ield .
Sampl es fro m th e field that had been
rotated (U rqu i I los ) averaged 6.4 pota to
tubers/ m 2 , but th ey we re not dam aged by 5.
tangolias (Tabl e 2). Observatio ns from th e
remainin g population in the field showed
the presence of larvae and pupae in dri ed
tubers and stems 3 months after harvest.

Moths
(N2/trap/wk)

Tubers
damaged(%)

160

100

11111!1 Tubers-+- Moths

80

120

60

80
40

40
0

20

J
I

N
Potato

1996

11995

Figure 2. Andean potato tuber moth store population and tuber damage, Carhuapaccha, Junin, Peru, 1997.

Moths (No/trap/wk)

1000
Flowering and maturity

Flowering

800

Hilling up
Harvest

Flowering

Harvest

400
200
0

Potato
M
A
M
I _M_i_no_r_s_ea_s_o_
n _cr_o~
p~
planting ~

. Main season crop

1996

Figure 3. Andean potato tuber moth field population in Santa Clotilde, Cajamarca, Peru, 1996-97.
Moths re maining in storage facil iti es
whe n there are no stored tubers ensure pest
persistence in stores. At higher altitudes,
the moth is introduced to potato fields
every year with the seed.
Evaluations in the fields and stores in
Carh uapacc ha showed that the pest had
spread from 3,020 to 3,800 m. Moths are
transported to the stores in harvested tubers
and returned to the fields in seed tubers.

Prevalence of the pest in the absence of a

crop suggests that the pest survives in crop
residu e between field plantings. Tubers kept
by farmers for their own consumption
usually serve as a food source for the moth
to complete its cycle.
Behavior
Egg-laying preference. The moth I ays
eggs almost exc lusively on the plant (92%),
independ ent of crop age (Figure 4). Egg-

CIPProgrom R'po1t 1997-98

157

Table 1. Life cycle (in days) and natural mortality(%) of And ean potato tuber moth in stores, Peru.

Department/Community
Av. Temp. ( Q
Stage
Egg
Larva
Pupa
Total
Longevity
Oviposition period
Egg/female
Mortality
Egg
Larva
Pupa
Total
Generations/year
0

Cusco/Urquillos
17.4

Junfn/Carhuapaccha
12.6

Cajamarca/Santa Clotilde
13.0

9
40
22
71
24
25
184

16
58
32
106
23
23
178

17
57
31
105
33
25
185

31
28
15
74
5

4
28
26
58
4

12
46
18
76
3

Table 2. Damage and remain ing population of Andean potato tuber moth in th e field and in stores, Peru.

Department/Community
Fields
Fields with da maged plants(%)
Plants withdamagedstems(%)
Damaged tubers(%)
Remaining tubers/m1 (no.)
Rema ining population/ m1 (no.)
In stems
In tu bers
In soil
Stores
Damaged tubers(%)
Remaining popu lation (no.)
In roofs
In wa lls

Cusco/Urquillos

Poto to

Cajamarca /Santa Clotilde

0
0
0
6.4
0
0
0
0

100.0
61.6
2.1
4.1
16.9
11.8
3.3
1.8

90.0
82.0
21.8
4.3
9.4
4.5
4.9

88.0
130
8.0
5.0

90.0
14.8
5.8
9.0

80.0

lay ing distri buti o n ana lys is showed that the

midd le th ird of th e p lant is th e preferred
area fo r ov ipos ition and that adu lts se lect a
m inohab itat fo r egg- lay ing. The preferred
m ic ro hab itat is the ax il la (Fi gu re 4). Eggs
laid in th e ax il Jae of the pla nt are protected

158

Junfn/Carhuapaccha

from adve rse env iron men ta I factors, hence

ensuring penetrat ion of the larvae into
stems.

Plant infestation preference. Si nee S.

tangolias moths prefer to lay eggs 0 11 th e

middl e part of the plant, th e analysis was

co nce ntrated on larvae released in that
area. Th e res ults show th e effects of plant
development stages on larva l distribution .
During vegetative growth, damage was
co nce ntrated on the middl e leve l of the
plant. Damage was limited to the lower part
of the plant during tuber initi at ion and tuber
bulkin g. (Fi gure 5).
Tuber infestation in the soil. Larvae
penetrated the so il to a depth of 5 cm.
Tuber damage decreased w ith depth, from
80% at ground level to 40 % at a depth of 5
cm. Tub ers at deeper levels are not dam aged by moths (Figure 6A), probably
because of the moth's limited penetration
ab ility and its low egg-laying rate in the
so il .
Pupation in the soil. Sixty-three percent
of larvae pupated in th e soil and 37% in th e
tuber (Fi gure 6B). Under th ese c ircumstances, damaged tubers remainin g in the
field are pest reservoirs and also th e means
of transferring infestation from field to store
and back to field. Pupation in so il occurred
from 0.1 to 3.5-cm depth, with 82% of
pupation in the first 2.5 cm (Figure 6C).

Oviposition preference
1 00~~~~~~~~~~~~~~~

80
60
40

20

Soil

Lower
Plan t

Conclusions

Adults of 5. tangolias were found in the

fields and in stores in th e three co mmuniti es und er study. Th e pest damaged potato
ste ms and tubers . Dama ge in harvested
tubers was more important in Santa
Clotilde, a community where potato is
grown twice a year. Tuber damage in stores
was significant in all study sites. Th e moth
population can survive in crop res idue in
the absence of a host plant. Thus, timely
san itation to collect and destroy plant
res idues and volunteer plants is an impor-

Ax illa

Leaf

Stems

Figure 4 . Oviposilion preference of Andean potato

tuber moth in potato, Peru, 1998. (Jn each

group, bars with the same leller are not
significantly different [P <0.05, LSD
lest] .)

Larvae boring stems (No.)

25

Lower filiilll Medium

Upper

20
15

10

Vegetati ve growth

Flight height. The number of ad ult moths

captured in sticky traps showed th at adults
could fl y to a height of 200 cm . But most
fli ght activity (about 67%) was co ncentrated
betwee n 20 and 80 cm. Th at co uld be
exp lain ed by the adult preference for
ovipositing on the plant.

Upper

Medium

Tuber initiation

Tuber bulking

Figure 5. Plant age preference of Andean potato

tuber moth neonate larvae for infeslalion,

Huancayo, Peru, 1998. (Jn each group,
bars with the same leller are not significantly different [P<0.05, LSD lest] .)
tant management option. In storage, the
moth survives in crev ices and in tubers
stored for hom e co nsumption. Henc e,
cleaning and disinfecting storage areas is
needed to break the moth cycle to preve nt
re-infestation of new ly harvested and stored
potatoes.
The adult preference for laying eggs in
the plant and larv al preference for lower
and middle parts of the plant explain th e
low percentage of tuber damage and the
high in cid ence of th e pest in stems. Thi s
could be a practical monitoring tool to
CI PProgrom Report 1997-98

159

this pest. Preventing access to tubers lim its

tuber infestation.

Depth of tuber (cm)

lab

:1
0

!b
10

20

40

30

50

60

70

References
1

BO

Tuber damage{%)

Pupation (%)

Soil Depth (cm)

0.1-0.5 f------~I

o.s-1.s ,___~_1a

c=:::J ab

o 6-2 .s

tJ b

0.6-3 .5

~----~--~-~

20

40

60

80

Pupation depth(%)

Figure 6. Ability of Andean potato tuber moth

larvae to reach tubers at different depths,
and their pupation inside and outside
tubers, Cajamarca, Peru, 1997. (Fig. A and
C: bars with the same letter are not
significantly different [P<0.05 , LSD test];
Fig. 8: bars with the same letter are not
significantly different [P < 0.05, Student's
I-test].)

detect the presence of the pest in plants and

assess potential damage. Th e limited
penetration capacity of larvae into the soil
demonstrates the importance of planting
depth and hilling-up in th e management of

160

Potato

Arenas M .R ., C.J. Herbas, and C.J. Arnold.

1998. Biologia, control preliminary
distribuci6n de la polilla de la pap a
Symmetrischema tangolias en Boli v ia.
Asoc iaci6n Latinoamericana de la Papa
(A LAP )_ XVIII Reunion de la ALAP:
Compendio de ex posiciones.
Coch abamba (Bo livia). 9-13 Feb . 1998.
Bacon, O.G., J.N. Seiber, and G.G.
Kenn edy. 1976. Evaluation of survey
trappin g techniques for potato tubeworm
moths w ith chemical baited traps. J.
Econ . Entomol. 69: 569-72.
Ewel l, P.T., H. Fano Rodriguez , K.V. Ram an,
J. Alcazar, M. Palacios, and J.
Carhuamaca. 1 994. Farmer management of potato insect pests in Peru:
Report of an interdisciplinary researc h
proj ect in selected regions of the hi ghlands and coast. International Potato
Center (CIP), Lima, Peru. 77 p.
Griepink, F.C., T.A. van Beek, J.H. Visser, S.
Voerman, and A. de Groot. 1995.
Isolation , identification and synthesis of
the sex pheromone of Symmetrischema
tango/las. J. Chem. Ecol. , 21, 2003-

2013.
Palacio s, M. and F. Cisneros. 1997. Integrated m ana ge m ent for the potato tub er
moth in pilot units in the Andean Reg io n
and the Dominican Republic. lntern ationa I Potato Center Program Report
1995-96. Lima-Peru.
Raman , K.V. and R.H. Booth. 1983.
Evaluation of Technology for Integrated
Control of Potato Tuber Moth in Field
and Storage. Technology Eva lu;:ition
Series No. 1 0, CIP. 1 8 pp.

Assessing Bl-Transformed Potatoes for Potato Tuber

Moth, Phthorimaea operculella (Zeller), Management
V. Canedo, J. Benavides, A. Golmirzaie, F. Cisneros, M. Ghislain, and A. L a gnaoui1

Th e potato tuber moth (PTM) Phthorimaea

operculella (Z eller) is one of the most
damagin g pests of potatoes , Solanum
tubero sum L., in w arm te mperate and
subtropi ca l clim ates (Fenemore, 1988) .
Durin g th e gro w in g seaso n, the potato tu be r
moth min es into foli age and stem s, w hil e at
plant senesce nce adults lay eggs in so il
cracks and on exposed tubers . At harvest,
tubers do not always show sig ns of damage
but may harbor eggs and ea rl y in star larvae.
Thi s may res ult in severe posth arvest losses
in the abse nce of ad equ ate control mea sures. Th e PTM is th e sin gle mo st si gnificant in sect pest of potato (fi eld and sto rage)
in North A fri ca and th e Middl e East (Fu gl ie
et al., 1992) . In th e abse nce of co ld stores,
farm ers mu st rely on chemi cal in sec ti cid es
to protec t tubers from tuber moth damage
durin g t he storage peri od (Roux et al. ,
1992). Storage losses reported from Indi a
vary from 30 to 70% (Saxe na and Riz v i,
1974). In Latin Am eri ca , a co mpl ex of three
tuber moth species is kn own to infest
potatoes . Farm ers routin ely appl y chemi ca l
in secti cides to protect th eir crop s (Raman,
1988). Thi s reliance o n chemi ca l in secticides not onl y increases producti on cos ts
but also has serious delete rious effects on
human hea lth and th e environmen t.
Fa ced w ith thi s se ri ous pestic ide
probl em, sc ientists started to seek safe r
altern ati ves su ch as th e use of Bacillus
thuringiensis to repl ace toxic chemi ca ls in
their integra ted pest manage ment (IPM)
strateg ies . Protein s of spore-forming so il
bacterium are tox ic to larvae of different
orders of in sects (Lepid optera , Diptera,

1 CIP, Lima, Pe ru.

Coleoptera , Hymenoptera, Homoptera,

Orth optera, and M al loph aga) and to nematod es, mites and protozoa (Feitel son et al.,
199 2) . Spo res of 8. thuringiensis are a Iready
w idely used as bi opesti cid e in co mm ercial
agri cu Itu re to contro l several key pests.
Howeve r, fi eld appli cat ions of th is bi olog ical
pest icid e have not alwa ys provid ed a costeffect ive control of in sect pests. A more
effic ient way to deli ve r Bt can be ac hi eved by
geneti ca ll y enginee ri ng c rop plants to
produ ce th e acti ve co mponent of thi s
biopesticid e. Th e cry stal protein Crylll A was
expres sed in potato and provided res istance
again st the most se ri ous defoli atin g pest,
Leptinotarsa decemlinea ta (Say) (Adang et al.,
1993; Perl ak et al., 1993). Simil arity, th e
express ion in potato of the crys tal pro tein
CrylA(b) res ulted in res istance to pota to tuber
moth (Pefe ro en et al., 199 0; Jansens et al. ,
1995) .
At CIP, we have deve loped potato tuber
moth res istant clon es suitable for most
agroeco log ies of deve loping countri es
w here PTM is a severe pest. Th e Bt techno logy (gene con stru cts) w as acquired from
the Belgian company, Pl ant Genet ic
Syste ms (PGS). Th e in se rtion and ex pression of th e cry!A (b) ge ne is reported here as
we l I as its effect on potato tuber moth in
laboratory and fi eld ex perim ents. Thi s lead s
us to suggest future steps th at should be
take n to effectively and safel y depl oy such
culti va rs in developin g country.

Materials and Methods

Transformation, regeneration,
and selection
Two Agrobacterium tumefaciens strains
w ere used for potato transform ation

CIPPrngram Report 1997-98

16 1

C58Cl Ri fR(p GV2260)(pTFD1 6) and

C58C l Ri fR(pGv2260)(pTRVA 1) co ntainin g
trun cated cry/A (b) gene fro m Bacillus
th uringiensis subs p. berliner 1 715 (H ofte, et
al. 1986 ). Th ese Bt genes are un der the
co ntro l of a sin gle (refe rred to as Bt-1 ) and
do ubl e (referr ed to as Bt- 2) 35s pro moter,
res pecti ve ly.
Potato pla ntlets we re pro pagated in
Mage nta boxes (S igma) by subcu ltu re of
single leafy no de cu tt ings on sol id
Muras hi ge and Skoog (M S) basa l medi um
(G IBCO ) suppl eme nted w ith v itam ins, 3.0%
sucrose and sol id ifi ed w ith 0.3% ph ytagel
(refer red to as MSA medium ). Th e mediu m
was sterili zed at l 20C fo r 20 m in. Cultu res
w ere grow n in an in c uba ti o n ro om under
1 6 ho urs ph oto-peri od at 20 00 lux, 22-24C
and 70% relati ve humid ity. Potato transforma tio n was cl o ne fol lowi ng th e meth od of
De Block (1988 ) w ith seve ral mod ificatio ns.

DNA extraction and hybridization

Ge no mi c D NA was iso lated fro m leaf
tiss ues fo ll ow in g th e meth od of Dell aporta
et al. (1 983). Prese nce of th e Bt ge ne was
determin ed by So uth ern bl ot analys is usin g
th e Eco RV fr ag ment of 650 bp fr om pTFD 16
p lasmid as a pro be (Sambrook et al. , 1989).

CrylA(b) protein determination

M o noc lonal anti bodi es specifi c aga in st
the 8 . thuringiensis sub sp. berliner 1715
c rys tal prote in , prov id ed by PGS, were used
to quant ify by ELI SA Cry lA(b) protei n in
so lu b le pro tein extracts of tube rs. Total
p rote in was determ in ed using a Bio- Rad
protein assa y.

Leaf and tuber bioassays

G ree nh ouse experime nts we re condu cted o n 3 d iffe rent cul tivars (Sangema,
Cru za 14 8, and LT-8) over thre e yea rs.
Trea tm ents co rres pond ed to th e transformed
lin es of eac h cu ltiva r. A tota l of 85 differe nt
lin es have bee n tested agai nst PTM, 31
transgeni c lin es of Cruza 14 8; 37 transge nic
lin es of LT- 8, and 17 lin es of San gema.
Trea tm ents we re ex posed to 25 neo nate
PTM larvae. Eac h trea tm ent w as re pli cated

162

Potcto

10 tim es . Eva luat ions we re made 7 and 10

cl ays after in festati on w here larva l morta li ty
was recorded.
In 1994, harvested tubers fro m d ifferent
Cruza 14 8 lines (25 ou t 31 lin es), we re
di v ided into three gro up s of 50 g eac h
represe nting three repl ica ti o ns fo r tu ber
assays. Eac h trea tment (SO g of tuberlets)
was ex posed to 10 neonate larvae.
In 1997, harves ted tube rs from the
di ffere nt Sa ngema lin es (14 o ut 17 lin es)
were used in tuber assays. Treat ments
cons isted of a sin gle tuber (Ca. 30 g)
rep licated six tim es. Each treatment was
exposed to five neo nate larvae.
Tests were ca rri ed out in an in sect massrearing room und er co ntro lled temp era ture
an d hu m idity conditio ns (25 2C and 70
5% re lative humi d ity) . Larva l mo rta lity
was reco rded after 25 days and ex presse d
in percentage.

larval development
Th e b ioass ay co nsisted of test in g th e
effects of transform ed potatoes o n larva l
deve lopme nt. Tub ers of three transge ni c
li nes of the c lo ne LT- 8/B t2 (44.9, 34 .1 , and
34.2) pl us an un tra nsformed co ntro l were
ex posed in a sli ced for m to 30 neo nate
larvae eac h. Each trea tm en t was repli ca ted
5 ti mes. Larva e were coun ted, meas ured
and weig hed dai ly un til they d ied o r
pu pate d.

Ston1ge
H arves ted tub ers we re eva lu ated fo r
stability of Bt gene ex press io n ove r th e
length of the storage pe ri od (4 months) in
CIP 's Sa n Ram on Stat io n (m id -eleva tio n
j ungle). A free -cho ice test was co nducted
und er d iffuse -li ght stme co nditi o ns, usin g
10 lin es in 1995 and 27 lin es in 1996 of th e
c ulti var LT-8. Treatm ents co nsisted of 5
tu bers re pl ic ated fo ur tim es in a co mpletely
rand om ized block des ign.
A ll treat ments in c lud ed ex pos ure to an
arti fic ial PTM in fe stat io n. New ly emerged

adult tuber moths w ere relea sed in th e

sto rage area to simul ate natural infestati on.
Eva lu ation s of tuber infestati o n we re made
12 0 d ays after infestation usin g a 1-5
sco rin g sca le (1 =non-d amaged tuber, 5 =
100% d am age) Tubers w ere c ut to assess
d amage.
Field trials
Se lected transgenic lin es we re fi eld
tested to eva luate res ista nce to potato tuber
moth and assess their morph o log ica l
c haracteri stics. Fi eld tes ts were co ndu cted
under stri ct bios afety reg ul ati ons (C IP,
1993) . Tri als w ere carri ed o ut in two
loca ti o ns, San Ramon and Tac na (winter,
main potato season) in ord er to eva lu ate
res ista nce to potato tuber moth and nonta rget pests.

In Sa n Ramo n, field tri als co nsisted of

several treatm ents with three rep lica ti o ns in
a co mpl etely randomi zed bl oc k des ign.
Eac h ex perim ental unit was fi ve rows (of 15
pl ants), 4. 5 m lon g, 0 .9 m ro w-s pac in g, and
0.30 m pl ant-spacin g. With th e tra nsform ati o n proce ss, plants from th e sa me c lone/
cul tivar may present different mo rphologica l c haracteristics. To as sess thi s potenti al
change, pl ant growth w as eva lu ated by
pl ant height and leaf ratio (lea f length/ leaf
w idth) acco rdin g to Vau ghan and
Ga teho use (1991) .
In Tacn a (1997 -9 8), tri als co nsisted of
nin e treatments (representin g 9 li nes of LT8) . Experiment al plots w ere of the sa me
size and des ign as the fi eld pl ots of th e
summ er crop in San Ramon. Evalu ations
fo r potato tuber moth w ere made in th e
sa me mann er. Data were co ll ected also on
th e in c id ence of non-target spec ies (Tuta

absoluta, Liriomyza huidobrensis,

Russelliana solanicola, and Myzus persicae)
o n transge ni c lines. At harve st, tubers w ere
eva lu ated for PTM d amage usin g sampl es of
25 tubers per plot. Yi eld data was also
reco rd ed.

Results and Discussion

Development of transgenic potato using
developing country cultivars
Transfo rm at io n ex periments we re
initi ated w ith three important potato
cultivars adapted to a relatively warm
climate: LT- 8, w hi ch is immun e to PVX and
PV Y; Sange ma, imp o rt ant in Rwa nd a; and
Cru za- 14 8, w hi c h has moderat e res istance
to bacteri al di seases and is al so impo rta nt
in Rw and a and Burundi . A large se ri es of
tran sge ni c pl ants have been obtain ed
throu gh Agrobacterium-m ediated transformati on. In total, 4 2 transge nic clon es w ith
Bt-1 and 7 w ith Bt-2 were generated fo r LT8, 46 and 4 2 transgenic clones of Bt-1 and
Bt-2, res pectively, for Sangema cultiv ar, and
114 transge ni c c lo nes of Bt-2 for Cru za148.
Gene insertion and assessment of
copy number
A sampl e of transge ni c c lon es was tested
for the prese nce of th e Bt gene by South ern
blot an alys is. Th e ba nd pattern indi ca tes
ind epend ent in se rtion events (Fi gure 1). In
other ex perim ents th e number of th e Bt
gen es w as es ti mated to range from 1 to 3
copi es per hapl o id genome (Table 1).
CrylA(b) protein quantification
Th e amount of Cry lA( b) protein in tubers
was determin ed by ELISA tests of 10
tran sge ni c lin es o f LT-8 and 4 of San gema
cultures . A range of 1 to 11 ng of Cry lA( b)
per mg of tota l so lubl e protein wa s found
(Tabl e 1). In ag ree ment with other studi es,
no correlati o n w as found between Bt gene
cop y num be r and amo unt of CrylA(b)
protein ex pressed in tub ers . Ho w eve r, th e
amounts of Cry lA( b) prote in in three o ut of
four transge ni c cl o nes transform ed with th e
Bt- 2 co nstru ct carry in g th e Bt gen e w ith
doubl e pro moter 35 S we re much hi gher
than th ose ca rryi ng th e Bt-1 con stru ct
(sin gle pro moter) (Tabl e 1). Thi s re sult
suggests th at th e do ubl e 35s promoter leads
to hi gher protein acc umulation in potato .

CIPProgram Repmt 199798

163

'

.
"

[ii

it'!'

'

10 11 12 13 1 4 1 5 1 6 1 7 18 1 9 20

.. -

..
'

"'

Figure 1. Hybridization of genomic DNA of transgenic LT-8 lines with the Bt gene :Lane C: DNA of non-transformed
LT-8. Lanes l to 19 : DNA of transgenic LT-8 lines . Lane 20: positive control (isolated bt probe) .

Table 1. Molecular characterization of transgenic lines by quantification of CrylA(b) protein {means of 2 assays in
ng/ mg soluble proteins) and by copy number of cry/A{b) gene.

Clone
LT-8 / Bt-l /DST 24-2
LT-8 / Bt-1/DST24-3
LT-8 / Bt-l /DST 34-l
LT-8 / Bt-1/DST34-2
LT-8 /Bt-1 /DST 34-3
LT-8 / Bt-1 /DST 34-5
LT-8 /Bt-2 /DST 36-1
LT-8 / Bt-2 /DST 36-4
LT-8 /Bt-2/DST 36-5
LT-8 /Bt-2 /DST 36-8
Sangema /Bt-1 I DST 39-1
Sangema /Bt-1IDST39-2
Sangema /Bt-1 I DST 39-3
Sangema /Bt-1 I DST 39-4
ng bt toxin/mg protein.
b Not determined.

164

Potato

CrylA{b) protein'
5.64
2.50
1.99
4.86
6.26
3.55
10.87
1.18
6.00
9.37
2.28
1.94
4.71
ndb

cry/A{b) gene copy number

3
l

l
2

nd
2

2
2

Leaf and tuber bioassays

Eva lu ation s of larva l mortality on
transformed Cruza 148, at seven days,
showed 100% mortality. Twenty-three out
of 31 lin es tested were sign ifi ca ntl y different
from the non-transformed check (P = 0.01)
(Table 2). The eig ht rema inin g lines were
not signif ica ntl y different from the con trol
(P = 0.01 ). M orta li ty ran ged from 36.30 to
49.48%. After o ne month, the susceptibl e
lin es still harbored live larvae, suggesting
that the Bt gene was not expressed properly
in the fo li age.
Eva lu at ion s of tubers, one month afte r
harvest, we re condu cted on 25 out of 31
lin es of Cruza 148. Larval mortal ity was
hi gh (100%) in 16 lin es of the 25 tested.
Eight out of the nin e lin es showed a
mortality around 50% as compared to the
chec k w ith 22.22% of morta lity (Tab le 2).
A high mortality of PTM in transgenic lin es
indicated the presence of Bt tox in s in leaves
and tubers.
Evaluations of larva l mortality on foliage
(10 days after infestation) on transformed
LT-8 indi ca ted exce ll ent results w ith 100%

mortality on all transfo rm ed lin es. The nontransformed check showed 37.2% mortality
(Tab le 2) . Evaluations of 17 transformed
Sangema at 10 days showed high mortality
(100%). All lin es showed better larval
co ntro l than the non-transformed check.
Larva l mortality in the chec k was 20%
(Tab le 2). The damage caused by larvae in
the transgeni c lin es was very low. The
neonate larvae ha ve to feed on the leaves
for th e Bt toxins to work. The effect of th e
tox in on th e larvae often ca uses stunted
growth and feedin g inhibition , leadin g to
death (F igu re 2).
Larval development
Infestat ion w ith neonate larvae on
transgenic lin es showed no effect for the
first 3 days. The fou rth day, morta lity
in creased drasticall y until all larvae had
died at the seventh day. Larvae fed on nontransformed plants we re still ali ve after 10
days (F igure 3).
Durability of tuber resistance storage
Transgenic potatoes res istant to PTM
larvae accord ing to leaf and tuber bioassays
were also tested for durability of th e tuber

Table 2. Mortality in potato tuber moth foliage and tuber assays of potato transgenic lines expressed as percent
of infested larvae.

Clone
Cruza l 48/Bt2/{8 lines)
CRUZA- l 48/Bt2/ (l Line)
CRUZA-l 48/Bt2/ (+ 22 lines)
CRUZA-148 (control)
LT-8/Btl/ {37 Lines)
LT-8 (control)
(Only foliage was tested)
Sangema/Bt2/ (+ 17 lines)
Sa ngema (control)

Leaf<'

Tuber'

7 days
50% be
l 00.00 a
l 00.00 a
47.55 be
10 days
All 100%
37.2

l month
50% cde
86.67 b
l 00.00 a
22.22 e

l 0 days
100.00
20.00

NT
NT

100.00
20.84

Means followed by the some letter ore not significantly different (Waller-Duncan test, p = 0.01).
NT = not tested.

(IP Progrom Re port 1997-98

l 65

we re se ri o usly dam aged . (Fi gure 4). H ence,

res istance ca n be cons idered durabl e in
storage.
Assessment of field resistance and
impact on non-target pests
Field trial at San Ramon, 1995. Ten
transgeni c LT-8 and 19 Cru za- 148 lin es th at
scored we ll fo r PTM resist ance und er
greenhou se co nditi o ns w ere eva lu ated fo r
PTM res istance in t he f ield under hi gh PTM
press u re. Th e trials w ere ca rried o ut at CIP's
ex perim ental stati o n durin g th e su mm er
season w hen no potato is grow n in th at
area . A ll transge ni c plants of LT-8 and
Cru za - 14 8 showed a hi gh res istan ce leve l
again st PTM attack (z ero larvae per pl ant)
und er th ese co nditions, w he1eas th e
nontransform ed o nes showed hi gher
numbers of PTM larvae .

Figure 2 . Damage produced by Phthorimaea

operculella on leaves of un tra nsforme d

(left) and transformed plants (right) of
San gem a clone. At the left corner shows
the initial PTM damage in transgenic
lines.
%

mortality

Morph olog ica l c hara cteristi cs (p lant

he ight and lea f L:W rat io) of LT-8 we re
scored and fo und , for th e m aj o rit y, to be
wi thin th e no rm al range of va ri ab ility
compared to co ntro l plants. H oweve r,
m ea ns of p lant height of six transge ni c lin es
of Cru za 148 out of a total of 26 we re
signi f ica ntl y lowe r th an co ntro l plants,
w hereas o ne was signi f icantl y hi gher th an
th e co ntro l C ru za 14 8.

100 80

-+- LT-8

--- 44.9
-+- 34.1
---- 34.2

60 40 -

20

Res ults fro m field data showed th e sa me

trends as those from laboratory and gree nhou se ex perim ents. Th ere we re so m e
ph eno typ ic va ri at io ns w ith res pect to p lan t
height and lea f length/ we ight (L:W) rati o.
Th ese field tri als all owed lin es to be
se lected th at were ph enotyp ica ll y simil ar to
th e no n-tr ansfo rm ed control.

Time

Figure 3 . Mortality of neonate larvae reared on

potuto transgenic lines. La Molina, Peru ,

1998.
res istance in storage. All transge ni c lin es
tested (3 7 of LT-8 and 4 of San ge m a)
rem ain ed uninfested over at least 6 mo nths
of sto rage w hil e the non-transfo rm ed pl ants

166

Potato

Field trial at Tacna, 1997-98. Nin e lines

of LT-8, showed adequate leve ls of res istance (no infestat io n) to PTM in fi eld tri als
in Tacn a. Th e transge n ic lin es we re fo und
to be res istant to th e to mato pin wo rm , Tuta
absoluta. Effects o n no n-target spec ies

Liriomyza huidobre nsis, Russe lliana

so/anico la, and M yzus persicae were al I
non-signifi ca nt (P = 0.05). H ence, this Bt
gen e affec ted both PTM Phthorimaea

an d for an exten ded duration in the field

present a situat ion hi ghl y conducive to the
development of resista nce . This co uld be
alleviated by th e use of tissue and timespec ific promoters of Bt gene ex pression
(va n Rie, 1991 ) and the inclu sion of
multiple Bt and other re sistan ce genes
(B rattsten , 1991 ).

Figure 4 . PTM damage on tubers from transformed

plants (right) and non-transformed plants

(left), alter 4 months of storage.
opercule/la and the tom ato pinworm Tuta
abso/uta. Both pests are commonly found in
potato fields.
At harvest, PTM infestation on tubers
was limited to co ntrol plots; no transformed
lines showed damage. Yield s were not
significantly different between transfo rmed
and non-transformed lin es. Jansens et al.
(1995) using the same cry!A(b) gene
obtained simil ar res ults. Ebora et al. (1994)
used a different cry/Ac ge ne and obtained
mu ch lower PTM mortality (1.03%).
Deploying Transgenic Potatoes:
Special Consideration
Potential for insect resistance
In sects have shown their ab ility to
overcome the va rious in secti cides-based
co ntrol meas ures by deve loping resistance
to all major classes. Res istan ce to Bt has
already been reported for Indian meal
moth, Plodia interpunctella (Hubn er)
(Mcca ughey, 1985) and for the Diamondback moth, Piute/la xylostella (L. )
(Tabashnik et al. , 1990; Shelton et al.,
1993). It is commonly accepted that
deployment of St-transformed crops should
be handled with utmost ca re to red uce or
delay th e potential of in sect resi sta nce to Bt
(Tabas hnik et al. , 199 1; Mccaughey and
Whalan, 1992, Wearing and Hokkanen,
1994). Bt tox ins deployed at a high rate

Resistance management
Based on past experience w ith i nsecticid es, there are seve ral option s available for
managing potential res istance to Bt.
Refugia. Thi s consists of providing a
"refuge" for susceptible in sects for dilution
of selection pressure by planting nontransfo rm ed cultivars or other host plants
within and/or aro und th e crop. Also, th e
use of time and tissue-specific promoters
allows th e gene exp ress ion in targeted area s
only.
Seed mixtures or mosaics. Mixing
susceptibl e and resi sta nt seed at planting
provides unprotected pl ants harboring
susceptibl e insects w ithin the field.
Rotation. Th e impact of crop rotation on
se lect ion intensity con tributes to the
persistence of the Bt gene in the ecosystem,
thus influ enc in g th e potential development
of resistance. This w ill have to be eva luated in relation to th e target pest population
dynamics and dispersal to other ho st pl an ts.
Evaluations of other systems indi cate that
under so und IPM prog ram s, the in co rp oration of Bt genes in potato is feas ibl e from
th e Bt resistance potenti al. Howeve r,
spec ial ca re w i 11 have to be taken to avoid
cros s-res istance as th e same genes are used
for Lep idoptera on several crops as biopesticide sprays. Also, broa d-spectrum
in sect ic ides should never be used aga in st
ad ults to prese rve susceptible immi grants
and natural enem ies in the system. Time
and tissue-specific promoters (i.e., ex pressin g Bt genes in tubers) may be ab le to
minimize th e m ajor drawbacks assoc iated

CIPProgram Report 1997-98

l 67

w ith the dep loy ment of Bt in th e potato

crop (persistence , gene flow).
References Cited

Adang, M.J., M.S. Brod y, G. Cardi nea u, N .

Ea ga n, R.T. Roush, C.K. Shewmaker, A.
Jones, J.V. Oakes, and K.E. M c Bride,
1993. The reco nstru ction and expressio n
of a Bacillus thuringiensis cry lll A ge ne in
protoplasts and potato p lants. Plant Mot .
Biol. 21:1131 -1145 .
Brattsten , L.B. 1991 . Bioengin eer in g of
crop p lants and res istant biotype evolutio n in insects counteract ing coevol uti on.
A rch. Insect Biochem. Physiol. 17:253267 .
CIP (In te rn at ional Potato Center). 1993 .
Biotec hnology and biosafe ty at CIP:
International guidelin es . CIP, Lima, Per u.
De Bl ock, M , 1988. Genoty pe-ind ependent
leaf disc tran sfo rma tion of potato
(Solanum tuberosum ) usi ng
Agrobacterium tumefaciens. Th eo r. A ppl.
Genet. 76 :767 -774.
Dellaporta, S.L. , J. Wood, and J.B . H icks.
1983. A plant DNA m ini preparation:
Version II. Plant Mot . Bio l. Rep. 1 :19-21.
Ebora, R.V., M.M. Ebora, and M.B. Sticklen.
1994. Transge nic potato express ing the
Bacillus thuringiensis Cry lA(c) gene
effects on the survival and food consumption of Phthorim aea operculella
(Lepidoptera: Gelechii dae) and Ostrinia
nubilalis (Lepidoptera: Noctuidae). J.
Econ. Entom o l. 87 (4): 1122-11 27.
Feitelso n, J.S. , J. Payn e, and L. Kim. 1992.
Bacillus thuringiensis: In sects and
beyo nd. Bio/ Technolo gy 10 :27 1-2 75.
Fenemore, P.G. 1988. H ost-pl ant locat ion
and se lection by adult potato mo th ,

Phth orima ea opercu/ella

(Lepidoptera:Ge lec h i id ae): A rev iew. J.
In sec t Physiol. 34:1 75-177.
Fu glie, K. , H . Ben Sa lah , M. Essamet, A.
Ben Tem ime, and A. Rahm ouni. 1992 .
Th e deve lopment and adoption of
integ rated pest man agemen t of the
potato tuber moth, Phthorimaea
operculella (Zeller) in Tuni sia. In sect Sci .
Appl. 14(4):50 1-509.

168 Pototo

Hofte, H ., H . De Greve, J. Seurinck, S.

Jansens, J. Ma hi l lon, C. Ampe, J.
Vanderkerc kho v. H. Vanderbruggen, M.
Van Montagu, M. Zabea u, and M. Vaeck .
1986 . Structura l and functional ana lys is
of a c loned delta-endoto xin of Ba cillus
thuringiensis berliner 17 15 . Eur. J.
Bio chem. 161 :273-280 .
Jans ens, S., M. Cornelissen , R. D e C lercq,
A. Rey narts, and M. Peferoe n. 1995.

Phthorima ea operculella
(Lepidoptera:Ge lec hiidae) res istance in
potato by ex press ion of the Ba cillus
th uringiensis Cry lA(b) insecticid al c rystal
protein. J. Eco n. Entomol. 88(5): 1469 1476.
McGaugh ey, W.H. 1985. Insect res ista nce
to th e biologica l insecticid e Bacillus
thuringiensis. Science 229:193-195.
McGaugh ey, W. and M. E. W halon . 1992.
Ma nagi ng in sect res ista nce to Ba cillus
th uringiensis to xi ns. Science 258: 14511455.
Peferoe n, M., S. Janse ns, A . Rey na erts, and
J. Lee mans, 1990. Potato plants w ith
engi neered res istance against ins ect
attack, In : Vay da, M. and W . Park (e ds.)
Mol ec ul ar and cel lular biology of the
pota to. CAB, Tu cso n, AZ. p. 193-2 04.
Perlak, F.J ., TB. Stone, Y.M. M uskopf, L.J.
Petersen , G.B. Parker, S. A . McPh erso n, J.
Wy man, S. Love, G. Reed, D . Bi ever, and
D .A . Fi sc hhoff. 1993. Geneti ca lly
imp roved potato: Protecti on from
damage by Co lorado potato beet les.
Plant. Mo l. Biol. 22:313 -321.
Rama n, K.V. 1988. In teg rated in sect pest
man ageme nt for potatoes in develop in g
countri es . In tern at ional Potato Cente r
Circular 16(1 ).
Roux, 0., R. Von A rx, and J. Baumgartn er.
1992 . Est imati ng potato tuberworm
(Lepi dopte ra:Gelech iida e) dam age in
sto red potatoes in Tu nisi a. J. Eco n.
Entom o l. 85:22 4 6-2250.
Sambrook, J., E.F. Fri tsc h, and T. Maniais.
1989. Mo lec ular c loning: A Lab oratory
manu al, 2"c1. Ed. Co ld Spring H arb o r
Laboratory Press , New York, USA.

Saxema, A.P. and S.M.A. Ri zv i. 197 4. In sect

pest problems of potato in Indi a. J. Indi an
Potato Assoc. 1 :45-30.
She lto n, A.M., J.L. Robertson , J.D. Tang, C.
Perez, S.D. Eigenbrode, H.K. Preisler,
W.T. Wi lsey, and R.J . Cool ey. 199 3.
Resistance of di amondba ck moth
(Lepidopte ra: Plutellida e) to Bacillus
thuringiensis subspec ies in th e field. J.
Econ. Ent. 86 :6 97-705.
Tabas hnik, B.E., N. Finson, and M .W.
John so n. 199 1. Managing res istance to
Bacillus thuringiensis: Lessons from the
diamondback moth (Lepid opte ra:
Plutellida e) . J. Econ. Ent. 84:49-55.
Tab as hnik, B.E ., N.L. Cushing, N . Fin so n,
and M .W. John son. (1990). Fi eld

deve lopm ent of res istance to Bacillus

thuringiensis in diamonback moth
(Lepidoptera: Plutellidae) . J. Econ. Ent.
83: 1671-1 676.
va n Rie, J. 1991 . In sect co ntrol w ith
transgen ic p lants: Res ista nce proof?
Trends in Biotech. 9: 177-179.
Vaugha n, A. H . and A.M.R. Gatehouse.
199 1. Ph enotypic cos t to plants of an
ext ra gen e. Transgeni c Res. 1 :5 4-60.
Wea ring, C.H . and H .M .T. Hokk ane n.
1994 . Pes t resistan ce to Bacillus
thuringiensis: Case studies of eco log ica l
crop assessment for Bt gene i nco rpor ation and strateg ies of ma nagement.
Bico ntrol Sc i. Tech. 4:573-590.

CIPProgromReportl997-98

169

Participatory Needs and Opportunity Assessment for

Potato IPM Development Planning: The Case of
Indonesia
E. van de Fliert1, Warsito1, and A. Lagnaoui 2

Potato (Solanum tub eros um) produ ct ion in

Asia has in creased tremendously over the
past decades as a result of in crea sed area
pl anted to potato and in creased yie lds (CIP/
FAO, 1995). Indon es ia is among the
countries showing th e hi ghest growth rates
for potato production. H arvested area of
potato almost tripl ed from an average of
19,442 ha ove r the peri od 1970-79 to an
average of 53,436 ha in the peri od 199097. Yi eld s almost doubl ed from 7.9 to 15.1
t/ ha (CBS, 1998).

Potatoes in Indonesia
Potato in Indon es ia is typically grown in the
upland s und er humid tropical conditions.
Production has increased, but at th e cost of
environm ental degradation due to unb alanced and heavy use of exte rnal inputs
(fe rtili ze rs and pesti c id es) and indi scr iminate forest clea ring as potato farm ers so ught
disease-free and nutri ent-ri ch, virgin soi ls.
Pest and di sease probl ems have become so
serious that fa rmers co mm o nly apply
mi xtures of pestic id es to th e potato crop at
intervals of only 2-4 d. Th e leafmin er fly
(LMF) (Liriomyza huidobrensis Blan chard ),
an exotic pest introdu ced into Indon es ia in
1994, is spreading rapid ly and has become
a major threat to potato cultivati on. Its
presence triggers frequ ent, mostly broad spectrum insecticid e applications, but with
little success (Rauf and Shepard, 1999) .
Thi s trend of increas ing pesticid e
ap pli cation, agg ravated by the mon etary
crisis th at hit Indon es ia in 1998, caused

farm ers to spend a large r proportion of their

potato culti va tion budget on pest co ntrol .
On the other hand, products wi th hi gh
pesticid e residues are being rej ected at the
export markets in Singapore and Malaysia.
Farm ers are beginnin g to rea lize that they
need to look for alternatives to chemi ca l
control. However, nat ional re sea rch and
extension efforts have been limited and
re sea rch res ults genera ll y do not reac h th e
maj ority of farmers. Seve ral non gove rnmental organi zations (NGOs) conducted potato
integrated pest manage ment (IPM) farme r
field sc hoo ls, but they lac ked effecti ve
tec hni ca l alternati ves to tac kl e the co mpl ex
on-farm problems. Throu gh th e implementati on of a particip atory needs and oppo rtuni t ies assess ment, CIP aims to estab li sh
effective IPM approaches for potato growers
in Indones ia. Those in vo lved in the needs
assessment we re farme rs, NGOs, Indonesian ag ri cultural resea rch institutes, and the
nati onal IPM program. W e are presenti ng
here the prelimin ary res ults of th e first
ph ase of this on-going study.

Participatory Needs Assessment

Methodology
Th e needs assessment consisted of a
participatory stud y w ith the following
obj ectives:
To understand th e importance and
constraints of potato cultivati on for farm
households in maj or potato grow in g
areas in Indones ia.
To inventory potato cu ltivati o n practices,
probl ems, and output relating to cro p
management in general and pest manageme nt in parti cul ar.

1 CIP ESEAP, Bogor, Indonesia.

2 CIP, Lima , Peru.
CIPProgrom Report 1997-98

l 71

 To identify opportunities and partners for

sustainable improvement of the potato
enterprise, prioritize needs, and set the
agenda for a collaborative potato IPM
research and development program in
Indonesia.
Major potato growing areas in Indonesia
are the provinces of West and Central Java
and North Sumatra, accounting for 77% of
national potato production (CBS , 1998).
Four sites in these three provinces were
selected for the needs assessment exercise.
Selection criteria were the coverage of
varied cultivation and marketing systems
and the availability of local partners, i.e. ,
NGOs, and organized farmer networks
(Table 1). The following seven methods and
activities were applied to achieve the
objectives.
Village profiling and problem ranking
using participatory rural appraisal (PRA)
methods including transects, group discussions, ranking exercises, seasonal calendar,
and informal individual interviews with
men and women farmers, village officials,
and traders.
Season-long record keeping of the
potato enterprise, including quantitative

technological and economic cultivation

data, by 45 farmers.
Individual interviews with potato
farmers on production and marketing.
Regular field monitoring throughout the
season using direct observation of pest,
disease, and natural enemy occurrence and
damage, and the use of sticky traps in fields
subject to the record-keeping activity.
Assessment of the rate of parasitism of
the LMF through rearing of larvae collected
from the field in funneled containers.
Data analysis meeting with participating
farmers and partner organizations at each
site at the end of the season.
National Potato IPM Program Planning
Workshop with representatives of al I
national stakeholder groups to present the
resu Its of the needs assessment and other
previous work, and to plan a collaborative
potato IPM development program.
Not all data were available at the time of
this writing. Therefore results are presented
in a descriptive manner; advanced statistical analysis will be done later.

Table 1. Profile of the four needs assessment sites, Indonesia, 1998.

Province

West Java

Central Java

District
Vicinity to city
Cultivation system

Bandung (BO)
Near
Intensive vegetable
crops

Potato marketing
Maior potato
cultivation constraints
(farmers' perception)
Local partner'

Local market
Leafminer fly,
late blight,
bacterial wilt
TP4 (farmer
trainer network)

Baniarnegara (BN)
Far
Intensive vegetable
crops, extensive
subsistence crops
Local market
Leafminer fly,
late blight,
viruses
YPPSE
(local NGO)

North Sumatra
Simalungun (SI)
Far
Intensive vegetable
crops, extensive
subsistence crops
Export
Bacterial wilt,
late blight,
leafminer fly
YCD/WE
(local/int'I NGO)

Tanah Karo (TK)

Near
Intensive vegetable
crops, extensive
subsistence crops
Export
Bacterial wilt,
late blight,
leafminer fly
JTIK/WE (farmer
network/int'! NGO)

a. TP4 = Pangalengan IPM Farmer Tminers Tearn, YPPSE = Foundation for Development and Socioeconomic Progress, YCD =
Foundation "Love the Village", JTTK = Tanoh Karo Formers' Network, WE = World Education.
172

Potato

The stud y covered th e 1998 dry seaso n

in Ce ntral Java and the 1998/99 wet season
in all fo ur sites. Farm ers appreciated their
involvement in the n eeds assess ment data
co ll ectio n, analysis, and plannin g process.
They were reported to have benefited from
the act iv iti es by having ana lyzed and
in c reased their und erstanding of their own
potato enterpr ise . Their parti c ipation in the
needs assessme nt is expected to fac il itate
farme r co ll abora tion in t he up com in g
tec hnology and trainin g deve lopment
phase.
Results
Potato cultivation opportunities and
constraints
Farmers at the study site in West Java
ra nked potato first in importance, w hereas
at th e ot her three sites potato was ranked
third , fo ll ow in g subs iste nce crops such as
mai ze and sweetpo tato, and other cas h
crops such as chi Ii pepp ers (Capsicum sp.).
Farm ers pointed out th at th e advantage of
potato over other commonly cu lti vated
c rops is its potential for high productivity,
easy and profitable m ark eting, and relatively stable mark et prices . Disadvantages
in c lude its susceptibili ty to a ra nge of pests
and diseases resulting in hi gh producti on
cos t, intensive c rop ca re, and high risk.
Pota to cu lti vat io n therefor e is limited to a
select group of farmers w ho can affo rd the
in ves tm ent and risk. A n economic analysis
of the data collected in this study suggests
the d iverse per spect ives of farmers. Some
obta in a relatively high net return from
potato, while others lo se money when local
cond iti ons are unfa vo rab le durin g a
part icu lar season. One way of reducing the
ri sk of crop loss is by cu lt ivatin g potatoes
onl y on a smal l plot. Th e field size pl anted
by farme rs in the study ra nged from 0.03 to
0.65 ha, w ith an average of 0.22 ha.
Consider in g total cu lti vated land area
averages about 0.5 ha per farm fami ly in
Java and 1-2 ha in North Sumatra, farme rs
obv io usly red uced the size of th eir potato
plots in res ponse to risin g costs of exte rnal
inputs due to the 1998 economic cr isis.

The mo st seve re production constrai nt

perceived by fa rm ers was LMF in West and
Central Java, and bacterial w ilt (BW) ca used
by Ralstonia so lanacea rum in North
Sumatra (Table 1). High LMF incidence and
damage was confirm ed in all areas through
field observat ions . Data are being ana lyzed
to assess crop lo ss . In addit ion to high
in cidence, w hi ch ca used considerab le y ield
reduction , th e LMF prob lem was cons idered most se riou s because farmers did not
ha ve adequate co ntro l measures. Th e
chemical meas ures taken we re ex pensive
but in effect ive. BW was more commo n in
North Sumat ra than in the other areas, but
in cidence was relat ive ly low compared to
LMF and late b li ght (LB) ca used by
Ph ytop hth ora infes tans. Farm ers, however,
ranked BW as th eir most se riou s problem
becau se th ere is no way to control it w hen
it occu rs.
The seco nd most important probl em
ment ioned in all areas was LB. LB inciden ce was general ly high, especially duri ng
th e wet season, but farmers considered it
rel at ively manageable by means of intensive fungicid e use. Other production
constraints co nside red important by farmers
in c lud ed thrips (Thrips spp.), espec iall y
under dry conditions, potato tuber moth
(PTM ) (Phth o rima ea operculella Ze ll er), and
the lack of good seed. The predom in ant
potato cult ivar planted in all areas is
Granola, wh ich, though more susceptible to
LB and LMF than other cu lti vars, is preferr ed by cons um ers. Furth ermore, most
farmers did not practice good seed se lection , nor was high quality seed alw ays
available on the market. Although farmers
initi all y neglected to mention it, we analyzed th e ser iou sness of so il fertility deg radation as a result of hi gh-level, unbalan ce d,
inorga ni c fertil izer app li ca ti o ns and poo r
soil conservation prac tices. We bel ieve that
cou ld be a major threat to su sta in able

potato cu lti vation, espec iall y in West Java

and North Sumatra.

CIP Progrnm Report 1997-98

l 73

Potato yields
Potato yie lds obtained by the respondents during the 1998 dry season and the
199 8/ 99 wet season varied greatly across
and within stud y si tes (Figure 1), ranging
from 1 .0 to 28.0 t/ ha with an average of
10.5 t/ha, 0.4 t/ ha of wh ic h was unmarketable due to tuber rot, mainly caused by R.
solanacearum. No rm ally, the wet season is
considered the best seaso n for growing
potatoes in the study areas in West Java and
North Sumatra. Farmers cu lti vate under
rainfed water supply, with a yield pote ntial
25-30 t/ ha . In the study site in Central Java,
w here rainfed wa ter supply is abundant
throughout the yea r, the dry season is
considered the best season for potato
cultivation, w ith a yield pote ntial of around
20 t/ha. Th e relati vely high incid ence of LB
during the wet season is tackled w ith h igher
frequencies of fungi c ide app lication,
although not always very effective ly. Figure

2 shows this ineffectiveness by the frequency of cases w here LB was id ent ified as
a major cause of low yields durin g the wet
season despite intens ive pesticid e use. The
1998/ 99 wet season was an extraordinary
season for LB , particularly in North
Sumatra, due to in creased rain as a res ult of
the La N ina phenomenon. Furth ermore,
LMF, commo nl y regarded as a dry season
pest and despite recent qui esce nce,
reo cc urred during the North Sumatran wet
season. In West and Central Java, th e 1998/
99 wet seaso n was co nsidered relative ly
fa vora ble for potato production. LMF
occu rre nce , however, continued to inc reas e
causing major damage to potato crops for
th e majority of participati ng farmers.
Several farmers who had planted before th e
onset of rai ns reported thrips damage. Earl y
crop growth was poor due to drought,
aggravated by th e high incidence of thr ips
during the first part of the growi ng seaso n.

Yield (t/ha)
30

25

-j

Unmarketable yield

D Marketable yield

I
I

20

15

I
~

10

>------

nM

Banjarnegara
'98 DS '98/'99 WS

Bandung
'98/99 ws

Simalungun
'98/99 ws

nnn~~
Tanah Karo
'98/99 ws

Figure 1. Potato yields in Indonesian study sites during the 1998 dry season (DS) and 1998/99 wet season (WS).

Each bar represents the crop of one farmer.

174

Pototo

Yield (t/ha)

30

25

t---

Leafminer fly
Late blight
Th rips
Baterial wilt

20

15

0
0

A
A

..

A
A

10

A
A

10

15

20

25

Pesticide applications per season

Figure 2. Relation between number of pesticide applications per season and yield. One data point represents one

farmer. The marker style for each data point indicates the major production constraint identified by the
farmers .
Farm ers' culti vat ion pract ices w ere
poo rl y related to y ields, indi ca ting th at
fa rm ers had no sound tec hn o log ies and
practi ces to in c rease yi elds and manage
th eir produ c ti o n constraints. Labor all oca ti o n to th e potato c rop , in c ludin g hired
labo r and an o pportunity cost fo r house ho ld
labo r, range d fr o m US $4 6 to 60 4/ ha. That
impli es low labor efficien cy for a large
po rtio n of fa rm ers, p arti c ul arly th ose w ith
small fi elds.
Ferti I izer m anage m ent practi ces va ri ed
w id ely, with total inorga ni c NPK doses
rangin g fro m 25 to ove r 2, 000 kg nutri ents/
ha w ith an ave rage of 60 9 kg/ ha. Ave rage
fe rtiliz er appli cat ion rates we re 245 kg N/
ha, 215 kg P/ ha, and 178 kg K/ ha. Farm ers
obtainin g th e hi ghest yi e lds appli ed fertilize r at moderate rates . In th e absence of
adapted, recomm ended rates, fa rm ers
depend on th eir own ex p eri ence and

perce pt io ns. Ye t th eir kn ow ledge abou t

plan t nutriti o n and crop health is ge nera ll y
poor. For exa mpl e, on e farmer repo rted
addin g extra fertili zer to compensate fo r
water sho rtage during th e earl y gro w th
sta ge of th e c ro p.
Pesti c id e use ave ra ged 12 appli ca ti o ns/
se aso n, rangin g from a low of 3 to a hi gh of
22. Farm ers giv in g onl y a few appli ca ti ons
fac ed total cro p fa ilure (Fi gure 2). Th e data
from th e stud y sites, howeve r, showed
neith er hi gher y ields nor better co ntrol of
pests at hi gher pesti c ide appli ca t io n
frequ enc ies . Thi s stron gly indi ca tes injudiciou s and erroneo us p est ici de app l icatio n
practi ces.
Practi ces perce ive d by farm ers to fa vor
high crop yi eld include 1) th e us e of
health y seed, 2) balanced fertili ze r man agement, 3) effect ive pest and di sease mana ge-

CI P Prog1amRe porl 199/.98

175

ment, and 4) good cultural practices. But

farmers lacked information and resources to
use these practices. They would like to see
them as topics for collaborative adapted
technology deve lopment.

Pest management practices

Farmers' primary alternative for controlling pests and diseases of potato was
pesticides. Only a few farmers reported
trying other methods, such as yellow sticky
traps , to control LMF albeit with undetermined success. Farmers spent an average of
US$378/ ha on pesticides or 23.8% of total
production cost for potato culti vation. Their
major sources of information regarding pest
management were pesticide retailers and
agrochemical salesmen who regularly visit
even the remotest vi I Iages where vegetables
are grown. Information is normally limited
to promotion of the type of chemical to be
applied, and does not include advice about
proper application methods. As a result,
farmers tend to mix seve ral types of pesticides for one application. One wo man in
North Sumatra was observed spraying a
cocktail of two insecticide and five fungicide brands, some with the same active
ingredient. Farmers normally do not
consider recommended doses or intervals
between applications.
Across areas, farmers began applying
pesticides to the crop 11-38 dafter planting, and continued at an interval of once
every 2-4 d until the crop was 40-83 d old,
depending on how soon it died off. Taking
into account the multiple doses given
through cocktails of several pesticides in
one application, an ave rage of 27 doses/
field per season were applied to the potato
crop, with a maximum as high as 58 doses:
Only one farmer reported monitoring pest
and disease status before deciding on a
chemical control measure. He never mi xe d
different types of chemicals in one application, and applied 20 doses throughout the
season . He obtained among the highest
yields in his area.
Not surprisingly, natural enemy populations in these ecologically disturbed
176 Potato

systems are extremely low, particularly for

new ly introduced pests such as LMF. Three
parasitoids of LMF we re identified in the
study areas: Hemiptarsenus varicornus
(Girau lt), Neochrysocharis formosa
(Westwood ), and Opius sp. The rate of
parasitism , however, was low (1 .6 % during
the wet season in North Sumatra). Farmers
were eager to develop more effective and
sustainable pest control methods, particularly for those pests and d ise ases for which
they have no adequate control measures,
i.e., LMF, thrips, and BW.

Economic return from potato cultivation

Potato cultivation in the study areas
during the 1998 dry season and the 1998/
99 wet season resulted in an average gross
income of US$2,621/ha. However, there
was large va riation between individual
farmers (see Figure 3) due to wide var iability in yields and market prices. Average
price recei ved for potato was US$0.24/kg,
but fluctuated from US$0.06 to 0.37/ kg,
depending on the season, type of market,
tuber quality (including size and shape),
and the distance from the field to the main
road.
Production cost va ried wi dely with
regard to total amount and to allocation of
budget items. The net return , measured as
the gross income minus production cost
(i ncluding inputs, hired and household
labor, and rent for land), averaged
US$1,032/ ha. Net return ranged from a loss
of US$2 , 151 /ha to a profit of US$5,842/ha
(see Figure 3). Although the average profit
was three to four ti mes that of what farmers
could earn from a good ric e crop, the large
v ariation emphasized the hi gh risk of potato
growing. Thirty percent of participating
farmers lost mone y on pota to culti va tion
during the study seasons, w hich the y
attributed to unfavor able weat her contributing to low yie lds and high production costs .

Discussion and Conclusions

The participatory needs and opportunities
assessment for potato IPM provided us with
an overall picture of potato growing

Gross income (US$/ha)

9,000
8,000
7,000
6,000

,_
D Labor
,_

Inputs

,_

D Net return

5,000
4,000
3,000

2,000

1,000

,_

.-

ii
LJ

ii

-1,000
-2,000
-3 ,000

Banjarnegara
'98 dry '98/'99 wet

Bandung
'98/'99 wet

Simalungun
'98/'99 wet

Tanah Karo
'98/'99 wet

Figure 3. Economic analysis of the potato enterprise in Indonesian study sites during the 1998 dry season (OS)
and 1998/99 wet season (WS). Gross income consists of the sum of production cost (labor and inputs)
and net return (US$1 = 8,500 Rp in 1998). Each bar represents the crop of one farmer.
systems in the major production areas in
Ind onesia, the problems farm ers face, the
effort they expend, and opportunities for
improving potato cu lti vation . Th e need to
deve lop more eco logica ll y and eco nomica ll y susta in able cultivation practices
beca me an apparent overarching concern.
In particular, farmers demand ed new, more
effectiv e pest management measures to (1)
provide better opportunities to benefit from
the potential of potato cu ltivation, and (2)
m ake potato cultivation less polluting. This
study provides the foundation for future
collaborative IPM development program
planning. We identified major resea rch
areas that need attention (LMF, thrips, and
BW management; soil nutrient management; and a hea lthy seed supply system). In
add iti on, the assessment triggered high
interest in participating in follow-up
adapted technology development act ivities
within the farming communiti es.

References Cited
CBS (Central Burea u of Statistics). 1998.
Statistica l yearbook of Indonesia . Centra l
Bureau of Statistics, Jakarta, Indon es ia .
CIP/ FAO (Intern ational Potato Center and
the Food and Agricultural Organi za tion
of the United N at ions). 1995. Potatoes in
the 1990s: Situation and prospects of the
wor ld potato economy. FAO, Rome,
Italy.
Rauf, A. and B.M. Shepard. 1999.
Lea fmin ers in vegetab les in Ind ones ia:
Surveys of host crops, species compos ition , parasitoids and control practices.
Paper prese nted at the CABl-FAO
Workshop on Leafminers in Southeast
Asia held 1-5 Feb. 1999 at Cameron
Highlands, Malaysia.

CIP Program Repml 1997-93

l 77

Interaction of Maternal Clones of Potato (Solanum

tuberosum) with Pollinator Clones of S. phureja for
Induced Parthenogenesis
M. D. Upadhya and R. Cabello 1

Induced parth enogenes is (pseud oga my) has

bee n used in potato, (5olanum tuberosum)
to extract matroclinous dihaploids (2x = 24)
from tetraploids usin g 5. phureja as the
po ll en parent (Ho ugas et al., 1964;
H ermsen and Verdenius, 1973 ). The us e of
dih aploid s for genetic res earch and breedin g is we ll-documented (Bradshaw and
M ac kay, 1994). More dihapl oid s from a
wid e rang e of cultivars or advanced
se lection s wou ld be des irable, and even
esse ntial to fu ll y rea li ze their potential for
breeding and genetic research (Kotch and
Peloquin, 1987) . At CIP produ ction and use
of dihaploids from culti va rs or advanced
c lones is bein g pursu ed as a breeding
approach to produce parental lin es for
hyb rid tru e potato seed (TPS) fam ilies.
A lthough parental lin es now ava il ab le
produce hybrid TPS fami li es, which are
fairly homogenous in tuber shape and
co lor, th ere is still a ce rtain deg ree of
heterogen eity in haulm morphology and
tuber composition . Furthermore, incorporating into parental lin es resi stance to late
blight (LB ), ca used by Phytophthora
infes tans, using major and minor genes
from divergent sources is requir ed to ensure
stability of res istan ce in hybrid fami lies.
Parental lin es should car ry gen es that res ult
in th e lea st seg rega tion for haulm and tuber
characteristics in hybrid famili es and
prov ide stabl e resistan ce to di seases such
as LB.
The first step is extracting a large number
of di haploids from a wid e range of clon es
belonging to S. tuberosum sub sp.

tuberosum and 5. tuberosum sub sp.

andigena, and selecting fo r de sired characters . The sel ections can then be used to
resynthesi ze tetraploid lin es carrying the
des ired genes in multipl ex co nditi on. That
is expected to provid e a high er order of
hom oge neity in the hybrid TPS fa milies.
A signifi ca nt influ ence on th e frequen cy
of haploids induced by both th e pistillate
and pol lin ator parents has been demonstrated (Hougas et al. , 1964). Monogeni c
dominance of low seed set in th e pollinator
has bee n po stulated along with th e idea
th at seed set/berry by pollinators shows no
correlation w ith hap lo id s/berry (Hermsen
and Verdenius, 1973 ). Literature dea ling
w ith haploid induction efficien cy of 5.
phureja clon es has not co nsid ered the
im portant ro le the mate rnal parent might
pl ay in the induction of parthenogenetic
seeds.
However, such a rol e has bee n noted in

Nicotiana (Pand ey and Phung, 1982) and

oth er speci es and strain s of plants (Asker,
19 79; Gustafsso n, 1947; Herm se n and
Ram anna, 1981 ), wh ere parthenogenesis
cou ld proceed w ith or wi thout ferti li zation .
A lso, there has been no stud y to eva luate
th e role of cytop lasm in the promotion of
parth enogenet ic seed deve lopm ent.
Th erefo re, we undertook to estimate the
foll owi ng:
Dihaploid inducti on efficien cies of two
5. phureja c lones thought to be efficient
inducers.
Dihap loid production abi lity of a numb er
of tetraploid varieties and parental lin es
as seed parents.

1 CIP, Lirna, Peru.

CIP P1ogrnm Repoit 1997-98

l 79

 Intera ction between maternal clones and

the pollinators for the induction of
parthenogenetic seeds.
Th e possible role of cytoplas m in
promoting parthen oge netic seed production.

Materials and Methods

The pollinator clones of 5. phureja used for
comparison in this in vestigat ion we re IVP 35 and IVP-1 01 . Both are homozygou s for
the dominant seed mark er embryo-spot.
The maternal clones consisted of two
groups: (1) 12 cultivars an d TPS parental
lin es Qf di ve rse background, and (2) 14
sister clon es selected from a single cro ss
(CFK-69.1 x CEW-69. 1) sharing common
cytoplasm.
Materna l and po lli nator clones we re
grown in pots in a protected nethou se at
CIP's Huancayo Station (3,2 80 rn) in Peru.
The nat ural photoperi od was extended by 4
h usin g sodium va por lamps prov idin g -50
lu x at the plant level to promote flowering.
Th e pl ants were drip irrigated and ferti l ized
w ith 5 split applications of N totaling 250
ppm w ith basal applications of 200 ppm P
and 150 ppm K.
The seed parents were tr im med to reta in
one infloresce nce with six buds/ stern and
were emasc ulated at th e ripe bud stage.
Two successive pollin atio ns at 4-h in terva ls
we re carr ied out w hen the flo wer ope ned
(Th akur and Upad hya, 1991 ). Pollen from
the pollinator clones was co ll ected and
preserve d until used follow ing the procedure of Th akur et al. (1994). Berri es were
col lected 8 wk after pollination and
all owed to ripen at ro om temperature for 2
wk before manual extraction of seeds.
Tru e seeds (TS) were surface treated with
0.5% sodium hypochl o ri te for 10 min,
rins ed, an d dried to <5 % mo isture content.
Seeds w ith em bryo spots and without
embryo spots (i.e., those not resulting from
a sexual cross) were separated and
counted. Seeds without embryo spots were
sto red at 30C for 2 rno to break dormancy.
180 Potato

Seed li ngs grown from these seeds we re

checked for nodal pigmentation at the 4-5
tru e leaf stage to seg regate the hybr id s in
case some seeds were undetected by th e
embryo marke r. Th e emb ryo spot in the
hybr id seeds denotes pigmentation of the
nod e between th e coty ledons and th e
rad ica l. The domin ant ge ne governing this
pi gmen tation also produces pi gmentat ion of
the nod es i n the plan t and the eyebrows of
the tubers (Hermsen and Verde niu s, 1973) .
Seed l ings without noda l pigmentation were
transplanted in 12-in. diam. pots and after
30 cl we re checked for ploid y, first by
stomata! chloropl ast coun ts and then by
chromosome counts using root t ip
squashes. The numb er of di haploid s and
tetraploids we re reco rded for each of the
seed parents from pol Ii nations with res pecti ve po llin ato r clones.

Results and Discussion

Number of seedlings tested, diploid plants
identified, and th e percentages for eac h of
th e po llin ator clones used for each of the
14 seed parents are given in Table 1. The
average dihaploid frequency was hi gher
when IVP-101 was use d as the pollin ato r,
compared w ith th e use of IVP-3 5. That
confirmed the res u Its of Hermsen and
Verde niu s (1973 ) th at IVP-101 exhibited
higher dihaplo id inducing efficiency than
other clones studi ed. In Tabl e 1, the seed
pare nts are arran ged in three group s
accord ing to their response to the two
pol lina tor clones w ith respect to percentage
of di hap loids reco ve red. The data clea.rly
show a differential maternal respon se to
parth enogenetic haploid induction by the
po l lin ator clo nes. Th ese observations
confirm prev ious reports that seed pare nts
used for the extraction of dihapl oid s, using
different pollinator clones of 5. phureja,
differ significant ly in thei r dihaploid
production ability. (Hutten et al. , 1994).
In maize, Chase (19 49 , 1952) postulated
that the differential behavior of the seed
parents for monoploid production efficiency could be du e to 1) frequ ency of
recessive lethal genes in the seed parent,

Table 1. Comparison of dihaploid induction efficiency of S. phureja pollinator clones IVP-35 and IVP-101 in
tetraploid clones/lines of S. tuberosum, Peru, 1997-98.
Males

Females

Nonspotted seeds

Plants grown to

Survival

Di haploids

Di haploids

sown

maturity

(%)

(no.)

(%)

(no.)

(no.)
IVP-35 IVP-101

IVP-35 IVP-101

IVP-35 IVP-101

IVP-35 IVP-101

IVP-35 IVP-101

Group l
Achirona

569

134

569

134

99.l

80.7

44

32

7.7

23.8

Ccompis

73

247

73

247

91 .2

68.0

39

15.8
75 .0

7.7

Desiree

97

81

83.5

100.0

12.3
4.9

MF-1

133

65

133

296

80.l

93. l

00

Serrano

70

74

70

74

39.3

44.8

11

8.6

14.8

Stirling

14

73

14

73

38.9

61.3

39

28.6

53.4

TPS-13

13

13

68.4

50.0

23. l

66.7

TPS-67

57

24

57

24

49.l

32.4

3
2

3.5

8.3

206

435

206

435

89.6

92.2

11

14

5.3

3.2

60

60

17.6

53.6

50.0

3.3

MF-11

21

15

21

15

84.0

100.0

00

0.0

TS-9

14

14

73.7

40.0

0.0

0.0

12.0

22.7

Group 2
Atzimba
TPS-7
Group 3

Average

and 2) th e tendency of particul ar genotypes

of the seed parent to favor mon oploid
deve lopment.
H ougas et al. (1964) later co nfirmed th e
two postul ates of Chase from their results
w ith tetrap lo id potatoes. O ur results also
support postul ate (1) (Tabl e 1). Here it is
clea r th at th e surv iva l of seedlin gs from
nonspotted seeds va ried by female parent
in vo lved and th e pollinator clon e used.
Low surviva l fro m certain female parents
may point to a hi gher load of leth al recessives. The data also show that, in certain
crosses, th e pol Ii nato r clone influ ences
surv iva l. In crosses w ith cv. Stirlin g and
TPS-7, surv iva l was markedly lowe r w ith
IVP-3 5 as th e po llin ator th an w ith IVP-101.
With TS-9 and Cco mpis, th e reve rse was
tru e. Thi s differenti al behavior ca nnot be
expl ain ed only o n th e bas is of Chase's
postulate (1). Poss ibly th e genotypes of th e

two po llin ator clon es differentiall y intera ct

through the endosperm (4x from th e seed
parent+ 2x from pollinato r (Hougas et al. ,
1964) to suppress lethal facto rs and enhance embryo viability, o r enh ance lethal
effects by syn ergy. Present data o n differenti al emergence/s urv iva l from no nspotted
seeds in vo lv ing th e tw o po llin ator cl ones
do not ag ree w ith res ults of Hutten et al.
(1994) w ho report a simil ar perce ntage of
nonemergin g seed s invol v ing 3 pollin ator
clones in crosses with 26 female parents.
Th e signifi ca nt differences in dihaploid
prod ucti on ab ility betwee n seed parents
cou ld be ex pl ained on th e ba sis of inducibility facto rs prese nt in th e female parent.
Th ey are simil ar to th e heritabl e fac to rs
found for th e culturability/rege nerati on
capac ity of po ll en for hapl o id produ ction
through anth er culture (see Wenze l, 1994).
Hence w e propose that such heritab le
CIPPmgmmRepo1t 1997-98

181

facto rs co ntro l I in g th e resp o n se o f pa rth enoge neti c seed produ cti o n are prese nt but
need co nfi rmation throu g h furth er in ves tiga tio n s.
Th e above discu ss io n stil l leaves th e
qu estion of w hat p art cy top las mi c influ e nce
pl ays o n th e i ndu c ib i I ity o f p arth enoge n etic
haploids. Th erefo re, we po l Ii natecl 14 sister
c lones fro m a cross ca rry i ng ide nt ical
cytop las m w ith th e two po llin ato r c lones.
D ata o n seed s set/berry, and th e pe rce ntage
of seeds w ithout th e embryo spot are g ive n
in Tab les 2 and 3. Th e res ults cl ea rl y show
th at seed s set/berry va ri es w id e ly betwee n
the siste r c lo nes in res p o n se to th e two
pol lin ator c lones (Tab le 2) . Simil arl y, th e
p erce nta ges of seeds w itho ut th e em b ryo
sp ot a lso va ry w id ely b etwee n th e siste r
cl o nes in res po n se to th e two p o l lin ato r
clon es (Tab le 3). Th at suggests th e ro le of
cytop las m probab ly is no t su c h th at it ca n
influ en ce nuclear facto rs d eterminin g th e
parth e noge netic indu cti o n abi lity of th e
seed p arent.
Tab le 3 indicates th at nu c lea r factors
govern th e i nclucti o n of pa rth en oge neti c

seed s in a ge notype. Th at is clu e to w ide

segrega ti o n am o ng sister c lones fo r th e
pe rce ntage o f seed s w ith out th e embryo
sp ot in resp o n se to th e tw o pollin ator
cl o nes.
Pandey and Phun g (1982 ) showed th e
presence o f th e H ertw ig effect in pl ants
throu gh indu ced p arth enoge nes is usin g
irradi ated po ll en. Th ey postul ated th at N.
alata has sp ec ifi c ge nes th at, w hen tran sferred to th e egg in seg mented fa sh io n,
promote parth en oge neti c dipl o id y. N.
g lutin osa has, in addition , ge nes th at
promote p arth en ogeneti c hapl o icly. Th erefore, it is reaso nab le to expect th e ge neti c
co n stituti o n of th e seed parent and th e
pollin ato r clon e to figure signifi ca ntl y in th e
success ful indu cti o n of parth enogeneti c
seed s. Th e p rese nt data demon strate th at
th e ge no types of bot h th e matern al clon es
an d th e po llin ato r cl o n es are cruc ial fo r th e
indu c ti o n o f p arth enoge neti c hap lo id s. If
w h at has bee n sh ow n in N icotiana ca n be
appl ied to th e prese nt ca se, th e fo ll ow ing
wo uld be tru e. In du ce r genes in th e po llin ato r clon es, w hen prese nt in th e 6x end osperm (4x of m atern al + 2x of pollin ator)

Table 2. Comparative seed set/berry with 5. phureja pollinator clones IVP-35 and IVP-101 involving
14 tetraploid female sister clones of
Males
Females

S. tuberosum, Peru , 1997-98.

Seeds/berry (no.)

IVP-35

Ratio
IVP-101

35 :101
4.2:1

C95LB-22.9

28

C95LB-22.l 4

49

12

C95LB-22.8

37

C95LB-22. l 5

39

11

3.5:1

C95LB-22 .3

63

20

3.2:1

C95LB-22.2

57

l8

3.1 :l

C95LB-22.6

61

21

2.9: l

C95LB-22.7

44

16

2.8:1

C95LB-22.5

24

2.6: 1

C95LB -2 2.ll

35

15

2.3:1
2.3:1

4.2:1
3.9: 1

C95LB-22. l 2

30

13

C95LB-22 .l 3

51

28

1.8:1

C95LB-22 .l

25

21

1.3:l

C95LB-22 .l 0

l0

0.6:1

182

Potato

Table 3 . Comparative percentage of seeds without embryo spot with 5. phureia pollinator clones IVP-35 and IVP101 involving 14 tetraploid female sister clones of 5. tuberosum, Peru, 1997-98.
Males

Total seeds

Seeds without embryo spot

(no.)
Females

(no.)

(%)

IVP-35

IVP-101

IVP-35

IVP-101

IVP-35

IVP-101

143

164

11.5

25.8

63

0.5

14.3

C95LB-22. l

l,247

636

C95LB-22.2

1,552

441

C95LB-22.3

1,315

237

30

0.7

12.6

C95LB-22.5

723
2,803
221
402
303
30
1,214
2,043
l , 112
1,181
426

101

73

30

1.0

29.7

415

31

60

14.4
20.5

C95LB-22.6
C95LB-22.7
C95LB-22.8
C95LB-22.9
C95LB-22.l 0
C95LB-22. ll
C95LB-22. l 2
C95LB-22. l 3
C95LB-22. l 4
C95LB-22. l 5

140

22

l.l
0.0

224

46

1.0

15.7

197

16

1.0

9.6

206

26
135

10.0

12.6

346

3
178

14.7

39.0

293

184

57

9.0

19.4

550

47

0.8

8.5

591

119

45

0.2
0.9

20.l

300

3.7

18.4

Average

in sy nergy w ith th e genes in th e egg, ca n

ini tiate parth enogeneti c deve loprnent of th e
2x/ 4x (unred uced egg ce ll) ernbryo and th e
seed.
Herrnsen and Verdenius (1973) reported
polyploid s from nonspotted seeds in th e
progen ies of cultivars x homozygous
pollinators of S. phurej a. They reported a
freq uency of 1.9% for rnatroclin ous
tetraploid s. Their data, however, clea rly
indi cate th e 13 phureja po llin ator clones
homozygous for embryo spot had differenti al parthenogenes is in duction efficiency.
The cu ltivars G in eke and Radosa produced
matrocl i nous tetraplo ids alth ough th e
frequency in Gineke (27) was hi gher than in
Radosa (14).
Such matrocl i nou s tetraploid s cou Id
orig in ate on ly from unreduced (4x) egg
ce ll s produ ced durin g megasporogenes is. It
has been shown that restitution , eith er in
the first meioti c di v ision (FDR) or in th e

15.0

seco nd rn eioti c division (SOR), in Solanum

dipl oid s and clihaploids is geneti ca ll y
contro ll ed (Mo k and Peloqu in, 19 75).
Hence, it is reaso nab le to ass um e that th e
gene or genes co ntrol Iin g restitution wi 11
also be operative at the tetraploid level. The
4x egg ce ll s wo uld res ult from FDR or
SOR, premeiot ic/postmeiot ic doubling, or
apospory. Such 4x egg ce ll s would then
give ri se to matroc lin ous tetraploids through
th e indu ce r effect of the phureja pollinator
coup led w ith the indu cibl e effect in the
seed parent. The frequ encies of such
matrocl in ous tetraploid s produced depend
on th e genotype of th e seed parent as wel I
as the po llin ato r.
Th at rn atrocl inous tetrap loids or iginate in
thi s rn ann er is supported by th e data in
Tab le 1. Chromosome counts co nfirrned th e
rem ainin g plants raised from nonspotted
seeds w ere tetrap loid s. There are marked
differences in the frequ enc ies depe nding on
the seed parent and th e po llin ator clon e.

CIP ProgromReport 199798

183

That co nfirm s parth enogeneti c seeds are

produced by the interaction of genotypes of
materna l clones (ind ucibility factors) with
the po ll inator genotypes (induc er factors ).
Th ese wo uld ha ve either a dihaplo id or a
tet rap loid embryo of mat rocl inou s origi n .
Conclusions

Th e results of these investigations show th at

the pollinator clones, IVP-35 and IV P-1 01
ha ve differential effic iency for dihap loid
indu ctio n. Th e overall avera ge for the two
po llin ators shows that IV P-101 is more
effic ient than IVP-35. H owever, th e
d ih ap lo id ind uction effic iency is also
governed by the genotype of the see d
parent.
Crosses in vo lv ing 14 sister clones with
th e two pollinators indicated th e absence of
cytoplasm ic factors influe nc in g the proportion of parthenogenetic seeds wi th o ut
emb ryo spot. Based o n thes e studies and
information o n other species, it is postulated that the indu cib ility of parth enogenetic seeds seems to be determin ed by
nucl ea r genes of th e seed parent in respon se to th e indu cer effect of the po llinato r c lon es. Moreover, at th e tet raploi d level
unreduced 4x egg eel Is are produced
throu gh restitutio n. That cou ld give ri se to
malroclinous tetraploids through parthenoge nes is.
References Cited

Asker, S. 1979. Progress in apomi xis

researc h. Hereditas 91 :231-240.
Bradshaw, J.E. and C.R. Mackay (eels.) .
1994. Potato genetics . CAB In ternationa l,
U.K. 552 p.
Cha se, S.S. 1949. Monoploid frequ enci es in
a com merc ial doub le cross hybrid maize,
and its compo nent single cross hybrids
and inbred lin es. Genetics 34:328-332.
Chase, S.S. 1952. Monoploids in ma ize. In :
Gowen J.W. (ed .). Heterosis. Iowa State
Col lege Press, Ames, IA, USA.
Gustafsson , A. 1947. Th e ca usal aspects of
apo mi xis. Lunds Un iv., A rss kr. Av d. 2,
43: 71-178.
184

Potato

Hermsen , J.G.Th . and J. Ve rdeniu s. 1973.

Select ion from Solanum tuberosum
group phurej a of ge nopypes combin ing
high-frequency haploid induction w ith
homozygos ity for embryo-spot.
Eup hytica 22:244 -259.
H er msen, J.G.Th. and M.S. Ramann a. 1981.
Haploid y and plant breeding. Philos.
Trans . R. Soc. (Land. ), Sec. B, 292:499507.
H ougas , R.W. , S.J. Pel oq uin, and A .C.
Gabert. l %4. Effect of seed parent and
pollinator on the frequenc y of hapl o id s
in Solanum tuberosum. Crop Sci. 4:593595 .
H utten, R.C.B. , E.J.M.M. Scholberg, D.J.
Huigen, J.G.Th. H ermsen, and E.
Jacobse n. 1994. Ana lysis of diploid
inducti on and production abi li ty and
see d parent x pollinator interact ion in
potato. Euphytica 72:61-64 .
Kotch, G.P. and S.J. Pel oqu in. 1987. A new
sou rc e of hap loid germp lasm for genet ic
and breeding researc h. Am. Potato J.
64:137 - 141.
Mok, D.W.S. and S. J. Peloquin. 197S. Three
mechani sm s of 2n po l len format ion in
diplo id potatoes . Can. J. Gen. Cyto.
17:217-225.
Pandey, K.K . and M . Phung 19 82. ' H ertw ig
effect' in plants: Ind uced parthenog nesis through th e use of irradiated poll en.
Theor. Ap pl. Genet. 62:295 -300.
Thakur, K.C. , and M .0 . Upadhya. 1991 .
H yb rid TPS production technolog y. In :
Proc eed ings of the Third Tri enn ial
Co nference of the Asian Potato Assoc iation (APA). APA, Bandung, In donesia. p.
147-1 58 .
Th aku r, K.C. , M.D. Upadh ya, S.N.
Bharga va, and A. Bhargava. 1994. Bulk
po lle n extraction proce dures and th e
potenc y of th e ex tra cted pollen. Pota to
Res. 37 (3) :245-248.
Wenze l, G. 1994. Tissue culture. In:
Bradsh aw, J.E. and G.R Mackay (ed s.).
Potato genet ics. CAB Internat io nal, U.K.
p.173-195 .

A Proposed Solanum tuberosum subsp. andigena

Core Collection
Z. Huaman\ R. Ortiz 2 and R. Gomez1

Although, potato is known wo rld w id e as

5olanum tuberos um, in Lat in Am eric a there
are thousands of farmer cu lti vars that have
been gro w n in the Andes for centuries.
They are c lass ified as Solanum tuberosum
subsp. andigena [h erein afte r called
andigena] and S. tuberosum subsp.
tuberosum (tuberosum) (2n = 4x = 48); 5.
stenotomum, 5. phureja, 5. gon ioca lyx, and
5. x ajanhuiri (2 n = 2x = 24); 5. x cha ucha
and 5. x ju zepczukii (2n = 2x = 36); and 5.
x curtilobum (2 n = 5x = 60) (Hawkes,
1990) . Th e primary center of diversity of
cultivated potato is in the hi ghland s of
A ndean South America w here all culti va rs
except 5. tuberosum sub sp. tuberosum are
native. The Chi loe Archipelago of Chile is a
seco ndary ce nter of dive1s ity. There hundreds of farmer culti va rs of 5. tuberosum
subsp. tuberosum adapted to lon g ph otoperiod s still grow und er coo l env ironm ents at
or ne ar sea leve l (H awkes 1990, Hu ama n et
al., 199 7).
CIP assembled a potato co llecti on of
more th an 1 5 ,000 farmer-selected accessions of culti va rs from Argentina, Bolivia,
Chile, Colombia, Ecuador, Guatemala,
Mexico , Peru, and Ven ez uela, large ly from
1971 to 1989. Since then, so me selected
donations we re rece ived from non governmental organizations (NGOs) wo rkin g to
co nserve o n-fa rm And ean cu ltivars. Many
of the accessions in the co ll ec tion were
donated to CIP by in st ituti ons in South
Amer ican countri es. A large number of
th em we re gath ered and shared by joint
CIP-NARS (natio nal agri cu ltural researc h
syste m) co ll ecting exped iti ons.

Since potatoes are clona ll y propagated,

the tubers collected in different fa rm er's
field s were ex pected to duplicate num ero us
accessions of the same culti var. In many
cases, the sa me cu Iti va r was collected
under different ve rna cul ar nam es along its
area of cultivation. In add ition, donations
from national or in st itut ional collections
add ed further dupli cat ion because the sa me
cultivar could be mainta in ed in more than
one potato col lection. Th erefore, it cou Id
have severa l identifi catio n co des . A process
of id entifyin g dupli cate access ion s of the
same culti va r, usin g morphological and
electrophoretic criteria of tota l proteins and
esterases, red uced th is potato collection to
about 3 ,500 different fa rm er cu Iti vars
(Huaman, 1986, 19 94).
Farme r se lected culti va rs of andigena are
pred om in ant in th e co ll ection and rep rese nt
about 75% of the total culti va ted accessions. With the exception of 5. stenotomum,
accessions that in clud e seve ral hundred
farmer culti vars, and 5. phureja with more
than one hundred, al l other cultivated
species have fewer cultiva rs and are more
restri cted geographi ca ll y (Huam an et al.,
1997) .
Th e core co llecti o n co ncep t was put
forward in Frankel and Brown (19 84). A
core was defined as a limited set of accession s derived from an ex ist ing germplasm
collection, chosen to represe nt the genetic
spectrum in the w hol e co ll ectio n. Th e co re
should co ntain the max imum genetic
di vers ity avai labl e in the co llecti on . Th e
rem ainin g access ion s in the collection are
call ed the rese rve co ll ection.

1 CIP, Lim a, Pe ru.

2 ICR ISAT, And hra Pradesh, In dia.
CIP Program Report 1997 -98

l 85

Brown (1989a 1 1989b, 1995) also

indicated that the accessions in the core are
chosen primarily to be representative, and
are ecologically or genetically distinct from
one another. Within the primary constraint
of covering the genetic diversity and
ecological range, the aim should then be to
maximize genetic diversity. This implies
that the core shou Id not contain duplicates
and should minimize similarity between its
entries.

accessions were converted to true seed

(Huaman and Stegemann , 1989). The
output of this work was to reduce the
andigena collection from 1OJ22 to 2A27
accessions. Further comparisons in 1998
showed 48 additional duplicates, reducing
the number of cultivars in the collection to
2,3 79. That represents about 22% of the
original collection. The number of duplicates of the same cu ltivar ranged from 1 to
276 accessions (Table 1).

The fundamental role of the core, then ,

is as a workable guide or point of entry to
the whole collection. Therefore, the
selection of a core subset of andigena
cultivars in the collection maintained at CIP
is essential to increase the use of the
genetic diversity available in the entire
collection. Core collections have been
defined for many crops in the i ast decade
(Ortiz et al. 1998). This paper reports the
process of selecting the andigena core
collection.

Morphological data were recorded from

these 2,379 andigena cultivars grown in the
field at Huancayo. Data were recorded
using a 0 to 9 scale for 25 key culti vated
potato descriptors selected from an internationally accepted descriptor list (Huaman et
al. 1977). Of the 25 descriptors, 9 are
related to tuber characters; 6 to plant habit,
sterns, and leaves; 8 to flo we rs; and 2 to
fruit characters (Table 2). Corolla, tuber
skin, and flesh colors were recorded using
two color charts that combine color and
intensity. They were developed to match
color diversity in flowers and tubers shown
by Andean potato cultivars. The codes used

Materials and Methods

The original number of andigena accessions collected in the highlands of Latin
America and assembled in the genebank at
CIP was 1OJ22. These accessions were
collected in eight countries from Mexico to
Argentina. Over the years, a process of
identification of duplicate accessions of the
same cultivar was undertaken in this
collection. Accessions w ith similar tuber,
stem, and floral characters were morphologicall y compared when grown side by
side in the field at CIP's highland experiment station in Huancayo, central Peru
(3,200 m). Accessions that were morphologicall y identical were also compared by
analyzing total proteins and esterases
extracted from their tubers by polyacrylamide gel electrophoresis. Those that were
morphologicall y and electrophoretically
identical were considered duplicates.
One clonal accession of each duplicate
group was selected to represent the group
for further comparisons w ith other accessions in the collection. The other duplicate
186

Potato

Table 1. Number of duplicate accessions found in

the original andigena potato collection
maintained at CIP, Lima, Peru, 1998.

Duplicates

Duplicate groups

within groups
None (unique)
1-2
3-5
6-10
11-15
16-20
21-30
31-50
51-75
76-100
101-125
126-150
276

1,204
747
482
194
60
27
37
19
11
7
4
l

Table 2. Means (x), standard error (SE), and x 2 tests of frequency distribution for morphological descriptors for
the whole (2,379 accessions) and core collection {306 accessions) of andigena potato cultivars. Descriptor
states transformed into a 0 to 9 scale' except for predominant flower and tuber skin color, which
combined color and intensity in a two-digit score. R1 values were calculated for each morphological
character in the whole collection considering a maximum of six clusters and the highest values are shown
in bold face, CIP, Lima, Peru, 1998.
Morphological
Character

Whole

Plant habit
2.77
Stem color
2.93
Stem wing shape
l.25
No. pairs primary
lateral leaflets
5.32
Pin
No. primary interjected
leaflets on rachis
12.82
Sin
No. secondary interjected
leaflets on petiolules
7.06
Pde
Pedicel color
4.70
Klx
Calyx color
4.57
Fwl
Predominant flower color 17.92
Fw2 Secondary flower color
4.45
Fw2d Distribution of secondary
l.01
flower color
Fwsh Corolla shape
5.80
Ant
Anther pigmentation
0.28
Pst
Pistil pigmentation
l.65
Frc
Fruit color
l.26
Frsh Fruit shape
2.17
Tskl Predominant tuber skin
16.17
color
Tsk2 Secondary tuber skin color 3.93
Tsk2d Distribution of secondary
3.76
tuber skin color
Tsh l Tuber shape outline
5.85
Tshx Odd tuber shape
0.96
Tey
Tuber eye depth
5.09
Tf1
Predominant tuber flesh
2.27
color
Tf2
Secondary tuber flesh
l.30
color
Tf2d Distribution secondary
0.55
flesh color
Phb
Ste
Stw
Pin

SE

Core

SE

Classes

x2

(no.)

colcu-

In clus-

lated

ters

P>x2b

R1

0.014
0.035
0.010

2.79
3.00
l.30

0.042
0.107
0.030

5
7
3

175
. 2.86
l.03

0.781
0.827
0.598

0.008
0.175
0.026

0.018

5.34

0.053

l.61

0.952

0.359

0.128

13.17

0.382

18.94

0.015

0.560

0.161
0.036
0.027
0.096
0.596
0.038

6.70
4.74
4.36
17.72
4.36
l.23

0.466
0.103
0.085
0.301
0.183
0.115

9
9
7
8
8
10

3.44
6.90
10.56
5.74
10.87
12.28

0.904
0.547
0.103
0.570
0.144
0.198

0.763
0.061
0.046
0.102
0.022
0.004

0.023
0.015
0.034
0.018
0.023
0.164

5.68
0.35
1.67
l.31
2.13
16.67

0.072
0.049
0.104
0.061
0.070
0.467

7
6
8
8
8
9

16.18
4.293
3.684
9.156
12.98
17.51

0.003*
0.508
0.815
0.242
0.073
0.025

0.003
0.008
0.229
0.003
0.008
0.859

0.065
0.053

3.84
3.75

0.193
0.157

l0
8

8.69
l.09

0.466
0.993

0.045
0.066

0.032
0.054
0.031
0.018

5.81
0.73
5.02
2.32

0.099
0.136
0.093
0.061

8
10
5
7

7.05
17.79
0.60
7.87

0.424
0.038
0.897
0.248

0.001
0.004
0.005
0.015

0.058

l.59

0.185

5.55

0.475

0.056

0.028

0.69

0.090

7.88

0.343

0.045

According to Huaman et al. (1977)

P>x2 means the probability of x2 greater than x2 calculated. If the value is 2'. 0.05, there is no significant difference between
whole and core collections.
CI PProgrom Report 1997-98

187

for tuber skin color were white-cream (1 =

pale, 2 =intermediate, 3 =dark ); yel low (4,
5, 6); orange (7, 8, 9); brown (10, 11, 12),
pink (13, 14, 15 ), red (16, 17, 18); redpurple (19, 20, 21 ); purple (22, 23, 24 );
purple-black (25, 26, 27 ). The codes for
flower color were white (1 =pale, 2 =
intermediate, 3 =dark); red-pink (4, 5, 6);
red-purple (7, 8, 9); blue (10, 11, 12); bluepurplish (13, 14, 15); lilac (16, 17, 18);
purple (19, 20, 21 ); and v iolet (22, 23, 24).
Cluster analysis of the morphological
data was done w ith NTSYS-pc vl .70 (Rohlf,
1992 ) using the simple matching coefficient
and the unwei ghted pair group method
arithmetic mean algorithm (UPGMA). The
phenogram produced was used to determine accessions with similarity coefficients
below 0.75. One accession from each of
these clusters was selected for the core
subset.
A proportional sampling strategy aimed
to maximize geographical representation
was adopted. The goal was to retain
potential differences in ecological adaptation to the niches where potato is grown.
Therefore, the number of accessions to be
selected for the andigena core collection
was based on the square root of the number
of different cultivars from the main geographical division (state, department, or
province) of each country (Table 3). The
goal for total sample size of the core was
10% of the original number of accessions
as suggested by Brown (1995) for clonal
crops. As much as possible, we intended to
include in the core those cu lti vars with
desirable genes for breeding. Therefore, we
used all available data from evaluations
conducted at CIP to identify sources of
resistance to diseases and pests as well as
other desirable traits (Huaman , 1987; other
data in the potato database at CIP).
The strength of association between the
22 morphological characters scored as 1-9
(3 characters were scored as actual numbers of leaf parts) was determined us ing
Cramer' s V coefficient

188

Potato

V=

* min(r-l,c -1)

whe re c 2 =chi square value from the

contingency table, N =total number of
accessions analyzed, r = row, and c =
column. R2 v alues we re also calculated for
the whole collection to determine the most
important morphological characters to form
a maximum of six clusters using the
FASTCLUS procedure (SAS Institute, 1985).
The mean phenotyp ic di ve rsity between
the collection of 2,379 (whole) and the
subset of 306 accessions (co re) was compared by Student's ttest for all descriptors.
Similarly, the homoge neity of the frequency
distributions between whole and core
collections wa s analyzed by chi-square test
as suggested by Ortiz et al. (1998).
Results
Analyses of morphological data
C1amer' s V coefficients between
pairwise comparisons of the 22 morphological characters scored as 1-9 showed
only 3.6% of them to be correlated at
values of 0.30-0.56. This means that these
characters are independent and therefore
suitable for these analyses of cultivated
potatoes (Table 4) . The highest values found
were V = .56 between secondary tuber
flesh color (Tf2) and distribution of secondary tuber flesh color (Tf2d ), w hich are
expected to have some association. The
same was true for predominant tuber flesh
color (Tfl ) and secondary flesh color (Tf2)
with V = 0.50; predominant flower color
and secondary flower color w ith V = 0.49;
secondary tuber skin color and distribution
of secondary tuber skin color with V =
0.48; and for primary and secondary tuber
flesh color (Tfl and Tf2d) with V = 0.40.
These associations could be explained by
the high percentage of accessions in the
collection tha t do not show secondary
colors in either the flowers or tubers.

Table 3. Number of accessions in the whole (W) and core (C) collection of andigena potato cultivars according to
geographic data from their collection site, CIP, Lima , Peru, 1998.

Country

State,

Square root

department, or
province
Argentina

Bolivia

Colombia

Ecuador

Guatemala

Catamarco
Jujuy
Salta
Unknown
Oruro
Potosi
La Paz
Chuquisaca
Cochabamba
Torija
Unknown
Boyaca
Cundinamarca
Quindio
Santander
Santander del Norte
Cauca
Nari no
Tolima
Valle
Bolivar
Caiiar
Chimborazo
Cotopaxi
Tungurahua
Carchi
lmbabura
Pichincha
Azuay
Loja
Unknown
Huehuetenango
Queza ltenango
San Marcos
Solola

3
42
41
6
50
59
35
4
8
3
161
39
13
3
6
4
13
46
2
4
18
11
34
15
10
25
31
6
8
16
20
6
9
10
2

1.7
6.5
6.4
2.4
7.1
7.7
5.9
2.0
2.8
1.7
12.7
6.2
3.6
1.7
2.4
2.0
3.6
6.8
l.4
2.0
4.2
3.3
5.8
3.9
3.2
5.0
5.6
2.4
2.8
4.0
4.5
2.4
3.0
3.2
1.4

2
6
0
7
10
6
2
3
24
6
3
l
2
3
7
l
2
4
3
6
3
3
5
6
2
2
4
l
3
2
3

Conti nued on next pog e.

CIP Program Report 1997-98

189

Table 3 (cont.)

Mexico

Limo

2
11
7
109
85
50

Junin

99

Posco

53
88
152
39
10
42
19
4
6
288
110
455
12
23
2,427

Mexico
Pueblo
Veracruz

Peru

An cash
Huonuco

Apurimoc
Ayocucho
Huoncovelico
Amozonos
Cojamarco
La Libertad
Lamboyeque
Piura
Cusco
Pu no
Unknown
Venezuela

Merida
Trujillo

Total

Clu ste r analyses showed the most

important morphologi ca l c haracters to form
the most different six clusters of access ions
were predominant tuber ski n co lor (Tsk l ),
number of seco nd ary interjected leafl ets on
the petiolules (S in ), number of primary
interjected leaflets on the rac hi s (Pin ),
number of pairs of prim ary latera l leafl ets
(P in ); pistil pigmentation (Pst), ste m co lo r
(Ste), and predominant flower co lo r (Fwl ).
The R2 values ranged fro m R2 = 0.001 to
0.859 (Tabl e 2).
Selection of a core subset
The frequency di stributi o ns of data from
the co ll ection sites of access io ns in th e
who le and core collections were determined by co untry and state (Tab le 3). As
much as possible, the number of access ions
in the co re subset for eac h state was eq ual
to th e square root of th e number of accessions in the w hole collection from that
state. Th e number of accessio ns from

190

Potato

1.4
3.3
2.6
10.4
9.2
7.1
9.9
7.3
9.4
12.3
6.2
3.2
6.5
4.4
2.0
2.4
17.0
10.5
21.3
3.5
4.8

1
3
3
10
9
7
10
7
11
12
6
4
6
4
2
2
22
13
23
4
5
306

Argentina and Ec uado r with unkn ow n

co ll ecti o n sites was reduced in the core to
favor access ions w ith known collection
sites. But, in the case of accessions from
Bolivia and Peru w ith unknow n co ll ectio n
sites, the proportions in the core we re
retain ed beca use of the relatively hi gh
number of those access ions.
From a dendrogram generated by c lu ste r
anal ys is of the morphological data, accessio ns for the core were se lected to assure at
least one representative accession from the
most different morphological clu sters (data
not show n). First priority was give n to se lect
accessions from the main geographi ca l
di v isio n of each country repres ented in very
low freq uenc ies. From those c lu ste rs
co ntainin g many access ions, th e representative access ions we re selected on th e basis
of their res ista nce to diseases and pests and
th eir dry matter co ntent. Another criterion
was to favor the most w idesp read cu lti va rs

Table 4. Matrix of measures of association between the morphological descriptors' recorded in the whole andigena collection using Cramer 's Vmethod ; values above 0.30 are in
boldface, CIP, Lima, Peru, 1998.

=o;
~

g
-0
-0
-0

--0

Phb
Ste
Stw
Pde
Klx
Fwl
Fw2
Fw2d
FwSh
Ant
Pst
FrC
FrSh
Tskl
Tsk2
Tsk2d
Tsh l
Tshx
Tey
Tfl
Tf2
Tfd2

Phb

Ste

Stw

Pde

Klx

Fwl

Fw2

Fw2d FwSh Ant

Pst

FrC

FrSh

Tskl

Tsk2

Tsk2d Tsh 1

Tshx

Tey

Tf1

0.07
0.07
0.08
0.08
0.13
0.08
0.05
0.06
0.06
0.03
0.05
0.09
0.10
0.07
0.06
0.06
0.06
0.06
0.07
0.08
0.06

l
0.07
0.21
0.17
0.17
O.ll
0.07
0.06
0.13
0.19
0.09
0.08
0.27
0.14
0.14
0.09
0.09
0.06
0.06
0.1 3
0.12

l
0.12
0.08
0.10
0.08
0.10
0.05
0.04
0.05
0.06
0.07
O. ll
0.09
0.09
0.09
0.09
0.13
0.06
0.05
0.06

0.37
0.26
0.15
0. 12
0.07
0.07
0.1 0
0.07
0.08
0.16
0.12
0.1 3
0.08
0.06
0.09
0.06
0.07
0.08

0.34
0.17
0.16
0.06
0.07
0.09
0.08
0.08
0.18
O. ll
0.09
0.07
0.07
0.08
0.06
0.08
0.07

0.49
0.23
0.11
0.15
0.12
0.09
0.13
0.19
0.19
0. 13
0.12
0.13
0.13
0.10
0.19
0.11

0.30
0.07
0.07
0.09
0.08
0.06
0.18
0.1l
0.08
0.08
0.07
O.ll
0.07
0.09
0.07

l
0.06
0.06
0.07
0.08
0.07
0.13
0.06
0.07
0.08
0.05
0.09
0.08
0.06
0.05

l
0.08
0.05
0.23
0.12
O.ll
0.06
0.04
0.07
0.05
0.12
0.13

0.37
0.11
0.08
0.06
0.06
0.04
0.07
0.06
0.06
0.06

0.12
0.07
0.06
0.08
0.08
0.12
0.05
0.03
0.04

0.32
0.28
0.12
0.15
0.13
0.18
0.24
0.18

0.48
0.09
0.08
0.09
0.09
0.13
0.10

0.09
0.08
0.07
0.07
0.09
0.09

l
0.30
0.10
0.09
0.09

0.26
0.05
0.06

0.50
0.40 0.56

l
0.04
0.07
0.09
0.13
0.1l
0.05
0.06
0.07
0.06
0.04
0.04
0.05
0.07

0.20
0.06
0.05
0.16
O.ll
0.09
0.06
0.06
0.05
0.09
0.18
0.17

0.31
0.20
0.08
0.06
0.07

Tf2

Tf2d

a. Phb = plant habit, Ste = stem color, Stw = stem wing shape, Pde= pedical color, Pin = pairs primary lateral leaflets (no.), Pin = primary interiected leaflets an rachis (no.), Sin = secondary interiected
leaflets an petialules (no.), Pde= pedicel color, Klx =calyx color, Fwl = predominant flower color, Fw2 = secondary flower color, Fw2d =distribution of secondary flower color, Fwsh = corolla shape, Ant =
anther pigmentation, Pst = pistil pigmentation, fr( = fru it color, Fr Sh = fruit shape, Tskl = predominant tuber skin color, Tsk2 = secondary tuber skincolor, Tsk2d = distri bution of secondarytuber skin color,
Tshl = tu ber shape outline, Tshx = add tuber shape, Tey = tu bereye depth, Tfl = predominant tu ber flesh color, Tf2 = secondary tuberflesh color, Tf2d = distribution of secondary tuber flesh color.

that comprised a high numb er o f duplicate

accessions in th e collection.
Th e co re collectio n of 306 accessio ns
represe nts 12 .9% of the whole . Th e
anal ys is of means and frequ ency di st ributions based o n data for morphological
characters reco rded in the whole co llect ion
and th e co re subset we re statistically similar
(P > 0.05 for 21 of 25 chara cte rs) (Tabl e 2).
Th e core co ll ect ion w as slight ly biased to
retai n more cu lti va rs w ith unu sual flower
shape (ste ll ate) in culti vated potatoes, rare
tub er shapes, or hi gh ly dissected leaves .
Th e who le collection has low frequencies
of cultivars wit h these c hara cters . In the
case of the predominant tuber sk in color
(Tskl ) c haracter, the core also has a slightly
hi gher proportion of red colors th an in th e
w ho le col lection.
Th e ana lys is of means an d fre quen cy
distributions, based on data for resistance to
di seases and pests and tuber dry m atte r
content, revea led similar res ul ts fo r 4 out of
16 characters (Table 5). Comp ared with the
whole co ll ect ion, the core subs et was
bia sed only to wa rd lo wer susc eptibility for
re sistance to Potato Vi rus X; cys t nematode
(C /ob odera pa/Iida) ra ce pa2 and pa3; and
A nd ea n potato weev i I (Premn otrypes spp. )

Discussion
As th e culti vated potato col lecti on was
assembled at CIP, the prese nce of num erous
dupli ca te access ions of th e sam e culti va r
was quite evident. Therefo re, duplicate
id enti ficat ion wa s the first step to rat ionaliz e th e size of co llection to be maintained.
Th at was also the onl y way to sec ure the
clon al conserv ation of th e most diverse
sample o f Andean potato culti va rs. Th e
maintenance costs of the fi eld ge nebank
and in vitro collection of andigena were
redu ced by about 78 %, th e sam e proportion as the number of duplicates id entified.
Earli er, effo rts to minimi ze redund ancy in
th e potato collection was one of the most
impo rt ant ste ps in selecting a co re co ll ec ti on. W ith such a high leve l of duplication,
it is hi ghl y unlikel y that any rand om sample
192 Potato

tak en fr om the ori gi nal collection wo uld be

represe ntat ive of the total genetic di ve rsity
in this co ll ection.
Th e procedure to c hoo se accessio ns for
the co re col lec ti o n w as ba sed on propo rti o nal samp lin g acco rding to p asspo rt data
and mo1-ph o log ical clusters. The compa risons of mea ns and frequen cy di stri butions
for the morp holo gical descri ptor s suppo rted
th e sampli ng strategy. Furth ermore, another
ind epe nd ent assessme nt ba se d on the
di ffere nt reactions to diseases and pests also
validated th is proportional sa mplin g
method . In sofa r as possible , we attempted
to in c lud e in the core subset tho se ac cession s that co mbin e the mos t desirabl e traits
for breeding .
The se lect ion of a core col lecti o n of
culti va ted potatoes from La t in America
offers man y adva ntages. Since thi s co ll ection is c lon all y co nse rve d, o nl y the co re
collection wi ll be fie ld planted . Th e rese rve
collection ca n be maintain ed in v itro . Th e
assessm ent o f new conser va tion m ethods,
e.g., cryoprese rvation , could be m ade using
ju st a few access ions representin g a b roa d
genet ic ba se . Pathoge n cleanup of clones
for distribution sho uld be a priori ty in the
core coll ection. For securit y, co re acces sio ns cou ld be duplicated in severa l
countri es. Most important of all is th at th e
full bree din g potential of th ese farm erselected potato cultivars, w hi ch have been
maintain ed for centuries in th eir center of
di ve rsity, remains unknow n. A th oro u gh
eval uati o n of the ir disease an d pest res istance and oth er desi rable traits is now
feasib le because th e core sub set cove rs the
broa dest ge neti c ba se th at is ava il abl e in ex
si tu co nse rva tion.

Acknowledgments
We are grateful to CIP scienti sts who
contributed data on the reacti o ns of va rious
accession s to biotic and abiotic stresses; to
F. de Mendiburu for ass istance in stati stical
analyses; and to D. Spooner and M .
Boni erb ale for their re v ie w and hel pful
co mm ents.

Table 5. Means (X), standard errors (SE), and XJ. tests of frequency distribution for reaction to diseases and pests
of accessions in the whole and core collection of andigena potato cultivars. Descriptor states transformed
into a 0 to 9 scale' except for tuber dry matter (%), CIP, Lima, Peru, 1998.

Resistance to

Whole

Core

Classes

xi.

P>xzb

SE

SE

(no.)

calculated

506

6.61

0.02

95

6.26

0.07

8.186

0.085

304

5.09

0.03

55

4.76

l.32

5.193

0.268

Synchytrium wort
Mocrophomina

745
37

6.70
5.32

0.02
0.03

85
10

6.7
5.5

0.06
0.10

4
3

0.821
l.289

0.844
0.525

charcoal rot
Erwinia soft rot
Potato virus X
Potato virus Y
Potato leofroll virus

361
1739
1726
2505

7.75
5.52
5.68
6.74

0.03
0.03
0.03
0.02

42
206
204
289

7.52
5.31
5.62
6.71

0.09
0.09
0.09
0.05

4
3
3
3

2.732
6.86
5.778
4.551

0.435
0.032
0.056
0.103

1107

6.74

0.02

107

6.57

0.07

8.689

0.034

1096

6.62

0.02

107

6.14

0.10

25.558

0.001

566

6.77

0.02

36

6.72

0.06

0.25

0.882

480

6.60

0.02

54

6.74

0.06

l.248

0.536

1124

6.01

0.02

106

5.85

0.07

0.21

0.981

1756

7.41

0.03

218

7.29

0.10

5.793

0.215

572

8.15

0.03

62

7.35

0.12

21.08

0.0001

785

24.14

0.06

114

24.06

0.16

0.944

0.815

Phytophthora blight
in leaves

Phytophthora blight
in tubers

Globodera pa/Iida
race po2

Globodera pa/Iida
race po3

Globodera pa/Iida
Cusco population

Globodera ros
tochiensis
Meloidogyne
incognito acrita
Phthorimaea tuber
moth

Premnotrypes
Andean potato weevil
Tuber dry matter
content
0

According to Huaman et al. (1977).

P>X2 means the probability of X2 greater than X2 calculated. If the value is greater or equal to 0.05, there is no significant
difference between whole and core collections.

References Cited

Brown, A.H.D. 1989a. The case for core

collections. In: Brown, A.H.D., O.H .
Frankel, D.R. Marshall, and J.T. Wil liams
(eds.). The use of plant genetic resources.
Cambridge University Press, Cambridge,
UK.

__ . 1989b. Core col lections: A practical

approach to genetic resources management. Genome 31 :818-824.
. 1995. The core collection at the
crossroads. In: Hodgkin, T., A.H.D.
Brown, Th.J.L. van Hintum, and E.A.V.
Morales (eds.). Core collections of plant

CIP Program Report 1997-98

193

genetic resources. John Wiley & Sons,

Chichester, UK. p. 3-19
Frankel, O.H. and A.H.O. Brown.
1984.Current pl ant genetic resources - A
critica l appraisal. In: Genetics: New
Frontiers (Vol IV). Oxford and IBH
Publishin g, New D ehli , India.
Ha wkes, J.C. 1990. The potato . Evoluti o n,
biodiversity and genetic re sources.
Belhaven Press , Lond o n UK. 259 p.
Hua man, Z. 1986. Conserva t ion of potato
genetic 1esources at C IP. CIP Ci1n1lar
14(2):1-7.
____ . 1987. In ventory of Andean potato
cultivars w ith resistance to some pests
and diseases and othe r desirable traits.
CIP, Lim a, Per u. 22 p.
___ . 1994 . Ex situ conse rvation of potato
ge net ic resources at CIP. CIP Circular
20 (3):1-7 .
Hu ama n, Z., J. T. Williams, W . Salhuana ,
and L. Vincent . 19 77. Descriptors for the
cu ltivated potato and for the maintenance and distribution of germplasm
collections. Intern at iona l Board for Plant
Genetic Resources, Rome, Italy. 47 p .

194

Pototo

Huaman Z. and H. Stegemann . 1989. Use

of electrophoretic anal yses to verify
morpholog ically identical c lones in a
potato collection. Plant Va rieties and
Seeds 2:155-161.
Hu aman Z., A. Go lmi1zaie, and W. Amoros .
199 7. Th e potato. In: Fuccil lo, 0., L.
Sears, and P. Stapl eton. (eds.).
Biodi versity in trust: Conservation and
use of plant genetic resources in CGIAR
Centres . Cambrid ge U ni ve rsity Press,
Cambridge, U.K. p . 21-28.
Ortiz, R., E.N. Rui z-Ta pia , and A. MujicaSanchez . 1998. Sampling strategy for a
co 1e collection of Peru v ian quinoa
ge m1pl asrn . Theor. A pp l. Genet. 96:475 483
Rohl f, F.J. 1992 . NTSYS -pc , numerical
taxo nomy and multivariate analysis
sys tem . Exeter Publis h ing Ltd., NY, USA.
SAS In st itute . 1985 . SAS user's gu id e.
Statistics, ve rsion 5 eel. SAS Inst. , Cary,
NC, USA.

Rustic Seedbeds: A Bridge Between Formal and

Traditional Potato Seed Systems in Bol ivia
G . Aguirre 1 , J. Calderon\ D. B uitra g o', V. Iriarte\
and A. D e va ux 2

Po tato (So lanum tuberosum ) is a tra diti onal

stapl e c rop produ ced by more th an
20 0,000 fa mili es in Bo li v ia w ith total
prod uct io n es tim ated at 677,000 to ns in
1993 -94 (Ze ball os, 1997) . It is cu ltivated
mainl y in the h ighl and s, above 3,000 m,
w ith a low average yield of 5.6 t/ ha . Th e
seve re pov erty of th e Bo li v ian hi ghl and s is
pa rtly clu e to th e ha1sh c lim ate th at limits
th e opti o ns for ag ri cultural pro du cti o n.
Produ ctio n r isk for potato is hi gh cl ue to
rec urrent drought, fros t, an d hail. Yiel d
losses of 40-60 % are com mon. In ad d ition
fa rme rs also face pes ts and d isea ses .
Hi gh produ ction ri sk lea ds not only to
direct y ield lo sses , but also to low in ves tments in agri culture. A dve rse c li mate, hi gh
pro du ct io n ri sk , and pove1ty have led to a
low- in p ut/ low-ou tpu t system . De sp it e th ese
co nditi o ns, th ere is poten tial fo r improve ment o f produ ct io n systems as show n
throu gh on-farm resea rch (D eva ux and
Ga nd arill as, 1998 ). Th ere is a need fo r
inn ova ti ve strateg ies to pro mote an effi c ient
use of appro pri ate tec hn o log ies and to
in c rease th e in co me of fa rm ers fro m
ag ri culture.

Potato Seed Systems

In th e hi ghland s, mos t farm s are small
w ith an ave ra ge of 5 ha o f land and 1 ha of
pota to (Thi ele, 1999) . D ata obta ined fr om
yie ld surveys by PROI N PA3 in th e ea rl y

1 PRO IN PA Found o! io n, Coc habamba, Bolivio.

2 CI P, Coc haba mba, 13o li v ia.
3 l'R O INPA, fo rm erl y !h e nali onal po talo resea rch progrom
o(

Bol ivi c.i im pleme nt ed thro ugh an Jg rccmcnt be tween the

governmen t o f Boli v ia, CI P, and the Swiss Deve lop men t

Coopcralion (S OC), is now an aul onomous Founda ! ion
deali ng w it h Andean crops wi th its mai n em pha sis on
pol aloes .

J. R amos1, J. Blaj os1, G . Thiele 2

1990s showed th e impo rtance o f seed tu be r

qu ality and seed suffic iency fo r improv in g
potato produ ct iv ity (Terr azas et al., 1998).
Farmers are awa re th at their seed is of low
qu al ity afte r yea rs of use and know th ey
need to chci nge it, bu t it is not easy fOI' th em
to get good qu alit y seed. T1a d itio nal seed
systems suppl y about 9 8'Yo of th e seed
potato demand in Bo li via (Thi ele, 19 99).
W ithin th ese sys tem s, farm e1s co mmonl y
save thei r ow n seed from o ne seaso n to the
next. Anoth er opti o n is to excha nge seed
wit hin the fa m i ly o r w ith fr iends and
ne ighb ors (loca l seed flow) . They may also
bu y seed fr om merchants and in loca l fair s
(lon g di stance seed fl ow). Th ey base th eir
cho ice o n th e reputation of th e trad ers, but
th e qu ality is not always guarantee d and
in forma l seed frequ entl y is di seased or
in fes ted by pes ts (Thi ele, 1999) .
Bo li via is one of th e few co untri es in
Latin America that has a fun ctionin g fo rmal
potato seed sys tem based o n seed ce rtifi cati on (Bentley and Vasq ues, 19 98) . Th e
for ma l seed is produ ce d main ly by a
semipr iva te co mpany, th e Potato Seed
Produ cti o n U nit (UPS/ SEPA), w hi ch wa s
develop ed w ith th e support of th e Swi ss
Deve lopm ent Coo peration (SOC). Th e
he alth y seed produced by thi s co mpany is
mul tip lied by no ngove rnm ental o rga ni zati o ns (NGOs) and seed produ ce rs throu gh
th e for mal syste m. In 1996, th e 3,60 0 t of
seed produ ced thro ugh the forma I syste m
represe nted abo ut 2% of th e seed volum e
used in th e co untry (Program a N acion al de
Semilla s, 1996 ). Most hi ghland farmers do
not make use of th e form al syste m seed
because th ey l ive in area s w ith poor access
away fro m so urces of suppl y, they lac k th e

CIP Program Repo111997-98

195

money or credit to buy certified seed,

environment constraints make heavy seed
investment risk y, and most nati ve cultivars
are excluded from the commercial seed
system.

of 10 to 1 5 grams, 3 to 5 kg of seed are

usually required for a rustic bed of around
17 m 2 , which normally produces from 40 to
60 kg of high quality seed to be multiplied
the next year in the field.

Looking for w ays to supply seed from the

formal seed system to highland potato
production systems, PROINPA began
research to adapt rustic nurseri es for potato
seed production. NGOs developed the
rustic nursery concept in the 1980s for
vegetable production in the highlands of
Boli v ia. The technology was first adapted
for seed potato tuber production in southern Peru by SEINPA, a national seed potato
project implemented wi th the collaboration
of CIP and SOC. Researchers in PROINPA
adapted the idea to the highlands of Boli v ia
specifically for areas where risks for crop
production are high due to adverse climatic
conditions (Aguilera, 1995). This article
describes the concept of rustic seedbeds as
a strategy to bridge formal and informal
systems and make good quality seed
available to farmers. It also assesses
progress made so far in va lidating the
system at the farmer leve l.

The theoretical advantages offered by

rustic nurser ies are:
Decreased risks: protect io n against frost
and hail.
Better control of pests and diseases.
Low construction and production costs.
Optimization of local resources, including water, nutrients, soil , and labor.
Higher efficiency: high rate of multiplication, may be used twice a year for potato
or other crop, small seed tubers may be
used wi th the possibility of using other
planting material (e. g., sprout cuttings).

The Rustic Seedbed

Based on the experiences mentioned
above, and its own research, PROINPA
developed a potato seedbed prototype w ith
recommendations for its management
(Aguilera, 1995 ). The rustic seedbed or
nursery is a small hothouse (big box) made
of local material , stones or adobes. It is
about 1 m high with a surface area of from
1 5 to 20 m 2 (Figure 1). It can be covered
and protected against frost and hail with a
cover made of a wooden frame w ith a
plastic sheet, and hand watered during
drought. Its small size and protection from
weather constraints means more intensive
technology is applied to the nursery: pest
free soi I brought from non-culti va ted areas,
high planting density (2 2 plants/ m 2), better
fertilization using manure and limited
amounts of chemicals fertilizers (ma inly
phosphorus), watering, and integrated pest
management. When using small size tubers
196

Pololo

A proj ect was begun in 1995 with SOC

financial support, to evaluate and adapt the
rustic seedbeds in pilot areas where
PROINPA was already coordinating its
research and technology transfer activities
w ith other institutions. The objective was
to test the seedbed system as a way to
deve lop links betwee n the formal and the
informal systems for the production of
healthy seed and new or inaccessible
cultivars, and to disseminate this material in
the highlands . The initial target of the
project was for farmers to install and
va lidate 500 seedbeds.

Materials and Methods

Th e seedbed system was tested and va l idated in collaboration with farmers'
communities and N GOs. The pilot areas
chosen for testing represen ted different
agroecological zones of potato production
in the highlands of Bolivia. Technical
assistance was prov ided b y PROINPA
w hich also proposed a prototype design for
the seedbeds as described above (Agu i Iera,
1995). PROINPA produced the seed for the
non-commercial cultivars, normally basic
seed, but for some cultivars prebasic seed
was also used. In the case of commercial
cultivars, certified seed from the commercial formal seed syste m was supplied to

.......
. .- ..

~,.. ~'.

. _,

'

ii--' .) "

. ,.,

...... - ...

;-

-./ ,.

Figure 1. Typical rustic seedbed in the highlands of Bolivia. Women and children take an active part in seedbed

management.
fa rm ers to plant their seedbeds. In this case,
because of the bigger tubers, between 8 to
12 kg of seed was required. Th e PROINPA
seedbed d es ign and management were
adapted by farmers according to local
co nditions in each area . Initially, the
monitoring process emphasized seedbed
management, i.e., recordin g data on
cultivars used, quantity of seed, yield,
production costs, and data for seedbed
const ruction, and separating investment
and labor costs.
Taking into consideration th e production
system and farmers' requireme nts, four
categor ies of cultivar were multipli ed.
Category 1. Commercial cultivars
required by the market and includ ed in the
formal seed system. Basic or certified seed
from seed organizations was used.
Category 2. Native and introduced
cultivars of local comm ercial importance
grown principally in the highl ands and not

included in the formal seed production

system. At th e req uest of farmers' communities and NGOs, PROINPA cleaned some of
these cultivars of v iruses through meristem
tip culture to produce healthy seed.
Category 3. Native cultivars for subsistence us e, which co ntribute to potato
di ve rsity and food security in local communities. Th ere are still hundreds of nati ve
cu ltivars of different species cultivated by
farmers in the highlands of Bolivia. Twentyfour native cultivars from different
ecoregions were selected for th ei r ag roeconomical and social characteristics . Th ey
were clean ed of viruses by PROINPA and
return ed to farmers.
Category 4. Selected clones from
PROINPA breeding and selection program.
Th e ru stic beds allowed validation and
multiplication of these cultivars in several
potato-growing areas to determine th eir
potential demand.

(IP Program Report 1997-98

197

This ana lys is of seedbed mana gement

factors and yield is first report ed fo r the
northern Potosi Department o f Bol ivia .
Situ ated at 3,50 0 m , this is one of the
poorest an d mo st isola te d re gions of the
country. There PRO INPA, UN ICEF, and the
loca l farme 1s' fe deration collaborated to
inst all aro und 900 se edbeds. Com p lete data
were collected in 19 96 an d 1997 on 63 0 of
these seedb eds . Yield was meas ured in
seedbeds with the help of exten sionists
wo rkin g for the collaborative p roje ct
im pleme nted with U N ICEF. Multiple
regress ion anal yses of facto rs such as si te,
cul tiva r, seed si ze , seed quality (basic and
certified ), and years vs . yi elds were run for
all of these see dbeds. Sim ilar ana ly ses we re
carried out on 190 ca ses for th e ye ars 1995 ,
1996, an d 1998 in more comm erc iall y
o ri ented and access ib le potato grow ing
area s of cent1al Poto si. Closer m onitmin g of
seedbeds was possible throu gh the col labo ration w ith loc al deve lopme nt projects that
introduced seedbeds in their rural d evelop ment program .
In Central and Nmth ern Potosi , the
followi ng cultiva rs we re mult iplied in th e
seedbeds:

Category 1. W ayc h a (5. ancligena) and

Desiree (5. tu berosum), co mm erci al
culti vars with seed from the forma l system.
Category 2. lmill a Negra (5. ancligena),
Revo lucion (5. tuberosum ).
Category 3. Gend arme (5. ancligen a), a
nati ve culti var se lected for its res istan ce to
the nematode Nacobbus aberran s and its
relati ve tolerance to d roug ht, an d several
nati ve cultivars includin g Luky ty pes, bi tter
potatoes (5. juzepczukii).
Category 4. Som e promising sel ected
clones.
A follow up of several mu ltiplication
cycles of the seed produced in the seedbeds w as don e with three ca se study

198 Pototo

farmei-s, each from di ffe rent pro duction

systems : (1) a market-oriented system in
Loyaza, La Paz, at 3,300 111 , (2) a subsis tence farm er w ith access to th e market in
Kollana, La Paz, at 3,700 m , and (3) a
farmer grow ing na tive culti va rs for his own
cons umption with limited access to the
m arke t in Tapacari , Coc habam ba, at 3,500
rn. In the f irst two cases, co111rne1cia l native
cultivars we re mult ip lied; in the third , a
lo ca l nati ve culti var th at had bee n c leaned
b y PRO INPA was used. A small quantity of
basic seed , 3-5 kg, wa s ini tial ly given to th e
farn1e rs fo1 seedbed multiplicat ion. Yields
obtained for the successive m ultipli catio ns
were recorded and the us e of the potatoes
produced analyzed w ith th e fa r111e1s.
In Lo yaza, La Paz, the performance of
seed produ ce d in the seedb eds was com pa1ed to the inform all y pro du ced seed of
nin e farmers. Each farmer planted the seed
from the see dbeds nea r his own seed. At
harve st, th e y ield of th e plo t pla nted w ith
the seedb ed seed was compared to yie ld
samples taken in the fa rm er' s seed plot
A part ial budget anal ys is was us ed
considering the co st of the seed bed seed as
th e margin al cos t. As, in most cases,
farme rs ' seed had been multiplied fo r man y
yea 1s, no cost was att ributed to the fa rn1ers '
trad it ional seed. Th e margin al rate of return
was cal cu lated usin g the increase in net
benefits obtai ned with the seed bed seed,
com pared to the existi ng tech nolo gy, in this
case farm ers' see d, div id ed by t he additiona l cost ge nerat ed by th e use of the
seed bed seed, exp1e ssed as a percentage.
Farme1s we re fully invol ved in research
and val id atio n to obtai n the ir ideas on how
to ada pt and improve the technolo gy. In
collaboration w ith UN ICEF an d th e loca l
farmers' federa tion , su rveys w ere carried
out th rou gh participatory meetings wi th 305
farme rs to eva luate the acc eptance of the
seedbed system an d factors lim itin g the
ado ption of the technolog y.

Results
Yield (kg/rn 2 )

After three year s of v alid at io n, th e initial

target of installin g and va lid atin g 50 0
seedbeds was almo st tripl ed. More th an
1,200 farmers wo rkin g with PROINPA,
helped by N GO s and ot her develo pment
in stitutions, have in sta ll ed nea rl y 1,400
ru st ic see dbeds . It is estimated th at in 1998
farme rs produced about 45 t of heal thy
seed in seedbeds d es pite the shortage of
irrigation wate r and hi gh temperatures
ca used by el Nino.

6 .-~--~----~~~~~~~---

5
4

i~i
~I~

o L_~~~~--~~--~~~~~~

N=

478

40

106

Waycha

Gendarme

!millia negra

Native

Variety
Yield (kgirn ')

6 .-~~~~~~~~~~~~~-

Yield evaluation
Yield data for 1996 and 1997 1 in Northern Potosi, were analyz ed by cultivar in the
630 seedbeds (F igure 2A). Th e h ig hest y ield
of 5.6 kg/ m 2 was obtain ed w ith th e com m ercial cul t ivar Wayc ha. Th e 1996 ave ra ge
y ield of 2.7 kg/ m 2 was signi ficantly h ig her
than the 1 .4 kg/m 2 obta in ed in 199 7
beca use of El Ni no. Multipl e reg ress ion
ana lys is of yield aga in st site, cu ltiva r, seed
size , seed qua I ity (bas ic and ce rti fied), and
years showed significance only for a site
th at appea rs to be high er than th e others
(average of 3,958 m). Th e R2 for the
reg ress ion was low, R2 = 0.371 1 w hich
m ea ns th at factors contributing to y ield
other than tho se in c lud ed were mo1e
impo rtant, e.g ., ge neral ma nagement of th e
seedbeds by th e farme rs. Th e y iel d va riab i lity betwee n farmer s un d erscores the
importance of seedbed manageme nt. Th ese
data show that by red ucing cl im atic ri sks
and im p rovin g c rop man age m ent p ractices
in the seedbeds, fa rm ers ca n obtain a hig h
production and mu c h better seed multiplica tion ra tes th an in open fi e ld s. Average
seedbed production was m o re than 25 t/ha
in 1996, and 14 t/ha in 1997. Fi el d produ ction in t his area in norm al years is between
5 and 7 t/ ha, and is even lowe r in bad
yea rs.
In ce ntral Potos i, average seedbed y ie lds
we re higher, 2 .4 kg/m 2 in 1995 1 3.6 kg/m 2
in 1996 1 and 3.1 kg/m 2 in 1997; hi ghest
yield was 6 kg/ rn 2 (Figu1e 2B). Multiple
reg ress ion analyses of facto rs such as

T
4

97

36

16

26

Revoluci6n
lrnillia
negra

Desiree
Selected
clones
Genda rme
Waycha
Native

Variety

Figure 2. Seedbed yields in Bolivian highlands. (A)

Average of 630 seedbeds by cultivar,
northern Potosi , 1996 and 1997. (B )
Average of 190 seedbeds by culti var in
central Potos i, 1995, 1996, and 1997.
cult iva r, seed size, site, and seed qu ali ty
m ainl y showed th e effec t of cu lti va r and
seed quality (p 1ebas ic seed) on yield. The
R2 for the reg ress ion was high er than in
north ern Potos i (R 2 =0.602), partially
because farn1 ers receive d better techni ca l
support in seedbed ma nage m ent. On
ave1age th e com m erc ial cult ivar
Re vol ucion y ielded 1 kg/m 2 more than
other cultivars tested (Fi gure 2B) . Th ere was
no sig ni ficant y ie ld d ifference between
Revo lu cion and the prom isin g se le cted
clon es from th e PROINPA breeding program (Catego ry 4). W hen farmers used
their ow n seed in th e seedbed s, their y ield
was 1 .7 kg/m 2 lower t han th e prebasic seed.

CIP ProgramReport 1997-98

l 99

Economic evaluation
The co nstru ct io n costs of th e seed beds
vari ed betw ee n d i ffe rent a1eas and w it h the
ad ap tatio ns m ade by fa rm e1s. The labo r
co sts to buil d the seed bed s va ried fro m
US $2 0 to US$ 3 1 by seed bed de pend in g o n
th e ca1e taken in th e con stru ction and the
d istance farm ers had to brin g clea n soil for
th e seed beds . Cas h costs fm m ateri als,
m ainl y fo r th e p lasti c co ve r, va ri ed fro m
US$25 to US $40 acco rd in g to far m er' s
ad ap tati o n of the system . In so me cases,
farme rs repl aced the p lasti c sheets w ith
em pty fe rti Iize r ba gs redu c in g th e cas h
co sts to aro un d US $ 15 . Th e effect of th e
ty pe of cover on y ield depends o n th e
w eather co ndi tio ns, th e pl ast ic cove r helps
to m aintain a hi gher ave rage m ax imu m
temp erature in the seed bed s w hi ch all ows a
fa ster deve lop ment of th e cro p and al so
som e protecti o n against frost. But it w as
ob se 1ve d th at th e fe 1ti I izer bags we re as
effic ient as th e p lasti c cove r and th ey also
protect th e crop fro m frost. W ith an extern al

It appeared fro m th e fo ll ow up of seve ral

multipli ca ti o n cyc les of potatoes prod uced
in seedbed s w ith th e three case stud y
fa rm ers th at sm all tubers fro m th e f irst
multipli cati o n in th e seedbed w ere kept to
be pl anted in the seedbed the next seaso n.
Large r o nes w ere set as id e fo r fi e ld pl antin g
o r fo r sa le as seed to other fa rm ers (F igure
3). Th e mul tip li ca ti o n rate in th e seedbeds
va1ied fro m 1 :6 to 1 :4 1 w ith an average of
aro un d 1 :1 5 . In th e f ield , the multi p li cat io n
rate ranged fro m 1 :5 to 1 :14 w ith an
ave rage of 1 :7. In Loya za and Tapaca ri ,
fa rm ers sta rted to se ll seed after o ne fie ld
m ultipli ca t io n. The fa 1m er fro m Tapaca 1i
had th e hi ghest multipli ca tion rates in
seedbeds and in the fi eld. H e so ld 300 kg of
seed after one f ield multi p li ca ti o n and
p rod uced 3 .5 t of pota toes afte r th e seco nd
field mul t ip li ca t io n. In Ko ll ana, th e fa rm er
kept all hi s m aterial until the third f ield
multipli ca ti o n and produ ced mo re th an 7 t
of potatoes fro m th e ini t ial 5 kg of seed
p lanted in the seed bed.

Year1

Year3

Year2

Year4

------- -- ------ --r ------ -- --- ----T--- - - -----------

Field
mutiplication

Seedbed
3 to
5 kg
of
seed

-r------.- Sale of seed

"h...
I

30 to 80 kg

:
:

:,
:
I

300 tO 700 kg

3 to 5 kg of seed

I~I

ci
40 to 70 kg

~-,-1#~--F-~--

Sale of seed and

ware potatoes
~!), --,--,,.or further
- ~-...
mutiplication

:, hfff_'~ ~ @. \~.,,

~~-

~:

845 tO 3,520 kg
3 to 5 kg of Seed

I~ I ~

mult=tioo

30 to 70 kg

Figure 3. Seed flow from the seedbed combining multiplication in seedbeds and in open fields. Results from three
case studies, 1994 to 1997.

200

Potato

temperature of -2.5 C, th ere was no

signifi ca nt difference in yield in th e seedbeds betwee n th e two types of covers
(Vill alobos, 1998).

Th e pri ce of the bas ic seed in 1996 was

US $0.70/kg in the ce nters of supply. Th e
rea l pr ice to fa rm ers is at least 10% more,
beca use of transport and transa ction costs
associated w ith purchas in g thi s seed in a
distant location.

Th e produ ction costs in the seedbeds

va ri ed betwee n US$ 6.30 and US$ 23.90 per
seedbed dependin g o n th e inputs used:
seed, ferti Iize rs and pesti c id es, and labo r.
Labor costs in some cases, are substantiall y
in creased beca use of the need for irri gation
during period s of drought.

Th e analys is shows that even fa rm ers

w ith ave rage produ cti vity ca n produ ce seed
in th e seedbeds at a cost equivalent to or
less than the pri ce of ba sic seed . Th e most
effi c ient fa rm ers ca n produce seed mu ch
more cheap ly than th e bas ic seed pri ce
(Tab les 1 and 2). In addition, th e seedbeds
make fa rm ers less depend ent on the fo rm al
seed system w ith all its un certainti es: it
guarantees them access to seed of th e
culti va r th ey wa nt w hen they need it.
Compared to th e fo rm al seed system, in
w hi ch most farm ers buy more th an 100 o r
200 kg of seed at a time for fi eld multiplicati on, th e seedbed system means less
fin ancial risks. Thi s is a great adva ntage for
small fa rm ers w ith limited fin anc ial resou rces.

Con siderin g th e co nstru ction costs and

the potato production costs in the seedbeds,
the cost of th e seed obtained in th e seedbeds ca n be co mpared to th e cost of the
seed from th e fo rm al system. An oppo rtu nity cost was give n to labor at th e rural
wage rate and co nstru ction costs were
distributed over th e fo ur-yea r life of the
seedbed and th e cover (Tab le 1). A more
rea li st ic seed cost was ca lcul ated by
va luin g labor at an oppo rtunity cost of 50%
of th e rural wage, refl ectin g th e sca rcity of
local employm ent opportunities (Tab le 2).

Table 1. Seed production costs in the seedbeds {US$).

Annualized construction costs

Labor

Material

Production costs
Labor

Total Costs

Inputs

Labor

Total

Inputs+
mat.

Minimum
Maximum
Average

5.0
7.8
6.4

6.3
10.0
8.1

6.3
23.9
14.4

11.3
31.7
20.8

5.0
8.0
6.5

11.3
18.0
14.6

22.6
49.7
35.4

Table 2. Cost of seed at different productivity levels and based on two labor opportunity costs in 630 seedbeds,
1996 and 1997 (US$/kg).

Productivity

Seed cost:opportunity

Seed cost: appartunity

kg/seedbed

cost {50% rural wage)

cost (0% rural wage)

8.75 (minimum)
98.3 (maximum)
34.5 (mean)

Average

Minimum

Maximum

2.9
0.3
0.7

l.9

0.2
0.5

3.9
0.3
l.O

Average

Minimum

l.7

l.3

0.1
0.4

0.1
03

Maximum
2.1
0.2
0.5

CIPProgram Report 1997-98

201

Th e eva luation ca rri ed out in Loyaza , La

Paz, o n the performan ce of seed produ ced
in th e seedbeds as compared to farmer's
seed , showed that seed fro m th e seedbeds
was more product ive than fa rm ers' seed. It
gave some fa rm ers an acce ptab le return on
the add iti o nal cash inves tment after ju st
one fi eld mu lti pl icat ion (Table 3). Th ere is
need to co ntinue to monitor the performan ce of th e seed from the seedbed s to
determ ine w ith more prec ision th e economi c adva ntages of th e system und er
different con dition s.
A noth er adva nta ge of the seedbeds is
th at fa rm ers use th em to grow other crops
and protect them from fro st w hen it is too
late fo r th ese crops in ope n field s. Th e extra
produ ction means extra in come an d better
food fo r the fami ly.

Farmers' evaluation
Of the 305 farme rs interviewed in
nmth ern Potos i, w here more bed s we re
in sta ll ed, more than 95 % accepted th e
seedbed tech nology, 74.5% are interested
in buildin g more seedbeds, and 27.5%
have alrea dy in sta ll ed a seco nd on e
(Loredo and Choqu ehu anca, 1997). Ei ghtyfo Lff percent had set aside a specific pl ot in
open fi elds to multiply th e seed from the
seedbed s. Some constraints we re also
menti oned. Access to hea lth y seed for th e
seedbeds was mentio ned by 80% of th e
farm ers. Th ey wa nt to be sure th at in the
fu tu re th ey w ill have access to good seed
fo r th eir nurseries. Anoth er aspect of

conce rn for 70% of fa rm ers is th e rep lacement of th e substrate of th e seedbeds as

fe rtility dec lines and pests build up after a
num ber of yea rs of use. The best tec hni cal
option ava i lab le to avo id conta mi nation is
to remove the substrate and repla ce it by a
clean one. But that imp li es a lot of wo rk
and c lea n so il, w hi ch is not always ava ilable . Th e plastic cove rs, w hich deteriorate
rapidl y and are costly, concerned 69% of
farmers. Some farm ers have already fo und
altern ati ves to th e plastic co ve rs as menti o ned above. Farmers are gene1a l ly
enthu siastic about seedbeds, beca use the ir
use gives the m th e oppo rtunity to use
improved seed , and they have a tec hnica l
too l th ey ca n manage themsel ves.
No rm all y, men bu i Id the seedbeds, but
th ey later often mi grate in searc h of sea sona l off fa rm wo rk leav in g the women in
charge of seedbed mana gement. Wome n
pl ay a signi f icant ro le. In 1997, PROINPA
or iented trai nin g act iv iti es to women; and
out of 790 participants in tra ining act iv ities,
252 we re w omen. Th ey w ere very keen to
lea rn new techniqu es or techno logies for
potato product ion and are requ est in g more
trainin g support.

Conclusions
Th e seed bed is a simpl e techno logy to
renew fa rmer' s seed and improve potato
prod ucti vity that has been used by many
fa rm ers in the hi gh lands w here the technol ogy was val idated. Farmers adapted the

Table 3. Agroeconomical evaluation of see d produced in seedbeds of 9 farmers , co mpare d with farmers ' see d,
after one field multiplication, Loyaza, Bolivia, 1995-97.

Seedbed
yield

Average

202

Pololo

Marginal rate

Seedbed

Farmer

of return

seed

seed

(%)

3.3

12.0

8.4

163

1.7
2.3

4.7

3.1
5.4

73
118

(kg/m
Maximum
Min imum

Field production (t/ha)

2
)

8.1

seedbed des ign to redu ce cons tru ction

costs and m ake them easi er to man;:ige.
Farm ers have al so explored growi ng other
cro ps in the seedbed s after potato such as
fl owers, vege tabl es, or fodder. Some of
th em have tri ed alternati ve prop aga tion
method s suc h as sprout c uttin gs to p lant
th eir nurse ries. Others had th eir see dbeds
ce rtifi ed by th e Seed Certifi ca tion System
and bega n producing cert ified seed of
nat ive culti va rs.
Seedb ed tech no logy is a vehicle for
helpi ng farme rs to learn tec hni cal p r;:ic tices
fo r mo re effici ent crop m an;:igement.
Bec;:iuse wome n and c hildren share
res pon sibility for seedbed m anageme nt,
th ere is sc ope for work in g in c lose r co ll aboration w ith them and trainin g th em in
potato and seed producti on tec hniqu es.
Based on the ex per ience gai ned in
va l id ating th e tec hnology in thi s study,
there are sti II some cons trai nts to be
reso lve d, in c ludin g fert i lity of th e sub strate,
co ntro l of so il pests (nem atodes) and
di seases , ;:iccess to healthy seed to p lant the
see dbeds, and marketin g of seed or igin ated
from th e seedbeds after fi eld multip li ca tion.
It w ill requ ire more research to d eve lop
more effic ient methods to co nt ro l so il pests
and ma intai n substrate ferti lity. Rese<nch is
alread y und erway to evaluate crop rotation
and compos tin g in the seedb eds.

References Cited
Agu ile ra, J. 1995 . Producc i6n de tuberculos
- Se rnill;:i de p apa en ca m a proteg id a.
Manual Tecnico 1/9 5. PRO INPA,
Toralap;:i Experim ent Stat ion,
Coch ab;:irnba, Bo li v ia. 11 p.
Ben t ley, J. and D. V;:isques. 1998. Th e seed
potJto syste m in Bolivia: Organizati onal
growth ;:ind m iss in g l inks. Network Paper
No. 85. Overseas Deve lopment Inte rn ation al (OD I), Lo ndon , En gla nd . 11 p.
D eva ux, A . an d A. Gandari l las . 1998 . Mas
y rn ejor papa, Logros d e PROI N PA.
PROCAMPO. Rev ista de l Desa rrollo
Rura l CID/Bo li v ia 8 1 (Feb.):31-35 .
Loredo, V. and L.Choquehuanca. 19 97.
Prorno c i6n y Producc i6n de Sern ill a de
Papa, utili z;:i nd o ln ve rnad eros RC1 sticos, a
ni ve l de hoga res campes in os en el nort e
de Potos i. Corn ite ln ter institu c ion al de la
Papa-Norte de Poto si (CI PA NP),
Ll al lag u;:i, Potosi, Bo l ivia. 39 p.
Programa Naciona l d e Sem illas-D irecc i6 n
Naciona l d e Semil las-Secretaria
Nacional d e Agricu ltura y Ganad eri a
(SNAG). 1996. ln forme Anua l 1996. La
Paz, Bo li v iJ. 94 p.
Terrazas, F. , V. Suarez, G. Gardner, G.
Thi ele, A. Devaux, and T. Wa lker. 1998.
Diagnosin g potato producti v ity in
farme rs' fiel d s in Bo li v ia. Social Science
D epartm ent Wo rkin g Paper No . 19 98 -5.
lnternat io n;:i l Po tato Center, Lim a, Peru.

7 p.
A strategy needs to be d eve loped w ith
deve lopm ent in st itution s to supp ly hea lth y
seed for th e seed bed s and to re;:ic h the
fa rm ers of the in formal sys tem. Indi ge nou s
mark etin g c hannels shou ld be take n into
co nsid eration and bette r use d to link th e
forma l to th e trad ition al sys tems (Bent ley
and Vasq ues, 1998) . There is also a need to
co ntinu e tec hni ca l fo llow up in co l labo ration w ith NGOs and farm ers' co mmunities,
emph as iz ing ge nder impli cat ions,
po stharv es t acti v ities, seed mu ltipli ca tion in
open fi eld s, and seed mark etin g fo r those
w ho w;:int to spec ialize in that Jrea.

Thi ele, G. 1999. Informa l potato seed

sys tems in th e Andes: Why are th ey
imp ortant and w hat should we do w ith
th em? World Deve lopment 27(1 ):83 -99.
Vill alobos, J. 1999. Eva luaci6n de tres t ipos
de cobe rtura para la producci6n d e
tubercu los se m ill a de papa en camas
orga ni cas protegidas. A lt ipl ano centra l.
Gr;:iduate th es is. Fa cul tad d e Agro no mi a,
U ni vers idad Mayor de San A ndre s, La
Paz, Boli v ia. 122 p.
Zeb all os, H . 1997. Aspectos Eco n6 mi cos
d e la Produ cc i6n d e Papa en Bo li v ia.
CIP-COSUDE, In ternational Potato
Cente1, Lim a, Peru. 178 p.

CIP P1ogrom Repo11l997-98

203

Intensification of Potato Production in Rice-Based

Cropping Systems: A Rapid Rural Appraisal in West
Bengal
S.K. Bardhan Roy1, T. Walker2 , V.S. Khatana 3 , N.K. Sah a1, V.S. Verma",
M.S. Kadian 3 , A.J. Haverkort5, and W. Bowen"

Sin ce th e beg inning o f th e gree n revo luti on

in the mid-1960 s, th e gro wth in ce rea l
produ cti on in South As ia has bee n spec tac ul ar. Rec en tly, th e lo nge r-term
sustain ability of these produ cti vity ga in s has
been qu estion ed. Emergin g ev id ence on
stag natin g yi eld s, declinin g fac tor produ cti v ity, and increas in g intensifi ca ti onin duced land deg rad atio n is used to make a
case fo r th e imp ortance of threats to
sustai nab il ity (Pin ga li et al. , 1997 ). Th e
rice -w hea t sequ enti al cro ppin g system ,
w hi ch is practi ced o n 12 milli on ha in
South Asia, has bee n a fo ca l po in t fo r
di ag nosti c research on sustain ability-related
iss ues (Fujis aka et al., 199 4).
Th e ra pid ex pansion of potato pro du ctio n has also bee n impress ive, if large ly
unnoti ced, in th e irri gated low land s of
South Asia. In 1997, Indi a alone prod uced
more th an 20 mi 11 ion tons of potatoes. Th e
bulk of thi s produ ct ion ta kes pl ace on th e
lndo-Ga ngeti c Pl ain w here pota toes o ften
fo ll ow rain y-seaso n ri ce . Ri ce-potato
sequ enti al croppin g systems are in creasin gly co mmon in Paki stan, N epal,
Banglades h, Chin a, and Vi etnam.
Thi s resea rch is part of a new CIP proj ect
which ai ms to assess the sco pe for and
enh ance th e po ssi bili ties o f the sustain abl e
intensifi ca ti o n o f potato produ cti on in

1 D irecto ra te o f Agricul ture, W est Bengal, Ind ia.

2 CIP, Lima, Peru.
3 CIP, New Delhi, In dia.
4 CP RI, Pat na, India.
5 AB -DLO, Nethe rla nds .
6 IFDC/C IP, Li ma, Peru.

cerea l-based croppin g systems in th e subtropi ca l low land s o f Asia. Th e proj ec t

add resses th e fo ll ow in g issues: (1) th e
potenti al for in te nsify in g potato produ ct ion;
(2) the extent of c rop rep lace ment or
displ ace ment effects w hen potato is
introdu ced into diverse cropping systems;
(3) in te racti ons betw ee n potatoes and oth ~ r
crops in th e syste m; (4) impli ca tions of
intensify in g po tato pro du cti on fo r natural
reso urces ma nage men t; and (5 ) th e id enti fication of researc h to ove rcome prob lems of
locati on specificity.
W e beg in our inqui ry with exploratory
diagnosti c resea rch in th e form of a rapid
rural appraisa l on th e sustain ability of the
ri ce- potato seq uenti al c ropp in g systems in
Wes t Be ngal. Thi s is the most impressive
exampl e of th e ex pansion of potato produ ction in th e subtropi ca l irri gated lowlands of
Asia. To set th e stage fo r th e rapid rural
app ra isal, we prov id e backg round in fo rma tion on the ra tio nale fo r foc uss in g on the
ri ce-potato-rice c roppin g system, on
produ cti on trend s in West Bengal, and on
the behavi or o f potato pri ces which w ere
forem os t in the mind s of fa rm ers at th e tim e
of the fi eld wo rk .
Rationale for Investigating the
Sustainability of Rice-Potato-Rice
In M ay 199 7, a " brain storming sess ion"
w as held in Shiml a at the headquarters o f
the Ce ntral Potato Resea rch In stitute (C PRI )
to obtain base lin e in fo rm ati on on potatoba se d di ve rsifi ca ti on in th e lndo-Gan geti c
plain. Th e meetin g wa s we ll attend ed not
onl y by potato sc ienti sts and directors from

CIP P1og1om Repo1t 1997-98

205

the Ind ian Council of Agricul tural Researc h

(ICAR ) but also by several ICAR scie ntists
from the rice and maiz e programs . Participants , w ho represented three broad regio ns
of th e lndo-G angetic Plain , des cribed
common and eme rgin g cropp in g systems.
In total , more than 15 diffei-ent croppin g
systems, featuring potato , were identified .
For most of th ese, wet-season rice prece ded
potato. Potato had also fo und a hom e in
other cerea l-b ased systems. For example,
an intercrop of ma ize/ po tato is growi ng in
popularity in Eastern Indi a.
No o ne cerea l-based po tato cropping
system appears to account for mo re th an 40
to 50% of potato produ cti o n in the lnd oGan ge tic Plain s of India . In relative terms ,
no si ngle cropping sys tem in co rporatin g
potatoes is as com mon as ri ce-wheat. On
th e othe r hand, two ric e- ba sed cropping
systems that include potato ex hibited good
potential for sys tems and natural resou rce
manageme nt implications. One was rice potato-sunfl ower, w hich is increasingly
popular in northwestern Indi a. The oth er
was ri ce- potato -rice that w as felt to be ve ry
producti ve, but w as viewed as in te nsive in
its demand for wa ter. Hardp ans from
puddling ri ce we re also thought .to compro mi se the productivity of the system over
tim e. Some anecdotal evi dence suggested
th at potato yiel ds were dec linin g. For these
reasons and bec ause of it s popu la rity acro ss
several countries, rice-potato-rice was
singled out for our in vest iga tion , w hi ch
bega n w ith a rapid rur al appraisal (RR A) in
West Bengal, w here this system is most
w idely practiced.
Production and Price
O ve r the past 20 years, irri gat io n
capacity ha s inne ased rap idl y fu elin g the
expansio n of boro (summer) rice an d
openin g up oppo rtuniti es for triple croppin g. In onl y o ne decade from 1985 to
1995, the area in boro ric e has doubl ed
from about 0.5 to 1.0 million ha (Direc torate of Agriculture, 1994 an d 1996).

206

Poloto

Othe r rabi (post rainy seaso n) and

summe r (non -r ain y) crops have also
benefited enormous ly from the in creasing
avai lJbi lity of irrigation. Whea t area
in creased fivefo ld in the Gree n Re vo luti on
Deca de from 1966-1976 (Fi gure 1). Yie ld s
more than d oubl ed in five yea rs (Figure 2) .
M ustard acreage rose dram at icall y in the
mid- l 980s in res ponse to a doublin g in
pri ce. Mustard yiel d has in creased slow ly
from a ve ry low base . Stead y expansion of
pota to area and progress in intensifyi ng
yiel d has res u lted in sustained growth in
producti on (Fi gures 1 and 2). Sesam e, a
summ er o i lseed crop, ha s also expanded
rap idl y.
This robust growth in potato produ cti on
ha s been achieved in spite of strong price
seasonality and cycli cal pri ce risk. Th e
main harves t on the Plain s occ urs in
Feb ruary and March. Tempera tures ri se
stea di ly until th e onset of the southwest
monsoon in Jun e. Tradition al storage is not
an effec ti ve o pt io n from mid-Apr il onwa rd s,
and co ld storage is costly. Prices ri se until
th e rainy season crop is harves ted in
October. Periodica lly, such as in 199 2, th is
typical seasonal pri ce mo veme nt, w hi ch
makes cold storage profitabl e, is not
manifested (see Fi gure 3). Supply substantial ly exceeds demand, and , in sub sequ ent
months, pri ces do not rec ove r from th ei r
ha rves t season lo w. Th e d ep ressin g effec t
on pr ices of a very large in crease in prod uction was dram ati c in 199 7. A rea planted to
potatoes in Wes t Bengal reac hed a record
hi gh of about 31 5,0 00 ha in 1996-97, up
from the previous high of 255,000 ha the
yea r before. With this back gro und , it was
not surpri sin g that production probl em s
we re not fore mo st in the m inds of farmers
w hen we car ri ed out the rapid rural appraisal in 1998. Whe n asked about potatorelated probl ems, the prev ious year's low
pri ces and Ii m ited co ld storage avai la bi I ity
dominated th e co nve rsation.
Th e 19 97-98 grow ing seaso n was a
retu rn to normal cy. Ind eed , potato pric es
ha ve rebound ed to set new highs for th e

Log base 2

Hectares ('000)

512

256

'

128

,...

64

6
~,.

5
4

...

32

16
8

2
Sesame

2
0

0
1946

1956

1966

1976

1996

1986

Figure 1. Area of selected post rainy season and summer crops in West Bengal, 1946-4 7 to 1995-96. (Source:

Directorate of Agriculture (1994 and 1996).)

Tons /ha

Log base 2

15

32.7
Potato

14

16.3

13

8.1

12

4.0

Wheat

11

2.0

10

1.0

0.5

0.2
0.1

1946

1956

1966

1976

1986

1996

Figure 2. Yield of selected post rainy season and summer crops in West Bengal, 1946-47 to 1995-96.

(Source: Directorate of Agriculture (1994 and 1996}.)

CIP P1ogrom Repo1t 1997-98

207

Price (Rs/t)
10,000
8,000

1991

1992

1993

1994

1995

1996

1 997

1998

Figure 3. Real prices {1990 = 100) by week in the Calcutta wholesale potato market, January 1991 to July 1998.

{Gaps indicate missing data .)

1990s (Fi gure 3). Th e 1996-97 ex perience w ill ca use a major hi cc up in th e
ex pansio nary path of potato produ cti o n in
West Benga l, b ut it sho u ld not sub stanti ally
affect the fund amen tal s res po nsible fo r th e
stea dy upwa rd trend in pro du cti on.
The Rapid Rural Appraisal
Th e rapid rural appraisa l wa s pattern ed
after simil ar diagnosti c resea rch o n ri cew heat, parti cularl y th e H arrin gto n et al.
(1 993) repo rt o n this cro ppin g system in
Karn al and Kuruksh etra di stri cts in Harya na.
The di ag nostic re searc h feat ured two
inte rdis c iplin ary researc h tea ms of NA RS
(nat io nal ag ri cultural resea rc h sys tem) and
IA RC (intern ational agri c ultural resea rc h
ce nter) sc ienti sts.

Potato production in W est Bengal is

co nce ntrated in Midn apo re, H oog hly, and
Burd wa n di stri cts. Vill ages in th ose di stri cts
we re v isited ove r a two-wee k per iod in late
Janu ary an d earl y Febru ary 1998 pr io r to
th e harves t. Group in te rviews we re
co mpl emented by v isits to po tato field s in
th e v i IIage. Seventeen vi IIages we re v isited
w ith th e ass istance of staff of th e Directo rate of Agriculture of th e State of West
Benga l. Th e area planted to ri ce-potato -ri ce
and prox imity to a Blo c k Deve lopm ent
offi ce we re th e criteri a fo r vill age select ion .

208

Potato

Findings and Prospects for

Sustainable Diversification
Th e fi ve c roppi ng sequ ences desc rib ed
in Figure 4 are broa d ly repr ese ntati ve fo r
the surveyed vil lages. Pul ses, ju te, and
vege tables ca n also figure in these ri cepotato sequ enti al cro ppin g system s.
Cro ppin g pattern s almost alwa ys start w ith
a cro p of aman (rain y se aso n) ri ce . Th e
aman ri ce-potato-ses ame sequ ence is mos t
po pul ar, pa rti c ul arly in Midn apur w here
heav ier c lay so il is not as com mon as in
Burdwa n and H oog hl y. A ny se riou s li mitatio n in wate r su pply o r in soi l-avai lab le
wa ter ca pac ity beca use of lighter, sa nd ie r
soi l co nfin es sum mer croppin g to sesame .
If wa ter is so mew hat li m itin g, aman ri cepotato-boro ri ce ca n still be ta ken, but an
ea rli er maturin g ri ce vari ety suc h as IR -36 is
gro w n in stea d of Swarna Ma hsuri . W hen
wa ter is no n-limi t in g, farm ers fa vor Swa rn a
Ma hsuri , w hi ch subseque ntl y delays th e
produ ction of boro rice by abou t 1 5 d ays.
Thi s cro pp in g seq uence is the farm er's
preference fo r maxi mizin g returns and
produ ctivity (D irecto rate of Agri cu lture,
1992). A few fa rm ers w ho pl ant potatoes
harvest th em ea rl y at 75 to 90 days fo r sa le
o n th e fres h marke t. W hen potatoes are
harves ted ea rly, bo ro ric e is pl anted o n
tim e. Potatoes do not enter into th e
cro ppin g sequ ence w hen soi ls are too

Jul

Sequence

Aug Sep Oct Nov Dec Jan Feb Mar Apr

May Jun
--

Aman riceAman rice

potato-sesame I Swarn a Mahsuri

(135 days)

Potato
Kufri Jyoti (120 days)

II

II

Sesame
B-67 (90 days)

Aman rice
IR-36(100 days)

Potato
Kufri Jyoti (120 daz:s)

I I

Aman rice

Aman ricepotato-boro
rice

Potato
Kufri Jz:oti (120 days)

I Swarn a Mahsuri (135 days)I

I

Potato Kufri
Jyoti (75 days)

Aman rice

I Swarna Mahsuri (135 days)!

Aman riceboro rice

I I

Aman rice

I Swarn a M ahsuri (135 days)!

11

Bora rice
IR-36 (100 daz:s)

Bora rice
IR-36 ~100 dalsl

Bora rice
IR-36 (1 oo days)

Boro rice
IR-36 (100 days)

I
I

I
I

Figure 4. Typical rice and potato cropping sequences in the surveyed villages.
heavy or in areas where drainage is problematic. These areas are restricted to the
double cropping of rice .
The most surprising feature of the
sequential cropping systems depicted in
Figure 4 is the dominance of a moderately
long-duration rice variety to start the
sequence. The cultivar of preference in the
rainy aman season is still Swarna Mahsuri,
a 135-to-140-day variety characterized by
high yields, good cooking quality, and wide
adaptabi I ity. Earlier ma tu ring varieties,
such as IR-36 are available, but farmers are
willing to accept lower boro productivity if
water for the summer season is readily
available. This strong preference for a
moderately long-duration rice variety to
initiate a triple cropping sequence suggests
that the conventional wisdom about the
catalytic role of high yielding varieties of
rice in facilitating the place of potatoes in
the cropping system is somewhat misplaced.
The optimal planting time for potato is
mid-November, and the turnaround time
between aman rice and potato is 10-15
days depending on soil texture. Therefore,

rice varieties of durations shorter than 120

days are not needed for the aman planting
if irrigation is readily available in the
summer season. Nonetheless, the preference for later maturing Swarna Mahsuri
does increase the premium for reducing
turnaround time in planting potato after
aman rice.
Replacing a cool dry season fallow with
potatoes to arrive at a triple cropping
sequence is characterized by both positive
and negative interactions for the following
crop. Positive benefits take the form of cost
savings in fertilizer use on boro rice and
sesame. On a negative note, farmers
believe that insect and disease infestation is
higher on both rice and sesame when these
summer crops are preceded by potatoes
than when they are sown after seasonal
fallows.
Over the past decade, potato yields have
increased from about 1 5 to 25 t/ha. Farmers attribute this growth in potato productivity to more intensive use of inorganic
fertilizer, better seed quality, and improved
crop management. Ferti I izer doses are
high . During the last 3 or 4 years, fertilizer

CIP Program Re port 1997-98

209

intensity has hit a plateau, and farmers feel

that gains from increasing fertilizer use are
exhausted. In the majority of villages, use
of farmyard manure is declining. To some
extent, oilcake is purchased to substitute for
manure.
Unlike the rice-wheat double crop,
weeds are not a problem in the rice-potatorice triple crop. Weed infestation was only
readil y visible in the village with the largest
holding size. There was scant evidence of
intensification-induced degradation.
Inserting potato befo re boro rice delays
the planting of boro rice by about 1 5 days
and substantially increases the demand for
water w ith rapidly rising temperature in
April and May. Groundwater is recharged
during the rain y season, but is perceived to
be seasonally declining during the warm
summer months when boro rice is cultivated. A declining groundwater table
appears to be less problematic in West
Bengal than in the Punjab where seasonal
recharge is considerably less. Nonetheless,
water for irrigation is heavily subsidized ,
and the resource cost of delayed sowing of
boro rice following potato is equivalent to
about five irrig ations .
Th e prospects are bri ght for planting
more area to potato after wet season rice in
the specialized potato-growing Midnapore,
Burdwa n, and Hooghl y districts of West
Bengal. As the 1997 slump in potato
prices suggests, cyclical price risk can
temporarily Ii m it area response , but such
periodic setbacks should not compromise
the strong growth in area that has occurred
in recent decades. Mustard is often grown
rather than potato because of the
government's protecti ve policy on oilseed
crops. This policy results in more area
being planted to mustard than would
otherwise be the case. Liberalization of
oilseed imports should stimulate area
response in potatoes . State and central
government regulatory policies on cold
storage and on the establishment of potato
processing plants may also adversely affect

210

Pototo

the demand for potatoes . Ade quate road

infrastructure should also contribute to the
strong growth in area response.
Increased area in potato wi ll not be
confined to the rice-potato-rice system.
Dec lining seasonal groundwater availability
w ill be an increasingly important constraint
to the stability and expansion of that
system. Prospects are brighter for the
expansion of rice-potato-sesame which is
less intensi ve in its use of w ater.
The expansion of potato area in Wes t
Bengal has not come at the expense of
alternative crops. Increasing potato area
has been contemporaneous w ith the
expansion of irrigation. The rice-potatorice sequence appears to be a more natural
agronomic fit than rice-wheat. In contrast
to w heat, crop stands of potato follo w ing
puddled rice are definitely not a problem.
The wet season rice crop does not unduly
delay the planting of potato. Potato in the
rice-potato-rice system resembles rice in
the rice-wheat system in the sense that
other crops in the seasonal rotation do not
compromise its productivity. By the same
token , boro rice takes on the role of w heat
in the rice- w heat system. W ith potato, the
sowing of the boro rice crop is dela yed and
pest and disease incidence is percei ve d by
farmers to be higher than in a rice-rice
double crop. Nonetheless, this rapid rural
appraisal reinforced the impression that th e
negative productivity effects of potato on
boro rice in the rice-potato-rice system are
sma Iler than the ad ve rse productivity
consequences of puddled rice on w heat.
Longer-term threats to sustaining the
expansion of potato production in these
specialized potato-growin g districts of West
Bengal are not imminent. No farmers
stated that potato producti v ity was declining over time. Evidence for intensificationinduced land degradation is scanty. Hardpans were only cited as a problem in one
v illage, and hardpans are desirable for
irrigation efficiency in the production of
puddled rice. Water for irrigation, particu-

larly seasonal groundwater availability, will

becom e sca rcer if over-utilization co ntinues . But the cas ualty from the fa ilure to
price wate r acco rdin g to its sca rcity va lu e
w ill be boro ri ce and not potato. Far mers
in several v ill ages no longe r produced boro
ric e wh en governm ent authoriti es imp osed
seaso nal irri gation restrictio ns. Nitrate
leachin g resulting in ground water pollution
could be a co nseq uence of the relatively
high doses of inorga ni c fertilizer app li ed in
the intensificat ion of potato produ cti on.
However, no one said that nitrate leac hin g
was a problem. (At the time of thi s rapid
rural app raisa l, arseni c poisoning was the
main co nce rn in th e co ntamination of
drinkin g water in so me restricted areas in
West Bengal.)
We obse rved that maintainin g the
histor ica l gro w th rate in yield w ill be the
mo st diffi cult propo siti on to sustain in the
intensifi cation of potato production in West
Bengal. Average on-farm y ields of 25 to 30
t/ ha may be ap proac hin g th e produ ctiv ity
poten ti al of grow in g potatoes in short-d ay
conditions. Biotic and ab iotic stresses are
not significant yield red uce rs. Seed quality
is reaso nab ly good and inorganic ferti I izer
is read il y ava il ab le. Improv in g irri gation
and crop management and seed qu ality
furth er co uld leve rage so me ga in s in
produ ct iv ity and there may be som e scope
for savings in fertilizer, but in general, we
we re hard pressed to identify produ ct ion
co nstraints w hose so luti on would translate
into sub stan ti al gains in productivity. W ith
limited scope for irri gation ex pansion , an
expected decline in the positive growth rate
in yield is a ca use for concern .
Th e productivity status of ric e-potato-ri ce
in W est Bengal contrasts sharply with
Ban glad es h, Vietnam, and South China
where the sa me cropping system is ob served. In these areas, so lvi ng th e seed
qu ality and ava ilability problem co uld
eas il y result in yield in creases of f ive tons
or more.

Priority Research Areas and Future

Directions
Following the im p leme ntati on of the
rapid rur al appraisa l, we arrived at a
consensus on spec ifi c topics for future
re searc h. Three or four ca ndid ate topics
w ere li sted for each of the more general
areas of productivity, susta inab ility, and
demand. Most ca ndid ate topi cs related to
th e need for focused diagnostic resea rch.
Thi s li st of priorities has guid ed app li ed
resea rch that started in the 1998-99 c roppin g seaso n.

Research and training in modeling so me

of the sustai nability dimensions of growing
potatoes in short-day irrig ated conditions
also began in 1998. Such mod elin g will
figure more promin ent ly in this project in
th e future as w ill a wide r reg ional participati on of co untries of the subtropi ca l low lands
of As ia w here potatoes increa si ngly con tribute to the cas h in come of small farm
hou seho ld s and to the diversification of the
diet of consum ers.
References

Directorate o f Agriculture. 1992. Cost of

production of principal crops in West
Bengal. Calcutta: Evaluation Wing,
Directorate of Agriculture, Government
of West Bengal, In d ia.
Directo rate of Agri cul ture. 1994. Estimates
of area and product ion of principal c rops
in West Bengal 199 1-92. Calcutta:
Evaluation Wing, Directorate of Agric ulture, Governm ent of West Bengal, Indi a.
Directo rate of Agri culture. 1996 . Est im ates
of area and prod uction of prin c ipal cro ps
in West Bengal 1994-95. Calcutta:
Evaluation Wing, Directorate of Ag ri culture, Government of West Ben ga l, India .
Fuji saka, S., L. Harrin gto n, and P. Hobbs .
1994. Rice-wheat in So uth Asia : Longterm priorities estab li shed throu gh
di agnosti c re sea rch. Ag. Sys. 46(2):169181.
H arrin gton , L.W., S. Fuji sa ka, M.L. Morris,
P. H obbs , H.C. Sh arma, M.K . Chaudhary,
and S. D . Dhim an. 1993. Wheat and rice

CIP Program Report 1997-98

211

in Kamal and Kuruk shet ra districts,

H arya na , India: Farmers' practice s,
probl ems, and an age nda for action.
H aryana Ag ricu ltural University, Indi an
Council of Agr icul tu ral Research,
Internatio nal Maize and Wheat Improvement Ce nter, and the Internation al Rice
Research Insti tute. CIMMYT, Mexico,
D.F., Mexico.
ICAR (Indian Council of Agr icu ltural
Research). 199 7. Summary of the

212

Potato

brainstorming sess ion on potato in the

rice-wh eat cropp in g reg ion of the lndoGangetic Plain . (Unprocessed) ICA R,
New Delhi, Ind ia.
Pin gal i, P.L., M. Hossain, and R.V.
Gerpacio. 1997 . As ian rice bow ls : The
returning crisi s? CA B Internation al,
Wa llin gfo rd , U .K.

On-Farm Profitability of TPS Utilization Technologies

A. Chilver1, T. Walker 2, V. Khatana 3, H. Fano 2 , R. Suherman4, and A. Rizk 5

Poor tuber seed quality is often cited as the

most important factor limitin g potato
productivity in developing co untri es (Rasco,
1994). Improving tuber seed quality and
ava il ab ility is institution ally complex in th e
tropics and sub-tropics. Progress of publicsecto r tuber seed programs has usually
been slow and disappointing even in
countries where sites for seed multiplication
are agroeco logically attractive (Crissman,
1987). True potato seed (TPS) is a ca ptivating technological alternative that offers
farmers an option to overcome the
abovementioned weaknesses of clonall y
propagated tubers as a source of planting
materials. Side-by-side comparisons of a
jar of TPS with 20 1 00-kg bags of clonal
planting material-the amounts needed to
plant one hectare-are visually compelling
abo ut the potential of th e technology.
Although the investment in agricultural
researc h and extension of TPS technologies
is sma l I compared with the amount spe nt
on clonal seed technolo gies and se lection
systems, sufficient experience ha s acc umulated, largely over the past 20 yea rs, to
review ac hievements, shortcomings, and
future prospects (Almekinders et al., 1996;
Simmonds, 1997). "At present, there are
few potato-producing countries in th e
developing world where TPS has not been
tried, or suggested as a means to aI lev iate
seed problems (p. 290, Almekinders et al.,
1996)." One of the defici encies in this
expe ri ence has been the lack of documentation of on-farm performan ce with farmer
man ageme nt (Simmonds, 1 997).

1 Delivery Prog rarnrne, Ind o nesia.

2 CIP, Lirna, Peru.
3 CIP, New Delhi, India.
4 Lernbang Hort ic ultural Research In st itut e, Lernbang,
Indonesia.
5 Ag ricultu ral Economic Research In stitu te, Gharbia
Goverrnorate, Egypt.

We address this large gap in the literature

by reportin g the results of on-farm resea rch in
Egypt, Indi a, Indon es ia, and Peru conducted
in th e mid-1990s. The research des ign is th e
same in the first three countries w here farmers
are experimenting with TPS technologi es in
the initial stages of acceptance or reje ction .
The Peru vian res ults are based on farmermanaged, on-farm trials.

Sampling and Data Collection

Sites in Egypt, India , and Indonesia were
chosen based on farmers' experience w ith
TPS utilization technologies. TPS farmers
were stratified by wealth group based on
asset holdings, mainl y land, and a random
stratified sample was taken.
Th e unit of observation was the fi eld.
One field planted using a TPS technology
was monitored during the growing season.
The point of reference for comparative
inference was a neighboring or nearby field
planted to co nvent ional seed tuber technology. For those TPS farmers without a tuberseed field available for comparison, a
nearby field of anoth er farmer was chosen.
An integ rated production questionn aire
combining survey responses and field
measurements was canvassed. Fi eld visits
were made at planting, 45 days after
planting, harvesting (tw ice), and postharvest
market ing. Seed rate, yield, and tuber size
were measured. Loc al market data were
collected week ly on seed and ware pri ces.

The format of data collection in Peru was

in the spirit but not to the letter of data
collected in the other three countries . Data
in Peru refe r to more structured on-farm
trials with two treatments in fields
CIP Prngrnm Repoit 1997-98

21 3

app roac hing one hectare in size. Fields in

ten v ill age locales were monitored.
Co mparati ve eva lu atio n was ca rri ed out
in nin e sites for TPS seedling tubers and in
two of th e three sites w here th ere were
suffi c ient observations to ca rry out an
analys is of TPS transpl ants. Sites are
desc ri bed in Table 1. A lm ost all sites
co ntained multipl e locat ions co mprising
seve ral or many vi II ages. Although a
minimum of 15 to 25 plots per site was th e
goa l, th e total number of TPS field s ranged
from 13 to 99. Experi ence w ith and
ava il ab ility of TPS in th e reg ion signi fi ca ntl y
influ enced sampl e size.

Results
Th e eco nomic assessment shows th at TPS
has mi xed prospects. Of the nin e generalized location s, TPS seedling tuber tec hn o logy was most profitable in Chacas, Peru ,
quite attracti ve in th e north eastern hill s of
Indi a and th e Ni le Delta of Egypt, and
marginally profitable in the northeastern
plains of India (Figure 1). In the other five
sites, fa rm ers lost the eq ui va lent of more
th an US$1 00 per hectare (relat ive to c Ion ally

propa ga ted materi al) by in ve stin g in

seedlin g tub ers. (Because o f th e sma ll
areas used to tes t TPS , no farmer s los t a
signi f icant am o unt o f mon ey o n th e
tec hn o logy.)
The compara ti ve profitability of TPS
transp lants hin ges on whether or not the
output is dest in ed for wa re or seed use.
Prod uct ion from TPS transplants was onl y
competitive if th e small er-sized output was
retained or so ld for seed (see th e ho ll ow
po int estim ates in Fi gure 1). Destining all
output fo r ware co nsumption in th e sa me
year was a los in g proposition (the so li d
point estim ates fo r Indi a NEH (TP) and NEP
(T P) in Fi gure 1). Demand for TPS as
tran spl ants should be strong in Indi a's
north eastern hills. There, TPS techn o logy
has a clea r eco nom ic advantage ove r
clonally propagated material by abo ut
$500/ ha. Transplants grown for tuber seed
offer an attractive be nefit. The fate of
transp lants o n th e northeastern pl ai ns is
more prob lematic because th e eco nomi c
adva ntage of TPS seed lin g tubers was onl y
abo ut $60/ ha. At this lowe r level of
profitab ility, it is unlikely that the derived
demand for transp lants could be susta ined.

Table 1. Description of recent on-farm research comparing TPS technologies with clonal seed tubers.

Country

Agroecology

Fields
Farmers

India
India
India
India
Indio
Egypt
Egypt
Indonesia
Peru
Total plots

Northeastern plains
Northeastern hills
Northcentral plains
Western arid zone
Deccan Plateau
Desert
Delta
Highlands
Highlands

groups comprised of 197 formers.

Source: Chilver (1997).

214

Potato

119
24
22
11
14
7
40
67
11

Clonal

TPS seedling

seed tubers

tubers

transplants

99
20
17
10
11
3
26
32
11
229

99
15
17
13
14
13
42
45

34
8
0
3
0
0
0
0
0
53

11

269

TPS

(US$/ha)
2,727

3000
3000

516

1000

484

474

0
-1000

-607

-771

-2000
-2 ,14 2

-3000

Peru
Chacas

Egypt
Delta

India
NEH

India
NEP

Ind ia
WAZ

India
NCP

India
DP

Egypt
NRA

In dia
Ind ia Ind ones ia
NEH(TP) NEP(TP)

Figure 1. Difference in net economic benefit (USS/ho) of TPS technologies from clonal tuber seed. (Source :

Chilver 1997.)
Yield
With a few notab le exception s, th e
co nventional wisdom about th e eco nomic
adva ntages and di sadvan tages of TPS
relat ive to clonal tuber seed was confirmed.
Th e symmetri ca l pattern in Figure 2 of yield
differences between th e two technologies
suggests that productivity on aggregate was
pretty much the same. Average seed ling
tuber yield s w ere mark edly hi gher in the
Peru site, significantly lowe r in the hi ghlands of Indones ia, and not signifi can tl y
different from average yields of clon ally
propagated mate ri al in th e oth er seven
location s. Th e Indones ian resu lts were

driven by hi gher rates of yield deterioration

over tim e for seedlin g tubers compa red
with th e domin ant clo ne Gran ola that
demon strated a slower rate of producti vi ty
dec ay. The large yield advantage of th e
var iety Chacas ina in Peru was attribu ted to
loss of sta nd in clonally propagated materi al in a dry year. In 1997 -98, these yield
differences in Peru disappeared in a wet
year in a more comm erciali zed settin g for
production across 14 locations.
TPS transpl ants y ielded fairly we ll at
leve ls between 15 and 20 t/ha in both the
northeastern plain s and the northeaste rn

(t/ha)
15
10

-10
-15

Peru
Chaca s

In dia
NCP

Egypt
Delta

India
NEH

In dia
DP

India
NEP

India
WAZ

Egypt Indonesia Ind ia

India
NRA
NEP(TP) NEH(TP)

Figure 2. Difference in yields (t/ho) of TPS technologies from clonal tuber seed. (Source: Chilver 1997.)
CIPProgrnm Repmt 1997-98

2 15

hill s of Ind ia . Howeve r, fa rmers in the same

regi ons were produ c in g 25-30 t/ ha w ith
clon ally propagated mate ri al. Cos t savi ngs
would have to be hu ge or output prices
substantiall y hi gher to comp ensate for a
relati ve lo ss o f 10 tons of yie ld.
Th e results stro ngly con firm that seedl ing
tubers of good hybrid s are hi gh yieldin g at
levels ex ceedin g 20 t/ ha; but so are competin g c lones . Th e estima ted hi gh yield s of
competing c lones are the most surprisin g
findi ng in thi s stud y. Th e lowest ave ra ge
yield of cl onal materi al acro ss th e nin e
locations was a re spectabl e 12 t/ha.

Seed costs
As ex pe cted, th e sav in gs in seed costs
w ere a ma rke d adva ntage for th e TPS
seedlin g tuber technology. The ave rage
savin gs in seed co st ranged fr om about
US $1 O/ ha on the northc entral pl ains of
Indi a to about US $400/ ha in the northe astern hill s. Th e d ifference in th ese two
re gions large ly re flec ts differenc es in see d
prices. Tuber seed in w estern Uttar Prad es h
on th e north ce nt ra l pl ain s is read i ly avail ab le . But to rea ch th e north ea stern hills, th e
mater ial is tru cked ac ro ss th e pl ai ns up into
the mountain s of Assam. The remotene ss of
the no rtheas te rn hill s and transport diffi culties makes goo d qu ality tub er seed a dea r
comm odity.
Lowe r seed rates co ntri buted pro porti o nally mo re to the sav in gs in seed cos ts th an
did lower seed prices. U sin g TPS tec hn o logies, les s seed per ha was need ed in all th e
locati o ns . But lower per unit pri ces for TPS
seedlin g tubers we re nota bl e on ly in three
settin gs, th e north eas tern hill s and the
Decca n Platea u of India and th e hi ghland s
of Ind o nesi a. Th e co mm erci aliza tio n of
seedlin g tubers by large dea lers in Egypt led
to sub stan t iall y hi gher pri ces per unit of
materi al for propagation; so mu ch so th at
the ad vantage of a 1.8 t/ha lower seed rate
w ith seedlin g tu be rs did not tran slate into
an eco nomi c reality. Howeve r, in three of
the nine sites, the savings of usin g TPS
seed lin g tubers over c lon all y propagated

2 16 Potato

mate rial appro ac hed 50% . In general , it

ap pears that 50 % is a rea li st ic estim ate fo r
th e cost sa ving po tentia l of TP S seedlin g
tu be rs.

Output price
Pa1ti al budgetin g res ults al so supp ort th e
hypo th es is that repla cin g cl onal tuber seed
w ith seedlin g tu bers is acco mp anied by a
fall in output pri ce . Bu t output pric e
d iscrim inat ion aga in st seedl in g tubers was
not as wid es prea d as ex pected .
Tw o of the expecte d co nsi deration s
poi ntin g to lower prices for th e outpu t of
TPS seedl ings did not mate ria li ze. Exce pt
for Peru w here the TPS progeny ma tu red
ea rlier th an the c lonal check , fa rmer s
harvested the produ ction fro m seedl in g
tu be r and the con ve nti onal see d techn o logy
at the sa me t im e; th erefo re, pr ice differences associ ated w ith seasonal ity we re not
a fa ctor. O nly in Indon es ia did the later
maturity of TPS prog eni es re sult in so mew hat lower (4%) prices relati ve to the ea rly
maturin g va ri ety Gran o la. No r d id we find
tang ibl e differences i n pri ces betw een TPS
and clon al output in th e sam e grade
c lass ifi ca ti on. Thi s po siti ve findin g fo r TPS
ind icates that th e proge ni es are suffic ientl y
homogen eo us to compete in th e marketpl ace in thes e co u ntries w here fr es h tab le
cons umptio n is th e do min ant fo rm of
ut i lization. Ind eed, farm ers co mmented
th at the sh in ier offs prin g of seedlin g tu bers
we re eve n slightl y prefer red in some
location s.
The ca use of price di spari t ies betwee n
TPS and c lonal seed tec hnol og ies cente rs
on the potenti al fo r TPS to re sult in a
sma ller di stributi on of tub er size . Thi s
d isadva ntage was most sal ient in th e
D ecc an Pl ateau in Karn ataka w here the
output of seedlin g tub ers fetch ed about
U S$ 15 pe r ton less th an the production of
c lonall y propagate d m ate rial. The bulk (or
85 %) of th e produ cti on of th e dom i nant
c lone Ku fri Jo yti belon ged to th e grad e 2
category; w h i le 75% of the output of
seed lin g tube rs was des tined fo r the

small er-s ized grades 3 and 4. Farm ers in

Karnataka di spl ay a stron g preference for
fewe r, but large r- sized, more mark etable
tub ers. An inferior di stributi on of tub er size
was also a disadvantage of seedlin g tuber
tec hn o logy in Gujarat in the wes tern ar id
zo ne and in Uttar Prade sh in the
north ce ntral plains of Indi a. In th ese hi ghy ieldin g reg ion s, pric e differences of 8 to
12 % may not see m that large, but they are
equi va lent to differences in net benefits of
US $300 to U S$400 p er hecta re.
Cha nges in ot he1 costs did not undul y
affect the o u tcomes of the comparat ive
profitabi lity o f th e two ty pes of potato
prop aga tion tec hnolo gies. On ly Ind ones ia
rea li zed la1ge sav in gs in fungi c id e cos t.
Th ere, TPS proge ni es w ere more res istant to
late b li ght th an G ranola, the dom in ant
culti var, but th e longer duratio n of the TPS
proge ni es in cre ased th e t ime of app lica tio n; thu s, th e sco p e fo r cost sav ings was
so m ew hat less than antic ip ated.
Sensitivity analysis
A simpl e se nsitivity ana lysis was co ndu cted to illu stl"ate the mo st imp ort ant
va ri abl es in co nditionin g th e compa rative
pro fitability of TPS seed lin g tub er tec hn o logies . Three po ss ib le improvements we re
tes ted: a redu ction of 1 0% in the cost of
seedlin g tubers; a 10% hi gher y iel d; and a
10% in c rease in th e size di stributi on of
output from th e small to th e large-s ized
ca te gori es . Th e analys is showed that
propo rti ona l ch anges in y ield and in th e
tuber size di stribut ion a1e mu ch more
effect ive in in neas in g TP S benefits th an
proportiona l c hanges in seed costs .
Empirical Rules of Thumb
TPS utili za tion tec hnologies presentl y are
not suffic ientl y we ll deve loped to rep lace
clonal propa gatio n syste m s in large potato grow in g reg io ns of deve lopin g co untri es,
altho ugh so m e places at so m e times wo uld
app ear to be rip e for their depl oyme nt. Are
th ere any empiric al rules of thumb th at
would predi ct success so as to enco urage
mo re effective targeting of resea rch and
extensio n effort? In oth er wo rd s, can we

find a good predi ctor of TPS profitab ili ty

so lely with info rmation on th e eco nomi c
perform ance of c lo nal prop aga tion systems?
The res pon se to this question appea rs to b e
yes: th e cost of propagation materi al
di v id ed by the va lu e of produc t ion is a
reasonab ly goo d pred ictor of TPS performan ce w hen signifi ca nt y ield effec ts
betwee n th e two types of tec hn o log ies are
absent.
Thi s res ul t is show n in Fi gure 3 for the
seve n locat io ns w here the yie ld d i ffe rences
in Fi gure 1 are not substa ntia l. Th e sca tter
of the ave 1age val ues in Fi gure 3 suggests a
linear assoc iation between th e qu oti ent of
seed cos t to va lu e of production and
chan ge in net benefit s between TPS and
clon al propa ga tion systems. Th e breakeven po int for the eco nomic profitabi li ty of
TP S utili za ti o n tec hn o log ies is equi va lent to
seed cost co mpri si ng 19% of the va lu e of
c lonal produ ct ion. A val ue of 22 % equ ates
to an eco nom ic ga in of US$200 per
hectare.
Bas in g a dec ision rule on so few ob se rvation s is ri sky, but th e intuition behind th e
relation ship depicted in Fi gure 3 is appea lin g. Th e qu oti ent is com pri sed of two
ratio s: in ve rse of the multiplicatio n ra t io;
and th e seed: o utpu t price rati o. Clo nal
prop aga ti on sys tem s w ith hi gher
seed:output pr ice ratio s and lowe r multiplication rat ios w ill ge nerate hi gher va lu es for
seed cos ts to th e va lu e of produ ct ion.
Neith er of th ese rul es is as c los ely co rrelated w ith eco nom ic p erforman ce as is th eir
comb ination in th e suggeste d qu oti ent.
As a point of refe rence, in th e Red River
Va ll ey of North Dakota in th e United States
the va lu e fo r this quoti ent is about 8 to 10%
indi ca tin g a ve ry effec tive seed system.
Hi gher va lu es signal more ineffici encies in
the c lo nal seed system . Advanc es in TPS
ut ili za ti on tec hnology w ill shi ft th e relat ionship depi cted in Fi gure 3 to th e left; improvements in th e c lona l seed sys tem is
tant ;:irnount to observations slidin g clown
the I in e.

CIP Progrorn Reporl 1997-98

2 17

6. in net benefits in US $/ha

600~~~~~~~~~~~~~~~~~~~~~~~~~~~~---,

Y = -1258.5+65 .6x

500
400
300
200

100

i--........................................................................................................~..........---...........,.............................,

25

10

30

-100
-200
-300

Seed cost
Value of production

Jin %

-400

Figure 3. Association between comparative profitability of TPS utilization technologies and seed cost as percent of
value of productio n in clonal tub er seed.
Apply in g even th is very simp le empirical
rule is not w ithout difficulty, but the
problems are not insurmountabl e. Est imates are needed on farmers' yie ld, seed
rate, price of tuber seed, and price of output
at harvest. There is no substitute to in vesting in well focused, albeit limi ted, diagnost ic researc h in farmers ' fields and in loca l
markets to obtain these estim ates . The
temptation to use national or FAO data on
yield levels and import data o n seed costs
shou ld be resisted. Yield sampling in
farmers' fie ld s show that officially published
estimates often und er repo rt per hectare
potato producti v ity by 30% to 100%
(Pak istan-Sw iss Potato Development
Project, 1991; Terrazas et al., 1998).
Imported propagation material is usuall y
multiplied seve ral ti mes before it becomes
cost effective to produce potatoes dest in ed
for the fresh market.
The Ind onesian highlands is an apt
example where th e use of the wro ng data
can give spu r ious resu lts o n the prospects

218

Potato

for TPS-re lated techno logies. Combin in g

national yield leve ls at 15 t/ ha with an
impo rted seed price at US$2 ,000/ha gives
an estimate of 46% for o u r empirica l rul e of
thumb. Th e farm -l evel data show a " tru e"
estimate of 18% which is less than ou r 20 %
threshold value indicating that TPS is not
nearly as promising a ve ntu re as published
data wou Id suggest.

Concluding Comments
These evaluations bring out two positive
d eve lopm ents in TP S technologies. First,
for a given d istribu tion of tuber size, price
differences between TPS progenies and
clones were not statistically different.
In deed, at times the sh in ier output of
seed Ii ng tubers fetched more attractive
prices than the output of clonal seed .
Second, t he fact that these evaluatio ns
cou ld be carried out indi cates that the
supp ly of TPS is not a problem or should
not be cons idered to be a constrain t in
tech nology assessment. Sufficient material

is available intern ationally for the technology to take off.

The immediate problems related to TPS
utili za tion technolo gies are on th e demand
side. Cost savings are not sufficient to
leve rage adoption. TPS technologi es have
to produce as much or more than clonally
propagated tuber seed. Seedling tub ers
more or less meet this production condition , but significantly lower yie lds of
transplants from botanical seed are a
binding constraint to technology acceptance. Moreover, th e distribution of tuber
size has been highli ghted as a critical
variable for improving the design of both
transplants and seedlin g tuber technologies .
Th e results of th ese recent evaluations
again confirm that utilization of seedling
transplants and tubers will only be economi ca lly viable in reg ions such as the Red
Ri ve r Delta of Vietnam where clonal seed
availability and quality is severely limiting.
But eve n in these id enti f ied region s one has
to rem ember that th e potential for clonal
prop aga tion systems is also changing. For
exa mple, in our earlier exa mple of th e Red
Ri ve r Delta of Vietnam, liberalizin g the
imports of clonally propagated materi al
from China or invest ing in cold storage
could substantially erode the attractiveness
of TPS technologies. Egypt is another case
where the economic fundamentals have
changed in favor of the importation of tuber
seed (C hilver, El-B edewy, and Rizk , 1997).
Of the nine multi-locational sites in
Egypt, India, and Indonesia, TPS appe ars to
hav e th e brightest prospects in the northeas tern hills of Indi a. Thus far th e exper ience with TPS is heav il y co ncen tra ted in
Tripura. Monitoring the extension of TPS
technologies from Tripura to other states,
particularly Assam, in the northeastern hills
is a resea rch priority.

References Cited

Almekinders, C., A. Chilver, and H. Reni a.

1996. Current status of the TPS technol ogy in the wo rld. Potato Res. 39:289303.
Chilver, A. 1997. Innovation paths in
developing country agriculture: Tru e
potato seed in India, Egypt, and Indon esia. Ph.D. thesis, University of East
An glia, UK. 229 p .
Chil ver, A. , R. El-Bed ewy, and A. Rizk.
1997. True potato seed: Research,
diffu sion, and outcomes in Egypt. CIP,
Lima, Peru. 29 p.
Crissm an, C. 1987. Id entifying strengths
and weaknesses of seed programs in
deve lopi ng countries. In: Report of the
Third Social Scienc e Planning Congress
held 7-10 September at CIP, Lima, Peru.
Paki stan-Swiss Potato D evelopment Proj ec t.
1991. Potato yield estimation survey of
autumn crop of Punj ab - January, 1991:
A study carried out jointly by Directorate
of Agriculture, Crop Reporting Services,
Punjab, Ministry of Food, Agriculture,
and Cooperatives, Government of
Pakistan, and the Paki stan-Swiss Potato
D eve lopment Project. PARC, Islamabad,
Paki stan.
Rasco , E.T. 1994. SAPPRAD on the third
year of Phase Ill. Coordinator's Report
July 1993-June 1994. SAPPRAD, Manila,
Philippines.
Terra zas, F. , V. Suarez, G. Gardner, G.
Thi ele, A. Deva ux, an d T. Walker. 1998.
Di ag nosing potato productivity in
farmers' fields in Bolivia. Social Sci ence
Department Working Paper No. 1998-5.
International Potato Center (CIP), Lima,
Peru.
Simm o nds, N. 1997. A rev iew of potato
propa ga tion by mea ns of seed, as distinct
from clonal propagation by tubers.
Potato Res. 40: 1 91-21 4.

CIP Program RepOlt 199798

219

Globalization Takes Root: Potato Trade

in Latin America
G.J.

Scott and L. Maldonado '

Much of the recent d ebate about the

implications of increased trade liberalization and globa li zat ion for d evelo pingcou ntry agric u ltur e has cen tered on cerea ls,
livestock and dairy products , and fresh
fruits and vegeta bl es. D espite its current
ranking amo ng the ten most imp orta nt food
crops in deve lop in g co un tries, potato is
rarely c ited for its trade potenti al-as
impo rts or expo rts. In fact, curre nt potato
trade is much more important than FAO
statistics indi cate. Such data on ly report
trad e in fres h potatoes for hu man consump t ion and seed , leaving out from pub li shed
figures trade in proc essed produ cts such as
frozen french fr ies, potato chips, and starc h.
After a b ri ef review of g loba l deve lo pments, th is pape r analyzes rece nt trends in
the trad e of potatoes and potato products in
Latin Amer ica. Re sults presented and
issues id en t ified for th is regio n are in tended
to promote a grea ter apprec iation for and
interest in similar trend s elsew here. The
overa ll intent of the paper is to d em o nstrate
that: (1) for a variety of reasons tra de
already ha s-a nd is li ke ly to becom e-an
eve n more impo rt ant in fluence o n th e
evo lu tion of potato produ ct io n in some
Lat in America n countries than has heretofo re been the case; (2) pol icymakers and
potato re sea rchers in those co untri es should
take approp riate action now to remain
compe titi ve in the yea rs ahead; and (3)
m easures requ ired to improve co mpetiti veness wi ll , of necessity, in vo lve seek in g new
partn ers b eyo nd tho se trad itiona 1ly engage d
in activi ti es aim ed at imp roving potato

produ ctio n fo r pure ly domesti c sa le and use

in fresh fo rm .

Materials and Methods

Results prese nted in th is paper synthesize
information co ll ected from a wide var iety of
sources . These include publish ed and
unpu b li shed FAO trade data; USDA trade
stat ist ics; rapid ap praisa ls carried out in
col labo ration w ith coope rators in a number
of PRECODEPA2 cou ntr ies; th es is resea rc h
in Peru; informa l interv iews with industry
and National Agricu ltural Re sea rch Institute
representatives in a number of Latin
Am eri ca n co untri es; and figures and
anal ys is glea ned from the "g ray literature",
e.g ., newspa p er articl es, unpubli shed
reports, and th eses co ll ected durin g vis its to
a series of Latin American countries ove r
the last severa l yea rs; and , propr ietary
information for reports comm issioned by
the pri va te sec tor.
Th e method com bines an histo ri ca l
anal ysis of tim e-series data on imports and
exports with brief country o r sub-regiona l
case studies. Th e intent is to provid e a
clearer se nse of the co nsid erab le differences that ex ist over time and by lo cat ion
as to the nature and likel y trajectory of this
comm erce for Latin America in the deca des
ahead.

Results and Discussion

Global trends
World exports of ware potatoes and seed
averaged 7.5 mil li on tons durin g the period

1 CIP, Lim a, Peru

2 The Programe Regional Cooperative de Papa (PR ECO DEPA) is a regional network for potato research for Mexico, Ce ntral
Amer ica, and th e Caribbea n and is comprised of scie ntisls in the respeclive nati ona l potato programs with backstopping
fromC IP.

CIP Pmgmm Report 1997-98

22 1

1995-97, or about 2.4% of global production for these same years (FAOSTAT, Jun e
1998). Recent estimates of potato exports,
in cludin g processed produ cts, place th is
tota l at about 4% of globa l produ ction , or
roughly double the percentage for fres h
potato and seed alon e (FAO, 1995). To put
these figures in perspect ive, globa l rice
ex po rts represent an estimated 3% of the
world's annua l rice produ ction (Th e
Eco nomist, 1993).
Expo rts of table and seed potatoes from
deve loping cou ntri es averaged 1.2 million
tons in 1995-97 or about one percent of
produ ction (FAOSTAT, June 1998). Imports
were at the same leve l. The volume of this
trade has tripled over the last three and a
half decades. During 1995-9 7, major
exporters were Egypt (32 1,000 t), Turkey
(191 ,000 t), Indones ia (91,000 t), and
Morocco (79 ,000 t) . For th e M editerranean
co untri es, these exports consist of ea rl y or
winter table potatoes shipped to European
Union markets under spec ial trade arrangements. Many of these same co untri es also
import seed potatoes from European seed
producers because co mparab le growing
cond iti ons and similar consumer tastes an d
preferences are highl y co mpl ementary.
Egypt alone earns we ll over US$25 million

from its annu al potato exports, and po li cy

refo rm s have spurred greater produ ction for
the loca l market and for export in rece nt
yea rs. As potato ac reage stead il y declin ed
in W este rn Europe over the last three
decades, Holl and now ships four tim es th e
vo lum e and rou ghly one quarter of its seed
exports to Afri ca n countri es-most ly in
North Afri ca-up from some 35,000 t and a
10% share in th e ea rl y 19 70s (Scott, 1994).
Latin America
Latin America is th e onl y deve lopingcountry reg ion w ith a trade deficit in
potatoes- imports of table potatoes and
seed exceeded ex ports by ove r 200,000 t
durin g 199 4-96 (FAOSTAT, March 1998) as
total imports reached 363,000 t, or some
2.5% of reg iona l product ion (Tab le 1 ).
Thou gh the statistics are sketchy and the
fi gures not always ava il able for th e same
time periods, rece nt information on imports
of frozen french fri es (Tab le 2), when added
to the tota l for tab le potatoes and seed,
suggests that this pe rcentage total is at least
double and growin g at a remark abl y fast
rate . Th e bulk-but by no means all-of
these imports ca me from industri ali zed
countr ies: Ca nada, the Netherland s, and th e
U.S. Th e hi ghly differentiated nature of
region al import and export markets mea n

Table 1. Import and export, in tons, of table and seed potatoes in Latin America, 1994-96.

Brazil
Venezuela
Cuba
Mexico
Trinidad and Tobago
Uruguay
Nicaragua
Others
Total'
Source: FAOSTAT, March, 1998.
Totals do not sum exactly due to rounding.

222

Potato

Imparts

Exports

(000 t)

(000 t}

95
72
32
31

27
19
13

73
363

Argentina
Colombia
Guatemala
Honduras
Mexico
Ecuador
Chile
Others
Total

79
36
25

3
145

Table 2. Imports of frozen french fries, in tons (t), into Latin America and the Caribbean from the USA, Canada,
and the Netherlands, 1991-92 and 1994-95.

1991-92

Southern Cone
Brazil
Chile
Uruguoy
Argentina
Andean Region
Ecuador
Colombia
Peru
Venezuela
Central America and
the Caribbean
Guatemala
Dutch Antilles
El Salvador
Jamaica
Honduras
British Virgin Isles
Bahamas
Dominican Rep.
Costa Rica
Othersb
Mexico
Total

1994-95

USA Netherlands Canada

Total

USA

Netherlands

Canada

503
218
134

2,833
1,780
252
159
653
2,863
20
24
83
2,737

18,722
12,159
3,726
930
1,908
3,587
947
191
220
2,229

10,651
5,986
612
3,483
572
42

21 ,799
13,874
2,192
1,016
4,717
8,898
16
495
1,608
6,778

51,173
32,019
6,529
5,429
7,196
12,527
963

12,512
447
1,676
227
480
316
36
2,752
189
314
6,192
15,158
33,366

14,599
3,942
1,543
1,492
1,399
1,210
302
801
794
609
2,506
33,316
70,224

1,842

15,518
3,140
2,715

31,959
7,082
4,258
1,492
3, 11 4
1,559
302
1,378
849
609
11,315
34,470
130,129

997
987
10
10

150
1,543
20
7

16
83
1,221

1,516
6,109
428
359
227
480
316
36
2,317
73
314
1,560
14,618
22,773

1,333
575
108
149
502
1,320

1,493

4,910
19
1,317

435
116
1,493
2,490

3,139
540
8,103

42

l,715
349

577
55
1,787
12,535

7,022
1,155
47,370

Total

729

1,828
9,007

Source: PacificVision (1996).

0
Based on fresh weight, using one pound of frozen potato as equivalent to two pounds of freshpotato, and converted to metric tons.
b In the case of the Netherlands, exports are mainly to Cura100, Aruba, Martinique, Guadeloupe, Haiti, Anguilla, and the Bahamas.
For Canada to Trinidad and Tobago, Jamaica, Bermuda, Barbados, and Cuba.

th at such agg regate trend s require grea ter

sc rutiny at th e sub-reg ion al and nati onal
leve l.
Multipl e factors contribute to thi s
ex pansio n in imports of potatoes and potato
produ cts. O n the dem and side for processed products, large urban popul ati ons
and ri sin g per capita in comes have generated changes in diets- ove r 70% of all
Latin Ameri ca n consumers now res ide in

urban areas and many household s ca n

afford to purchase mo re processed foods.
Growing female parti c ipation in the form al
w orkforce and the shift to shorter lun ch
hours have stimul ated demand for eas ier-toprepare foods. Th e exp los ion in touri sm
has brought new influ ences on loca l food
dem and both directly by non-residents and
indirectly by th e effect th eir eating habits
has on loca l co nsum ers. Rapid expansion
of th e fas t food indu stry- both domesti c

CIPPrngrnm Report 1997-98

223

and foreign chains, with their associated

aggressive advertising ca mpaigns (Table 3)
also merits mention as a factor influencing
the demand for processed potato products.
Similar factors have in c reased demand for
fresh potatoes, albeit at a slowe r rate, in
virtually al I countries except Bolivia and
Peru w here per capita consumption is
alread y quite high.
On the supply side, a combination of
factors has made imported processed potato
products more competitive. Imports are
more affordable for the following reasons .
Prices are falling due to tariff reductions.
Economies of scale in large processing
plants for french fri es and starch in
industrialized, exporter countries
influence cost of production.
There is an abundance, if not oversupply,
of raw material in industrialized countries (New York Times, 1997) and
selective export promotion schemes.

Containerized shipping (The Economist,

1997) lowers long distance tran sport
costs. At the same time, locall y produced
processed products have lost market
share from high unit costs for domestically produced potatoes and poorly
articulated I in kages betwee n producers,
processors, and consumers. Seizing on
these trends in neighborin g countries as
commercial opportunities , a number of
Latin American countries have also
entered the potato trade as exporters for
many of the same reasons.
Imports vs. Importers. Only 6 of the 37
countries in Latin America that imported
table and seed potatoes during 1994-96
accounted for 70% of all the imports: Bra z il
(95,000 t), Venezuela (72 ,000 t), Cuba
(32,000 t), Mexico (31 ,000 t), Trinidad and
Tobago (29, 000 t), and Uruguay (19,000 t).
Most (27 of 37) of the oth er countries
import less than 5,000 t annually. They are
found in the Caribbean; the Guyanas, e.g. ,

Table 3. Distribution of McDonald 's restaurants by region and country, 1987 vs. 1995.
Region/Country

1987

1995

Tota l Pacific
Total Europe/Africa
Total Latin America
Argentina
Brazil
Colombia
Costa Rico
Chile
Guatemala
Mexico
Panama
Puerto Rico
Uruguay
Venezuela
Others
U.S.A
Canada
Total

951
755
99
3
37
0
4
0
3
5
8

2,735
2,710
665(l,160)
75
243 (505)

22

14
21
15
132
15
75

0
3
14
7,567
539

25
286
11,368 (12,406)
902 (l,062)

9,911

18,380 (23,726)

Source: McDonalds Corporation {1992, 1995) (Figures in parenthesis are for 1998 and are taken from the McDonald's web site)

224

Potato

Surin ame; o r in Central A meri ca, e.g.,

Beli ze, El Sa lvador, Ni ca ragua.

(bi g prod uce rs and sm all produ ce rs),

exporters, and neg li gibl e trad ers.

Impo rts of processed potato produ cts are

simil arl y d iffu se in term s of numbers o f
countri es and are equ ally co nce ntrated in
term s o f vo lumes imported, with so me
ove rl ap betwee n th e two . For exa mpl e,
Braz il , M ex ico, and Ve nez uela im po rt
relati ve ly hi gh qu antiti es o f both fres h and
processed potatoes. Th e run -up in th e
vo lum e of imports of processe d potato
produ cts is ce rtainl y notewo rth y, as is th e
evo luti on of th e combin ed to tals. In rece nt
yea rs, Braz il 's tabl e, seed, and processed
potato imports exceeded 214,000 tin
1994- nea rly 8% of annu al output, alth ough imports of fres h potatoes in Braz il
and elsew here have fall en o ff sin ce th en
(FAOSTAT, M arch 1998) .

Importers: Big producers. Th e fo remost

exa mples of thi s ca tego ry are M ex ico and
Braz i I. Eac h has a good-sized potato sector,
alth ough thi s co mm odity is not th e prin cipal crop. Eac h of th e sectors is di v id ed
between a relati ve ly few, large-sca le,
tec hni ca ll y adva nced potato fa rm ers and a
mu ch larger number of small -sca le, peasa nt
potato house ho lds. Eac h co untry has a
rel ati ve ly large pop ul ati on, rel ati ve ly hi gh
per cap ita in co mes, and has w itn essed a
rapid expansion in the fast food trade in
rece nt yea rs (Tabl e 3). Eac h has a so phi sti cated major potato producer as a neighbor
that is res ponsibl e for nea rly al I its imports.
Each has join ed a reg ional tradin g bl oc (i.e.,
NA FTA, MERCOS UR) and has see n impo rts
of fresh and processed potatoes ri se as
tari ffs have co me dow n as a co nsequ ence
of th eir tradin g bl oc parti c ipati on.

W ith co ntinu ed growth in total pop ul ati on, urban area popul ati ons, touri sm, and
hopefull y per ca pita in co mes in the deca des ahead, the qu estio ns th at emerge are:
Wh at is the most Ii kely scenari o fo r the
future evo luti o n of th e po tato trade in Latin
Am eri ca? And perhaps of more importance,
w hat ca n countries w ith limited resources
ava il ab le fo r research and deve lopment of
th e po tato sector do about it ? Given th e
w id e range of parti c ipants, a typo logy of
co untri es m ay help cl ari fy th e releva nt
opti ons ava il abl e to th e d ifferent types .

The M ex ica n case is parti cul arl y notew orthy beca use current tariffs on fresh
potatoes are still above 200% but are
sch edul ed to fa 11 to ze ro by 2004 . Al th ough tariffs on processed potatoes are
onl y a fracti on o f thi s leve l, qu otas di sco urage unlimited imports. Imports in M ex ico,
like more rece ntl y in Braz il , have also
tempo raril y stag nated fo r ph ytosa nitary
reaso ns and deva lu ati o ns durin g th e 199 0s.
N oneth eless, potato imports have co ntin ued to ri se throughout th e deca de (Tabl e 4).
Rece nt studi es suggest th at, given current
differences in th e costs of produ cti on and

Typology of Countries. For th e purposes

of thi s analys is, potato traders ca n be
grouped into variou s ca tego ri es: impo rters

Table 4. Mexico : Potato imports from the U.S.A . by type of product, in 000 t, and% distribution of product,

1990-92 and 1994- 96.

Fresh
Frozen
Chips
Other
Total

1990-92

1994-96

15
11
9
5
40

36
29
23
12
100

22
34
16
15

25
39
18
18

87

100

Source: USDA (1991-97)

CIP Program Report 199798

225

the growth in demand for potatoes, potato

imports could rise considerably in the years
ahead. At the same time, however, cons iderable improvements in productivity co uld
be made to raise competitiveness in the
fresh market, take advantage of local
preferences for varieties not grown north of
the border, and fully exp loit year-round
harvesting. Moreover, these initiatives
co uld be linked to efforts to strengthen
particular compone nts of the processed
potato industry (e.g., medium-scale chipping plants) so as to exp lo it emergi ng
segments in the dom est ic regional market
for processed potatoes, thereby sat isfyi ng
grow in g demand with both imports and
loca ll y produced produ cts.

Importers: Small-producers. Th ese

countries are largely those in the Caribbean , the Guyanas, Central Amer ica except
Guatemala, and Ve nezue la. Potatoes are
typically not a major part of the local diet
and , with the excepti on of Cuba and
Venezue la, are produced by relatively smal l
numbers of lo ca l growers for sale on the
fresh market as an expensive, luxury
vegetable. Imports of processed potato
products are on the rise in many instances
simp ly because the lo ca l potato sector is
typically too small and producer prices for
fresh potatoes too hi gh to economica ll y
ju stify processing on even a cottageindu stry basis. Demand has mushroomed
w ith risi ng incomes, the advent of more
tourists, and fast food restaurants cater in g to
loca l consume rs and visitors. Lowe r tariffs
and ship ping costs ha ve made imports
c he aper in many instances.
Some countries-most notably Costa
Rica, Panama, and Venez uela-have seen
imports of table potatoes sp ike at particular
times w hen a shortage of local supplies
drive up prices on the domestic market.
Th e response, in some instances, has been
to reinstate hi gh tariffs on a temporary basis
to discourage flooding the local market
with imports, but such ad hoc measures
have been controversial because they
provoke threats of retali ati on by trading
partners.

226

Pololo

Seed represents a major share of potato

imports in on ly a few instances, e.g., Cuba ,
Nica ragua, Ve nezue la.
In too many cases, local potato programs
in these countries have responded to the
rise in potato imports with a cal I for an
indi scr imin ate expans ion of the lo cal
processing indu stry for french fries, chips,
starch, and even flakes . Or, they have
issued appeals to governmen t auth orit ies to
re-i nstate ta ri ff barriers. An alternative
approach wou ld be to channel th e l im ited,
lo ca l resources avai lab le on improv in g
performance of those producers engaged in
tabl e potato production. Their comparative
advantage v is-a -vis imports is that some
consumers will always prefer high quality,
fresh ly produced loca l tubers and in many
of these locations agro-climat ic condit ions
allow continuous production throughout
the year. Th is work m ight the n be comp lemented with efforts foc used on one segmen t of the processing indu stry such as
potato chips where local spices, brandname affi I iation, and batch-style production
techniques might be used as leverage to
capture a signifi cant, albeit not dominant,
segment of the lo ca l market.

Exporters. This grou p of countries

consists of Argent in a, Colombia, and
Guatemala. They export to Brazil, Venezuela and Ecuador, and El Salvador and
Nica ragua, respectively.
A rgentina's exports of table potatoes to
Brazil topped 160,000 tin 1994 with the
opening of MERCOSUR and the red uct ion
in tariffs. Many of these potatoes are
shipped from northe rn Argentina into
southe rn Brazi I when th e ear ly potato
harvest on one side of the border co inc id es
w ith hi gh seaso nal prices and scarcity in
major markets suc h as Sao Paulo. With the
opening of a large multi-natio nal french fry
processing plant in Balcarce in early 1995
and a second lin e to the same facility
recently, Arge ntin a's exports have also
included frozen french fries. Large, irrigated, highly technically advanced

produ cers in Argentina w ill continue to put

pressu re on sma ll , peasa nt produ ce rs in
so uth ern states lik e Para na w ho are dependent on rain-fed agricu lture for th eir
li ve lih oods. Braz il 's recent 30% currency
devalu at ion wi ll give its growers brea thin g
room , thou gh perhaps o nly temp orarily.
Colombia 's expo rt s of ta bl e potatoes and
co m mo n seed to Ecuador and Ve nez uela
are gro und ed in its:
hi gher yi eld s and lower unit costs of
produ ction , althou gh thi s is exchange
rat e-se nsiti ve;
we ll-organ ized federation of growers,
trad ers, and shippers; an d,
well-developed network o f mark etin g
infrastru cture: roads, wholesale m ark ets,
and teleco mmuni cat io ns (Scott et al.,
1997).
Argentina and Co lo mbi a, in partic ul ar,
have emb ark ed o n agg ressive campaigns to
ca pture a grow in g share of th e sub-reg io nal
ex port mark et. Th ese effo rts have in c luded:
closer co ll aboration in research between
the publi c and private sector (Rev ista
Papa, 199 2);
policy measures to encou rage expa nsion
of in fras tru cture by processo rs, e.g., tax
and in vestm ent in ce nti ves; an d, per haps
most fund am entall y,
an approach that sees the search to
develop or acquire im proved potato
va ri et ies for processing as but a part of
an integ rated initi at ive intend ed to
strength en com petiti ve ness in al I aspects
of the food system for potato: varieties;
seed; raw m ateri al production; pl ant
equipment and facilities; m ark et infrastructure (e.g. , ports, telecommuni cations) ; and trade and in ves tm ent policy.

Negligible traders. Thi s category

includ es Bo li v ia, Chile, Peru, and Paragu ay.
These co untri es traditional ly have not
imported (because th ey are self-suffi c ient)
or exported (because th ey produ ce so few
and what th ey ship abroad are simp ly transshipments from oth er co untries ) potatoes in
any appreciable quantiti es. Th ey have also

explo1e d ex port po ss ibili ties (Fano, et al.,

1998; Muchnik and Tejo, 1997).
In rece nt yea rs, howeve r, these co untries
have w itn essed a ris e in potato imports of
both fresh and processed potatoes. Though
th e quantities have remain ed small , th ey
have attracted co nsid erab le attentio n in the
press give n thei r strateg ic pre se nce in the
fastes t growi ng seg men t of the potato
market as in Peru or due to Bolivia's recent
ascens ion to associate member statu s of
MER COSU R. Howeve r, in both th ese
cases, give n th e hi gh leve ls of per ca pita
co nsumpt io n and the large nu mbe rs of lo w
income urban consumers, th e threa t of
much grea ter imports of processed o r fresh
potatoes that are mu c h chea per than loca ll y
produ ced material ha s to be taken se riously
in th e med ium term . In th at rega rd , a
number of meas ures merit consideration.
In the production sp here, these co untries
can improve yie ld s of local ly preferred
va ri et ies and more effect ively ex ploi t yea r
round harves ting of fresh potatoes. In the
po stharves t area, th ey ca n take full er
advantage of rec ent inn ova tions in processin g such as chill ed, in li eu of frozen, potato
produ cts. As poli cy ini tiat ives, they can
promote up-grading of processing in fra stru cture by both domes ti c and foreig nba sed firms through tax and investm ent
incenti ves. These various initi atives wou ld
prov id e st rateg ic co mpl eme nts to lo ngstandin g wo rk to deve lop improved va ri et ies as we ll as more effic ient and su sta inable seed sc hem es .

Conclusions
Gl oba li za ti o n has tak en root in th e Latin
Am eri ca n potato trade. A ll indi ca tors are
th at such trade is hi ghl y sustain abl e and is
lik ely to ex pand rap id ly in the deca des
ahead. Some Latin American co un tr ies
have w itnessed faster exp ansion in potato
imp orts o r ex ports than others have. From
thi s co ll ec tive experi ence the following
acti o n reco mmendations are suggested for
Latin American gove rnm ents.

CI PProgram Report 1997 -98

227

 Become mo re in formed abou t rece nt

trends and trade com mitm ents in terms
of tariff reductions for di fferent types of
potato products so as to bette r ground
loca l strategy.
Target researc h and developm ent efforts
to focus o n a particu lar segment of th e
domestic (import substitution ) or export
market for processed potatoes, rath er
than propose across-the-board initi atives
in volving seve ral segments si mu ltaneousl y o r b lanket pro posa ls (e.g.,
reinstat e tariffs).
Seek out more in formation abo ut how
neighborin g o r other Latin A merican
countr ies have addressed trad e issuesinc lud in g deve lopme nts in indust ri alized
co untri es w ith respect to such trend s as
the move to wa rd c hi I led potatoes,
the success and failure rate of suc h
ini tiat ives.
Co nsider trade in particular produ ct lines
from the pers pective of th e who le
system. This in cludes not ju st efforts
aimed at rais in g prod ucti v ity but also
lower in g processin g and marke ting costs,
and improvi ng qu ality. In evitab ly, such
an approach involves deve lopi ng
link ages that extend beyond the searc h
for better va rieti es pe r se . It en tai ls
participation by al l those engaged in
improv in g system performance includin g
pri va te compa ni es (e .g., inpu t suppli ers,
potato w holesalers, and pro cess ors),
gro we rs' assoc iati ons, and public
policymakers outside agriculture, e.g. ,
ministries of trade and indu stry.
References Cited
FAOSTAT, 1998. Statist ics Database
(O nlin e) . Jun e and Ma rch. Ava ilabl e
HTTP: http:// apps.fao.org.
FAO, 1995. Potatoes in the 1990s. Situation and pros pects of the wo rld potato

228

Potato

eco nomy. International Potato Center

and FAO. Rome, Italy.
Fano, H. , G. Ca rm ona, M. Ordinola , and G.
Scott. 1998. Expe ri encias de ex porta c i6n
de la papa amari l la pe ruana. Socia l
Sc ience Wo rkin g Paper No . 1998-3. Copublished by CIP and ADEX-AID. CIP,
Lima, Peru .
Th e Econ om ist. 19 93. Rice prod uct io n:
Sta rs and stripe s Sus hi. 27 November
1993. p. 36 .
Th e Economis t. 19 97. Deli ve ring the
goo ds. 15 November 1997. p. 85 -86.
McDonald s' Co rp orat io n. 1992 and 19 95.
A nnu al report to share holders.
Oakb rook, IL, USA.
Muchnik, E. an d P. Tejo J. 1997. La pap a en
el co mercio re gio nal yen los acue rd os
co merc iales . Docum ent prepared fo r
Com isi6n Eco n6m ica para Ame rica
Lat ina y el Carib e (C EPAL). CEPA L,
Sa ntiago, Chi le.
New Yo rk Times. 1997. Anxious days in
Potatol an d: Competitive forces threaten
to knock Id aho from top. 1 2 M arch
1997.
Pac ific Vision. 1996. Report prepared fo r
the U.S. Nation al Potato Board. Trade
Stats Northwest/Pac ific Vision, Portland ,
OR, USA.
Rev ista Pap a. 1992. Ava nza la
in dustr alizacio n de la papa en Colombia .
Special issue (June) 1992.
Scott, G.J. 1994 . Th e emerg ing world
market for pota toes and potato produ cts
w ith partic ul ar reference to developing
co untr ies. Eco nom ie et Gesti o n Agro A l imen taire 30:19 -27.
Scott, G. , R. Basay, and L. Ma ld o nado.
1997. El comercio exte ri or de papa en
America Lat ina . Come rci o Exte rior
(Decemb er) p. 988-996.
USDA (U .S. Dept. of Ag ri cu lture). 199 1-97.
Potato facts. Econo mi c Research Serv ice,
USDA, Was hington , D.C.

The Economics of Generating International Public

Goods from Investing in Potato Plant Breeding
T. Walker1 and K. Fuglie 2

Th e words "scarc ity" and "fa tigue" figure

promin ently in the rec ent liter atu re on
gove rnm ent-fin anced agr icu lture res ea rch
(A lston, Norton, and Pard ey, 1996; And erso n, 1 998). After two deca des of growth ,
th e rea lity of tig htenin g resources fo r publ ic
sector agr icu ltural research became
apparent in th e rnid-1980 s for both developed and develo pin g cou ntri es as w ell as
for the CGIAR (A lston , Pard ey, and Smi th,
1998).
The eco nomi cs of invest in g in crop
impro ve ment programs a Iso have been
known fo r some tim e (Bin swa nger, 1974).
Re leva nt cons iderations include the
potenti al for ga in s in we ll -defined environ ments, eco nom ies of sca le, size, and scope
in research, dimin ishin g returns to search
from th e same distribution , and th e scope
for tech no logy borrow in g, i. e., spi ll -in s. In
sp ite of a grow in g consensus abo ut sca rc ity
and of the ex iste nce of estab li shed econo mi c knowl edge, app lications of econom ics to commodity crop improvement have
been slow in corn in g.
In th e 1990s, eco nomists at the Intern ational Center fo r the Improveme nt of Maize
and Wheat (C IMMYT) mad e seve ral
p ioneer ing appl ica tion s to bring economi cs
to bear on decis ion-m ak in g in wheat
breeding in deve lop in g countries (Brennan ,
1992; Maredia and Byerl ee, 1998) . Tho se
app li cat ions w ere motiv ated by two
emp iri ca l facts that chal lenge d th e conven ti ona l w isdom o n th e locat io n specifici ty of
var ietal tran sfe r and o n the reported
und erinv estment in p lant breeding resea rch.

1 CIP, Lima, Peru.

2 CIP, Bogo r, Ind onesia .

First, analys is of the C IMMYT/ NARS

(nati onal agricultura l rese arch sys tems)
intern<Jtional spring w he<Jt data showed that
w heat var ieties are highly tran sferable
wi thin some mega-e nv ironments (Ma redia ,
Ward, and Byer lee, 1 998). Th is findin g
rin gs es pecially tru e in th e important high
rain fa ll and irri gated mega-e nviro nments
where CIMMYT-bred materials y ield as
mu ch as nationa ll y bred mater ials. Second,
man y small w heat- prod ucing count ries
have invested in plant breeding progr<Jms
char<Jcteri ze d by hi gh resea rch intensiti es in
term s of investm ent per unit of production
(or val ue) of the crop (Bo hn, Byerlee, and
Maredia, 1998). Th e res ul ts of eco nom ic
mode ling showed th at 21 of 71 programs
were in vesting too much in wheat breeding
given the small size of their natio nal
produ ct ion and ampl e potential for technology bo rrow in g from abroa d.
To w hat exte nt do the CIMMYT resu lts
for wheat breed ing app ly to potato breed ing
in deve lopi ng countr ies? W hat is a " bes tbet" spat ial allocation of prioriti es by CIP to
gen erate public good s in deve lop in g
countr ies from potato breed ing? We begi n
to answer the se qu es tions by comparing
re se<Jrch intensiti es in potato Jn d whe<Jt
improvement programs.
Research Intensiti es in Potato and
Wheat Crop Improvement

In 1992, CIMMYT surveyed 71 w heat

improvement programs in 3 7 deve lop in g
coun tri es. Th e res ul ts showed that ave rage
rese<Jrch intensity was inversely and
strong ly co rrelated with the size of NARS .
Th e smallest systems we re in vesting the
equi va lent of mo re than 100 full -tim e

CIP Progrorn Report 1997-98

229

equivale nt scientists per million metric tons

of production. For the largest six NARS,
eac h producing more than five million to ns,
the ratio fel l to four full-time equiva lent
scientists per million tons.
About 1,200 full-time equivalent sc ientists with a B.S. or hi gher were working in
the wheat breeding programs in countr ies
which accounted for 97% of developingcountry wheat production . Comparab le
esti mates from similar su rveys undertaken
by CIP in 1993 ind ic ated a total of abo ut
350 sc ientists work in g in 21 countries
accounting for about 70% of deve lopin gcountry potato production .
We are assembling more comp lete data
for potato so that a ri gorous comparative
analys is can be ca rri ed out for th e two
crops . But we can make four preliminary
observat ion s based on the 1993 survey and
ou r knowledge of potato breeding in
developing country agriculture. First, the
weighted average res earch intensities
across all developing countries between
impro vement programs for the two crops do
not appear to be significantly different.
Similar investments, in relative terms, in the
largest NARS in China and India probably
explain this finding. Second, the improvement programs are qualitativel y different.
Wheat research features a hea vy concentration in breeding; potato researc h is more
diverse wi th an emphasis o n pathology,
tissu e culture, viro logy, and entomo logy.
Third , more w heat-produ cing developing
countries have invested in full adaptive
breeding programs that not only test
material from other programs but also make
their own crosses . Sixty-one of the 71
programs had invested in breeding; onl y ten
programs had re stricted their act iv iti es to
testing elite imported material. In potatoes,
the majority of countries with a comm itment to crop improvement have invested in
testi ng programs that stop short of making
their own crosses from parental material.
And fourt h, ve ry high research in te nsities in
sma ll NARS seem to be more of a source of
inefficiency in the gen etic improvement of

230

Pototo

w heat than in potatoes . Urban consumer

subs idi es probably exp lain the i nterest of
small NARS, particularly those in SubSaharan Africa, in investing in agricultural
research to increase their domestic wheat
production. Potatoes have not benefited
from such urban biases nor are they as
traded internationally as is wheat.
Threshold Levels of Production

The economic mode lin g of whea t breeding

by C IMMYT focuses on the size of co untry
production that warrants a given level and
type of in vestment. Based on costs , three
mutual ly exc lu sive o utcome s are possible:
(1) production is insufficient to justify an
investment in testing, (2) producti on is
suffic ient to make testing profitable but
insufficient to es tablish the economic
superior ity of breeding, and (3) production
is suffic ient to justify an in vestment in
breeding the most expensive alternative .
Note, howeve r, th e profitability of a breeding program cannot be considered without
quantifying the opportunity cost of in vesting
in a testing program. The structure of the
model is given as equat ions (1 ), (2), and (3)
in Figure 1. An equat ion identical to (3) can
also be written for the breeding alternative.
Applying the C IMMYT model to potatoes
requires one ma jor modification on the
assumptions about the parameters (see
Table 1). Patterns of varietal cha nge are
very different in potatoes and wheat
(Wa lker, 1994). Maredia's and Byerlee's
adoption scenar io in their base model
features a rapid rate of varietal turnover.
Ann uall y, a new variety is released that
reaches peak adoptio n in five years. After
20 years, the mean age of va rieti es in
farmers' fi eld s approaches five yea rs. An
average var ietal age of five years ma y be a
reasonable target for w he at, but it is not for
potatoes (eve n if one wa nts to err on the
sid e of optimism). Th e incidence of
successfu I released var ieties is also substan tially lower in potatoes than in wheat,
although the odds for success (contrary to
intuiti on) favor deve loping countries over

(3) NVPT

=I.

i=l

t=l

(1 + r )

Where

.r

Q*B

= Minimum size of production for which testing is profitable;

= Minimum size of production for which breeding is more

profitable than testing;

NVPr

= Net present value;

Pi
Q

= Price of output in year t;

Ys

= Number of full-time equivalent scientist for program

Ci

= Cost per scientist in year

r
v; 1

g;

= Size of production;

s;

t;

Discount rate;
Proportion of area sown to variety i in year t;
= productivity gain attributed to variety i.

Figure 1. Equations (1) , (2), and (3).

developed countr ies that have mature
production and processing sectors. For
these reasons, a mean varietal age of 15
yea rs for potatoes is used to anchor th e
analysis in a rea li stic setting for potatoes.
Th e est im ated threshold leve ls of
producti on and in cidence of inefficient
programs are co mpared for the two crops in
Table 2. These results are generated from a
best-bet set of assumptions. Threshold
leve ls of 25,000 and 200,000 ton s in

potatoes are about twice as large as the

amounts needed for w heat.
Using these threshold levels, we make
the fo ll owing country breakdown for the 94
developin g nations producing potatoes
from 1995 to 1997 (FAOSTAT, 1998).
Thirty-three countri es produce less than
25,000 tons annua ll y. Of these, on ly five
have in vested in testing programs, and
three have released va ri eties that we re
se lected from CIP m aterial. Four have been
CIP Program Report 1997-98

23 1

Table 1. Comparing parameter assumptions on the economic fea si bility of wheal and potato breeding programs.
Parameter assumptions
Benefits
Rate of genetic gain
Testing
Crossing
Price (US$1/t)
Adoption
Varietal success
Average varietal age (years)

Wheat

6.35/kg/yr
l 0/kg/yr
Real price trend

0.635%/yr
1%/yr
150

Annually
5

Once in a decade
15

5
12

5
12

3
35,000

3
50,000

12%
70

12%
50

Research lags (years)

Testing
Crossing
Costs
Number of scientists per year
Testing
Crossing
Cost per scientist (US$)
Other
Discount rotes
Time horizon of the project (years)

partners in larger region al netwo rks .

Ove rall , th e eco nomic loss attributed to
inves tin g too much in test in g in ve ry sma ll
producing cou ntri es does not appea r to be
sizeab le.
Twenty-four countries have leve ls of
product ion between 25,000 and 200,000
tons w hi ch eco nomi ca ll y warrants a test in g
program but does not ju stify the additional
investment in a breedin g program . Severa l
of th ese cou ntri es do have act ive test in g
programs, but onl y one, Uruguay, has a fu ll
adapt ive potato-breedin g program.
Th e remaining 37 deve lop in g countri es
produ ce more than 200,000 tons annually.
On ly 11 of these in vested in fu ll potato

232

Potato

Potato

breedin g program s in the 1990s. Th e two

sma ll est of these co untries are Cuba w ith
about 350,000 tons and Ecuador w ith
around 500,000 tons. These leve ls are
considerab ly above the threshold ju stifyin g
a breedin g program. About half of th e
remainin g 26 co untries have testin g
program s. A sizable number of countries,
mainl y those in North Africa and th e
M iddl e East, do not even in vest in test ing
and se lecting imported material. In summary, th e picture emergin g in Tab le 2 is
markedl y different from Mared ia's and
Bye rl ee's ana lys is of w heat. Only a handful
of countries ca n be described as over in vesting in potato breeding on a size-of-produ ction cr iterion .

Table 2. Size of threshold production to justify investing in a testing or a more expensive crossing program and
the number of national programs that have over-invested relative to this economic criterion for wheat
and potato.

Threshold production (t)

Inefficient programs
(number of programs)

Program

Wheat

Potato

Wheat

14,800
111 ,400

25,000
200,000

2
19

Potato

capacity
Testing only
Crossing

International Public Sector Investment in

Potato Crop Improvement in Developing
Countries
Findin g an absence of ove rin vestment
makes th e effic ient all oca tion of intern ati ona l reso urces destin ed for potato plant
breedin g all th e mo re imperati ve. W alker
and Fu gli e (1999) set fo rth som e eco nom ic
co rn erston es for estab li shin g a strategy to
generate intern ati onal publi c goods fro m
co ll ec ti ve action in potato pl ant breedin g.
Sp ill -in ben efit s from va1ietal change are
we ll documented in North Am eri ca and
elsewhere. It is cl ea r th at altern ati ve
so urces of suppl y are important from NARS
in oth er develop in g countries, as w ell as
from NARS in deve loped countr ies. Typi ca l
of th e ex peri ence w ith va rietal change in
North Am eri ca, eco nomi c return s to potato
breed in g are likely to fa ll as th e produ cti on
and consumpti o n env ironments imp rove.
Over time dem and for di sea se res istance
dec rea ses and demand for traits related to
mark et qu ality in creases. Hi gher mark et
qu ali ty begets econom ic ri giditi es th at favo r
va ri eties alrea dy in th e market. Lastly, as
we saw in th e prev ious section , nati onal
product ion be low 25,000 ton s does not
wa rrant an in vestm ent in potato plant
breeding.
Armed w ith th ese facts, we now tack le
our dec ision-m akin g problem: Wh at is a
"best-bet" spati al all oca tion of pri oriti es by
an internation al agricultural resea rch center

(IARC) to generate in tern ation al pub li c

good s in potato crop im prove ment? Rath er
th an use a mathemati ca l mod el to so lve th e
prob lem, we employ a more intuiti ve
approac h draw ing on our ea rli er findin gs
and on additional in format ion related to
techn o logica l chan ge in potatoes in spec ifi c
reg ions.
Ou r popul ation of interest is th e 94
deve lop ing co untri es th at produ ce potatoes.
We brea k our dec isio n-making probl em
down into its pa rts by spec ifying criteri a in
th e fo rm of qu estion s th at fac i I itate th e
sortin g of countries into groups fo r differential trea tm ent (Figure 2).
Th e produ cti on thresho ld of 25,000 to ns
is our first criterion. About one-third of the
countri es fa ll below thi s economic thres ho ld fo r in vestin g in a genetic im provement
progra m fo r potatoes. In th ese very small
prod ucin g countries, potatoes are neith er a
stap le food crop nor a crop th at is impo rtant
to th e rural and urban poor. Countries
close to th e 25, 000-ton threshold wa rrant a
watc hin g bri ef, but overall , the prior ity fo r
in ves tin g in plant breed ing is neg li gible.
Th e nex t cr iterion gives priority to
countr ies th at have stab le breedin g programs and do not appea r to have reac hed
th e po int of diminishing return s to in ves tment as exe mpli fied by co nstant or even
ri sing va ri etal age. Onl y o ne country fits
th at bi ll : Chin a. Chin a is th e largest potato

CIP Progrnm Reporl 1997-98

233

94 developing countries with 107 million tons of production in 1995-1997

----------.
G
6
------- e

<Is production less than 25,000>

Very small producers
(33 countries, 0.17% production)

...

<Has the country been able to maintain a stable breeding program?>

...

<Are prospects for varietal change b right ?>

JI' '..
Other strong NARS
(9 countries, 26.6%
production)

China (45.5%
production)

<Are prospects good for

alternative sources of supply
from developed countries?>

------e

Middle East and North Africa (14

countries , 13.75% production)

...

<Are prospects bright for technology

borrowing from strong NARS in
developing countries?>

G----. ------ e

Neighboring countries (12

countries, 4.2% production)

<Is the country located in a relatively contiguous region

whose share of production exceeds 1 %?>

West Africa
(2, 0.1%)

Central Ame rica

and the Caribbean
(4, 0.2% )

Andes
(5, 6.2%)

East and Central

Africa (1 0, 2%)

Southeast
Asia (4, 1 .3%)

Figure 2. Spatial and institutional priorities for generating international public goods from potato breeding.

234

Potolo

produ ce r in th e world acco untin g for 40 %

of developing country output.
Demand for va ri eta I resis tance to di sease
is strong in China. Provin cial seed program s are improving, but on average, th ey
are not as developed as those in other
stro ng NARS. Severa l o ld c ulti va rs, such as
Mira, are st ill cult ivated in Chin a. Th ere
appea rs to be ample scope for varie tal
chan ge, howeve r, and China is sti ll lik ely to
be a major recipient of spi ll ove r benefits
from other strong NARS . Therefore, input
of CIP to China should span th e ga mut of
material from elite c lones for testing to
parents for cross in g.
Th e other strong NARS th at have mature
plant breeding programs ha ve for the most
part suffe red from a slow in g in th e pace of
va ri etal change. For th ese stro ng NARS, the
role of an IARC such as CIP should foc us on
the provi sion of appropriate parental
material for crossing. Stron g NARS may
sti 11 benefit from se lectio n of outstanding
individual s from progeny of cros ses made
by CIP, but a more up strea m, strategic
research role should enhance th e prospec ts
fm th e attainm ent of genet ic pro gress by
stron g NARS. Equ ally im portant, CIP needs
to fi nd ways to max imize th e chances that
strong NARS w ill generate sp il lover effects
in neighboring reg ions or in kindred
agroecologies.
Many of the conditions tha t give ri se to
stron g NARS may al so stimul ate th e
developm ent and enforcement of intell ectua l property rights whi ch will enhan ce
greater private sector part ic ip at ion particularly in the area of transge ni c va ri etal
change. Deve lopment of intellectual
prope rty ri ghts has not been sufficie nt to
engend er much priv ate secto r parti cip at ion
in potato cro p improve ment in deve lop ing
countries. But tran sge nic varieta l cha nge
could res ult in more pri va te sector participation. M o re needs to be lea rn ed about the
prospec ts for and impli ca tion s of transgenic
vari etal change in stron g NARS in develop ing countries.

Our next cr iteri on focuses on altern at ive

sources of supp ly from developed co untri es. Th e countries of the Middl e East and
North Africa (MENA) reg ion, are characterized by some uniqu e opportunities for
potato produ cti on (parti cularly in N orth
Afri ca and to a lesse r extent in th e Middl e
East). They export potatoes to Europ e in the
w inter seaso n and re ly on imports of tuber
seed from Eu rope. Th e seaso nal expo rt
market in fl uences loca l co nsum er preferences. Europea n varieties, particul arly
Dutch varieties, do wel l in this setting, and
seed co mpani es aggress ive ly promote
them. Ten countries in this region are large
enou gh to warrant a national potatobreeding program, but none have invested
in one on a sustained basis. In this relati ve ly hom oge neo us reg ion where a handful
of clo nes acco unt for th e bulk of prod uction , transge ni c var iet ies wi ll li kely ha ve a
greater role to play than conventionally
bred va ri et ies . Private sec tor biotechnology
companies, most li kely in collaboratio n
with publi c sector in stitutes in Europe, are
wel I placed to provid e transgeni c varieties
for th e MENA region .
We are not say in g th at investing in
potato cro p improvement in a wider sense
w il l not pay in the MENA region . Indeed,
success sto ri es of appl ied research o n
integ rated pes t ma nage ment and on seed
produ ct ion and suppl y to farmers are wel l
do cum ented. We are say ing, howeve r, that
interna tionall y ass isted breeding wi 11 not be
produ ctive.
This brings us to co untries bordering
stron g NARS. Deve lopin g country alternati ve suppli ers is our nex t fi lter (Fi gure 2).
Does th e co untry bord er a stron g NARS or
is it located in a pro ximate and shared
agro climati c environm ent? An affirmative
respo nse to thi s qu estio n applies to 12
countries that accou nt fo r about 5% of
deve loping country production. At times,
rece ivin g sp i 11-i n benefits from neighboring
countri es may not be politically popul ar,
but it is eco nomi ca ll y co rrect particul arly in
tim es of sca rc ity. Th ese cou ntri es should

CIP Program Report 1997-98

235

inve st in testing programs targeted on

adaptat ion of elite c lona l material and not
wait for fa rme r seed from the donor cou ntry
to sp i l l across the bord er. The rol e of an
IARC is to prov ide an alternati ve so urce of
materi al.
Th e above four criteria allow us to so rt
70 of th e 94 countries into five group s. Th e
rem ai nin g co untri es produce more than
25 ,000 to ns, have not bee n able to esta blish a stable potato breed ing program
without donor assistan ce, do not bord er a
stron g NARS, and do not have prod uct ion
characte ri stics w hi ch all ow for the transfe r
of deve loped country material w ithou t the
need fo r co nsi derabl e adaptation. Fro m the
persp ecti ve of potato breeding, th e optimal
con figuration is centralized bre edin g with
decentral ized selecti on (Simmonds, 1997).
This obse rva ti on suggests a final filter: Are
there co nti guous countries that account for
a siz eab le share of deve loping country
produ ct ion ? Finding such groupings co uld
wa rr ant a direct in vestme nt by CIP in a
regi ona l breeding progra m. We find three
such groupings : the Andes, the East and
Central African Highl and s, and South east
Asia.
Regio nal gro upin gs of the remainin g
countries account for less than one percent
of global production in our population of
interest. These countri es come from two
sources: large pockets of potato produ ct ion
in West Africa , specifically north wes t
Cam eroo n and th e Jo s Plateau of Ni ge ria,
and in the highest elevat ions of Central
Ameri ca and the Caribbea n. Traditionall y,
CIP has serviced the needs of th ese cou ntries through donor-ass isted special
proj ects, and this wo uld appear to be the
most effect ive modu s ope randi in the
future.
References Cited

Alston , J.M., P.G. Pard ey, and V. H . Smith.

1998 . Financing agricultural R&D in rich
countries: W hat's happe ning and w hy.
Aust. J. Agr ic. Res . Eco n. 42 (1):51-82.

236 Potato

A lsto n, J.M ., G.W. Norton, and P.G. Parde y.

1996. Science und er scarcity. Cornell
University Press , Ithaca, N Y, USA.
An derson, J.R. 1998. Selected po l icy issues
in internati onal ag ri cu ltural researc h: On
stri vin g for international publi c goo ds in
an era of donor fat igue. World Dev.
26(6):1149-1161.
Binswan ger, H .P. 1974. Borrow in g technology, ada pti ve resea rch , and researc h on
home tec hnol ogy. Pakis. Econ. Soc. Re v.
12 (S ummer):l 44-156.
Bren nan, J.P. 1992. Economic cr iteria for
esta bli shing pl ant breeding p rogra ms.
CIMMYT Econ om ics Work in g Paper 9201, CIMMYT, El Satan, Mex ico.
Bohn, A., D. Byerl ee, and M .K. Maredia.
19 98 . In vestment in wheat improve ment
in deve lop ing countries. In: Maredia, M.
K. and D . Byer lee . (eds. ). Research
effic iency in the prese nce of techno logy
spi ll ove rs: The case of nation al an d
international w heat improvem ent
research. CIM MYT Research Report.
CIMM YT, El Satan , Mexico. p. 1 6-29 .
FAOS TAT. 1998 (November). FAO (Food
and Agriculture Organization of th e
United Nations) Statistics Database
(O nlin e). Acc essed No vemb er 1998 .
Avai lab le HTTP : http://apps.fao.org.
Ma redia, M .K., R. Wa rd , and D.B yer lee.
1998. Assessme nt of the potential
intern at ional transferability of wheat
va ri eti es across co untries. In: M aredia,
M. K. and D. Bye rl ee. (eds.) . Resea rch
efficiency in th e presence of tec hnology
spill ove rs : Th e case of national and
intern atio nal whea t improve ment
res ea rch. CIMMYT Research Repo rt.
CIMMYT, El Satan , Mexico. p. 45-64.
Mared ia, M .K ., and D . Byer lee, D . 1998.
Efficiency of whea t improve ment re search inves tm ents in th e of sp ill-ins. In :
Maredia, M. K. and D. Bye rlee. (eds.).
Research effi c iency in the presence of
tec hn ology spillove rs: Th e case of
nati onal and intern ational w heat im prove ment researc h. CIMMYT Research
Report. CIMMYT, El Satan, Mexico. p.
99 -118.

Simmonds, N.W. 1997. A rev iew of potato

propagation by mea ns of seed , as di stin ct
from c lonal prop agat ion by tu be rs.
Potato Res. 40:191-214.
Wa lker, T. S. 1994. Patterns and implications of var ietal cha nge in potatoes .
So c ial Science Department Workin g
Pap er No. 1994-3. CIP, Lim a, Peru.

Walker, T. and K. Fu gli e. 1999. Th e econom ics of generatin g internation al public

goods from in ves tin g in pot ato pl ant
bree ding . ETPD Di sc ussion Pap er, IFPRI,
Washington, D .C.

CIP Prngrom Report 199798

237

Research on Sweetpotato
Impact on a Chang ing Wo rl d

The Causes and Control of Virus Diseases of

Sweetpotato in Developing Countries: Is Sweetpotato
Virus Disease the Main Problem?
E.E. Carey1, R.W. Gibson 2 , S. Fuentes 3 , M. Maclunud4, R.O.M. Mwanga 5,
G. Turyarnureeba 5 , L. Zhang 6 , D. Ma7, F. Abo EL-Abbas 8 , R. E L-Bedewy1
and L.F. Salazar3

Virus di seases ca n po se a significant

constraint to sweetpotato prod uction
(Mukibii, 1977; H ahn 1979). As a resu lt of
the histori ca l neglect of sweetpotato
research in most developing countries, our
know ledge of sweetpotato viruses, th eir
importance, distribution, and contro l are
still rather limited. Howeve r, recent stud ies
on th e occurrence of sweetpotato v i ruses in
a num ber of cou ntri es are helping to c larify
our und ersta nd in g of the globa l importance
of sweetpotato v iruses, and to defin e
contro l st rateg ies. An array of comp lementary techn iques are bein g used. Th ey
in clud e diagnostic methods fo r th e detection of spec ific v iru ses, pathogen-tested
planting materials for the assess ment of
y ield loss due to v iru ses, and host p lant
resistance to virus diseases.
More than 14 virus diseases of
sweetpotato have been reported (Moyer
and Salazar, 1990; Brunt, et al., 1996
onward). Study of severa l of these diseases
has been hampered by th e lack of simp le
detection techni qu es. Antisera are now
avai lable aga in st a numb er of sweetpotato
viru ses including sweetpotato chlorotic
fleck virus (SPCFV), sweetpotato latent v irus

1 CIP, Nairobi , Kenya.

2 Natural Resources Institute, UK.
3 CIP, Lima, Peru .
4 Research In stitute for Food Crops Biotechno logy, Bogor,
Ind onesia.
5 Namulonge Agricultural and Anima l Production Research
Institute, Uganda.
6 Shangdong Academy of Agricul tural Sciences, China.
7 Xuzhou Sweetpotato Research Cen ter, China.
8 Uni versi ty of Ainsha ms, Egypt.

(SP LV), sweetpotato mild mottl e v irus

(SPMMV), C-6, sweetpotato mild speck ling
virus (SPMSV), sweetpotato feathery mottle
virus (SPFMV), and sweetpotato chlorot ic
stunt virus (SPCSV) previo usly known as
sweetpotato sunken vei n v iru s (SPSVV).
Antise ra for all but SPCSV are produced
at CIP and are routin ely used for nitrocellulo se membrane-en zyme- Ii nked
immunosorbent assay (NCM-EL ISA).
Antisera (monoc lonal antibodies) spec ific to
SPCSV are produced by J. Vetten at the
In st itute for Bioch em istry and Plant Viro logy, Braunschwe ig, Germany (Hoyer et al.,
1996). These antisera have been used to
survey the inc id ence of sweet potato v iru s
diseases at a number of sites.
Sweetpotato virus disease (SPVD) is the
name that has becom e associated with a
serious disease of sweetpotato, symptoms of
which typica ll y in c lud e leaf distortion ,
chlorosis, disco lorat ion, and stunting. Work
initi ally in West Africa and later in East
Afri ca showed it to be caused by comb in ed
infect ion with aph id-borne SPFMV and
wh itefly-borne SPCS V (Sc haefers and Terry,
19 76; Winte r et al. , 199 2; Hoyer et al.,
1996; Gibson et al. , 1998a) Despite the
apparent broad mean ing of the nam e
SPVD, the symptoms are so characteristic
that the name has become restricted to the
disease with these sympto ms and ca used by
these v i ruses.
Th e first report of SPVD may have bee n
in eastern Belg ian Congo (now DR Congo)

CIPProgram Report 1997-98

241

in 1939. SPVD was first described in East

Africa by Sheffield (1953).
SPVD ca n red uce y ield s of infected
plants by up to 80% (H ah n, 19 79; Mukiibi,
1 977). It has recentl y bee n show n to be the
princip al v iru s diseas e of sweetpotato in
East Afr ica whe re hi gh disease incidences
are often recorded in farmers ' f ields (Gibso n
et al. , 1998b; A ritua et al. , 1998a). In
Ugand a, disease incidence and severity in
farmers' fields have been show n to be
c losely assoc iated with preva len ce of
w hiteflies (Aritua et al., 1998a, b, c). There
are large var ietal differences in susceptibility to SPVD . Good sou rces of resistance are
prese nt in local germplasm, but ofte n are
associated with later-m aturing, lowe ry ieldin g genotypes (Aritua et al. , 1998b,
Carey et al. , 199 7). SPCSV isolates from
eastern and southern Africa have been
shown to be se rolo gically different from
those from West Africa, Brazil, USA, and
Israel (H oyer et al., 1996; Gibson et al.,
1998a). These different isolates may have
biologica l sign ificance in asmuc h as cult ivars resistant to SP VD in N igeria we re
susceptible to SPVD in Uga nda (Mwanga et
al., 199 1 ).
Th is pape r presents a brief ove rview of
recent co l laborative sweetpotato vir us
rese arch co nducted by our project. We
start w ith ev id ence from surveys conducted
in a number of countries usin g antisera
available aga in st sweetpotato vi ru ses,
inclu d in g SP FVM and SPCSV, th e co mponents of SPVD. We th en assess the effectiven ess of the use of plant in g materi al fr ee
of kn own pathogens (pathoge n-tested) as a
v iru s co ntrol measure in China and
Uganda. Th e impli cat io ns of these resu lts
for vi ru s con trol strategi es are di sc ussed.

Materials and Methods

Surveys of sweetpotato v iruses we re
conducted in China, Egypt, Ind onesia, Peru,
and Uganda. Vi rus es assayed usin g anti se ra
and NCM-ELISA k its from CIP were SPCFV,
SPLV, SPMMV, C-6, SPMSV and SPFMV.

2 4 2 Sweetpototo

SPCSV was assayed by di ffe rent means at

di fferen t sites. In Uganda, tripl e antibody
sandwich -ELIS A was used as described by
Gibson et al. (1998b). Thi s method used
two sets of monoc lo nal antibodies to
discriminate between the East African and
West Afr ica n serotypes of SPCSV. In
Ind ones ia and Peru , SPCSV was detec ted
using NCM-EL ISA. In Egypt, SPCSV was
detected usin g po lymerase chain react ion
(PCR) assay and tests we re co ndu cted by
Dr. S. Winter, Braun sc hwei g, Germany.
Deta il s of sa mplin g procedures at eac h
survey locati o n are given in the references
c ited in Ta ble 1. In Ke nya an d Egypt, both
symptomless and symptom -ex pressing
plants we re sampled. At other sites, o nl y
symptom-expressing plants we re samp led.
Pathogen-tested p lantin g mate ri al fo r
tria ls in China and Ugan da was obtai ned
using standard tec hn iq ues of th ermothe rap y
and meristem tip c ul tu re, fo ll owed by
ind exing to assure v irus-free status (Love et
al, 1989). Fi eld trials we re co ndu cted to
compare y iel d s of the pathogen-tested
p lanting materia l w ith fa rm-d eri ved planting mater ial of the sa me cultivars. In
China, trials were conducted durin g the
sp rin g crop ping seaso n of 1 997 at Xuz hou ,
Ji angs u Province, and Jinan , Shangdong
Prov in ce. Fi ve cu ltiva rs were used. Tria ls
consisted of 100-pl ant plots (5 rows of 20
cutt in gs eac h) pl anted in a randomized
comp lete block design (RCBD) with three
replica tio ns. Tr ials we re harvested 170 d
after p lanting.
In Uganda, three cultivars were used and
tr ials were co ndu cted at three si tes:
Kachwekano in the so uth weste rn hi gh land s,
and Nam u longe and Nakabango in the tall
grass lands agoecologica l zone in the so uthce ntra l part of th e co untry. Trials were
co ndu cted at Kachwekano in 1995 , and at
Namulonge and Nakabango in 1996.
Plant in g mater ial from the 19 95 tr ials was
used to p lant th e 1996 tr ials. Each tria l was
pl anted using a RC BD w ith three replicati ons . H arvest plot sizes were 16 plants for
trials co nducted in 1995 and 54 plants for
the tr ials conducted in 1996.

Table 1. Frequencies of detection of sweetpotato viruses from surveys conducted at various sites.

=;;

Location

SPCFV

SPLV

SPMMV

C-6

SPMSV

SPFMV

SPCSV

SPFMV + SPCSV

Uganda (six sites sampled)

Kenya (four sites sampled)
Canete, Peru
San Romon, Peru
W., C. and E. Java, Indonesia
(20 villages)
Irion Joya germplosm
collection, Lembang,
Indonesia
Kafr El Zayat, Egypt
(6 cultivars from the
Egyptian seed
program)
Shongdong, China
(5 infected farmers'
cultivors)

5/106
0/ 182
0/183
2/87
no

0/106
6/182
13/ 183
0/87
no

6/106
13/182
4/183
0/87
no

no'
0/182
0/183
0/87
no

no
no
12/183
2/87
no

58/l 06
40/182
43/183
72/87
256/890

105/lQbb
no
59/183 b
61/87 b
237/890'

57/ l 06

11/381

8/381

5/381

7/381

5/381

18/381

no

Pro in et al., 1998

4/382

5/382

6/106

no

no

67/382

+d

yes

Abo El-Abbas et aI., 1998

0/115

15/ 115

0/115

0/115

0/115

93/115

no

Zhang and Ma, 1998

no = not assayed.
E. African strain detected by use of monoclonal antibodies.
' NCM-ELISA test, SPCSV strain not yet identified.
d + signifies detected by PCR, but frequency not quantified.
0

=
~

~
~

~
./;>.

34/183
59/87
155/890

Reference
Gibson et al. 19980
Corey et al., 1998
Solazar, 1998
Salazar, 1998
Proin et al., 1998

Results
Results of virus surveys conducted in
China, Egypt, Indon es ia, Ken ya, Peru, and
Uganda are presented in Tabl e 1. SPFMV
was assayed at all sites and was detected
w ith hi gh frequen cy. SPCSV was not
assayed at all si tes, but w here assayed , it
was detected at hi gh frequencies , similar to
SPFMV. In Uga nd a, 54 % of th e plants
sam pl ed had SP VD, bein g infec ted by both
SPFMV and SPCSV. At si te s in Java,
Ind onesia, about 65 % of th e SPCSVinfected sa mpl es were also infected with
SPFMV. In Egy pt, th ere w as also combined
infecti o n, but the frequency of SPCS V was
not quantified. In Peru, the fre qu ency of
comb in ed infecti o n by SPFMV and SPCSV
was 34% of the sampled plan ts . Th e
serotype of SPCSV was only determined
durin g surveys in Uganda and Peru. In
both countries, th e so-cal led East A fri ca n
serotype was detec ted , not the West Africa n
seroty pe. Th e remaining v iru ses, except for
C-6, w hi c h was only detected in the lri an
Jaya ge rmpl as m collection at Lemba ng,
Ind ones ia, were ge neral ly detected at a low
frequency but on eac h of th e co ntin ents
su rveyed.
Yield loss assessment trials in Chin a
(Tabl e 2) demonstrated a significant benefit
to usin g pathoge n-tested planting m ater ial
at both Jin an and Xuz ho u. Th e effect of
using pathoge n-tested planting material was
greate r at Jin an, w ith a signifi cant ly hi gher
y ield from pathogen-tested pl ant in g
m ater ial than from far m-deri ved material for
eac h cultivar. At Xuzhou the pathogentes ted pl antin g m ater ial o nl y yie ld ed
signifi ca ntly more fo r two of the cul tiva rs
eva lu ate d . Ho weve r, the mean y ield from
pathoge n-tested pl antin g materi a I was
cons istentl y high er th an from farm-derived
material.
In the Ugandan tria ls (Table 3), no
sig ni fica nt effect of using pa thoge n-tested
pl antin g material was detected. The mean
yield of pathogen-tes ted plantin g m ater ial
was not co nsistentl y hi gher. Howeve r, in
three of fou r trials, th e m ea n y ield from
2 4 4 Sweetpototo

pat hogen -tested cul ~ i va r Tanza nia was

higher than th e y ield from farm-derived
pl ant in g material. Tanza nia was also th e
hi ghest y ieldin g cul t iva r in three out of fo ur
tr ia Is.

Discussion
Th e find ing that SPFMV was predom in ant at
each site sam pled is in line w ith previous
in for m ati o n abo ut th e cos mopo litan
di stribution of t hi s virus. In th e countries
w here it was assessed-Egypt, Ind ones ia,
Peru, and Uganda-SPCSV was present in
as hi gh a propo rti o n of v irus symptomexpressi ng pl ants as SPFMV, often in
association wi th th e latter. This appea rs to
indi ca te a broader distribution and importance of SPV D than h as previously been
repo rted. Th e detecti on of th e East Afr ica n
strain of SPCSV as th e predominant form in
Peru also indicates a w id er di str ibuti o n o f
this serotype th an was previo usly re ported
(H oye r et al. , 1996). Prelimin ary resu lts
have indi cated the presence of SPCS V in
China (L. Salazar, CIP, Lim a, Peru , 1999,
pers. com m .).
The results of th e presen t stu d y indicate
th at SPVD may be the most important v irus
disease of sweetpotato globall y. Until now,
its globa l impact has been overlooked. Th at
is largely because diagnostic to o ls have not
been readily ava ilabl e fo r SPCS V, so only
SPFMV was ubiquito usly detected. Furth er
survey work is required to confirm th e
occ urrence and prevalence of SPVD and its
causal agents (SPFM V and SPCS V) aro und
th e wo rld. Further wor k is al so required to
clarify the biological sign ifican ce of
di fferen t seroty pes of SPCSV. Because of a
lack of ant ise ra against a number of
addit ional sweetpo tato v iru ses reported in
th e literature, these have not yet bee n
wide ly surveyed. One o r more of th em
co uld also be important.
Resu lts from Chin a of yiel d com par iso ns
of sweetpotato vari eti es using pathoge ntested and fa rm-deri ved planting materials
(Table 2) indi ca te th e large y ield gai ns th at
ca n be m ade usi ng this tec hn iqu e. Ind eed,

Table 2. Fresh storage root yields (t/ha) of five major sweetpotato cultivars grown from pathogen-tested and
farm-derived planting materials at two sites in China in 1997.

Cultivar

Planting

Yield and difference {%) between planting

material

material sources for each cultivarb

source'

Jinan

Xuzhou

Lashu No. 7

PT
F
PT

Lushu No. 8

PT

Beijing 553

PT
F
PT

43. l (36%**)
31.6
46.4 (35%**)
34.6
39.9 (23%**)
32.4
39.4 (47%**)
26.9
37.7 (75%**)
21.5

41.5 (22% ns)

34.0
39.0 (17% ns)
33.3
4l.6 (48% ns)
28.0
39.0 (64%*)
23.8
45.7 (92%*)
23.8

Xushu 18

Fenghsoubai

PT = Pathogen-tested, F = Farm-derived.
bSignificant according to LSD (0.01) = **or LSD {0.05) = *,not significant = ns.

Table 3. Fresh storage root yields (t/ha) of three sweetpotato cultivars grown from pathogen-tested
and farm-derived planting materials at three sites in Uganda during two seasons.

Cultivar

Planting

Yield and difference{%) between planting material

sources for each cultivarb

material
source'

Tanzania

PT

Bwanjule

PT
F
PT
F

Wagabolige

Nakabango

Kachwekano

Namulonge

(1995)

(1995)

(1996)

(1995)

15.0 (19%)
12.6
5 9 (-40%)
9.9
6.3 (-40%)
10.5

39.8 (39%)
28.7
28.2 (28%)
22.l
20.8 (-13%)
23.8

l l.5 (-14%)
13.4
19.3 (7%)
18.0
20.4 (l 6%)
17.6

13.8 (50%)
9.2
11.0 (29%)
8.5
9.4 {-8%)
10.2

PT = Pathogen-tested; F = Farm-derived, symptomless.

bNo significant effect of planting material source was detected.

in Shangdong, the use of pathogen-tested

planting mater ial has been widely adopted
in th e past few years and extended to all
parts of the Prov ince (Fuglie et al., 1999).
There has been a very high payoff to this
research and extension effort. Increased
production has supplied the increas ing

dema nd for sweetpotato as an industrial

and feed so urce, while fa rm ers' production
costs have actually been redu ced due to
improved seed hea lth. Fo llowing initi al
investment of public mon ey in the cl ea nseed system, the costs of operation of the
seed system are borne by the farmers w ho

CIP PJOgram Report 1997-98

245

purchase the planting material , result in g in

an econom ic ally v iab le business. Th e
experie nce in Shangdong is being extended
to other provinces in China, and may
cont inu e to have a large impact on
sweetpotato production there. A successfu l
clean planting material program has also
been established in Egypt, made highly
effective by the high virus pressure there
and the susceptibility to viruses of the major
commercial cultivars (Abo EL-Abbas et al.,
1998).
Results of our assessment of the potential
for using pathogen-tested planting material
in Uganda were in marked contrast to the
results from China. In Uganda, no benefit
to using pathogen-tested planting material
was demonstrated. That was probably due
to the relatively high leve ls of virus resistance in the landrace cultivars that we used
for our test in Uganda. Also, we fo ll owed
the normal farming practice of using only
symptom less (esse ntial ly health y) plants as
sources of planting material (Gibson et al. ,
1997).
H ad we chosen more susceptib le
genotypes for our comparison in Uganda,
or had we chosen to use symptom-expressing farm -d er ived planting material fo r our
experiments, it is likely that we wo uld ha ve
detected differences. Ind eed, for cultivar
Tanzania, th e most vi ru s-susceptib le of the
genotypes evaluated (rated moderately
resistant by Mwanga et al. (1995)), there
was a tende nc y for pathoge n-tested planting material to y ield hi gher than fa rmderived planting materi a I. We w i 11 continue
to investigate the potenti al for using
pathogen-tested plantin g material in
Uganda and elsewhere in eastern and
southern Africa, particularly for more virussusceptib le but high y ieldi ng cultivars. It
should be noted , however, that sweetpotato
production systems in Sub-Saharan Afr ica
are less commercially oriented than systems
in China and elsewhere, with sweetpotato
primarily a food security crop for lowincome farmers. That might pose a constraint to the wi despread impl ementation of

246

Sweetpototo

pathogen-tested planting material schemes

as viab le enterprises.
Given the relatively unprom isin g
prospects for pathogen-tested seed schemes
in Uganda, and the high leve ls of res istance
to SPVD that ex ist in the African germplasm
(Carey et al., 1997), breeding resistant
varieties wi ll continue to be a principal
component of SPVD control here. As
already mentioned, there is a tendency for
local farmers' cu ltiva rs w ith high leve ls of
resistance to be rather low yielding and late
ma tu ring compa red with earlier maturing,
high yielding, yet susceptible local cu lti vars
or exotic introductions (Aritua et al. , 1998b;
Carey et al., 1997). Th e higher mean yield
of the more v irus susce ptible cultivar
Tanzania compared with the other two
cultivars in our trial is in line with this
observation .
A principal aim of the sweetpotato
breeding program at Namulonge, Uganda,
where the Ugandan program and CIP are
concentrating breeding efforts, is to com bine SPVD resistance with desirabl e
agronomic traits such as yield, earl in ess,
and acceptable culinary quality. Considerable progress has already been made, and
the program has released 12 cultivars
combining good leve ls of field resistance
with hi gh yiel ds and other desirable traits
(Mwanga et al., 1995; Turyamureeba et al. ,
1998). However, improvements in breedin g effic iency are required, including the
development of screen ing techniqu es to
rapidl y id entify resistant genotypes. Cu rrently, very large progeny population s (as
many as 100,000 seed lings) are required to
id entify a handful of resistant, agronomically superior genotypes. The inh er itance
of host plant re sistance is und er stud y. It is
anticipated that considerable progress wi ll
soon be made in breed ing for resistance .
The international value of resistance
breeding wil l depend on the degree to
whic h resistant genotypes fro m one site
maintain their resistance at other sites.
Recent ev id ence of the widespread globa l

occurre nce of SPV D lea d s us to anti c ip ate

that res istan ce from Africa may ho ld up in
oth er enviro nm ents w here v iru ses are a
probl em. H o w ever, difference s in v iru ses
or vi ru s strain s betwee n si tes co uld ca use
res ista nt ge no types fro m one site to be
susce pti b le in others. Intern ati o nal studi es
are und erway to in vesti gate thi s, usin g
stand ard sets of ge notypes w ith va ri ous
leve ls of res istance to SP V D .
Literature Cited
Abo El-Ab bas, F, J. Ishak, A. Hewezy, A.
Kand i !, and R. El- Bedewy. 1998. Co ntro l
of sweet po tato vi ru ses . A nnu al Subproj ect Prog ress Report (080305) fo r
Egypt. CIP, Lim a, Peru.
Aritu a, V., E. Adip ala, E.E . Ca rey, and R. W
Gib so n. 1998a . Th e in c id ence of sw ee t
potato v irus di sease and v irus res istance
of sweet pota to grow n in Uga nd a. A nn.
Appl. Bi o l. 132 :399 -411.
Aritu a, V., T. A li ca i, E. A dip ala, E.E. Ca rey,
and R.W. G ib so n. 199 8b . A spec ts of
res istance to sweet potato viru s di sease
in sweet potato . Ann . Appl. Bio l.
132:387 -398 .
Ari tu a, V., J. Legg, P. Sse ru wag i, N. E.J.M.
Smit, and R.W. G ib so n. 1998c. W hitefli es and assoc iated v iru ses on
sweetpotato in Ugand a: Prevale nce,
di stributi o n and importance . Prese nted
at th e 7111 tri enni al sy mpos ium of the
ISTRC-AB held 11 to 17 Oc t. 1998 in
Cotonou , Benin.
Brunt, A.A., K. Crabtree, M.J. Dall witz, A. J.
G ib bs, L. Wa tso n, and E.J. Z urc her. (eds.)
1996. Pl ant v iru ses onlin e: Desc rip tions
and l ists fro m the V IDE database.
Version: 20th Au gust 19 96 . URL http ://
bi o logy.a n u .ed u .a u/ G rou ps/M ES/v i de/.
Carey, E.E., S.T. Gi c huki , P.J. Nd o lo, G.
Tur ya muree ba, R. Kapin ga, N. B.
Lu taladi o, an d J.M. Teri . 1997. Co llabo rative sweetpotato breedin g in easte rn ,
ce ntral and so uth ern Afri ca. In: CIP
Prog ram Report 1995 -96 . Lim a, Peru . p.
49-57.
Carey, E.E ., S.T. Gi c huki , R.0.M . M wa nga,
S. Kas ul e, S. Fu entes, C. M ac hari a, and

R.W. G ibs on. 1998. Sweetpo tato

viru ses in Uga nda and Kenya: Res ults of
a survey. In : A ko rod a, M .0 and 1.J .
Ekan aya ke (eds.). Proceedin gs of th e 6111
trienn ia l sympos ium o f the Intern at io nal
Soc iety of Tro pi cal Root Cro ps - A fri ca
Branch held 22 to 28 Oct. 1995 in
Li lo ngwe, Malawi, spo nso red by Governm ent of M alaw i and In te rn ation al
In sti tute fo r Tropic al Agri cul ture (llTA ).
llTA, Ibadan, Ni geria. p . 4 57-461 .
Fu gli e, K.O., L. Zhang, L. Sa laza r, and T.
Wa lke r. 1999 . Eco nom ic Im pac t of v iru sfree sweetpotato pl antin g mate ri al in
Shangdong Prov in ce, Chin a. In : Impact
on a Changi ng W orld . C IP Prog ram
Report 1997 -98 . CIP, Lim a, Peru. p. 24925 4.
G ib so n, R.W ., R.0. M . M wa nga, S. Kas ul e,
I. Mpemb e, and E.E. Ca rey. 199 7.
Ap pare nt abse nce of v iru ses in most
sy mptomless field- grown sweet potato in
Uga nda. A nn . Appl. Bio l. 130: 4 81-4 90.
G ibso n, R.W, G.C. Kaitisha, and J.M .
Ra ndri anaivoa ri von y. 1998a . Id entifi c atio n of th e Eas t Afri ca n strain o f swee t
potato chl o roti c stunt viru s as a maj o r
co mpo nent o f sw eet potato viru s di sea se
in So uthern A fri ca. Pl ant Di s. 132:1063.
G ib so n, R.W., I. M pembe, T. A lica i, E.E.
Ca rey, R.O .M. M wa nga, S.K . Sea l, and
H.F. Vetten. 1998 b . Symptom s, aeti o logy
and sero log ica l anal ys is o f sweetpotato
v iru s di sease in Ugand a. Pl ant Path.
47:95- 102.
Hahn, S.K . 1979. Effect o f viru s (S PV D) o n
growth and y ield of sweetpotato. Exp.
Agr ic. 15:253-256 .
H oye r, U., E. M aiss, W. Jekm ann , D.E .
Lesse mann , and H .J. Vetten. 1996 .
Ide nti ficat io n of coa t pro tein ge ne of
sweetpotato sunken ve in c lote roviru s
iso late from Kenya and ev id ence for
se rol og ica l relation ship amo ng geo gra phi cal ly di ve rse clos terov irus iso lates
fro m sweetpotato . Ph yto path. 4 7: 58258 7.
Love, S.L ., B.B. Rh odes, and J.W. Moyer.
19 89. M eri stem-tip cul ture .a nd v iru s
ind ex in g o f sweet potato. IBPG R, Ro me,
Italy.

CIP Ptogram Report 1997 -98

24 7

Moyer, J. and L.F. Salazar. 1990. Viruses

and v irus-like diseases of sweet potato.
In: CIP. Control of virus and virus- like
diseases of potato and sweet potato.
Report of the 3rd Planning Conference
held 20-22 Nov. 1989 in Lima , Peru. CIP,
Lima, Peru. p. 13-19 .
Mukiibi, J. 1977. Effect of mosaic on the
yield of sweetpotatoes in Uganda. In:
Cook, J., R. Macintyre, and M. Graham
(eds.). Proceedings of Tropical Root
Crops. CIAT, Cali, Colombia. p. 169-170 .
Mwanga, R.O.M., C.N.O. p'Obwoya, G.W.
Otim-Nape, and B. Odongo. 1991.
Sweetpotato improvement in Uganda. In:
Alvarez, M.N. and R. Asiedu (eds.). The
role of root crops in reg ional food
security and sustainable agriculture.
Proceedings of the 4h Eastern and
Southern African. Root Crops Wo rkshop.
llTA, Ibadan , Nigeria. p. 59-67 .
Mwanga, R.O.M., B. Odongo, N. Smit, C.
Ocitti p'Obwoya, and R.W. Gibson.
1995. Submission to the Variety Release
Committee for release of sweetpotato
varieties. Application for inclusion of a
crop/ variety in the National Cultivar List
No F00004-9. The Ugandan Ministry of
Agriculture , Animal Industries and
Fisheries, Uganda.
Prain, G., M. Machmud , and Rusmadi.
1998. Evaluation of v irus diseases.
Annual subproject progress report
(080304) for Indonesia. CIP, Lima, Peru.

248

Sweetpototo

Salazar, L. 1998. Id en tification and characterization of sweetpotato vi ruses and

search for resistance. 1998. Annual
Subproject Progress Report (0801) for
Peru. CIP, Lima, Peru.
Schaefers, G.A. and E.R. Terr y. 1976.
Insect transmission of sweet potato
disease agents in Nigeria. Phytopath.
66:642-645.
Sheffield, F.M.L. 1953. Virus diseases of
sweet potato in parts of Africa. Empire J.
Exp. Agric. 21 :184-189.
Turyamureeba, G., R.O.M. Mwanga, B.
Odongo, C. Ocitti p'Obwoya, and E.E.
Carey. 1998. Application for inclusion of
a crop/v ariety in the National Cultivar
List. Submission to the Va riety Release
Committee for release of sweetpotato
va rieties. The Ugandan Mi nistry of
Agriculture, Animal Industries & Fi sheries, Uganda.
Winte r, S. 1 A . Purac, F. Leggett, E.A. Frison,
H.W. Rossel, and R.I. Hamilton. 1992.
Partial characterization and molecular
cloning of a closterovirus from sweet
potato infected with the sweet potato
vi rus disease complex from Nigeria.
Phytopath. 82:869-875.
Zhang, L. and D. Ma. 1998. Evaluation of
sweetpotato viruses. Annual Subproject
Progress Report (080302/ 03) for China.
CIP, Lima, Peru.

Economic Impact of Virus-Free Sweetpotato Planting

Material in Shandong Province, China
K.O. Fuglie1, L. Zhang2 , L. Salazar 3 and T. Walker3

In many developing countr ies, y ield s of root

and tuber crops are significant ly reduced
due to diseases and pests in the planting
material. (In sweetpotato, the sources of
plantin g material [wh ich are either v in e
cuttings or roots] are equ iva len t to c lona l
seed for vegetative propagation.) The
development and transfer of new methods
and technologies for producing disease-free
clonal seed ca n overcome this constraint
and help unlock the significant yield
potential of these crops. In potato, the use
of virus detection and tissue culture techniques in tuber seed programs is common
espec iall y in developed countr ies. In
sweetpotato, such techniques have not
been used even in the Un ited States where
propagation material is periodically renewed to maintain c lon al purity. Virus
el imin ation is not an objective or consequence of these programs.
In the late 1980s, the International Potato
Center (CIP) in co ll aboration with Chinese
agricultural sc ient ists began a project to
develop and transfer new methods for
propagating v iru s-free roots and vines for
sweetpotato production in China. Such
systems are now being developed in the
main sweetpotato growing areas of China.
The most advanced program can be found
in Shandong Province, where v iru s-free
roots were first distributed to farmers in
1994. This paper briefly describes the
sweetpotato propagation program, its
impact on yie ld, and its economic consequences.

1 CIP, Bogor, Indonesia.

2 SAAS, Jinan , China.
3 CIP, Lima , Pe ru.

Development of the Virus-Free

Sweetpotato Multiplication System
Sweetpotato is the th ird ranking crop in
Shandong Province after maize and wheat.
In the farmers' traditional system,
sweetpotato roots are stored from the fal l
harvest to the following sp ring planting.
Sweetpotatoes are typica lly grown on
poorer-quality rainfed land in rotation w ith
maize or grou ndnuts, or in a w inter wheatsweetpotato-maize rotation that produces
three crops every two years . The spring
sweetpotato crop is planted in late March
or Apri l and harvested in September or
October. When following w inter wheat,
sweetpotato is pl anted as a summer crop in
Jul y and then harvested in October after the
spring crop is harvested. The spring crop
accounts for about 60% of the total
sweetpotato area in Shandong and exhibits
a higher average yie ld than the summer
crop because of its longer growing season.
Planting material for the summer crop
comes from vines cut from the previously
estab li shed spr in g crop .
In the late 1980s, propagation of virusfree roots was identifi ed by international
and national sweetpotato scientists as a
possible means of improving fa rm yields of
existing sweetpotato var ieties in China.
Field observations together with laboratory
testing indicated high leve ls of virus
infection in sweetpotato plants in farmer's
fields, espec ially the sweetpotato feathery
mottle virus (Zhang, 1996). In 1988 with
technical assistan ce from CIP, training
courses were organized to demonstrate
ho w to produce and multiply virus-free
sweetpotato roots. The technique involves
tak ing meristem tips from selected plants

CIP Prngrom Report 1997-98

249

and then regenerating the entire plant using

a culture medium (Sa lazar, 1996). Heat
treatment is used to reduce vi ru s concentration in the plant' s growi ng tips up to a point
w here so me are freed of v i ruses. ELISA
tests are cond ucted to confirm that plantlets
are free of virus infection. Virus-free
plantlets are then grown out in heated
greenhouses an d used as mother plants for
multiplication.
Following the CIP training courses,
scient ists from the Shandong Academy of
Agricultural Science (SAAS) co ndu cted
applied research in tissue cu lture and
multiplication techniques to estab li sh the
most appropriate medium fo r the cultures,
the best tim in g for transplanting, and other
multiplication methods. Beginning in 1992,
SAAS cond ucted fie ld trials in different
locations of the provin ce to compare y iel ds
of virus-free roots with yie lds from farmers'
roots. These trials showed an average y ield
increase of aroun d 40 % in plots using v irusfree roots of popular new var ieties and eve n
larger increases w ith virus-free roots of
o lder va rieties (Zhang, 1995). Th e success
of th e trials led to substantial financial
support from provincial and loc al govern ments for extension and mu ltipli cat ion of
v iru s-free material for propagation . In 1993
and 1994, large-scale exte nsion and
demonstration trials were carried out in al l
of th e major sweetpotato producing
counties of the province. The first v irus -free
roots we re exten ded to farmers in 1994 and
extensio n activities were intensified in
1995.
Th e organizatio n of the v iru s-free
sweetpotato multiplication system in
Shandong is depicted in Figure 1. Because
of th e possibility of getting two multipli cations of propagation material per year (the
spring and summer crops), the Shandong
program has ac hi eved exceptionally high
rates of multiplication: one ha planted the
first year can provide suffic ient clean roots
to plant 250-400 ha the fo llow in g year. In
1998, net houses in the province had a
capacity to produce 10,000 plant cuttings
per year, enough to produce first generat ion
250

Sweetpototo

Tissue culture lab with

Elisa testing

Province

Virus -free plantlets

District
First year
Winter: JFM

Heated greenhouse

500 virus-free
cuttings
County
First year
Spring Summer:
A-M-J-JA-S

Net house (670 m2)

April: 500 cuttings
July: 4,000 cuttings
2 tons of virus-free
pre-original seed

Township
Second year
SpringSummer: AMJ-J-A-S

Open field in iso lated

area
April: 2.7 ha
July: 26. 7 ha
800 tons of
original seed

Farmer
Third year
Spring-Summer:
A-M -J-J-A-S

Farmer's field
April: 1,000 ha
July: 1o,ooo ha

200,000 tons of 1st

generation seed roots
stored for next season

120,000 tons
of commercial
roots

Figure 1. The virus-free seed propagation system in

Shandong Province.
production roots for more than 500,000 ha
two years later.

Data and Methods for Impact

Assessment
Two main sources of information were
used to exami ne th e impact of v irus -free
sweetpotato plantrn ate rial on fa rm y ield s,
production , and income: (1) a survey of
sweetpotato farmers in 30 vi l lages condu cted in 1998; and (2) data on yield
demonstration trials, plant material production, and costs of researc h, extension, and
plant multiplication provi ded by SAAS.

The village survey was designed and

conducted by a multidisciplinary team of
researchers from SAAS and Cl P in August
and September 1998. Thirty villages in
Shandong Province were selected from the
seven districts with the greatest area
planted to sweetpotato.

Results and Discussion

Market structure
Responses from the surveys suggest that
demand is highly elastic. Not only is
production impact regional, but more than
80% of the sweetpotatoes produced in the
province are either fed to livestock (where
they compete with locally grown and
imported maize-based feed) or are processed into starch and/or noodles for export
to other provinces or countries. In the
survey, farmers were asked whether the
sweetpotato market prices had been
affected in recent years by the increase in
production attributed to virus-free planting
material. In all cases, farmers responded
that market prices for sweetpotatoes had
been stable over the past several years and
that demand from the processing industry
was strong and could absorb whatever they
produced.
Survey responses also provide evidence
of an inelastic supply, at least with respect
to the adoption of virus-free planting
material. Inelastic supply coupled with
elastic demand is our preferred "baseline
characterization" of market structure for
estimating the economic impact of virusfree planting material in Shandong Province.

The rapid diffusion of virus-free roots can

probably best be explained by its significant
and noticeable effect on yield. Both largescale demonstration plots conducted by the
Shandong agricultural extension service
during 1993-1994 and the village survey
show similar levels of average yield improvement from virus-free roots. For the
most widely grown variety, Xushu 18, virusfree roots are estimated to have increased
yield of the spring crop by 11 t/ha, or by
more than 30% over the yield obtained
from farmers' own roots. Yield gains from
other important varieties (Lushu 7, Lushu 8,
Beijing 553) for the spring crop ranged from
6.7 to 10 t/ha according to the village
survey and by between 13 and 16 t/ha in
the demonstration plots.
For the summer crop, yield of Xushu 1 8
was estimated by farmers to have increased
by an average of 10.6 t/ha, or by 41 %.
Summer crop yields of other varieties were
estimated to have increased by from 8.5 to
10 t/ha.
Gross benefits are estimated by multiplying the estimated area planted with virusfree roots by the average production
increase per ha by age of the planting
material. For new roots, the production
increase is estimated to be 10.35 t/ha , for
first generation roots the production
increase is estimated to be 9.6 t/ha, and for

Percent of Sweet Potato Area

100

- - - - - - -- - - - -

90
BOf----------7''------~----o

70 1---------r'---=,.,..c_-~--1

60f---- - - - - T - - -+----,,.-<'--------j

Estimation of benefits
According to the sweetpotato multiplication program, virus-free plants were
estimated to have reached 84% of the
sweetpotato area in Shandong Province by
1998. This estimate includes areas planted
to new and first and second generation
planting material. Among the 30 villages
surveyed, virus-free roots were first used
in 1995 and by 1998 had spread to 78%
of the sweetpotato area in the vi IIages
(Figure 2).

sor---- ----;"--T--+----------<
4or------ -+--,,_- + - - - - --- -----<
30 f - - --

--+----,;<---,,------1 ->- Spring crop

20
10

Summer crop

Total Crop

'-----"~~---~--~

1995

1996

1997

1998

Figure 2. Diffusion of virus-free sweetpotato seed in

the 30-village survey.
CIPProgramReport 1997-98

251

second generatio n roots it is estima ted to be

6.9 t/ha. Th ese figures are derived fro m the
v illage survey. Area planted to virus- fr ee
roots is deri ved from data on v irus-free
p lanting mater ial prod uct io n that was
provided by th e provincial seed program
and ass um es a seedin g rate of 750 kg/ha.
Th e benefit est im atio n also assumes that the
tota l area planted to sweetpota toes wi ll
rema in at about 533,000 ha in the future
and that th e diffusion of v irus-free roots
reac hes an up pe r limi t of 78% of the tota l
sweetpotato area. Both ass um pt ions are
probably co nserva ti ve.
Th e program is estimated to in crease
sweetpotato production in th e prov in ce by
3.965 million metric tons annu all y eq ui va lent to a va lu e of gross benefits of US$1 67
m i lli on/y r. Be nefits are assumed to rema in
at thi s leve l until the end of th e act ivity in
2020. The diffusion of v irus-free roots
betwee n 199 4 and 1998 resu lted in a 22 %
in crease in sweetpotato production in the
prov ince, equivale nt to a 2.6% increase in
globa l prod uctio n.

Costs of research, extension and

seed multiplication
Research and extens ion costs were
incurred early in the project and were soon
dwarfed by expe nses for th e multipli cation
of planting mate ri al. Research was the
principal acti vity fo r the fi rst four years of
th e program. As virus-free roots we re being
tested and provided to farmers, govern ment
subs id ies-espec iall y fo r techn ica l train in g,
construct ion, and materials-pla yed an
imp orta nt ro le in gettin g the multipli cat ion
system estab li shed and prov id in g fa rm ers
with re lative ly low-cost material. Once
farmers we re conv in ced of th e benefits of
v irus-free roots and adopted them, revenues
from sa les of virus-free roots soon supported most of the costs of the program. By
the eigh th yea r of the progra m, about 90%
of the an nu al US$19.6 million cost of th e
program was derived from seed sa les.
Th e initial subs idi es provid ed to the
program help ed kee p the price of improved
planting materia l low, and thu s promoted
2 52 Sweetpototo

its rap id d iffus ion. One factor that furthered

rapid d iffu sion was that virus-free roots
were relatively cheap compa red w ith the
ben efits of higher yiel d. Part ial budgeting
shows that adopt ion of first ge nerat ion
virus-free roots fo r the sp rin g sweetpotato
crop produced on average nearly 7 Yuan in
add itiona l va lu e of yie ld for each 1 Yuan of
added costs. A lth ough it may have been
poss ibl e to fund a larger share of the
program from root sa les, it wou Id have
re quired charg in g fa rm ers higher prices for
improved seed and co u ld have dampened
diffusion . After large-scale ado pti on has
occurred, it shou ld be possib le to recover
the initial subsidie s through a sma ll ,
temporary tax on virus-free roo t sa les. Such
a tax cou ld provide revenues fo r fund in g
the d eve lopme nt and diffusion of im proved
planting material in othe r regions.

Benefit-cost measures and

sensitivity analysis
Under the baseline assumpt ions descr ibed above, benefi t-cost analys is shows
that th e virus -free roo ts program in
Shandong Province had an internal rate of
return of 202%, and, ass umin g a 10%
disco unt rate, yielded a net present va lu e of
$550 mi ll ion. O nce the program was fully
establ ished (1998 and beyond), virus-free
roots provided net benefits of $145 mi l li on
per year.
Other sce na rios in Tab le 1 show the
sensitiv ity of the results to the estim ates of
yield ga in s and project costs. Even under
grossly co nservati ve assumptio ns, the vi ru sfree roots proj ect appears to have resulted
in impressive ly high returns.

Poverty and the distribution of

benefits to producers
Vi l lages located in the more industria liz ed pa rts of Shandong Provi nce, suc h as
those in Yantai and Weifang districts, are
noticeab ly better off than vi llages in other
parts of th e provin ce. Vill ages in the
mountainous ce ntral and so uth ern districts
(i ncludin g RiZhao, LinYi , and Jin zin g
distri cts) are significantly poorer. In the
poorer districts, sweetpotato area per

Table 1. Benefit-cost analysis of virus-free sweetpotato seed in Shandong Province, China.

Assumptions

Internal rate

Net present value

Annual net

of return

(million US$)

benefits at full
diffusion

(%)
l. Baseline assumptions
2. Adoption peaks at 90%
3. Costs of research, extension, and
seed multiplication doubled
4. Yield improvement estimate halved
5. Costs doubled and yield improvement
halved

(million US$)b

202
202

550

145

620

168

170

467

124

170

234

62

132

151

40

' Net present value is calculated assuming a real discount rate of l 0%.
b Figures in 1998 US dollars.

household is larger and sweetpotatoes

make up a larger component of household
income. Survey responses from the 12
villages in Yantai and W eifan g indi cated an
average total income for 1997 of 3,064
Yu an/cap ita (or 9,817 Yu an/ household), and
an average agricultural income of 1,798
Yu an/cap ita. In these villages, average area
planted to sweetpotatoes was 0.07 ha/
househo ld, and sweetpotatoes contributed
15% of agri cultural income. Among the 18
ot~er vi I Iages in the survey th at were
loca ted in poorer areas, average in come
was reported to be 2,091 Yuan/ca pita (or
7,3 16 Yu an/ household), with 1,394 Yuan/
cap ita from ag riculture. Sweetpotato area
per household was 0.14 ha, and
sweetpotatoes contributed 25% of ag ricultural in come. The virus-free roots diffused
eq ually well in poor villages as in relatively
ri ch villages and yield effects were not
significantly different between rich and
poor villages.

and Weifang area, improved sweetpotato

planting materi al in creased household
incomes by an average of 160 Yuan/
household/yr. In the poorer villages of the
hilly regions in the center and south of the
province, adoption of virus-free roots
increased household incomes by an
average of 265 Yu an/household/year
because of th e large r area planted to virusfree roots per household. In the rich er
areas, virus-free roots only increased
average hou sehold in come by around
1.6%/yr, and in the poorer region s the
increase in total hou se hold income from
virus-free roots was 3.6%/yr. The income
effects were thus progressive, with a larger
share going to households in relativ ely poor
regions. The effects on household income
do not appear to be large because the
benefits are so widely distributed among
sev.en million sma ll farmers of th e province,
each of which, on ave rage, plants about
0.10 ha to sweetpotatoes annually.

First generation virus-free roots gave a

net benefit of 3,500 Yuan/ ha for the spring
crop and abo ut 1,600 Yuan/h a for the
summer crop. Multiplying these figures by
th e average sweetpotato area planted to
virus-free roots suggests that in the Yantai

Summary and Conclusions

The rapid diffusion of virus-free sweetpotato
planting mate ri al in Shandong Provin ce,
reaching 80% of the province's small
growers in only 4 years, can be explained

CI P Piogiom Report 1997-98

253

by seve ral factors. Most im portan tl y, users

of the new ro ots saw yields in crease by 10
t/ ha, or 30%. Further, the tec hn ica l package was simple and required only one sma ll
change in the farmers production system:
the repl ace ment of th e sou rc e of planting
materi a I. Strong demand for sweetpotatoe s
fr om the food processing industry also
co ntr ibuted to rapid diffu sion by keeping
prices fro m falling in the face of increased
supply. This enab led farmers, including
late ado pters, to continue to capture th e
gai ns from techn ical change. Fin al ly,
govern ment subsidies fo r th e estab lishm ent
of the multiplication program made largesca le production possib le in a short time
and helped keep the price of improved
planting material low.
Shandong Prov ince represents only
abo ut 9% of the total area planted to
sweetpotatoes i n China , and it would seem
that v irus-free roots would have co nsid erable potential fo r increasing yield in other
provinces as we ll . In fact, vi ru s-free roots
programs are currently und er development
in al I of the major sweetpotato-producing
prov inces in the country. A stra ight-lin e
extrapo lation of the net va lu e of productivity in creases achieved in Shandong in 1998
(when 80% of the sweetpotato area was
estim ated to have been planted to viru s- free

254

Sweetpototo

roo ts) would impl y potential ben efits to all

of China of arou nd US$1.5 billion pe r yea r.
It is not yet known whether China 's
success with v iru s-free sweetpotato planting
materi al can be extended to other co untri es . Experiments with virus-free seed in
East Africa, for examp le, have not resu lted
in much yield gain. The reaso ns for this are
not yet we l I understood, but they may
im pl y that th e success of the mul tipli ca tion
program in Shandong Province mi ght not
be repl icable in some other important
sweetpotato-growi ng areas, such as those
fo und in Sub-S aharan Africa.
References

Salazar, Luis. F. 1996 . Potato viruses and

thei r con trol. International Potato Center,
Lima, Peru.
Zhang, Limin g. 1995. Progress of research
and appl icat ion of v iru s-free sweetpotato
seed in Shandong. Proceedin gs of the 1"
Chinese-Japanese Symposium on
Sweetpotato and Potato. Beijin g Agric ultural University Press, Beijing, China.
Zhan g, Liming. 1996. Propagation system
and production techniques of vi ru s-free
seed sweetpotato. Chinese Sweetpotato

7:67-71.

Biological and Selective Control of the Sweetpotato

Whitefly, Bemisia tabaci (Gennadius) (Hom.:
Aleyrodidae)
F. Cisneros and N. Mujica1

Sweetpotato (/pomoea batatas) whitefly

(SPWF), 8emisia tabaci (Gennadius) is
wide ly spread throughout the tropics and
subtropics. The pest infests many food
crops, ornamentals, and weeds. In temperate countri es, whitefly infestations often
occur in greenhouse crops.
For the most part, infestations have been
trivial. Then, in the 1970s and 1980s,
extraordina ry whitefly outbreaks occurred
in suc h diverse areas as Sudan, the Middle
East, and California and Florida in the
United States.
In most cases, whitefly outbreaks have
been associated with the intensive use of
insect icid es, greater tolerance of the pest
for insecticides and destruction of natural
enemies. In sectic ides have been reported to
stimul ate whitefly fertility (Cast le, 1998) .
There are cases, however, of outbreaks
before the appea rance of modern insecticid es.

Compared w ith SPWF, SLWF is becoming one of the most important crop pests
because of its higher rate of increase
(Bethke et al. , 1991 ), wider range of
thermotol erance (Salvucci et al., 1998),
efficient transmission of a group of
geminiviruses (Gilbertson et al., 1998), high
levels of resistance to insecticides (Byrne et
al., 1998), more honeydew production
(Byrne and Miller, 1990), and a wider range
of host plants (Byrne et al., 1990).

Occurrence of SPWF in Peru

Th e occurrence of SPWF in Peru was
overlooked until 1982 when it was reported
for the first time in a review of sweetpotato
insect pests (Cetraro an d Ortiz, 1982).
Rodri guez and Redolfi (1993) reported the
first evaluations of SPWF population
densities in cotton, string beans, and
sweetpotato. Pop ul atio n densities were low
(av 1.0 nymph/ leaf for Rimac Va ll ey and 3
nymph s/leaf for Canete Valley). According
to Gerling (19 85), whitefly infestations are
considered low when nymph/leaf is < 5.

A new whitefly species

In the 1990s, widespread whitefly
infestations occurred in the tropics and
subtropics, including Central and South
America and the Caribbean. Th is new
aggressive whitefly was initi all y thought to
be a new stra in of 8 . tabaci (strain B) . It
now has been described as a new spec ies,
8. argentifolii Bellows & Perring. It is
commonly called silverleaf whitefly (S LWF),
an apparent reference to the symptoms
produced in cucurbit plant leaves (Cohen et
al., 1992).

This situation chan ged dramatically in

1997 and part of 1998 with climatic
changes brought about by El Nino. H eavy
infestations were even higher in areas
w here in secticides we re applied. That
could be interpreted as a result of the
destru ction of the w hitefly's natural
enemies.
Thes e heavy infestation coincided with
reports of SLWF in the Canete Valley. High
infestations are seen on sweetpotato,
cucurbits, and cotton, leading to specula-

1 CIP, Lima , Peru.

CIPP10g10m Reporl 1997-98

255

tion that the pest ma y be replacing SPW F as

a sweetpotato pest.
Natural enemies of SPWF
Numerous predators, parasitoids, and
some entomopathogen ic fungi attack
whiteflies. The y are usually able to effectivel y regulate w hitefly populations. That
was so much the case that for years outbreaks were linked to mismana gement of
the agroecosystem, particularl y the destruction of natural enemies. General surveys in
sweetpotato fields conducted from 1987 to
1991 indicated 65 species of natural
enemies that prey upon or parasitize
sweetpotato pests. For decades, entomologists considered sweetpotato fie lds that had
not been spraye d with insecticides as
natural refuges for biological control agents.
Effects of nonselective compounds on
the upsurge of SPWF
More than 50 con ve ntional insecticides
are registered for use against Bemisia
whiteflies, in some cases for reducing v irus
transmission. All kinds of products are
invol ved including organophosphates,
carbamates , and pyrethroids. Most treatments are directed to wa rd controlling
adults , because immature life stages are
usuall y less susceptible and more difficult
to reach. Whiteflies at those stages are
immobile and located on the underside of
leaves.

Additional applications are required as

new adults emerge from undamaged
puparia. Nine to 15 applications may be
necessary to keep adult populations lo w.
Under these conditions w hitefl y readil y
develops resistance to insecticides.
In 1997, under the influence of high
temperatures brought on by El Nino, most
affected farmers applied insecticide 2-4
times, some up to 6 times, against the West
Indian sweetpotato weevil (Euscepes
postfasciatus), w h itefl ies, and w hite grubs.
As many as 30 commercial insectic ide
formulations we re used. One of the most
commonly used against the West Indian
sweetpotato weev il was carbofuran. Fields
2 56

Sweet potato

treated with carbofuran showed the highest

populations of whiteflies. This effect was
not only due to the destruction of natural
enemies, but to the trophobiotic effect of
the insecticide on the whitefly.
Effects of bioinsecticides and other
selective compounds on whiteflies
A search for alternative compounds led
to the use of rotenone, pyrethrum, a chitin
inhibitor, and entomopathogenic fungi
against the whitefly pest.
Plant-derived insecticides and mineral
oil. Plant-derived insecticides, particularly
rotenone and pyrethrins, are coming back
into use to control crop pests in response to
the need for safer products. Mineral oils act
as a suffocating agent for immature stages
of w hiteflies.
Selective compounds. Several chitin
inhibitors have been introduced in recent
yea rs (buprofezin, pyriproxyfen,
diafenthiuron, and imidacloprid) to control
Bemisia whi teflies on cotton and other
crops. Most of them have low toxicity to
mammals and to biological control agents.
Buprofezin shows low toxicity to Encarsia
formosa Gehan and Cales noacki Howard,
parasitoids of whiteflies (Me ndel et al.,
1994; Garrido et al., 1984). It affects
embryogenesis in adults and moulting in
immature stages.
Entomopathogens. Pathogens most
commonly recorded infecting Bemisia
w hiteflies in nature are Paecilomyces
fumosoroseus, Verticillium lecanii, and
Aschersonia sp. Occasional pathogens
in fe cting Bemisia w h itefl ies, also under
natural conditions, are Beauveria bassiana
and Entomophthora sp.

There is not much information on the

occurrence of epizootic diseases caused by
pathogenic fungi. But under high humidity,
and if sufficient numbers of pathogenic
spores are present, disease should develop.
Experiences with the greenhouse whitefly,
Trialeurodes vaporariorum (Westwood), and
the citrus woolly w hitefly, Aleurothrixus

flo ccosus, have shown th at path oge nic

fungi ca n be effect ive (Fran se n, 199 0; va n
der Shaaf et al ., 1991 ).
Project goal
A st rategy has been de veloped for an
integ rated pest management (IPM) program
for th e management of Bemisia wh itefl ies in
sweetpotato . The activiti es reported here
we re aim ed at eva lu ati ng th e occurrence of
natura l enemi es and th e effec ti ve ness of
se lect ive co mpounds, includ ing
b io in sect ic id es , plant-d eri ved prod ucts, and
a chitin inhibi to r to id entify an altern ative
to chem ica l in sect icid e use.

Th ese studi es were und ertaken befo re

th e in cid ence of SLWF in th e Canete Valley.
It is hoped that the meth ods reported here
w ill have simil ar efficacy in th e management of SLWF.
Materials and methods
Natural enemies of the whitefly
CIP, in col laboration w ith th e Department of Entomo lo gy of th e Agrari an
University, La Molina, Peru, has studi ed
sweetpotato pests and th eir natura l enemies
alo ng the ce ntral coast of Peru si nce 1986.
In 1997 and 1998, the researc h em pha sis
was on stud yi ng th e na tural enem ies of
w hi tefly.
Soil-inhabiting predators. Pitfa ll traps,
1-L pl asti c cups containing 500-cc formaldehyd e (5% conce ntrati on), were placed in
sweetpotato field s. Trap s were co ll ected
week ly and the number of predato rs
reco rd ed.
Foliage-inhabiting predators. Predators
we re co ll ected directly from th e ca nopy of
sweetpotato plants and their presence and
numbers reco rded . Samples (25 pl ants/ field)
were taken week ly.
Parasitoids. Whitef ly-infested leaves of
sweetpotato were gathered at 15-day
interv als during th e sweetpotato-g row in g
season. Co ll ected leaves were pl aced in

plastic conta in ers and cove red w ith a fin e

mesh unti I th e emergence of paras itoids.
Selective compounds
Several compo und s from different gro up s
(pl ant-derived in sect icides, entomopath ogens, a chitin inhibi to r, and min eral
oil) were tested fo r their effect iveness
again st whitefly (Table 1).
Preliminary trial. A preliminary stud y
was conducted in a co m merc ial
sweetpotato fiel d, 30 d after plantin g, with
sweetpotato cu Iti var I NA- 1 00, at San Lu is,
Can ete, Peru, May 1998. Five compounds
we re eva lu ated (rotenone, buprofez in,

Verticillium leca nii, Entomophthora

virulenta, and detergent). Carbofuran was
used as a farmers' check. The fi eld was set
up in a randomiz ed comp lete block des ign
(RCBD), w ith seve n treatments and four
repli cates, in p lots 30 ro ws (Im betwee n
ro ws) w ide and 50 m long (1,500 m 2).
Dosages are prese nted in Tabl e 1. Th e pl ots
were sprayed on 27 May wi th a motori zed,
backpack sprayer at a rate of 300 L/ ha.
Trials with bioinsecticides and other
selective compounds. Four tri als were
condu cted to eva luate the effective ness of
th e fi ve com pound s li sted in the pre l iminary ex periment. Th e compo un ds we re
teste d ind ividu al ly and in m ixtures. Tri als
were ca rri ed out in a sweetpotato fie ld 60 d
after plantin g w ith sweetpotato cu lti var
INA-1 00 in Sa n Lui s, Canete, Peru, MayJul y 1998.
Trial 1 - Plant-derived insecticides.
Pl ant-d eri ved insecticides sele cted were
three comme rcial formulations of rotenone,
an emulsifiable co nce ntrate (EC), a wettab le
powder (WP), and a co mmercial mi xture of
rote none and pyrethri n (EC).
Trial 2 - Chitin inhibitor. Th e chitin
inhibitor, buprofez in , was tested in a
mi xture w ith p lant-d eri ved in sect icid es .
Trial 3 - Entomopathogens. Three fun gi
pathogens were se lected and tested in the

CIP Prng1om Repo1t 1997-98

257

Table 1. Treatments used in 5 different trials with selective compounds (entomopathogens, plant-derived
insecticides, mineral oil , and a chitin inhibitor) against sweetpotato whitefly. Checks included are
absolute checks and insecticide-treated checks. Canete Valley, Peru , 1998.
Technical name

Trade name'

Concentration/
dosage

Preliminary
experiment

Tl

Buprofezin

Aquitin

0.1%

T2
T3
T4
TS

Rotenone

Extracta 50 WP
Vertisol
Vektar
Ace
Carbo for 75 WP

0.1 %
l xl 09 conidias/bottle
l xl 09 conidias/bottle
0.2%

Verticillium leconii
Entomophthora virulento

Tb
Tl

Detergent
Treated check: corbofuran
Absolute check

Tl
T2
T3
T4

Rotenone + buprofezin
(Rotenone and pyrethrum) + buprofezin
Rotenone + buprofezin
Check

Rhotenox l 0 EC + Aquitin
Rhotenox-SP + Aquitin
Extracto 50 WP + Aquitin

0.4% + 0.1 %
0.4% + 0.1 %
0.1 % + 0.1%

Experiment 2 Tl
T2
T3
T4

Rotenone
Rotenone and pyrethrum
Roten one
Check

Rhotenox l 0 EC
Rhotenox-SP
Extracto 50 WP

0.4%
0.4%
0.1%

Experiment 3 Tl
T2
T3
T4
TS

Oil mineral
Oil mineral + rotenone
Treated check: endosulfon
Treated check: corbofuran
Check

Triona -5
Triono-5 + Extracto 50 WP
Thiodan 35 EC
Carbofor 75 WP

1%
1%+ 0.1 %
0.15%
0.2%

Experi men! 4 Tl
T2
T3
T4

Verticillium leconii
Entomophthora virulento
Beouverio bossiana

Vertisol
Vektor
Bouveril

l xl 09 conidias/bottle
l xl 09 conidias/bottle
Sx l 011 conidias/bottle

Experiment l

Check

' WP= wettable powder; ED =e mulsifioble concentrate, SP = supension powder.

fi eld : Verticillium /eca nii, Entomophthora

virulenta, and Beau veria bassiana. Formulations of the fungi (deh ydrated conidia) w ere
commerciall y-produ ced. Dosages used
were those recomm end ed by th e manufacturer.

Trial 4 - Other products. Selecti ve

produ cts we re an emulsifiable min eral oil
and min eral oil mixed with rotenon e. Two
nonselective products were used as farmers' ch ec k (e ndosulfan and carbofuran).
Eac h tri al w as an RCBD w ith four
repli ca tes. Experim ental plots w ere 14 rows

2 58

Sweet pololo

wide and 25 m long (350 m 2) . Pl ots were

treated twice at 14-d interva ls (24 Jun e and
10 Jul y) using a motori zed, backpack and
sprayed at a rate of 400 L/ ha. Sprayers used
spec ial nozz les to ensure wettin g of the
underside of leaves whe re wh iteflies li ve.
Sampling methods
At week ly inte rva ls, 20 leaves/p lot were
co ll ected from among the most heav il y
infested leaves on the bottom or midd le
part of the p lant. N ymp hs and pupae we re
counted under a stereoscope in the laboratory. Morta lity of nymph s and pupae was
determ in ed w hen they were dry. Popu lat ion
dens ity was exp ressed as number of
imm atu re stages/c m 2 du e to th e differences
in lea f size .
Estim ates of the w hitefl y populations
were done 0, 7, and 14 d after app li cation
in the pre limin ary tri al. In the four other
tria ls, estimates of the w hi tefly popu lation
were done 0, 7, and 14 d after the first
app li cat ion; and 7, 14, and 21 d afte r the
seco nd application. Adu lt popu lations were

not eva luated because of the high mi gratory

movement of w hitefli es from hi ghly infested
surround ing fie lds.
Statistical analysis
Data were transformed (squ are root)
before statist ica l analysis and w ere analyzed usin g 1-way ANOVA . Means were
separated by LSD (P = 0.05) . To separate
the effects of pestic ide treatment from those
caused by natura l facto rs, the percent of
population reduction was corrected using a
mod ified Abbott' s fo rmul a (F lemi ng and
Retnakaran, 1985).
Results and Discussion
Natural enemies of the whitefly
Predators. Predators collected in th e
ca nopy of sweetpotato plants (Tabl e 2) were
as varied as the group s reported worldwide
by Nord lund and Legaspi (1995), except for
mites, wh ich were not samp led . The most
abun dant predators we re sp iders of two
genera, Anyphaena and Pardosa, and th e
fl y, Condylostylus similis. The predatory

Table 2. Ocurrence of canopy inhabitant predators' in sweetpotato infested with white fly Bemisia tabaci in the

Canete valley, Peru.

Spiders

Chrysoper/o externo
Orius sp.
Hyo/aides sp.
Nobis punctipennis
Rhinocloo sp.
Aknisus sp.
Scymmus sp.
Coccinelids

Condylostylus similis

1995

1994

Predators

No

101
3

57.4
l.7

5
18
4

1996
%

No

15

45.5
7.3

113
9
7

2.8
10.2
2.3

8
24
9

3.9
l l.7
9.4

18
26

10.2
14.8

18
30

8.8
14.6

176

0.6
100

8
205

3.9
100.0

38.4
3.1
2.1
1.4
2.4
2.4
0.3
4.3
10.l
32.l
1.3
3.2
100.0

No
93

Syrphidae

Geocoris sp.
Total
0

3
13
30

93
8
8
292

Number of predators in 1.25 m2

CIPProgram Report 1997-98

259

acti vi ty of nonw ebbin g spiders on w hitefli es

is not we ll know n, but A nyphaena spp.
efficientl y trapped w hitefl y adults in their
webs. Th e coccinellids (Cycloneda
sanguin ea, Scymnus sp. , Hippodam ia
con vergens, Eriopis connexa, and
Co leom egilla maculata) were fairly abundant and preyed upon w hitefl y immature
stages . Other, less abundant species, that
also preyed upon the imm ature stages we re
Chrysoper/a spp. Hya liodes sp ., Rhinacloa
spp. , Nabis punctipe nnis, and C eocoris

punc tipes.
Th e most abundant so il inh ab iting
predators w ere a carabid beetl e
(Pterostichus sp.) and one species of
ea rw ig (Labiduria riparia) (Table 3).
Parasitoids. Three genera of paras itoids
we re recorded in th e Canete Va ll ey (Table
4). As in other parts of South Am eri ca and

th e wo rld , Enca rsia spec ies were the most

abundant, w ith E. portieri being th e most
frequ entl y found . Parasitism used to be the
most important mortality factor of w hitefli es
in th e Canete Va ll ey. In 1987, an ave rage of
41 % paras iti sm was recorded , ran gin g from
4% in fi eld s with insecticide treatm ents to
70% in untreated fields (Table 5).
Paras itism has been reduced w ith
increased in secti c ide use. At the sa me tim e,
infestati ons of w hi tefl y have steadil y
in creased to present levels. Don et and
Ve rgara (199 4) re corded the destru ction of
sp ider populations in sweetpotato fi eld s
treated w ith in sect ic ides against th e W est
Indi an sweetpotato weevil. Velapatino and
Sanchez (1996) recorded low popul ations
of pred ators in Canete fields commonly
treated with in sect ic ides.

Table 3. Soil inhabitant predators captured by pitfall traps in sweetpotato fields in Canete Valley, Peru, 1998.

Predators /trap/season (No)

Pterostichus sp.

Labiduria riparia

(Coleoptero: Carobidae)

(Dermaptero: Labiduridae)

Year
1994
1996
1998

94.8
44.9
350.7

306.4
100.0
26.0

Table 4. Parasitoids' of sweetpotato whitefly nymphs, central coast of Peru.

Parasitoids

Locality
La Molina

Platygasteridae
Amitus sp.
Encarsia sp. prob. bicalor De Sontis
Encarsia porteri (Mercet)
Encarsia (Prospaltel/a) sp. (A)
Encarsia sp. (B)
Eretmocerus sp.
0

260

VS= very scarce; S= scarce; A=abundant.

(A) and (B) non-identified species.

Sweelpololo

Canete

vs

vs

vs
vs

vs
vs

Table 5. Variation in the level of parasitism of the sweetpotato whitefly Bemisia tabaci, on sweetpotato during
the last 11 years, Canete Valley, Peru, 1987-1988.

Year

Parasitism (%)

Nymph/leaf
Minimum

1987
1989

1.9
2.4

1994
1995
1996

20.5
39.8
46.4

1998

853.5

Selective compounds
Preliminary trial. Buprofez in was very
effecti ve aga in st nymph s, redu cin g the
w hitefly popu lat ion of 32. 1 nymph/c m 2 to
0.9 nymph/c m 2 in 14 d (98% mortality). No
signi f ica nt differences in nymph number
were observed between other trea tments
(rotenone, entomopathogens, and detergent) and the chec k, but a dec rease in the
adult popul ation in the field was observed.
Plots treated w ith carbofuran had an
increase in whitefly popul at ions from 62.7
nymph s/c m 2 to 102.9 nymph s/c m 2 in 14 d.
In the tri als discussed below, rotenon e and
buprofez in were mixed for bette r wh itefly
co ntro l.
Identifica tion of nymph s affected by the
different trea tments was one of the main
difficulti es faced durin g th e ex periments.
Becau se nymphs and pupae are immobil e
and fixed to th e undersid e of th e lea f
surface, movement could not be used as a
sign of li fe. In addition, most products
(e ntomopathogens, rotenon e, min eral o il ,
and buprofez in) are intrin sically slow
actin g. Mortality does not occur as fast as
w ith co nventional in sect ic ides. It was also
difficult to sp ray the und erside of th e leaf
surface eve n w ith the spec ial nozz les used.

Trial 1 - Plant-derived insecticides. No

significant d ifferences (P = 0.05) in nymph
number w ere observed unti I 2 wk after the

Maximum

Average

4.0
5.0

70.0
27.0

41.2
28.0

0.0
0.0

(few, not quantified)

2.6
0.0

0.3
0.0

(few, not quantified)

second appli ca tion (Tab le 6). Roten one +

pyrethrin resulted in an avera ge of 1.2
nymph/c m 2 compared w ith 5.2 nymph/c m 2
in th e chec k, represe nting 73 % mortality
(Tabl e 6). Th ese formu lat ions ca n be used
under moderate leve ls of w hitefl y in festation .

Trial 2 - Chitin inhibitor. Buprofez in

mi xed w ith other compounds (rotenone or
pyrethrin) did not in crease control of
imm ature stages (Tabl e 6). The effect of
buprofez in was ev id ent (more th an 90%
mortality) 7 dafter th e second appli ca tion
w hen the av was 0.6-0.8 nymph/c m 2
compared with 13.3 nymphs/cm 2 in the
check. Th e res idu al effect of buprofez in in
the field resulted in nymph mortality of 7688.5% over 2 1 d.
Trial 3 - Entomopathogens. To sepa rate
hea lth y imm ature stages from th ose in fected
by the entomopathogens was diffi cult,
parti cul arly durin g th e first 10 d afte r th e
first app li ca tion when no externa l changes
in nymphs and pupae we re ev id ent. After
th e second ap plication, only 8. bassiana
was effective, w ith 69% of nymph popu lation dead (dried) after 2 1 d (Tab le 6). We
were un ab le, howeve r, to detect th e fungus
before dehydration.
Trial 4 - Other products. Min eral o il
acts as a suffocating agent for immature
CIP ProgramReport 1997-98

261

Table 6. Leaf average nymph population per square cm at the indicated days after application per treatment,
Canete, Peru, 1998.

Days after application

Before

treatment

Treatment
Exp. l

Rotenone
Rotenone and pyrethrum
Roten one
Check

Exp. 2 Rotenone

First application

+ buprofezin

Second application

14

3.7
12.3

a
a
a

5.8
5.l
7.5
7.3

5
11.2
11.7
11.7

21.9 a

7.5

4.9

0.7

36.5 a
33.4 a
35.7 a

6.4
7.5
8.5

5.7
7.2

9.l

0.6
0.8
13.3

34.9 a
30.6 a
34.6 a
20.9 a

13.9
11.7

13.3

14.3

8.9
8.2
11.8
10.2

16.2 a

6.1

4.6

21.3 a
24.2 a
15. l a
20.9 a

5.6
9.9
6.6

6.4
ob

14.3

15.5
23.6
34.6
27.6

0
0
0

2.7

9.3

14

2.4
l.2
4.7

ab

5.2

21
ab
b
ab

2.03 a
1.2 a
3.2 a
2.73 0

0.3

0.26

0.1
0.4
3.5

0.5
0.83
3.7

0.85
0.83

(Rotenone and pyrethrum)

+ buprofezin
Rotenone + buprofezin
Check
Exp. 3

Verticillium /econ ii
Entomophthora virulenta
Beauveria bassiana
Check

Exp. 4 Oil mineral

Oil mineral

+ rotenone

Treated check: endosulfon

Treated check: corbofuron
Check

b
b

6.6

2.8
2.3
1.8
2.1

4.4

6.9
5.7
0

6.7
9.7
9.8

2.2
2.3
2
7.6
4.5

be

ab

3.8
3.1
2.7
7.8
4.2

b
ab
ab

0
0

0.4

0.78

3.48 ob
2.6 ob

2.45 ob

ab

5.35 a
3.6 ab

Within a column, means followed by a common letter do not differ significantly at P=0.05 by DMRT.

stages of whitefly. Its effect was noticeable

7 dafter the first application. The number
of nymphs was reduced by 45% compared
to the check (6 .1 nymph/cm 2 compared to
14.3 in the check) (Table 6). No residual
effect w as observed. When rotenone was
added to the mi nera I oi I treatment, no
significant increase in the effect (P = 0.05 )
was observed throughout the evaluation
dates (Table 6).
Mineral oil treatment was as good as the
farmers ' check (endosulfan) or better in the
cases w here farmers used carbofuran. Plots
treated with carbofu ran showed the highest

262

Sweetpototo

whitefly population when , 14 dafter the

second application , the population increased 157% compared w ith the check.
This illustrates the trophobiotic effect
of carbofuran on the whitefly nymph
population.

Conclusions
Al I treatments- entomopathogens, rotenone, mineral oil, and buprofezin-contributed to controlling whitefl y by killing
nymphs. Mortality probably was higher
than the level recorded due to their slow
action. The compounds tested are consid-

ered fairly safe for natural enem ies (except

ca rbofur an and endos ul fa n, used as treated
chec ks).
Unfortunate ly, under the experimental
co ndition s, it was not poss ibl e to ver ify th e
reaction of natural enemi es . Th e area had
bee n und er such intensive chemical co ntro l
fo r severa l seas ons that no paras itoid s were
recove red. Und er o ther co nditi o ns we
mi ght have expected th e action of natural
enemi es to co mpl ement trea tm ents.
Th e w hitefl y outbreak in Ca1iete Va lle y,
which coi ncid ed with th e identificati on of
SLWF, a new spec ies of w hitefl y displ ac in g
SPWF in Canete, is a rep eat of w hat has
happe ned in other p art s of the wo rld .
Co nditi ons were similar, in parti cul ar the
intensi ve us e of in secticid es. Th e effects
were also simil ar: hi gher populations, w ider
range of crops affected, more seve re
damage, and res istance to insecticides.
The use of rotenone, chitin inhibitors,
mineral o ils, and entomop atho ge nic fungi
we re fairly effecti ve in our ex per iments.
Th ese products can be integ rated with oth er
IPM components to man age SPWF. A lso,
th ese treatments are ex pected to all ow the
recove ry of paras ito ids and preda tors
dec im ated by the use of co nve ntion al
in secti cid es.
Acknowledgment
We are gratefu I to th e Depa rtm ent of
Entom ology, Ag rari an University, La
M olina, Peru, for id entifyin g captured
preda tors and parasitoids.
Literature Cited
Bethke, J.A., T.D. Paine, and G.S. Nuessly.
1991. Comparative bi o logy,
morph ometri cs, and development of two
populations of Bemisia tabaci
(Hom optera: A leyrodid ae) on co tton and
poinsettia. Ann. Ent. Soc. America

84: 40 7-411 .
Byrne, D.N. and W.B. Miller. 1990. Ca rb ohydrate and amino acid comp os iti on of

phl oe m sap and hon eydew produced by

Bemisia tabaci. J. In sect Ph ys. 36:433439.
Byrn e, D.N., T.S. Bellows, Jr. , and M .P.
Parrella. 199 0. Whiteflies in agricultural
systems. In: Gerlin g, D. (ed .). Whiteflies:
Th eir bionomics, pest status and manageme nt. Intercept Ltd ., Andover, H ants,
UK. p. 227-261.
Byrne, F.J., M. Cahill , I. Denholm, and A.L.
Devons hire. 1998. Understanding
in sect ic id e resistance in Bem isia.
Sil ve rfeaf W hitefly. National Researc h,
Act ion, and Technology Transfe r Plan,
1997-2001: First annu al review of the
second 5-Year Plan. Proc. of the Ag. Res.
Serv. U nited States Department of
Agriculture, Wa shin gton, D. C., U SA .
p. 3.
Castle, S.J. 1998. Primary pest or syntheti ca ll y indu ced ? The ro le of in secti c id es
and other factors in the pest statu s of
Bemisia tabaci. Silverfeaf W hitefl y.
Nati onal Research, Act ion, and Tec hnology Transfe r Plan , 1997-200 1: First
annu al revi ew of the second 5-Year Plan .
Proc. of th e Ag . Res. Se rv. United States
Department of Agriculture, Wa shin gton ,
D.C., USA.
Cetraro, L. and M. Ort iz . 1982. Ocurrencia
estac ional de in sectos de camote
(lpomoea batatas), en la costa central del
Peru. Rev. Per. Ent. 25(1 ): 17-3 2.
Cohen, S., J.E . Duffu s, and H. Y. Liu . 1992.
A new Bem isia tabaci biotype in the
so uth weste rn United State and its ro le in
sil ver lea f of squ as h and tran smissi on of
lettu ce infectiou s ye llow v iru s. Ph ytopathology 82:86-90.
Donet, C.J. and C. Verga ra. 199 4.
Fluctu ac i6n pobl ac ion al de Bemisia
tabaci Gennadius (Homoptera:
A leyrodid ae) y de sus co ntroladores
biol6gicos en tres va ri edades
co mercial es de camote en el va ll e de
Can ete. ln fo rm e de ln ves ti gaci 6n
Co nve ni o UNA LM -CIP. 39 p.
Fleming, R. an d A. Retnakaran. 1985.
Evaluating si ngle treatment data usi ng
Abb ott's formula wi th refer ence to
insecticides. J. Econ. Entomol . 78:1 179 1181.
CIPProgrnm Report 1997-98

263

Fransen, J.J. 1990. N atura l enemies of

whitefl ies. In: Gerling, D . (ed.). W hiteflies : Their bionomics, pest status and
management. Intercept Ltd ., Andover,
Hants, UK. p.187-210 .
Garrido. A., F. Beitia, and P. Gruenholz.
1984. Effects of PP618 on inmatures
stages of Encarsia formosa and Cales
noacki (Hymenoptera:Aphlel inidae).
Proceedings of the British Crop Protection Co nference - Pests and Diseases. p.
305-3 10.
Gerling, D . 1985. Parasitoids attacki ng
Bemisia tabaci (Hom.: Aleyrodidae) in
eastern Africa. Entom ophaga 30: 163165.
Gilbertson, R.L., M.R. Sudarshana, H.
Wang, Y. Hou , R. Salati, E.G. Ramirez,
and W.J. Lucas. 1998. A n update of the
status of white fl y-tran sm itted
geminivi ru s: The good , the bad and the
ugly. Silverfeaf Whitefly. Nationa l
Research , Action , and Technology
Transfer Plan, 199 7-2001: Fi rst annua l
review of the second 5-Year Plan . Proc.
of the Ag. Res. Serv. United States
Department of Agricu Itu re, Was h ington,
D.C., USA. p. 1-3.
Mendel, Z., D. Blumberg, and I. lshaaya.
1994. Effect of insect growth regulators
on natural enemies of scale insects
(Hom.:Cocc idae). Entomophaga 39:199209 .

264

Sweetpototo

Nord lund, D.A. and J.C. Legzspi. 1995.

W hitefly predators and the 'r potencial for
use in biological co ntrol. In: Gerling, D.
and R.T. Mayer (eds.). Bemisia: Tax onomy, biology, damage, control and
management. Intercept Ltd., Andover,
Hants, UK. p. 499-513.
Rodriguez , C. and I. Redolfi. 1993. Bemisia
tabaci (Homoptera:A leyrod id ae) y sus
parasitoides en la costa central peruana.
Rev. Per. Ent. 35:77-81.
Salvucci , E.M. , G. Wolfe, and D .L. Hendrix .
1998. Polyol synthes is as a mechanism
for thermotolerance in Bemisia.
Silverfeaf W hitefl y. National Research,
Action, and Technology Transfer Plan,
1997-2001: First annual re v iew of the
second 5-Year Plan. Proc. of the Ag. Res.
Serv. United States Department of
Ag ri cu lture, Wash ington , D.C., USA. p. 1.
Va n der Shaaf, D. A., W.J. Ra vensberg , and
M. Malais. 199 1. Verticillium lecanii as
a microbial in secticide against w hitefly.
IOBC/W PRS Bulletin 14:120- 12 3.
Velapat in o, J. and G. Sanchez. 1996.
Fluctuaci6n poblacional de Bemisia
tabaci Gennadius (Homoptera :
A leyrodid ae) en dos variedades
co mercia les de camote en el valle de
Canete. lnform e de lnvestigaci6n
Convenio UNALM-CIP. 30 p.

An Efficient System of Embryogenic Suspension

Cultures and Plant Regeneration in Sweetpotato
Q.C. Liu1, H. Zhai1, D .H. Lu1, Y. Wang 2 , and D.P. Zhang'

Genetic transfo rmati on offers great potenti al

for improving the di sease or pest res istance
and nutrition quality of sweetpotato
[fp o moea batatas (L.) Lam.]. On-going
researc h in these areas includ es developing
transgenic sweetpotato w ith res ista nce to
sweetpotato v irus di seases (Beachy et al.,
1992; Murata et al., 1998), sweetpotato
weev il and othe r in sect pests (Newe ll et al.,
1995; Zhang et al., 1997), and enhan ced
protein content in th e storage root and
leaves (Prakash et al., 1998). In spite of the
va ri ous successful cases in sweetpotato
transformation, the lack of an efficie nt
system for rege nerati on has bee n a bottl eneck for th e appl ication of transge nic
technology in sweetp otato. The regenerati on frequency in sweetpotato is ofte n
ge noty pe dependent, rang in g from 0 to
85% in tested culti vars (Gosukonda et al. ,
1995, Zhang et al., 1997). So far onl y a
sma ll numb er of genotypes have been
regenerated, based on the inform ati on
gathered from the related publi cat ion s.

Th e devel opme nt of a system of em bryogenic suspensio n cultures has been a focal

point for imp rov in g the rege neratio n
freque ncy in sweetpo ta to. Chee and
Cant liffe (1988, 1989), Chee et al. (19 90),
and Bieni ek et al. (1995) succeeded in
establishing emb ryo gen ic susp ens ion
cultures of sweetpotato cu ltivar White Star.
Liu et al. (1997) investi gated embryogenic
suspension cu ltures of a Japan ese culti va r
Kokei No.14 and a Chin ese cultivar
Xindazi. An im proved frequency of plant
regene ration from th e embryoge ni c suspension cul tures was obtained.
Th e objective of thi s research is to
develop an effic ient system of embryoge ni c
suspension cu ltures fo r a w ide range of
sweetpotato genotypes, w hich w ill eve ntually lead to a universal rege neration
protoco l for sweetpotato.

Materials and Methods

Plant materials

In sweetpotato, shoot api ces of many

genotypes ca n produ ce embryoge nic ca llu s
w ith so matic emb ryos on th e medium
supplemented with 2,4-D. Somatic embryos
can also develop in to w ho le plants on the
basal med ium (Liu and Ca ntliffe, 1984;
Jarret et al., 1984; Liu et al., 1992, 1993,
1997; Tan et al. , 199 3; Desamero et al. ,
199 4; Otani and Shim ada, 1996). However,
in most cases the frequencies of so mati c
embryogen es is were low.

1 Departmen t of Plan t Genetics and Breeding, Chi na

Agri cultu ra l University, Beijing 100094, China.
2 CIP, Beij ing, China.
3 CIP, Lim a, Peru .

Ei ght culti vars of sweetpotato, Gaozi

No.1, Koga nesenga n, Kokei No.14,
Li zixiang, Nongdahong, Tamayutaka,
Xindazi, and Xushu 18 we re used in this
study. They are important culti vars in China
or Japa n. Th e sto rage roots we re grown in
pots in a greenhouse. Th e youn g sp routs
obtain ed were used for the isol at ion of
shoot apices.

Induction of embryogenic callus

Young shoots abo ut 30 mm-l ong we re
excised, was hed w ith tap water, and
sterilized wi th 70% ethanol for 20 sec and
2% sod ium hypoc hlorite so lution for S min.
Th e ste rili zed materi als we re fully rinsed

CIPProgram Report 1997-98

265

with steri le distilled water. Shoot apices of

about 0.5 mm in length were isolated w ith
the aid of a dissecting microscope and
cu ltured on MS medium supplemented with
2.0 mg/ l 2,4-D, 3.0% sucrose (w/v) , and
0.8% (w/v) agar, pH 5.8, at 28( in the
dark. Th e cultures were exam in ed periodical ly under a dissecting microscope.
Initiation and maintenance of
embryogenic suspension cultures
Embryogenic calli we re obtained from
shoot apices of 8 cultivars. Six- to eightweek-o ld embryogenic cal li were used for
the initi at ion of embryogenic suspension
cul tu res. Embryogen ic ca l Ii were crushed
into cel l-aggregates and free ce ll s through a
stainless steel microsi eve with mesh size of
0.46 mm. The cells were then placed into a
100 ml Erlenmeyer flask contain in g 20 ml
liquid MS medium w ith 2.0 mg/ l 2,4-D and
3% sucrose (p H=5.8). The cultures were
in cubated on a reciprocal shaker (100
strokes/ min) at 28( und er 13 h of coo lw hite fluo resce nt light at 500 lux. Embryogen ic suspe nsion cultures were maintained
by subcu lture at a 2-week interval. At each
subculture, cell-aggregates over 0.5 mm in
size were crushed throu gh a stain less stee l
microsieve w ith mesh size of 0.46 mm and
then transferred to the fresh medium.
Regeneration of plants from
embryogenic suspension cultures
Twenty weeks after the initiation protoco l and at 4-week intevals cell-aggregates
about 1.0 mm in size were transferred to
solid MS medium supplemented with 2,4-D
(2.0 mg/U. This condition is suitable for the
proliferation of cell-aggregates and the
formation of somatic embryos at 28( in
the dark. Four to six weeks after transfer, the
emb ryoge nic calli obtained w ith somat ic
embryos were further cultured on MS
medium supp lemen ted with 1.0 mg/ l ABA
for 4 to 5 weeks to induce the germ in ation
of somatic embryos at 28( under 13 h of
cool -wh ite fluorescent light at 3000 lu x.
Plantlets from somatic embryos developed
into w hole plants on the basal medium.

266

Sweetpototo

Results
Formation of embryogenic callus
The in cubated shoot ap ice s responded
actively to the cu lture conditions. After 3 to
4 days, most of them started to produce
wh ite, friable ca llu s. This callus was nonembryogenic and grew rapidl y. A few of
shoot apices directly formed embryogenic
callus. This type of embryoge ni c cal lu s
grew mu ch slower than the non-em bryogenic cal lu s. Four to five weeks after
in cubation , emb ryogen ic cal Ii were formed
on the surface of the non-embryogenic cal Ii
(F igure 1 ). Th ese two kinds of embryogen ic
cal Ii (the one directly formed from the shoot
apices and the one fo rmed on the surface of
the non-embryogenic cal Ii ) we re easi ly
distinguishable from the non-embryogenic
cal Ii because of their bright-yellow and
compact appearance. Both embryogenic
calli were used to initiate embryogenic
suspension cu ltures . A ll of the eight tested
cu Iti vars successfu 1 ly formed an embryogenic ca l lus, alth ough there is variation in
the frequency of embryogenic callus
formation (Ta ble 1 ).
Establishment of embryogenic
suspension cultures
Using embryogenic cal Ii derived from
shoot apices of eight cultivars, we have
deve loped a system of embryogen ic
suspension cu ltures, w hich has hi gh
potential for regeneration. When maintain ed in liquid MS me dium containing 2.0mg/ l 2,4-D, the cel ls proliferated rapidly
and dispersed we ll (Figure 2). The embryogenic suspension cu ltures consisted of
bright-yellow and com pact embryogenic
cell-aggregates and free cells. After 20
weeks of initiation, approximate ly 20,000
embryogenic ce ll-aggregates about 1.0 mm
in size were obtained from a single embryogenic callus derived from a shoot apex in
all eight cultivars.
Regeneration of plants from
embryogenic suspension cultures
After 20 weeks of the initiation protocol,
the re generation ability of embryogenic

Table 1. Formation of embryogenic callus from shoot apices in eight sweetpotato cultivars.

Cultivar

Origin

Shoot apices

Shoot apices forming

embryogenic callus

Incubated
Gaozi No. l
Koganesengan
Kakei No.14
Lizixiang
Nongdahang
Tamayutaka
Xindazi
Xushu 18

China
Japan
Japan
China
China
Japan
China
China

suspension cultures was tested eve ry 4

weeks. By the transfer of eel I-aggregates
(about 1.0 mm in size) to solid MS medium
supplemented with 2.0 mg!L 2,4-D, the
ce ll-agg regates proliferated rapidl y into
typi ca l embryogenic ca lli with large
numbers of somatic embryos (Figure 3).
Wh en th e obtained embryogeni c ca lli with
somatic embryos were further cu ltured on
MS medium supplem ented with 1.0 mg!L
ABA, somatic embryos germin ated and
formed plantlets (Fi gure 4).
The reg eneration freq uency peaked at
about 24 weeks after initiation when 100%
of the cell-aggregates regenerated into
pl ants in all eight cultivars. After that, the
regeneration frequency gradually decreased, as some of th e ee l I-aggregates
beca me non-embryogeni c. However, th ere
is cultivar difference in this rega rd . For
examp le, Xindazi still maintained a frequ ency of up to 100% as lon g as 36 weeks
after initiation.

Discussion
The structure, co lor, and pattern of the
embryogenic ca lli formation from shoot
apices of the eight culti va rs were very
similar to th ose reported by Liu et al. (1992
and 1993). The formed embryogenic calli
were used to initiate embryogen ic suspension cultures. The embryogenic suspen sio n

No.

No.

57
27

16
5
98
61
29
14
5
7

28.1
18.5
85.2
68.5
51.8
26.4
16.7
21.9

115

89
56
53
30
32

cultures proliferated rapidly and dispersed

well in MS m ed ium containing 2.0 mg!L
2,4-D. By using the suspension culture for
20 weeks, approximately 20,000 embryogenic ce ll-aggregates about 1.0 mm in size
can be produ ced per embryogeni c ca llu s
derived from shoot apices. The suspension
cultures can maintain their embryogenic
abi I ity for at least 24 weeks.
The long-term maintenance of embryogenic suspension cultures can be ac hieved
by the re-culti va tion of embryogeni c cellaggregates iso lated from the suspension
cultures. Additionally, seve ral facto rs are
importa nt for successfull y maintaining an
embryogenic culture.
First, the emb ryogeni c culture needs to
be subcultured in a timely fashion. Our
previous study shows th at delayin g the
subculture diminished the embryogenic
ability of th e ce ll -agg regates. Liu et al.
(1997) reported that at 21 weeks after
initi at ion, only 61.1 % of cell-aggregates of
Kokei No.14 were still emb ryogeni c if the
suspension cultures were not subcultured in
a tim ely way.
Second, the pro Iiferation of ee l I-aggregates on solid MS medium with 2.0 mg!L
2,4-D is of critical importance for ac hievin g
a hi gh frequ ency of somatic embryo
formation and plant regeneration.

CIP Progrom Report 1997-98

267

Figure 1. Embryogenic callus derived from a shoot

apex of sweetpotato cultivar Lizixiang on
MS medium supplemented with 2.0 mg/L
2,4-D (scale bar = 1.0 mm).

Figure 2. Embryogenic suspension cultures of

sweetpotato cultivar Lizixiang proliferating in MS medium containing 2.0 mg/L
2,4-D.

Figure 3. Somatic embryos formed on embryogenic

callus derived from a cell aggregate of
sweetpotato cultivar Lizixiang on MS
medium supplemented with 2.0 mg/L 2,4D (scale bar = 1.0 mm).

Figure 4. Germination of somatic embryos and

formation of plantlets of sweetpotato
cultivar Lizixiang on MS medium
supplemented with 1.0 mg/LABA.

Third, the size of cell-aggregates, which

were transferred to 2,4-D solid medium, is
also important to obtain a high regeneration
frequency. Chee and Cantliffe (1989)
reported that the percentage of somatic
embryo formation decreased as the size of
the cell-aggregates decreased. They found
that sweetpotato somatic embryos only
formed on eel I-aggregates at least 0.18 mm
in size and cell-aggregates 0.71-1.0 mm in
size gave the maximum percentage of
somatic embryo formation (up to 70%). Our

resu Its indicated that 1 00% of eel I-aggregates about 1.0 mm in size produced
somatic embryos and plants in all 8 tested
cultivars.

2 6 8 Sweetpototo

We have succeeded in the development

of an efficient system of embryogenic
suspension cultures and plant regeneration
in eight sweetpotato cultivars. This system
of embryogenic suspension cultures has an
excellent application potential in genetic
transformation for cultivar improvement, as

well as in cryopreservation of sweetpotato

germplasm.
Literature Cited

Beachy R.N., C. Malpica and G. Clark.

1992. Coat protein mediated res istance:
Strategy for genetic engin ee rin g virus
res istance in sweetpotato. In: Hill , W.,
C.K . Bonsi and P.S. Loretan (eds.).
Sweetpotato for the 21" ce ntury. Proc.
Int. Symposium, June 2-6, 1991 ,
Tuskegee Uni versity, Tu skegee, AL, USA.
p. 507-513.
Bieniek, M.E., R.C. Harrell, and D.J.
Cantliffe. 1995. Enhanc ement of so matic
embryogenesis of lpomo ea batatas in
solid cultures and production of mature
somatic embryos in liquid cultures for
application to a bioreactor production
system. Plant Cell Tiss. Org. Cult. 41 :1-8.
Chee, R.P. and D.J . Cantliffe. 1988. Selecti ve enhancement of lpomoea batatas
Poir. embryogenic and non-embryogenic
ca llus growth and production of embryos
in liquid culture. Plant Cell Ti ss. Org.
Cult. 15:149-159.
____ . 1989. Composition of embryogenic
suspension cultures of lpomoea batatas
Poir. and production of individu ali zed
embryos. Plant Cell Tiss. Org. Cult.
17:39-52.
Chee, R.P., J.R. Schultheis, and D.J.
Cantliffe. 1990. Plant recove ry from
sweetpotato somatic embryos. Hortsci.
25(7):795-797.
Desa mero, N.V., B.B. Rhod es, D.R.
Decotea u, and W.C. Bridges. 1994.
Picolinic acid induced direct somatic
embryogenesis in sweetpotato. Plant Cell
Tiss. Org. Cult. 37:103-110.
Gosukonda, R.M ., C.S . Prakash , and A.P.
D essa i. 1995. Shoot rege neration in vitro
from diverse genotypes of sweetpotato
and multiple shoot production per
expla nt. HortSci. 30:1074-1077.
Jar ret, R.L ., S. Salazar, and R.Z. Fern andez.
1984. Somatic embryogenesis in
sweetpotato. Hortsci. 19:397-398.
Liu, J.R. and D.J. Cantliffe. 1984. Somatic
embryogenesis and plant rege neration in

ti ss ue cultures of sweetpotato (lpomoea

batatas Poir.). Pl ant Cell Rep. 3:112-115.
Liu , Q.C., T. Kokubu, and M. Sato. 1992 .
Varietal differences of somatic embryogenesis in shoot tip culture of
sweetpotato, lpomoea batatas (L.) Lam.
Japan. J. Breed. (Suppl. 2) 42:8-9.
Liu, Q.C., J.Q. Luo, H.Y. Zhou, and S.Y. Lu.
1993. Hi gh frequency somatic embryogenesis and plant regeneration in
sweetpotato, lpomoea batatas (L.) Lam.
J. Agr. Biotech. 1 (1 ):84-89.
Liu , Q.C. , K.X. Mi, D.H. Lu, H.Y. Zhou , and
Z. Fu.1997. Establishmentofembryoge nic cell suspension cultures in
sweetpotato, lpomoea batatas (L.) Lam .
Acta Agr. Sinica 23(1 ):22-26
Murata, T., Y. Okada, A. Saito, T. Kimura,
M. Mori, M. Nishiguchi, K. Hanada, J.
Saika, and H . Fukuoka. 1998. Transformation by direct gene transfer in
sweetpotato [lpomoea batatas L. (Lam)].
In : Proceedings of International Workshop on Sweetpotato Production System
toward the 21" Century held 9-10 D ec.
1997 at Miyakonojo, Miyazaki, Japan . p.
159-180.
Murashige, T. and F. Skoog. 1962. A revi sed
medium for rapid growth and bioassays
with tobacco tissue cultures. Physiol.
Pl ant 15:473-497.
Newe ll , C.A., J.M. Lowe, A. Merryweather,
L.M. Rooke, and W.D.O. Hamilton.
1995. Transfo rm ation of sweetpotato
(lpomo ea batatas (L.) Lam. ) w ith
Agrobacterium tumefaciens and regeneration of plants expressing cowpea
trypsin inhibitor and snowdrop lectin.
Plant Sci. 107:215-227.
Otani, M . and T. Shimada. 1996. Effi cient
embryogenic ca llus formation in
sweetpotato (lpom oea batatas (L.) Lam.) .
Breeding Sci . 46:257-260.
Prakash , C.S, M . Egnin, and J. Jaynes. 1998.
Biotechnology touches a Third World
crop: Improv in g protein content in
sweetpotato. Agricultural Biotec h
International Conference held 9-12 June
1998 at Saskatoon, Saskatchewan,
Canad a. p. 42. (Abstract)

CIPProgrom Report 199798

269

Tan, F., K.P. Li , L.Q. Lan, and Q.T. Zhang.

1993 . Somatic embryogenesis and plant
regeneration in sweetpotato. Acta Agr.
Sinica 19:372-375.
Zhang, D.P., A. Golmirzaie, and G.
Cipriani. Developing weevi l resistance in

2 70

Sweetpototo

sweetpotato with genetic transformation.

In : Internationa l Potato Center Program
Report 1995-96, CIP, Lima , Peru. p. 205210.

Expression of Soybean Proteinase

Inhibitor in Sweetpotato
G. Cipriani\ D. Michaud 2 , F. Brunelle 2 , A. Golmirzaie1, and D.P. Zhang1

Transformation of sweetpotato Upomoea

batatas) with serine-type proteinase inhibi tors (P is) offer a new strategy for developing
host resistance aga inst sweetpotato weevil
(Cy/as spp.), the major threat to sweetpotato
wor ld w ide. Foreign genes encoding Pis
ha ve been introduc ed into tobacco, tom ato,
strawberry, potato, and m any other crops
for control of a range of ins ect pests. Th ese
protein s bind to proteinases in the midgu t
of the insect thu s interfering with the
insect's metaboli sm . Tran sge nic tobac co
plants ex pressin g recombin ed serine Pis
have demonstrated enhanced resistance to
H eli ot his vires cens (Hi Id er et al. , 1987),
Manduca sexta (Johnson et al., 1989), and
Spodoptera litura (Yeh et al., 1997).
The predominance of ser in e-type
proteases in th e midgut of three Cy/as spp.
was co nfirm ed by inhibition assays. Serinetype inhibitors, rega rdless of the species or
deve lopmental stage of th e in sect, inhibited
th e proteinases detected (Rety et al., 199 8) .
Th e inhibitory spectrum of soybean (Glycine max) trypsin inhibitor (SBTI) against
the midgut proteases was apparently large r
than that of endogenous inhibitors in
sweetpo tato. It was apparently stable in the
presence of insect insensitive proteinases,
suggesting the potential usefulness of this
inhibitor in controlling weevil larva e and
adults (Z hang et al., 1997) .
Newe ll et al. (1995) d eve loped the first
transgenic Ii nes of Pl-express in g
sweetpotato with a cONA seq uence
encoding a cowpea (Vigna ung uiculata)
tryp sin inhibitor (CpTI). Feed ing tests with
th e West Indian sweetpotato weevil
1 CIP, Lima Peru.
2 Lavol University, Quebec, Canada.

(Euscepes postfasciatus) were carried out

rec entl y at CIP by performing a no-ch o ice
test in a screenhouse-based bioassay. Two
transformed clones were found moderate ly
resi sta nt in consecutive tests over 2 ye ars,
suggesting th e actual usefuln ess of the
trypsin inhibitor in weevil control.

This paper desc ribes th e Agrobacteriummed iated tran sfo 1mation of sweetpotato
with a soybean kunitz tryps in inhibitor
(SKTl-4) cONA c lone and th e analysis of
recombinant SBTI produced in sweetpotato
cell s aga inst trypsin and sweetpotato lea f
proteina ses .

Materials and Methods

Sweetpotato transformation with
SKTl-4 gene
Complete leaves with peti o les of in vitro
plants of sweetpotato cv. Jewel, Pl-3188463, and Jonathan , and leaf segments of cv.
Maria A ngola we re us ed as exp lants.
Explants were inoculated and co-cultivated
with Agrobacterium tumefaciens LBA4404
carrying plasmid pKTl-4. Th e T-ONA reg ion
of thi s plasmid harbors the SKTl-4 ge ne, the
b-glu curonidase (GUS) reporter gene, and
the NPTll marker gene (Figure 1).
Th e first three ex plants went throu gh two
successive Uan sfe rs: First to Fl 5 medium
(Murashige-Skoog (MS) salts, 0.2 mg/L 2,
4-D; 0.2 mg/ L zea tin; 50 ppm kanamyc in sulfate; 200 ppm cefo ta x ime; 30g/L sucrose;
and 3.6 g/L phytagel) for 3 d; and second to
F9 medium (same as Fl 5 but without 2,
4-0) . All calli formed were analyzed for
GU S ac tivit y, and positive ca lli were
transferred to G240 medium (bas al medium supplemented with 0.2 mg/L 2,4-0

CIP Program Report 1997-98

2 71

pKTl4
LB

RB
PHB

I
NOS

GUS

2'

EE; .H

I
35S NPTll

PHB

ocs

NOS KTI

35S

'

500*bp

Figure 1. Diagram of T-DNA region of plasmid pKTl-4. Coding regions for n-glucuronidase (GUS} , neomycin
phosphotronsferase II (NPTll}, and Kunitz trypsin inhibitor (KTI) genes are transcribed between TR 2'
(2 '} or Cauliflower mosaic virus 35S (35S} promoter sequeneces and nopaline synthase (NOS) or
octopine synthase (OCS} terminator sequences . The direction of transcription is shown by the arrowhead . Relevant restriction enzyme sites are indicated: Pac I (P}, Hind Ill (H}, Barn HI (B}, Eco RI (E}.
Left border (LB) , right border (RB}, base pairs (bp}.
and 0.05 mg/L gibberellic acid ) for multiplication. After 2 mo they we re tran sferred to
auxin-free media for regeneration.
Leaf segm en ts of Maria Angola, followin g inoculation and co-culture, were
culti vated in medium G24D for 2 mo for
embryoge nic cal Ii induction. Th ese ca l Ii
were evaluated for GUS activity and
positi ve cal Ii we re transferred to MS30-Vit
medium for embryo deve lopm ent.
GUS test of SKTl-4 transgenic lines
GUS expression in transform ed plants
was evaluated by hi stoc hemical assay as
desc ribed by Jeffe rs o n (1987). Leaves, roots,
and stem segm ents were incubated in
phosphate buffer pH 7 with X-gluc (5bromo-4-ch loro-3-i ndo Iy 1-beta-D-glu curon i c acid ) substrate. Blue stainin g of the
tissue denoted positive reaction.
Kanamycin resistance evaluation of
SKTl-4 transgenic lines
Three petridishes w ith five leaf segments
(each segment co rre spo nding to a different
leaf) we re cultivated for each putative
transgenic line. In cluded in each petridish
were two leaf segments of a kno w n trans-

272

lweetpototo

formed line as a po siti ve co ntrol. In addition, 2-4 leaf segments of an untransform ed

plant of the same cultivar we re includ ed as
a negati ve control. Medium used for calli
indu ct ion was 303-64 (MS salts, 1 mg/L
zeatin , 0 .1 mg/ L naftal en- acet ic acid, 30 g/L
sucrose, an d 3.6 g/L ph ytage l) containing a
hi gh concentration (100 mg/ L) of the
antibiotic kanamycin sulfate. As a control
for th e effect of the physiological status of
the pl ants, or the effect of the medium , the
same expe riment was repeated w ith each
lin e wit hout kanam yc in , and also w ith th e
untransformed cultivars w ith the same
concentration of kanam yci n. Calli deve lopment was evaluated after 1 mo in culture.
Southern blot test of transgenic lines
The prese nce of SKTl-4 se quence was
confirmed by Southern blot analysis. DN A
was iso lated from 4 g of froze n leaf tissue
from in v itro plantlets (De llaporta et al.,
1983 ). Ten micrograms of DNA were
di gested w ith restriction enzy me BamHI
(th e SKTl-4 gene has BamHI restriction sites
at left and right borders as shown in
pla smid pKTl-4, Figure 1), and separated by
electrophoresis in a 0.9 % agarose gel. The
ge l was treated for 1 0 min w ith a solu t ion

co ntainin g 0.25N H CI, tw ice for 15 min in

0.5N NaOH, l .5M NaCl , and tw ice for 30
min in lM tris pH 7.4, 20 mM N aO H and
was blotted to a nylon fi lter (Amersham) by
ca pill arity.
One hundred nanograms DNA of SKTl-4
gene isol ated by restri ct ion of th e pl as mid
pKTl -2 was radiolabeled w ith 32 P (dATP)
and used as a probe for fi lter hybridi zation.
After hybridi zation , the membrane was
washed twi ce, for 5 and 15 min in a buffer
so luti on containing 2X SSC,. 0.5 % sodium
dodecy l sulfate (SOS) at room temperature.
Th e membrane was wash ed twi ce more for
30 min eac h in 0.1 X SSC and 0.5% SOS,
once at 37C and once at 65 C. Filters were
ex posed for autoradiography with an
intensifying scree n at -70C for up to 2 d.

PCR analysis of transgenic lines

Th e SKTl-4 gene was targeted w ith
spec ifi c prime rs and amp Iifi ed by po lymerase chain reaction (PCR) from DNA of
in v itro transgenic plants. These 18 base
pairs (bp) forward and reverse primers
where spec ially synthesized to fl ank the
653 bp of the SKTl -4 gen e accord in g to the
comp lete PUC-derived pKTl-2 pl as mid's
sequ ence provided by C. Newel l of Pestax
Ltd. , UK (Table 1 ).
Th e PCR amplification co nditi on was as
fo ll ows:
94C for l '
1 cycl e:
94C for l '
35 cycl es:
52C for 2'
72C for 3'
72C for 7'
1 cyc le:
Fo r th e amp I ification of eac h transge nic
line, 40 ng of DNA extracted from in vitro
pl ants were used. One nanogram of
pl as mid pKTl-2 (PUC-derived pl as mid

harb oring the SKTl-4 gene) was used as a

positi ve co ntrol. DNA of untransformed
plants (C-) and the amplification mi xture
without DNA (M ) were used as nega ti ve
controls.

Protease and protease inhibitor assays

The protease assa y was adapted from
Michaud et al. (1993). Thirty-fi ve L of leaf
protein extracts from 8 transgeni c lin es o r
from an untransfo rm ed con trol of genotype
Pl-318846-3 (1 g protein/ L extract) we re
incubated w ith 5 ml of 0.5m g/ml (w/v)
trypsin di sso lved in 50 mM tris, pH 8.0. Th e
samples we re in cubated for 15 min at 37C
to all ow protease inhibition, and 80 L of
2% (w/v) azocase in dissolved in th e sa me
buffer w as added as substrate to each
sampl e. After a 180-min incubation, 300 L
of 10% (w/v) tri chloroacetic acid (TCA) was
added to th e mi xture, and residual
azoca se in was removed by centrifu ga ti o n
for 5 min at 14,000 rpm . The supernatant
(350 L) was then added to 400 L of 1 N
NaOH , and the absorbance was meas ured
at 440 nm usin g a Spectronic spectrophotometer (Mi lton Roy, Rochester, NY, USA).
The abso rb ance of blanks, which cons isted
of compl ete mixtures without trypsin or
plant extract submitted to the same treatment, was subtracted from each va lu e. One
unit of protease acti vity was defin ed as th e
amount of enzyme required to produ ce an
absorbance change of 1.0 h-1 in a 1 cm cuvette, under th e co nditions of the assay.
Th e resu lts obtained were compared w ith
control s co nsisting of (1) trypsin without
plant extract (maximum hydrolysis), and (2)
trypsin plus extract from an untransfo rm ed
plant.

Electrophoretic analyses
Express io n of proteinase inhibitors in th e
leaf ti ss ues was v isualized by pol yac ryla -

Table 1. Primer's sequence for PCR reaction.

Primer name
SKTl2.1(forward)
SKTl2.2 (reverse)

Melting temperature (C )

Corresponding sequence

Sequence

SKTI leader sequence

Cterminal peptide

ATGAAGAGCACCATmc

52

TCACACACTGCGAGAAAG

54

CIP Program Report 1997-98

273

mide gel electrophoresis (PAGE ) on gels

treated with SOS , a mildl y denaturing
agent. Plant extracts from trans ge nic lines
or from untransformed (control) plants we re
first incubated w ith trypsin for 15 min at
37C to allow protease inhibition by
recombinant SBTI. The samples we re then
fractionated into 10% (w/v) pol yac rylamide
gel slabs containing 0.1 % (w/v) gelatin. The
SOS-PAGE procedure is more fully described in Michaud et al. (1996). After
electrophoretic migration at 4C using the
MiniProtean II electrophoretic unit (BioRad , Richmond , CA, USA), the gels were
transferred into a 2.5 % aqueous solution of
Triton X-100 for 30 min at room tempe1ature to allow proteinase (trypsin) renaturation. The gels were then placed in an
activation or proteolysis buffer (SO mM tris,
pH 8.0) at 37C for 3 hand finall y transferred into a protein staining solution [0.1 %
(w/v) Coomassie Brilliant Blue in 25% (v/v)/
10% (v/ v) acetic acid]. For storage, the gels
were dried using Gel Dry ing Film
(Promega, Madison, WI, USA). Trypsin
activity was visualized as a clear (lysis)
band against the blue gelatin background;
trypsin inhibition was visualized as a
weaker band compared with noninhibited
controls, i.e., those with no plant extract
and/or with extracts from nontransformed
plants.

GUS test of SKTl-4 transgenic lines

Good expression was observed in the 72
SKTl-4 lines of the 4 cultivars. Because
GUS and SKTl-4 genes are driven by
CaMV35S, it was possible to obtain positive
reaction in different parts of the transgenic
pl ants.
Kanamycin resistance evaluation of
SKTl-4 transgenic lines
Eighteen lines of Jewel, Pl-318846-3,
Maria Angola, and Jonathan have been
tested. All lines had the same performance
in the medium with 100-mg/ L kanamycin
and without kanam yc in, showing good calli
formation. Untransformed leaf segments
we re necrotic in the presence of kanamycin
in all cases . Figure 2 shows the response of
Maria Angola's transgenic line and the
negative control to the presence of kanamycin antibiotic in the medium.
The resistance to the antibiotic is due to
the expression of the NPTll gene, w hich
correlates w ith the expression of the
in serted foreign gene. These lines are in
propagation for bioassay (Tab le 2).
::;o3(<><>-.)
eX"f": v:rz
G;.r.tt: .CO'(

'"

"'- u.:;,,

J'

Results
Transformation
Calli were obtained in different parts of
the exp Iants, mainly at the cut edge of the
petiole . Most positive cal Ii were small,
green, and round. After 2 mo in G240
medium, all calli proliferated. All established calli lines were confirmed GUS
positi ve. Most developed cal Ii were light
green or yellowish, friable cal Ii. Embryogenic orange cal Ii were also obtained in
Jonathan and Jewel. In regeneration media
without auxins , a total of 50 transgenic
lines have been obtained. Through somatic
embryogenic regeneration of leaf segments,
22 transgenic lines of Maria Angola have
been obtained.

2 74

Sweelpoloto

Figure 2. Kanamycin test. Calli formation on leaf

segments of in vitro plant of transgenic
line of cultivar Maria Angola in presence
of high concentration of the antibiotic
kanamycin (100 mg / L). Note dried
segments and no calli formation of
untransformed leaf segments of the same
cultivar (negative control).

Table 2. List of transgenic clones with SKTl-4 gene ready for bioassay.

Name of transgenic lines ready for bioassay

Pl-318846-3 41 (1.1)
Pl -318846-3 73 (1.1)
Pl -318846-3 89 (1 .1)
Pl-318846-3 89 (1 .2)
Pl -318846-3 89 (1.3)
Pl-318846-3 89 (1 .4)
Pl-318846-3 89 (1 .5)
Pl-318846-3 89 (1.7)
Pl-318846-3 92 {l .l)

Pl -318846-3 92 (1 .3)
Jewel 6 {1.1)
Jonathan 3R (1.1)
Jonathan 21 (1.7)
M. Angola 63 {1.1)
M. Angola 104 {l .2)
M. Angola 104 (1 .4)
M. Angola 104 {1.7)
M. Angola 106 {l .l)

Southern blot
To date eight transge ni c lines of Pl 3 18846-3 have bee n confirmed through
South ern bl ot. Total DNA of transfo rmed
and untransform ed control was di gested
with BamHI to release th e intern al fragment
(65 3 bp) corres pondin g to th e SKTl-4
cDN A. The SKTl-4 probe did not hybridi ze
with DNA of th e co ntrol pl ant (non transgeni c) but did with DNA of all Pl -3 1884 6-3
transge ni c lin es tested, show in g th e insertion of the SKTl-4 gene in th e tran sformants.

PCR analysis
Th e SKTl-4 gene w as targeted w ith
spec ifi c primers and amplifi ed by PCR from
DNA of in vitro tran sgeni c pl ants. Twentyfour transgeni c lin es of Jewe l, Pl- 3 18846-3,
Mari a Angol a, and Jo nath an have been
co nfirm ed for the in se rtion of SKTl -4 gene.

C-

-! ...

mix

....

Jl

Th e amplifi cati o n band is show n for 16

sa mpl es (Fi gure 3).

Activity of recombinant SBTI

The inhibitory effect of re co mbin ant SBTI
on trypsin acti v ity w as analyzed in v itro
with protein extracts from upper (youn gest),
middle, and bottom (o ldest, nonsenescent)
leaves of sweetpotato gree nh ouse pl ants
(Fi gure 4). For eac h assay, th e data w ere
analyzed by comparin g protease acti v iti es
meas ured in th e modifi ed extracts with
those meas ured in extracts from
untransfo rm ed pl ants. Th at took into
acco unt the age-d epend ent acti v ity of pl ant
leaf endogenou s protease and trypsin
inhibitor activiti es (data not sh own).
Beca use th e parental leaf materi al w as
identical for all cl ones analyzed, it co uld be
assumed th at th e vari ati ons obse rved fo r a
give n stage of leaf deve lopment resulted

10

11

12 13

14

15 16

--

Figure 3. PCR amplification of 16 putative transgenic plants with SKTl-4 gene. The absence of the band in
negative control (C-) and the band of the same weight (653 bp) in the positive control (P) .

CIPProgram Report 1997-98

275

Trypsin inhibition(%)

140
1201---------1 1---;::====;--1
1001-------j

sol-----,..--

O Young
Ill Middle
O Mature

601-------1

-40 L___

2--3--4--=--5--=--6-7::----;8;-----;9;--

Figure 4. Differential expression of SBTI inhibitor in

transgenic sweetpotato lines divided by

young, middle, and mature leaves. Clones
1-8 are lines of transgenic Pl-318846-3;
Clone 9 {NT) is the nontransgenic control.
from various expression levels of the
inserted DNA. As shown in Figure 4, a net
increase in antitrypsin activity was noted for
some clones regardless of leaf age (e.g. ,
clones 3 and 4), when compared to the
activity measured in the corresponding
extracts obtained from nontransgenic clone
9. Collectively, these observations indicate
the presence of added antitrypsin activity in
the transgenic lines, due to the insertion
and expression of the SBTl-encoding cDNA
sequence into the sweetpotato genome.
Electrophoretic analysis
In parallel, mildly denaturing gelatin/
SOS-PAGE allowed the occurrence of SBTI
activity in extracts of transgenic plants
expressing SKTl-4 to be seen. Because SBTI
is a strong inhibitor of trypsin, the complex
formed between the enzyme and the
recombinant inhibitor may remain stable in
the presence of SOS. That prevents restoration of trypsin activity during migration and
allows the identification of transgenic lines
expressing SBTI under an active form
(Michaud et al. , 1996). In the present case,
noninhibited trypsin activities appeared as
clear (lysis) bands against the blue gelatin
background following electrophoresis.

276 Sweetpototo

Conversel y, under the electrophoretic

conditions used, the enzyme was irreversibly inactivated when mixed with an extract
from a transgenic plant expressing active
SBTI, resulting in no or only a weak lysis
zone in the gel following gelatin/SDSPAGE. Figure 5 shows the difference
between the lysis bands obtained w ith
extracts from the untransformed control and
from two samples of transgenic lines with
different levels of trypsin activity. Compared
with the bands observed when trypsin was
previously mixed with leaf extracts from the
untransformed Ii ne, the bands observed
with extracts from the transformed Ii nes
were weaker. The resulting intensities (and
inhibitory levels) varied with the expression
of the recombinant inhibitor.
Discussion

The present study was aimed at assessing

the potential of an SBTI cDNA clone for the
production of wee v il-resistant sweetpotato
plants. It is clear from both in vitro and.
electrophoretic assays that the transgen 1c
I in es possess increased antitrypsi n activities
compared with the parental control
(nontransgenic) line. The most obvious
increase in antitrypsin acti v ity in the
transgenic lines was noted for the upper

Figure 5. Electropherogram showing the

gelatinolytic activity of trypsin

preincubated with sweetpotato leaf
extracts. Wells were loaded with 5 mg
sweetpotato leaf proteins. Lane l:
untransformed control extract; lanes 2-3
Pl-318836-3 transgenic lines extract.

(young) leaves (Figure 4, clone 4). Increased inhibitory activity was also noted
for older leaves, suggesting th e presence of
recombinant SBTI in mo st of the leaves.
Given the occurrence of SBTl-sensitive
trypsin activity in sweetpotato weevils
(Zhang et al., 1997), clones 3 and 4 (Figure
4) may rep resent interesting candidate lin es
for the development of sweetpotato weevil
mana gement strategies.
These two Pl-318846-3 lines (89 1.3 and
89 1 .4) are part of a group of transgenic
lin es expressing active SBTI that are in the
multiplication process for bioassay. Continued efforts are need ed to find the most
appropriate co m bi nation of protea se
inhibitors, for enh ancin g the exp ression of
pl ant defense proteins in sweetpotato, and
for designing strategies to introduce protease inhibitor-expressing plants into
integrated pest managem ent systems.
Acknowledgements

The authors thank Pestax Ltd ., UK, for

supplying the SKTI gene construct, M. Jusurf
of Lembang Root & Legume Research
Institute of Indonesi a, th e Sweetpotato
Labo ratory of Guangdong Academy of
Agricultural Science of China, Paul
Kakande of the Biochemistry Department,
Makerere University, and Justin e Nanteza of
th e Namulonge Agricultural Research
Station, Uganda, for collecting the weevil
samples. This work was partially supported
by a grant from th e Canadian Intern at ional
Development Agency.
References Cited

Dellaporta, S., J. Wood, and B. Hicks. 1983.

A pl ant DNA mini prepa ration: Version II.
Plant Mol. Biol. Reporter 1:19-21.
Hilder, VA., A.M .R. Gatehouse, S.E.
Sheerman, R.F. Barker, and D. Boulter.
1987 . A novel mechanism of insect
resistance engineered into tobacco.
Nature 330:161-163.

Jefferson, R. 1987 . Assaying chimeric genes

in plants: The GUS gene fusion system.
Plant Mol. Biol. Rep. 5(4):387-405.
Johnson, R. , J. Narvaez, G. An, and C.
Ryan. 1989. Expression of proteinase
inhibitors I and II in transgenic tobacco
plants: Effects on natural defe nse against
Manduca sexta larvae. Proc. Natl. Acad.
Sci. USA. 86:9871-9875.
Michaud, D., L. Cantin, D.A. Raworth, and
T.C. Vrain. 1996. Assessing the stability
of cysta tin/cystei ne protei nase complexes using mildly-denaturing gelatin/
polyacrylamide ge l electrophoresis.
El ectro phoresis 17:74-79.
Michaud, D., L. Faye, and S. Yelle. 1993.
Ele ctro phoretic ana lysis of plant cysteine
and ser in e proteinases usin g gelatincontaining polyacrylamide ge ls and
class-specific protein ase inhibitors.
Electrophoresis 14:94-98.
Newe ll , C., J. Lowe, and A. M erryw ea th er.
1995. Tran sfo rm atio n of sweetpotato
(lpomoea batatas (L.) Lam.) with
Agrobacterium tumefaciens and regeneration of plants expressing cowpea
trypsin inhibitor and snowdrop lectin .
Plant Sci. 107:215-227.
Rety I., G. Cipriani, D.P. Zhang, and D.
Michaud. 1998. Soybean kunitz an d
Bowman-birk inhibitors strongly inactivate the majo r digestive serine protein ase
of sweetpotato weevils. Paper prese nted
at the Annual Meeting of the American
Society of Plant Physiologists held 27
Jun e to 1 July in Madison, WI, USA.
Yeh K.W., M .I. Lin , S.J. Tuan, Y.M. Chen,
C.Y. Lin , and S.S. Kao. 1997.
Sweetpotato (ipomoea batatas) trypsin
inhibitors ex pressed in transgenic
tobacco plants confer resistance against
Spodoptera litura. Plant Cell Reports
16:696-699.
Zhang, D., A. Golmirzaie, G. Cipriani, A.
Panta, M. Ghislain, N. Smit, I. Rety, and
D. Michaud. 1997. Developing weevil
resistance in sweetpotato with genetic
transformation. CIP Program Report
1995-1996. p. 205-210.

CIP Progrnm Report 1997-98

277

Starch Content and Properties of 106 Sweetpotato

Clones from the World Germplasm Collection Held at
CIP, Peru
C. Brabet1, D. Reyno so1, D. Dufour 2 , C. Mestres 2 ,

Poverty al lev iation throug h in creased

postharvest cro p use is the maj or thrust of
CIP's researc h o n sweetpotato (lpomoea
batatas) in Asia. Much of the effort ha s
focused on China, which produ ces roughl y
85% of the wor ld 's sweetpotato (CI P, 199 8).
A series of col laborative dia gno sti c studies,
and technology and market assess ments
identified sweetpotato starch production,
espec ially for noodle processing, as an
im portant in co me- generat in g acti vity in
poorer areas of China (Scott and Wheatley,
1997; Scott et al ., 1992 ).
Rea li zin g the full income-genera tin g
po tential of this activity dep end s on an
integ rated approach that includ es better
raw mater ial and improved procurement,
processing, and packag in g (W heatley et al.,
1997; Zhang, 1999). From a var ietal
perspective, increa sin g starch co ntent in
fres h roots , extraction rate, and qua li ty are
criti ca l co mponents . A better kno w led ge
and und ersta nding of sweetpotato starch
propert ies, and how th ey compare with
those of other starch sources, is necessa ry
to enh ance the potential va lu e of
sweetpotato starch in ex istin g and novel
uses. Latex, the resin produ ced by
sweetpotato latificers, is another important
trait to be considered by th e starc h indu stry
as the latex may contaminate the sta rch and
adhere to equipment (Woolfe, 1992) .
Th ere is sign ifi cant cu ltivar difference in
th e content and properties of sweetpotato

1 CIP, Lima, Peru.

2 Cen lre de Coope rali on ln lernali ona le en Rec herche
Agronomiqu e pour le Developpemenl (C IRAD),
Monlpellier, Franc e.
3 Universiclacl Nacional Agraria, La Molina, Lima, Peru .

J. Arredondo3,

and G. Scott1

starch (Tian et al. , 1991 ; Woolfe, 1992 ),

w hi ch suggests that ge net ic improveme nt of
th ese traits may be achieved . Howeve r, th e
variability of sta rch content and properties
in accessions from the sweetpotato gene
bank held at CIP has not been com prehensively evaluated. The gene bank maintain s
a co ll ection of about 5,5 00 culti va ted
accessions from 57 countries.
The obj ec tives of thi s stud y were to (1)
evaluate the va riabil ity of starch content
and properti es (a mylose content and
pastin g properti es) in advanced swee tpotato
clones se lected from th e gene bank held at
CIP, (2) estim ate the most rele va nt co rrelation betwee n the evaluated variab les, (3)
id en tify clones with potential fo1 in corpora tion into CIP's breeding program s for starc h
prod uction and use, and (4) provi de
recomm end at ions for future eva lu ation and
use of the sweetpotato gene bank.

Materials and Methods

Plant material
One hundred and six sweetpotato c lones
were eva lu ated in this study (Tab le 1 ). Th e
set was chosen to represen t di verse geographi cal origin, high var iation in predominant root fl es h color, and high dry matter
(OM) content(> 25% for orange-fleshed
clones and > 30% for others). Clones were
also se lected for adaptation to the
agroecological condit ions at th e trial site
ba sed on CIP's root y ield data. In add iti on,
performance (high root y ield s) and relative
importance of the clones in Cl P's reg ions
(widely grown) were co nsidered.

CIPProgrnm Repori 1997-98

279

Tablel. Origin and predominant root flesh color of 106 sweetpotato clones selected from the gene bank
held at CIP, Lima , Peru.
Origin

Clones {N)

Predominant root flesh color

White/cream

North America
Mexico'
United States
Central America and
Caribbean
Cuba
Puerto Rico
South America
Brazil
Peru
Sub-Saharan Africa
Burundi
Kenyo
Nigeria
Rwanda
Uganda
Middle East and
North Africa
East and Southeast Asia
and the Pacific
Chino
Japan
Korea
Papua New Guinea
Philippines
Taiwan, China
Thailand
Tonga
South and West Asia
Bangladesh
Sri Lanka
Oceania
Total

(10)
l
9
(8)

7
(18)
3
15
(11)

Yellow

(1)

(9)
l
8

l
(7)

(l)

6
(6)

l
(4)

(8)

Sd

2
2
(3)

8
(3)

(5)

(57)

(27)

5
5
2
9
5
15
15
l
(2)

106

(11)

2 80

Sweetpototo

(19)

7
l

2
10

3
2

3
10
3

(l)

(l)

.l

46d

20
Assessed visually ofter cross sectioning fresh roots according to the method described by Huaman (1991 ).
b Including yellow/orange- to orange-fleshed clones.
cIncluding bred lines developed at CIP.
d One clone was strongly pigmented with anthocyanins.
0

Orangeb

40

Sweetpotato c lones we re grown und er

sta ndard cu ltural prac tices durin g the dry
seaso n (av tem perature rang e 18.9 -32 .5 C,
tota l rainfall 437 mm) at CIP 's expe rim ent
sta tion in San Ramon , Peru (tropica l
mid land located at lat. 11 06 ' S, long .
75 18' W, and 800 m above sea leve l).
Nitrogen fertilizer was appli ed at the rate of
80 kg/ ha and sp rink ler irrigati on was
supplied as needed. Clones we re pl anted
on 26 May 1997 and harvested after 164 d.
Two- row pl ots were used with 10 pl ants per
row spaced 0 .25 m with in ro ws and 0.90 m
between row s. Plots w ere sep arated by one
unpl anted row and not repli cated. Ten to 20
heal th y roots (> 12 5 g each) per clo ne were
sa mpled and imm ed iately pl aced in paper
bags und er shade. Th ey were th en taken to
CIP-Lima and stored at l 3C until processin g. Because of the large number of clon es
and the lon g di sta nce between field and
laboratory, the time between harvest and
processing was abou t 15 d.

Sample preparation
Washed and unpeeled sweetpotato roots
we re cut longitud in ally in on e hal f and two
qu arters after removin g the extremities. One
half was used for starch extract ion, one
quarter fo r fl our prepa ratio n, and the other
quarter for dete rminin g DM co ntent.
Starch. Root hal ves were sliced and
thoroughly mi xed. A subsample of approximate ly 1 kg was mace rated in a kitch en
blender w ith tap water (1 :1 v/v) fo r 2 min at
max imu m speed and filtered th ro ugh a
muslin cloth. Th e res idu e was resuspended
in tap water (1 :2 v/v), mace rated, and
filtered in th e sa me way. The two filtrates
were pooled, passed through a 250 m
sieve and adjusted to 4 L with tap water.
Starch was all owed to settle for 3 hat room
te mperature (20-2 4C) and th e supe rn ata nt
was di sca rd ed . Th e sta rch was resuspended
in 2 L of tap water, filtered throu gh a 75 m
sieve and left to settle in a tra y for 2 h. This
last step was repea ted three tim es w ithout
th e sievin g step, and using deioni zed water
instea d of tap water for the last two
wash ings. Th e recove red starch was dried

in a fo rced-air ove n at 40 -45C fo r 24 h,

grou nd w ith a mortar an d pestle to pass
th rough a 250 m sieve, and stored in
sealed pol yethyl ene bags at 6C.

Flour. Root quarters we re cut into 1 mm

thi ck sli ces and mixed. A sub sa mpl e of
app roximate ly 400 g was freeze-d ri ed, th en
ground in a mi ll to pa ss th ro ugh a 425 m
sieve. The res ul ting fl our was stored in
sealed polyethylene bags at -20C.
Root analysis
Dry matter. Root qu arters were cut into
about 0.5 cm 2 cubes and mixed. Three
subsam ples of approximately 200 g we re
dri ed in a forced -air ove n at 90C fo r 48 h
(i. e., until co nsta nt weight).
Extractable starch. Extra ctab le starch
wa s ca lculated as the ratio of g starch at
standard 14% moisture con tent (MC)/100 g
fres h roots.
Total starch. Total starch in flour was
determin ed using a Total Starch Assay Kit
(Cat. No. K-TSTA, M egazyme, Ire land). Th e
method consisted of hydrolyzin g starch to
glucose by an enz ymic proced ure. Gluco se
was meas ured colorimetrica l ly wit h glucose
oxidase-peroxidase reage nt. Fl our MC was
determin ed using the AOAC Officia l
M eth od 925.10 (AOAC, 1995). Both
analyses were done in dupli cate.
Latex. Latex produ ction was assessed
visua lly after cross sectioning fi ve fres h
ro ots usi ng the ratin gs 0 = latex not di sce rnibl e, 3 = li tt le latex, 5 = so me latex, and
7 = ab und ant latex.
Starch properties
Moisture. MC was determin ed by oven
dryi ng two represe ntati ve starch sam pl es of
abo ut 4 g eac h at 1 05 C for 24 h.

pH. 5 g o f sta rch on a dry we ight basi s

(dwb) were di spersed and stirred in 50 ml
of di st ill ed water at room temperature for
30 min. After filtering, the pH of th e

CIP Program Report 1997-98

28 1

Statistical analysis

so lution was measured. Th e ana lys is was

done in duplicate.

PROC M EAN and PROC CORR procedures (SAS, 1992) we re used to ca lcul ate
the descriptive statistics and Pearson's
correlation coeffic ients betwee n var iables.

Amylose. A my lose content was determined usin g th e differential scanni ng

ca lori metry method described by Mestres et
al. (1996).

Results and Discussion

Root dry matter and starch content

Pasting properties. Viscos ity profiles of

starc h suspensions of 9% (dwb/w) in
distilled wate r throu gh a heat in g and
cool in g cyc le were obta in ed usin g a Rapid
Visco -An alyzer (RVA) model 3D (N ewport
Scientific, Austra I ia) as describ ed by
Co llad o and Corke (1997). Pasting temp erature (Tp) at w hi ch viscos ity started to
in crease, maximum or peak v iscos ity (PV),
v iscos ity at the end of th e ho ld time at 95C
or hot-paste viscosity (HPV), and viscosity
at the end of the hold ti me at 50C or coo lpaste v iscos ity (CPV) we re recorded (F igu re
1). The stab ii ity (HPV/PV) and setback rat ios
(C PV/HP V) we re ca lcul ated.

Root DM an d total starc h co nte nt of the

106 sweetpotato clones va ri ed w id ely. Root
DM ranged fro m 19.9 to 45.4 % and starch
conte nt from 11 .1 to 33.5% on a fresh
we ight ba sis (fwb), w ith an average total
starc h co ntent of 21.6 % (Tab le 2) or 61.5 %
dw b. These va lu es are w ithi n the ranges
repo rted in the literature (Woo lfe, 1992) .
Th e 106 c lones we re initia ll y se lected for
hi gh DM con tent based on CIP data, wh ich
exp lai ns the hi gh average of 34.9% (Tab le
2). Most of the w hi te/c ream- (58%) and
ye ll ow-fl eshed clon es (85%) had a DM
content> 35% (Fi gure 2). Sixty pe rcent of
th e oran ge-fles hed clones (i ncluding ye ll ow

Temperature (oC)

Viscosity (RVU)

700

~~~~~~~~~~~~~~~~~~~~~~~~~~~

100

Temperature profile

600

80
500
CPV

400

60

300

40

200
20

Tp

100

10

15

20

25

30

Time (min)

Figure 1. Typical RVA viscosity profiles of sweetpotato starch suspension of 9% (dwb/w) in distilled water
observed in 106 clones selected from the gene ban k held at CIP, Lima , Peru . Position of posting
temperature (Tp), peak viscosity (PV), hot-paste viscosity (HPV), and cool-paste viscosity (CPV) is
indicated. Profiles A (CIP440029, Feng Shou Bai, China , white /cream-fl eshed clone) and B (CIP440068 ,
llTA-TIS 5081 , Nigeria, yellow/orange-fleshed clone) correspond to the lowest and highest PV values.
282

Sweetpototo

Table 2. Variability of root dry matter and total starch content, extractable starch, and starch physicochemical
properties' among l 06 sweetpotato clones selected from the gene bank held at CIP, Lima, Peru.
Variables'

Dry matter (%)

Total starch(%, fwb')
Extractable starchd(%, fwb)
Starch pH
Amylose (%)
Tp' (C)
PV' (RVA)
HPV' (RVA)
CPV' (RVA)
Stability ratio (H PV/ PV)
Setback ratio (CPV/ HPV)

Range

19.9 - 45.4
ll.l - 33.5
6.5 - 25.7
5.1 - 7.0
18.6 - 27.l
70.2-76.6
410 - 600
161 - 243
242 - 342
0.31 - 0.52
l. 22 - l.62

Mean

34.9
21.6
17.3
6.1
21.8
73.8
494.0
204.0
285.0
0.41
1.4

Standard

Coefficient of

Deviation

variation{%)

4.5
4.5
3.6
0.5
l.3

1.4
38.0
16.0
23.0
0.04
0.08

13
21
21
9
6
2
8
8
8
10
6

Literature rongeh

13. 6 - 48.2
5.3 - 28.4

/-/
5.9 - 6.8
8.5 - 38
66.5 - 86.3
329 - 428
127 - 203
208 - 284
0.35 - 0.55
l.40 - 1.67

Analyses were done at least in duplicate, except for extracta ble starch, and starch amylase content and posting properties.
Dry motter and total starch (Woolfe, 1992); Amylase and Tp (Tian et al., 1991); Storch pH and posting properties, except
Tp (Collado and Corke, 1997).
' Fresh weight basis.
d 14% moisture content.
' Storch posting properties: Tp= Posting temperature, PV = Peak viscosity, HPV = Hot-paste viscosity, CPV = Cool-paste viscosity,
RVA = Rapid Visco-Analyzer unit.
0

o ran ge) had a DM co ntent > 30%, w ith 10

clon es > 35% . Th e Peru v ian cl o ne
(CIP4 20053, Ca padito), w hi c h is stro ngly
pi gmented w ith anth ocya nin s, had a DM
co ntent o f 39 .1 % . Th ese res ults suggest th e
potenti al fo r in c reas in g DM co ntent in
sweetpotato ro ots by usin g these cl o nes in
breeding prog ram s. M any va ri eti es now
culti va ted have a DM co ntent too low (2530%) to be used in th e process in g indu stry
(Mok et al ., 1997), w hi c h prefers a DM
co ntent > 30-35% .
Extractabl e starc h ranged from 6.5 to
25.7% fwb, w ith an average of 17.3%
(Tabl e 2) . Th at represe nts an average
recove ry rate o f 80.3% (rat io of starch at
stand ard 14% MC to total starc h co ntent in
roots), w ith va lu es rangin g fro m 54.6 to
9 1.4% . Wh ea tl ey (1996) fo un d a hi ghl y
signifi ca nt pos itive co rrelati o n (r = 0.90,
P < 0.01) between labo rato ry and indu stri al
extracti o n rates . Th at dem o nstrates th e

usefuln ess o f the labo rato ry meth od used in

thi s stud y fo r sc ree nin g c lo nes fo r th e starc h
indu stry. W hite/crea m - and ye ll ow-fl es hed
c lo nes ave raged hi gher extracta bl e starc h
(1 8.3% and 19. 1%, res pecti ve ly) th an
orange-fl es hed cl o nes (15.2%). Sixty-two
percent of th e w hite/c rea m-fl es hed cl o nes
and 70% of th e ye ll ow-fl eshed cl o nes had
extractab le sta rc h va lu es betwee n 1 5 and
20%, w hereas abo ut half of th e o rangefl es hed c lo nes had va lu es< 15% (Fi gure 2) .
Approx im ately 29% of th e w hi te/c rea m fl eshed cl o nes, 30% o f the ye ll ow, and 18%
of th e o range were in the hi ghest ex tractabl e starch 1a nge of 20-26%.
A small amo unt of latex (3 rating) was
produced in fres h roo ts o f approx im ate ly
40% of th e 1 06 cl o nes; 12% produ ced
abund ant latex (7 rating) (Fi gure 2). Co mpared to the orange-fl es hed cl o nes o f w hi c h
28% produ ced littl e latex and 20% abundant latex, a hi gher pro po rtion o f w hite/

CIP Program Report 1997-98

283

Number of clones
60 ~~~~~~~~~~~~~~---,

50

A O White-cream D Yellow

QOrange Purple

40

21

30

16
3

20

14

10
0

25

20

30

35

40

45

50

Dry matter(%)

Caroteno id pigment responsible for the

orange flesh color of sweetpotato roots and
anthocyanin pigment, which is responsible
for purple flesh color, make the production
of a white starc h difficult. Hence white/
cream- and ye llow-fleshed clones are the
most suitable for the starch industry. Seven
light-fleshed clones with high extractable
starch and low latex production in fresh
roots have been id entified from the 1 06
sweetpotato clones (Table 3).

Number of clones

60
50

28

40
30
3

20
10
0

15

10

14

12

14

6
7

20

25

30

Extractable starch (%, fwb)

Number of clones

60
50

20

40

20

30

20

21

10

10
0

Starch properties
8

Latex

Figure 2. Frequency distribution of (A) dry matter

content, (B) extractable starch at 14% MC,

and (C) latex production in fresh roots for
l 06 sweetpotato clones selected from the
gene bank held at CIP, Lima, Peru . Latex
(C) was assessed visually using the scale
0 = latex not discernible, 3 = little latex,
5 = some latex, and 7 = abundant latex.
Anaiyses were done at least in duplicate,
except for extractable starch. (a. Including
yellow/orange- to orange-fleshed clones.)
cream- (47%) and ye ll ow-fl eshed clones
(50%) produced a sma l I amount of latex,
and a lower proportion of these clones (9%
and 5%, respectively) produced an abun dant amount.

284

Sweetpoloto

A highly significant positive correlat ion

was found between extractab le starch and
both root DM and tota l starch content (r =
0.92, P < 0.001 for both ). A similar result
was reported by Mok et al. (1997). Thu s,
root DM content, which is simple, fast, and
cheap to determine, can be used to select
sweetpotato clones w ith high extractab le
starch. That wo uld be particularly useful in
breeding programs in the first or second
generation w hen a large number of c lones
are eva lu ated. Extractable sta rch in advanced clones cou Id then be measured in
the laboratory and co nfirm ed in the field.

Amylose content and pasting propert ies

are among the most important quality traits
of starch. W hen an aqueous suspension of
starch is heated above a critica l temperature, granu les swell irreversibly and
amy lose leac hes out into the aqueous
phase, resulting in increased viscosity
(past in g). At this stage the granu les are
highly susceptibl e to therma l or mechanical
breakdown, whic h leads to a decrease in
starch paste v iscosity. Upon cooling, the
starch paste forms a ge l (gel ification ) along
with increased viscosity. Pasting and
ge li fication are important properties in
determ ining starc h behav ior in various food
and industrial applications. Th ey affect
starch-based produ ct quality suc h as
texture, stabi lity, and digestibility. The RVA
prov ides as good a method for measuring
these functional properties and describing
starch potential end -u ses as the Brabende r
viscoamylograp h, wh ich is usually used but
consumes more time and sample.

Table 3. Sweetpotato clones selected for high extractable starch and low latex production' in fresh roots from the
gene bank held at CIP, Lima, Peru.

CIP

Name

Origin

Number

Predominant

Latex,h

Dry matter

root flesh color

(%)

Total starch Extractable

(%,fwb')

starch
(%,fwb)

187016.l
18701 6.3
188004.3
440041
440046
440341
440376

Caplina
Peru
TN89.315
Peru
LM88.113
Peru
Papota
Puerto Rico
Viola
Puerto Rico
101273
Thailand
Woksaken Papua New Guinea

White/cream
White/cream
Yellow
White/crea m
White/cream
Yellow
White/cream

3
3
3
3
3
3
3

41.0
39.8
39.4
39.3
38.9
38.4
39.2

27.3
25.6
24.8
25.9
25.5
24.4
26.8

21.4
20.0
22.6
23.3
21.0
20.2
21.7

' Analyses were done at least in duplicate, except for extractable starch.
bA
ssessed visually using the scale 0= latex not discernible, 3= little latex, 5=some latex, and 7=abundant latex.
' Fresh weight basis.

Amylose content. Amylase content

varied from 18.6% in the yellow-fleshed
clone Cl Pl 8700 1.2 to 27. 1% in the ora ngefl esh ed clone CIP420012, w ith an average
of 21 .8% in all c lo nes (Tab le 2). These
va lues are within th e ranges reported in th e
literature (Ti an et al., 199 1). Over 50% of
th e 106 clones had an amy lase content of
between 20 and 23%.
Pasting properties. Pastin g properti es
also vari ed among th e 106 sweetpotato
clon es (Tabl e 2). Th e RVA viscosity profil es
were all of the A type (using th e starch
classification of Schoch and Maywald,
1968) normally observed for root and tuber
starch. They w ere characterized by mode rate to hi gh PV with a maj or breakdown and
low CPV with respect to PV (Figure 1).
Starch pH typi ca Ily fel I between 5 .1 and 7
(Table 2), w ithin the range w here it usuall y
does not affect pastin g properties, as
reveal ed by nonsignifi ca nt Pearson's
correlati on coefficients (not shown).
Collado and Corke (1997) obta in ed a
simil ar RVA pattern for 14 Philippine
sweetpotato clones usi ng the same operating conditi o ns, but with a lowe r ran ge of
variation in pasting parameters, espec ially
in PV (Tab le 2).

Although the impo rtance of amyl ase has

been estab lish ed, vi scoa mylograp hy wa s
found to be a more practical and reli ab le
meth od for predi cting starch nood le quality
(Collado and Corke, 1997). These auth o rs
found a hi gh and positive correlation
betwee n the firmn ess of sweetpotato sta rch
noodl es and the RVA pasting paramete rs of
stab ility ratio (r = 0.95), CPV (r = 0.83), and
HPV (r = 0.73), P < 0.01. Twenty clon es
with the lowest and highest starch paste
stability and amylase co ntent, as well as 40
additi onal clones se lected at rand om from
th e 106, are being eva lu ated for sta rch
noodl e production in As ia at Hong Kong
Uni versity, to gain better informati on o n the
rel ations hip between sweetpotato sta rch
properti es and noodl e quality, and to
identify suitable cl o nes for thi s purpose.
Conclusions
Thi s study is the first eva luation of starc h
content and properti es of a large samp le
from the sweetpotato gene bank. Th ese
studied traits can also be affected by
env ironmental factors such as site, seaso n,
and year, as well as root storage time (Tian
et al., 199 1). Results reported in thi s study,
th erefo re, ca nnot be considered as abso lute

CIP Program Report 199798

285

val ues for each clon e. Nevertheless, th ey

are use ful for gene ral co mpariso n and
prel imi nary screenin g of sweetpota to
ge rmplasm collection s for sui tab le root
qu ality traits. Th e second pha se stud y,
alrea dy being initiated at CIP, ai ms to
eva lu ate the stabi lity of starc h co nte nt and
prop erti es across d iffe ren t envi ro nm ents in
Peru.
Acknowiedgments
Th e authors are gratefu l to th e French
Mini stry of Foreign Affairs for fi nancial
support. They also th ank Olivi er Gibert and
M ari e-Christine Lahon for th e analyses of
starc h amylose content and RVA prop erties
at CIR AD, Montpelli er, France, and Eve line
Boul eje for her contribution to th e va rious
tri als at CIP, Lima , Peru.
References Cited
AOAC. 1995. Official Methods of Anal ys is
of AOAC In ternat io nal, l 6'h Edit ion,
Vol ume II. Food Compos iti o n; A dditi ves;
N atural Contaminants. Cereal Fo ods.
Chapter 32 , p. 1. www.aoac.org.
CIP (International Potato Center} 1998.
Sweetpotato Facts. A compe ndium of key
figu res and ana lys is fo r 33 important
sweetpotato-produ ci ng countries.
Internatio nal Potato Ce nter (CIP), Li ma,
Peru. (Brochu re).
Co ll ado, L.S . and H . Corke. 1997 . Properties of starch noodl es as affected by
sweetpotato genotype. Ce real Chem.
74 (2): 182-1 87.
Hu aman, Z. 1991. Desc riptors fo r
sweetpotato. CIP, AVRDC, IBPGR.
International Board for Pla nt Genetic
resources, Rome, Italy.
M est res C. , F. Maten c io, B. Pon s, M. Ya jid,
and G. Fliede l. 1996. A rap id method for
the dete rmination of amy lase content by
using differential- sca nnin g ca lor im etry.
Starch 48 (1 ):2 -6.
M ok, l.G., Tj intokohadi , L. Ni ngs ih , and
T.D. Hoan g. 1997. Sweetpotato breeding
strategy and germpla sm tes tin g in
Southeast Asia. Internati o nal Potato

286

Sweelpololo

Center Prog ram Report 1995 -1996. p.

104-1 09.
SAS In stitute. 1992. SAS/ STAT users guid e.
Ve rsi o n 6.09, 4'h ed ., Vol. 2. SAS Inst.,
Cary, NC, USA.
Schoc h, T.J. and E.C. Maywald. 196 8.
Preparati o n and properties of va riou s
legume starc hes . Cereal Chem. 45:564 573 .
Scott, G.J. and C. Wheat ley. 1997. Rece nt
Adva nces in CIP's Strategy for Col labora tive Po sth arves t Res earch on
Sweetpota to. In te rnationa l Potato Ce nter
Program Report 1995 -1 996. p. 264-269.
Scott, G. , S. Wi ersema, and P.I. Ferguson .
1992. Prod uct dev elopment for root and
tub er crops. Volume I - Asia. Proceed in gs of the In ternational W orkshop , held
A pril 22-May 1, 1991 , at Visayas State
Col lege of Ag ri culture (V ISCA), Ba yba y,
Leyte, Phi I ippi nes, sponsored by th e
In ternat ional Potato Center (CIP), the
Centro In tern acio na l de Agr icul tura
Trop ica l (CI AT), and the Intern atio nal
Institute fo r Tro pi ca l Ag riculture (llTA).
CIP, Lim a, Peru.
Tian, S.J. , J.E . Rickard , and J.M.V. Blan shard.
1991 . Physicochem ical properti es of
sweet potato starc h. J. Sci. Food Agr ic.
57:459 -491.
W heatl ey, C. 1996 . Produ ct deve lo pm ent
for sweetpota toes in As ia . Subp roject
A nnua l Progress Report October 1 995September 1996. Internationa l Potato
Center. Li ma, Peru. (M imeo).
W hea tl ey, C., L. Lipi ng, and S. Bo fu . 1997.
Enh anc in g th e role of small-sca le
sweetpo tato starch enterprises in
Sichu an, Ch in a. International Potato
Center Prog ram Report 1995-1996. p.
270-275.
Woo lfe, J.A. 1992 . Sweet potato - A n
untapped food resource. Publi shed in
co l laborat ion wi th the Internationa l
Potato Cente r, Peru . Cambridge U n iversity Press, Cam bri dge, UK. 643 p.
Zh ang, X. 1999 . Ag ri cultural process in g,
marketing and co nsumer beha v io ur in
tran sition: Sweetpotato products in
Sichu an Provin ce, P.R. China. Ph.D.
th es is. Wage nin ge n, The Netherland s.
(Unpublish ed .)

Potential of Sweetpotato in Reducing Vitamin A

Deficiency in Africa
V. Hagenimana1, L.M. K'osambo 2 , and E.E. Carey1

Recent studies have associated th e consumption of foods rich in carotenoids with

decreased incidence of certain cancers in
humans (Gester, 1993). That and the
possible role of carotenoids in immunity,
fertility, and early prophylaxis of cardiovascular diseases in livestock , have generated
interest in these compounds (Pfan der,
1992). Carotenoids represent the most
widespread group of naturally occurring
pigments in nature. They are primarily of
plant origin and G-carotene, with few
exceptions, predominates. G-carotene
se rves as an important nutritional component in foods, as a major precursor of
vitamin A, and provides pleasant yelloworange colors to foods (Simon, 1997).
Dietary vitamin A deficiency causes
debi I itati ng health problems such as
xerophthalmia, corneal lesions,
keratomalace, and, in many instances,
death. The World H ea lth Organization
(1995) reports these problems affecting
young children in Africa.
Sweetpotato has been receiving increasin g attention in part because it grows on
so ils with limited fertility, is relatively
drought tolerant, provides good ground
cover, and is often cultivated without
fertilizer or pesticide (Ewell, 1990). Those
qualities are attractive to agriculturalists
and ecologists interested in developing
sustainable food production systems in the
tropics. Also, it has remarkable pro-vitamin
A qualities (Woolfe, 1992). In parts of
West, Central, and East Africa, sweetpotato
is an important staple source of calories and

1 CIP, Nairobi, Kenya.

2 University of Nairobi , Nairobi, Kenya.

is consumed by all age groups. Children,

the group most at risk of vitamin A deficiency, particularly Ii ke sweetpotato (Low et
al., 1997). Widely consumed cultivars,
howeve r, have white or pale yellow flesh
and contain very little G-carotene (Ameny
and Wilson, 1997) . Orange-fleshed
sweetpotato storage roots hi gh in carotenoid s and vitamin A-active G-carotene are
eaten less . Compared to light-colored, low
G-carotene sweetpotato, consumption of
orange-fl es hed sweetpotato roots and
sweetpotato-based processed foods cou Id
provide sustainable, cost-effective, and
much needed vitamin A. Th erefore, the use
of orange-fleshed sweetpotatoes as a food
source of pro-vitamin A merits further
attention.
Fres h swee tpotato roots are bulky and
highly perishable. In Africa they are
commonly consumed fresh, usually boil ed.
These roots are generally not harvested and
stored for extended periods. That is, farmers
keep unharvested mature sweetpotatoes in
the field until they are needed for consumption or sale, a practice called piecemeal
harvest in g (Rose, 1970; Onueme, 1982;
Smit, 1997). In semi-arid areas with a lon g
dry season, in-ground storage is limited by
attacks from sweetpotato weevils (Cy/as
spp. ). Farmers have traditionally chipped
or crushed sweetpotato roots and sun-dried
them for storage and year-round use.
Chippin g, drying, and storing orangeflesh ed sweetpotato can overcome season a I
shortages of pro-vitamin A in the diets of
many low-income African households
during the dry season when there are no
fresh , gree n vegetables. CIP has been
working to make more nutritious
sweetpotato varieties available to African

CIP Progrnrn Report 1997-98

287

countries (C ichuki et al. , 1997 ). Little is

known , however, about the effect on
carotenoid con tents of drying and processing sweetpotato into fl our and flour-related
products.
Some inconclusive reports say that
caroteno id content changes during
sweetpotato storage root growt h and
development (Data et al., 1987). Studying
var iations in carotenoid co ntent, especial ly
pro-vitamin A carotenoid content, during
storage root deve lopm ent is relevant in the
pro cess of maximizing th e ava il abi lity of
that nutrient. Recommendations could then
be made to farmers to start practic in g
piecemeal harvestin g of newly intro du ced,
hi gh-y ielding, orange-fleshed c ultivars .
This work is part of a larger study
reported in part elsewhere. K'osambo et al.
(1998 ) report o n the H PLC carote noid
analys is. In th e HPLC analysis, one ml of
total carote noid extract fro m 2 g of grated
sweetpotato sample was freeze-dried and
reco n stitu te d in hi gh performance I iqu id
chromatog raph y (HPLC) mobile phase of
90% methanol:l 0% tetrahydro furan . Th e
reconstituted samples we re ultr afi ltered
th rough 0.5 m mi crofil ters before inj ect ion
into the HPLC syste m. HPLC analyses were
done as d escribe d by Ruddat and Wi ll
(1985).

profiles of carotenoid stand ard s,

sweetpotato extracts , and co-chromatography indi cated that three sweetpotato
carotenoid peaks at 5.36, 8.6 1, and 14.60
min. correlated to the co-chromatographic
peaks produced by G.-carote ne-5,6 ,5',6'diepoxid e; G.-carote ne-5 ,6-monoepox id e,
and all-trans-G.-carote ne as indicated by
simi l ar rete ntion times (K'osambo et al.,
1998).
Pl predominated in wh ite- or creamfleshed cu ltivars such as KSP 20, Na veto
(CIP44013 l ), and Kemb 10, w here it
formed a sig n ifican t portion of total carotenoids. Th e probable id entity of this
ca rot eno id was postulated on th e basis of
its eluti on pattern. Craft (1992) repo rted that
du rin g non-aqueous reversed-phase H PLC,
xanthop hy lls, that are more po lar, partition
more preferentially in th e pol ar mobile
ph ase and , therefore, elute earlier than the
less pol ar ca rote nes. Early elution of Pl , as
we ll as that of P2, stro ngly suggests Pl is a
xa nth op hy ll , possibly lute in, and P2 a
zeaxanthin (K'osambo et al., 199 8).
Thi s paper reports on the part of the
study undertaken to identify high-y ieldin g
sweetpotatoes that are high in G.-carotene
<rnd study th e effects of dry in g and p ro cess ing o n carotenoid con tent.

Materials and Methods

HPLC resu lts reported in K'osambo et al.
(1998) indicated that a good number of
carotenoi ds occur in sweetpotato root
extracts. Six of these were present in
signifi ca nt amounts with all-trans-G.carote ne predominant, acco untin g for more
than 80% of total carotene in most orangefl es hed sweetpotato roots ana lyzed. A lltrans-G.-carotene, G.-carotene-5 , 6mo noepox ide, G. -carotene-5,6 ,5 ' ,6' diepoxide, and 2 unidentified carote noids
(des ignated Pl and P2 ) we re present in all
cu lti vars analyzed. Th e amou nt depe nd ed
o n the c u Iti var. Chrom atog raph ic profil es of
sweetpotato extracts from different cu Iti va rs
were compared with co -c hromatograp h y
with pure carotenoid. Com par iso n of HPLC

288 Sweelpololo

Plant material
Selected sweetpotato cultivars from
Kenya and Ugand a and others from CIP's
pathogen-tested li st (CIP, 1998) were grow n
in Kabete, Kenya (a ltitude 1,800 m , amb ient
temperature 20.3 3C, and annu al
rain fa ll , 1,046 mm ) using a sta nd ard
rand o m co mpl ete bloc k design. Storage
root flesh co lo rs of th ese cultivars ranged
from w hi te to ora nge. Samples of eac h
cultivar we re taken for carotenoid analysis
in a two-stage random sampling. Three
plants were rand om ly selected and three
medium-sized storage roots taken at
random from eac h plant. These were used
for carotenoid extract ion and analysis.

Roots w ere pee led, and about 2-c m-thi ck

medial transve rse slices were taken from
each root. Th ese were fin ely grated length w ise usin g a c heese grater. Th ese samp les
we re th o ro ughly mixed, packed und er N
into pl asti c bags, and stored at -20( until
used for ca rotenoid extraction.

Effect of root age

Th e effect of root age on ca roteno id
content was assessed by sa mplin g
sweetpotato roots at 12 , 16, 20, and 24
weeks after planting. At eac h stage, three
p lants were randoml y selected and the
largest sto rage root s pi ece meal harvested
from eac h rand o ml y sel ec ted plant. Different pl ants w ere sampled at eac h harvest.

Total carotenoids in processed foods

Sweetpo tato chips we re dri ed at 65 ( in
a fo rced-a ir ove n to a mo isture co ntent of
6-8% . Th e process of producin g dri ed
sweetp otato c hips is th at describ ed by
Hagenim ana et al. (1999). To chec k fo r
changes in tota l carotenoid co ntents during
storage, dri ed c hips from eac h culti va r were
stored in an opaque paper bag and samp led
for flour processing after 3, 6, and 11
month s of sto rage und er ambient air
conditi ons. Buns, c hapatis (fl at unl eave ned
bread) , and man dazis (dou ghnuts) we re
prepared as desc rib ed by H agenimana et
al. (19 98). Ca rotenoid was extracte d
followin g th e procedure described by
Khachik et al. (1992). Tota l caroteno icl s and
r!.-carotene w ere determ i necl spec tro photometri cal ly as descr ibed by lmun gi and
Wabul e (199 0).

Results and Discussion

Carotenoids and vitamin A values of
sweetpotato roots
Large variation was obse rv ed in ca ro tenoicl co ntent of th e cu ltivars studi ed. This
was a refl ection of th e w id e spectrum of the
root fles h co lo r of sweetpotato. W h iteflesh ecl roots such as th ose fro m culti va rs
Mugand e, TIS 2534 (CIP44006 2),
LM88.014 (CIP188001.2), and KSP 20 had
the lowest total ca roteno id . Orange-fleshed

cultivars such as Camote Amarillo

(CIP400014), Japon Tresmesino Se lecto
(CIP420009 ), Kakam ega 4 (SPK 004), and
Zap all o (C IP420027) had the hi ghes t (Table
1). Our res ul ts agree wi th th e co nc lu sio n
th at ca rotenoids, especial ly r!.-carote ne, are
large ly res ponsible fo r the oran ge fl es h
col o r in sweetpotato storage roots (P urce ll ,
196 2; Purce ll and Walter, 1968; A lm eid aMuraclian et al. , 1992; Taka hata et al. ,
19 93). Th e depth of ora nge fl es h co lor was
mainl y a function of the co nce ntration of
all-trans- r~ - ca rot e n e, as was simil arl y
re po rted by Simonne et al. (1993) . Th ese
res ults indi cate th at caro teno icls from
orange- fl es hed sweetpotato are hi ghl y
vitamin A ac tive and are consistent with
tho se of Jalal et al. (1998). Th at work
showed that co nsumption of r!.-carote neri ch foods, mainly in the form o f o rangefl es hed sweetpotato, inc reased serum
reti no l leve ls in Ind ones ian children and
all ev iated signs of v itamin A defici ency.
Sin ce th e ea rl y 1990s, th e main strategy
for co mbating vitamin A defic iency in
Afri ca has been to distribute mass ive-close
cap sul es using donati o ns from UNI CE F
(Kennedy and Oni ango, 199 3). Moth ers
w ho do not iden tify adverse ph ys ica l
sympto ms in th ei r chi ld ren often neglect
returni ng for add iti o nal dos es . Th e sa me
benefit co uld be ac hi eved if children
con sum e sufficient qu antities of r!.-caro tene
and v itam in A-rich foo dstuffs . Thi s is one
of th e most reliabl e, rea dil y ava il abl e, and
most sustainable ap proac hes to co ntro lling
v itam in A defici ency in many rural areas of
Afri ca w here chronic defic ienc ies are still
comm o n (Roels et al. , 1958). Foods such
as dai ry and meat produ cts contain ing
prefo rm ed v itamin A are often too ex pensive for most peo pl e in African co untries.
Th erefore, it is important to make available
more potent and susta in ab le food so urces
of pro-vitamin A carote no ids, such as
oran ge- fl es hed sweetpotatoes, and to
improve th eir produ ct ion, sh elf life, and
consum er acce ptance . Th at co uld make a
tremend ous contribution to improvin g
Afri ca n nutrition and hea lth.

(IP Program Repo rt 199798

289

Table 1. Relationship between the flesh color and carotenoid contents of different sweetpotato cultivars.
Cultivar name
CIP440057 (Tl Bl l)
Cl P400014 (Ca mote amarillo)
Un known
CIP420009 (Japan Tresimesino Selecto)
CIP440015 (W-220)
CIP420010 (Teoboza)
CIP420027 (Zopollo)
CIP420004 (Momolo)
SPK 004 (Kokomego 4)
Cl Pl 87004. l (LM87.009)
CIP440377(NG 7570)
CIP440186 (Tainan No 15)
Cl Pl 87004.2 (LM87.045)
Kemb 10
CIP420047 (Estrella)
Kl48
CIP440224 (CNl 517-139)
CIP440154 (Xiong shu 6)
CIP420053 (Copodito)
CIP420031 (Cascajo morado)
CIP440228 (CARI 9)
CIP400002 (Morado moravi)
CIP440025 (Xushu 18)
CIP420008 (Maria Angolo)
KEMB 33
CIP440162 (Mabrauko)
CIP440078 (TIS70357)
CIP440023 (W-228)
KSPl l
Ex-Diano
CIP440062 (TIS2534)
CIP440160 (Philippine)
CIP440066 (lhuanco)
Cl Pl 88001.l (LM88.002)
CIP440094 (NC 1582)
KEMB 20
CIP440144 (IRA502)
CIP440131 (Noveto)
Con tinued on next page.

290

Sweetpototo

Visual flesh color

Orange (deep)
Orange (deep)
Orange (pale)
Orange (deep)
Orange (deep)
Orange (deep)
Orange (deep)
Orange (light)
Orange (light)
Cream
Cream
Cream
Cream
Yellow(deep)
Cream/ purple
Cream/yellow
Cream/yellow
Cream/white
Purple/white
Purple/ cream
Cream/yellow
Purple/white
Yellow (light)
Orange (light)
Yellow (light)
Yellow (light)
Cream
White/yellow
White
White
White
Cream
White
White
White
White/yellow
Yellow/ purple
White/yellow

Carotenoids (g/100 g fresh root tissue SD)

Total

13-carotene

8823.5654
7195.4 2169
7451.5 718
6931.7 4620
8389.9446
8942.5 2021
4692.0 550
2671.9 83
2715.5 212
1700.3 70
1009.0 500
1700.3 213
1289.2 109
1006.9 402
604.l 71
834.6 80
736.2 83
460.9 70
722.5 72
784.8 50
473.3 49
290.2 13
168.2 23
390.9 28
336.3 20
137.0 11
248.8 39
239.2 50
159.6 9
191.7 24
138.3 2
158.5 5
112.l 10
106.8 11
99.7 15
93.4 21
69.8 27
68.5 29

7983.9 339
6271.0 1312
6234.3 92
6175.0 4251
6023.l 465
5144.5 1001
4085.0 10
2217.2 311
2130.0 109
1071.3 101
847.0 351
832.l 286
747.4 76
627.0 219
485.8 14
473.3 lll
448.4 90
274.0 42
249.l 78
199.3 124
174.4 93
161.9 17
lll.8 45
110.6 10
99.7 37
49.8 9
39.3 7
37.4 7
19.6 2
19.1 2
16.7 2
Traces
Traces
Traces
Traces
Traces
Traces
Traces

Tobie 1 (cont.)

CIP440164 (K5 1/3251)

ClPl 88001.2 (LM88.0l 4)
KSP 20
Mugande

White/yellow
White
White
White

W oo lfe (1992) and Low et al. (1997)

suggested th at cu lti va rs hav in g more th an
100 pg retin o l equi va lent per 100 g fres h
roots we re good so urces of v itamin A.
H agenim ana et al. (1999 ) reported th at
cul tivars suc h as Kakamega 4 (S PK 004),
Zapa ll o, Japo n Tresmes in o Se lecto, un known, W -22 0 (CIP44001 5), and TIB 11
(C IP440057) have suffi cient levels o f retin o l
equivalents to meet thi s criteri a. Their
culti vation, co nsumption, and use in
different d ishes shoul d be enco uraged in
combating nutritional v itamin A deficiency
in Afr ica.

Effect of root age

Sixtee n to twenty-week-o ld roots
co ntain ed hi gher caroteno id co ncentrat ions
th an younger roots (Tabl e 2), except the low
P.,-ca rotene co ntent cultivar TI S 253 4, w hi ch
reac hed its hi ghest ca roteno id co ntent at 24
wee ks. Th ese d ifferences in total caroteno id co ntent betwee n you ng and o lder
roots depended o n the cul ti va r. Sixteenweek-o ld roots from Kakamega 4 (S PK 004)
were twofo ld hi gher in tota l ca rote no id
content than 12-week-o ld roots. O rangefl eshed cultivar Japo n Tresmes in o Se lecto
had two-third s of its total ca roteno id
co ntent ava il ab le after ju st 12 wk. Co ncentrat ion of to tal ca rotenoid co ntinu ed to

62.3 22
56. l 22
Traces
Traces

Traces
Traces
Traces
Traces

increase up to th e 24th wk in low-carotenoid -con tent culti vars TIS 2534

(CIP44006 2) and Kemb 10 (Tabl e 2).
Th erefore, to rece ive the max imum prov itamin A benefits from th e sweetpo tato,
piecemea l harves tin g and consumption of
roots from Japo n Tres mes i no Se lecto co u Id
beg in aft er 1 2 wk, after 16 wk from
Kakamega 4 (SPK004), and after 20-24 wk
from th e lower ca rote no id content cu ltiva rs.

Total carotenoids in processed

sweetpotato products
Addi ti on of o range-fl es hed sweetpotato
roots to bun s, chapati s, and mandaz is
increased th e total ca rotenoid s and co ntent
of th ese produ cts (Fi gure 1). Us in g
sweetpotato fl our in the rec ipe was one of
the best mea ns of in creas in g total caroten oid s and P.,-ca rotene, fol lowed by usin g
boil ed and mashed sweetpotato, and th en
ra w and grated roots. Add in g sweetpotato
also improved the produ cts' co lo r, giv in g
th em an attracti ve ye ll ow appea rance.
Boi li ng th e roots redu ced th e total
carotenoid content by 20%, w hereas drying
roots at 65"( fo r 12 hours redu ced tota l
caroteno id co ntent by 30%. Storin g dried
chips for 11 mo nth s reduced tot a I ca rotenoid co ntent an add iti ona l 11 %, from

Table 2. Total carotenoid contents of 4 sweetpotato cultivars grown at Kabete, at different root ag es.

Total carotenoid content (mg carotene equiv./100 g fresh root SD)

Cultivar
KEMB l 0
Japan Tresimesino Selecto
Kakamega 4
TIS 2534

12 weeks
0.43
3.39
0.85
0.05

0.02
0.14
0.16
0.0 1

16 weeks
075
5.47
2.58
0.08

0.04
0.34
0.15
0.03

20 weeks
0.91
6.91
3.09
0.06

0.04
0.39
0.38
0.01

24 weeks
l.13
5.17
3.03
0.16

0.43
0.08
0.08
0.01

CIP Program Report 1997-98

29 1

Tota I pro-vitamin A
(~ car o tene

equivalent g/100 g pro duct)

D

2500

Raw & grated

Coo ked & mashed

Sw eetpotato flou r
Wheat !lour

2000

Buns

Chapatis

Mandaz is

Processed products

Figure 1. Improvement of pro-vitamin Ain

processed foods by incorporation of

sweetpotato cv. CIP420027.
70% to 59 % (Hagenimana et al. , 1999).
Although fresh sweetpotato carotenoid
levels decreased during the drying process,
carotenoid concentration was much higher
in the dried products than in fresh roots .
Good agronomic performance and
consum er acc eptabi I ity of processed
products combined with the high carotenoid levels of orange-fleshed swee tpotato
culti v ars provide an elegant and welladapted mechanism for combating vitamin
A deficiency among resource-poor urban
and sweetpotato farming communities in
Africa.

Conclusions
Orange-fleshed sweetpotato roots
contained higher total carotenoid and ~
carotene content than white- and creamfleshed roots. Carotenoids from orangefleshed sweetpotato are highly vitamin
A-active and their consumption in Africa
where vitamin A deficiency is prevalent
should be e ncouraged.
Twel v e weeks after planting, th e y ield
and amount of pro-vitamin A present in
roots of orange-fleshed culti v ars were
high enough to potentially provide

292

Sweetpototo

adequ ate dietary pro-vitamin A. Piecem eal harvesting should be started then.
Incorporation of flour made from orangefleshed sweetpotato roots into buns,
chapatis , and mandazis si gnificantly
enriched the products in pro-vitamin A.
Result s of this study suggest that incre ased consumption of orange-fleshed
sweetpotatoes in either fr esh or processed form can contribute to all eviating
dietary deficienc y of vitamin A.

Acknowledgement
V. Hagenimana ' s research is partially
funded by the Department for International
Development (DFID ) of the United Kingdom, Crop Postharvest Research
Programm e [R7036]. DFID accepts no
respon sibility for any information provided
or v iews ex pressed. HPLC carotenoid
analyses we re conducted at the International Li ves tock Research Institute Nairobi
Kenya.
'
'

References Cited
Almeid a-Muradian , L.B. de, MVC. de
Penteado , and V.L.P. de Ferreira. 1992.
Relationship betw een carotenoid content
and Hunter colour param eters of Brazilian sweetpotato. Revista Espanola de
Ciencia Y Technologia de Alim entos
36:611-619.
Ameny, M.A. and P.W. Wilson. 1997.
Relationship between Hunter color and
~-carot e n e content in white-fleshed
African sweetpotato. J. Sci. Food Agric.
73:301-306 .
CIP (Int ernational Potato Center). 1998.
Pathogen-tested sweetpotato germplasm
for distribution. Fourth Edition. CIP, Lima,
Peru.
Craft, N .E. 1992 . Carotenoid reversedphase HPLC methods: Referen ce compendium. Methods Enz y mol. 213:185205.
Data, E.S., F.G. Villamayor, Jr. , Reoma, and
D.R. Anzano, 1987. Yield and chemical
composition of cassava and sweetpotato

at different ages of harves t. Philippin e J.

Crop Sci. 12: 44.
Ewell, P.T. 1990. Sweetpota toes in Eastern
and South ern Africa. Pape r prese nted at
the workshop on sweetpotatoes in th e
food systems of Eastern and Southern
Afri ca held in 1990 in Nairobi , Kenya.
Gester, H. 1993. Anticarcinogenic effect of
com mon ca rotenoids . Int. J. Vitamin
Nutr. Res. 63:93-121.
Gichuki , S.T., V. H age nimana, J. Low, M.A.
Oyunga, E. Carey, and J.N . Kabira.
1997. In vo lv in g women group s in th e
eva luation of introduced high 5-carotene
sweetpotato vari et ies. Proceedings of the
4'h Triennial Congress of the African
Potato Association held 23-28 Feb. 19 97
in Pretori a, South Africa. p. 228-233.
Hageniman a, V., E.G. Karuri , and M.A.
Oyunga. 1998. Oil content in fri ed
sweetpotato pro cessed produ cts. J. Food
Proces. Preserv. 22:123-137.
Hagenimana , V., E.E. Carey, S.T. Gichuki,
M.A. Oyunga, and J.K. lmun gi. 1999 .
Carotenoid contents in fresh, dried and
process ed swee tpotato produ cts. Ecol.
Food Nutr. 37(5):45 5-474 .
lmun gi, J.K. and M.N. Wabule. 199 0. Som e
chemical chara cteristi cs and availability
of vitamin A and vi tamin C from Kenyan
varieties of papaya (Carica papaya L.).
Ecol. Food Nutr. 24: 115-120.
Jal al, F., M.C. Nesh eim, Z. Agus, D. Sanjur,
and J.P. H ab icht. 1998. Serum retinol
co nce ntrations in children are affected
by food sources of 5-caroten e, fat intake,
and anthelminti c drug treatment. Am. J.
Clinical Nutr. 68:623-629 .
Kenn edy, E.T. and Oniango, R. 1993 .
Hou sehold and presc hool er Vitamin A
co nsumption in southwestern Kenya. J.
Nutr. 12 3:84 1-846.
Kh ac hik, F. , C.R. Beecher, M.B. Coli, and
W.R. Lusby. 1992 . Sep aration and
qu antitation of carotenoids in foo ds .
Meth. En zy mol. 213:347-359.
K'os ambo, L.M., E.E. Carey, A.K. Misra, J.
Wilkes, and V. Hagenim ana . 1998.
Influence of ag e, fa rmin g site, and
bo iling on pro-vitamin A content in

sweet potato (ipomoea batatas (L.) Lam.)

storage roots. J. Food Comp . Anal.

11 :305-321.
Low, J., P. Kinyae, S. Gichuki, M.A.
Oyunga, V. H age nim ana , and J. Kab ira.
199 7. Combating vitamin A defici ency
throu gh the use of sweetpotato. Res ults
from Phase I of an action resea rch
project in South Nyanza, Kenya. Internation al Potato Ce nter (CIP), Lima , Peru.
Onueme, l.C. 1982. Th e tropi ca l tuber
crop s: Yams, cas sav a, sweet potato, and
cocoyams. En glish Language Book
Society and John Wiley and Sons,
Chichester, UK.
Pfander, H . 1992 . Carotenoid s: An
ove rview. Meth. En zy mol. 213:3-13.
Purce ll , A.E. 1962 . Carote noid s of goldrush
sweetpotato flakes. Food Technol. 1 6: 99-

102.
Purcell, A.E. and M.W. Walter, Jr. 196 8.
Carotenoids of centennial va riety
sweetpoato, lpomoea batatas. J. Agric.
Food Chem. 16:769 -770.
Roe ls, O.A., M. Trout, and R. Dujacqui er.
1958. Carotene balances on boys in
Rwanda where vitamin A defici ency is
prevalent. J. Nutr. 65:115-127.
Rose, C.J. 1 970. Comparison of sin gle and
progress ive havestin g of sweetpotato
(lpomo ea batatas (L) Lam.). Papua N ew
Guinea Ag ri c. J. 30 :61 -64 .
Ruddat, M. and 0.H. Will Ill . 198 5. Hi gh
performance liquid chrom atog raphy of
ca roteno ids. Meth. Enzymol. 111 : 1 89-

200.
Simon, P.W. 19 97. Pl ant pi gments for
co lor and nutrition. HortScience 32:12-

13.
Simonn e, A. H, S.J. Kays, P.E. Koehler, and
R.R. Eilenmiller. 199 3. Assessment of 5ca rotene content in sweetpotato breeding lin es in relation to di etary requirements. J. Food Comp . Anal. 6:336-345.
Smit, N.E. 1997. Th e effect of the indi genous cultural practices of in-ground
storage and pi ece mea l harvesting of
sweetpotato on yield and qu ality losses
ca used by sweetpotato weevils in
Ugand a. Agri c. Environ . Sys. 64:191 -

200 .

(IP Prngrnm Repo11199798

293

Takahata , Y., T. Noda, and T. Nagata. 1993 .

H PL C determination of P.i-carote ne
co nte nt in sweetpotato cul tivars and its
relationship wi th color va lu e. Japa nese J.
Breed. 43:421-427.
WHO. 1995. World H ea lth Orga nis ation
Mi cro nutrient Defici ency In formation

2 9 4 Sweelpolalo

System. G lobal prevalence of v itam in A

deficiency. M DI S w orking paper 2.
W HO/N UT/95.3, Geneva, Switzerland.
Woo lfe, A.J. 1992. Sweetpotato: A n
untapped food resource . Cambrid ge
University Press, Cambridge, U.K.

Development of a Sweetpotato-Based Instant Weaning

Food for Poorly Nourished Children Six Months to
Three Years Old
N. Espinola\ H. Creed-Kanashiro 2 , M.E. Ugaz 2 , M. van Hal1, G . Scott1

In Peru , as in man y oth er deve lopin g

countri es, infant malnutriti o n is a se ri ous
pro bl em . A rece nt n ation al hea lth and
demog raphic survey .found th at ove r 25%
of children und er 5 yr of age suffe r fr o m
chro ni c m alnutrition (END ES, 1996) .
U sually th e percentage of m aln ouri shed
c hildren is hi gher in rur al areas, but du e to
th e greater urban popul ati o n in Peru th e
majority o f und ernouri shed chil d ren li ve in
th e peri -urban area s.
Th e criti c al period for a c hild 's gro w th
and deve lo pm ent is betwee n 6 an d 24 mo
durin g the transition from a di et based on
mo th er's milk to one akin to t he fa mily di et.
Durin g this period the c hild is susce ptibl e
to malnutri t ion and defici enc ies in se veral
mi cronutri ents that can lea d to gro wth
reta rd atio n and suscepti b i I ity to in fecti ous
di seases (Creed-Kanashiro et al., 199 0).
Th e situ ation is the res ult of a nutriti o n
ga p: th e differen c e betwee n th e qu antity
and qu ality of food th at th e infa nt rece ives
vs hi s/her d aily requirem ents. Thi s is th e
res ult of infant feeding pattern s. Th e custom
is to se rv e liquid or semiso lid preparations ,
su c h as soups and broth , th at moth ers
co nsid er to be the b est for th eir c hildren.
Food s that fi 11 them but do not ad equ ately
nouri sh th em. Thus, a c hild rece ives food
low in energy and nutri ent d ensity. A lso, the
target population has limited purc has in g
powe r, lacks nutrition kn ow ledge, and does
no t necessa ril y orient fo o d purc hases
towa rd m ax imum nutriti o nal benefit.

1 CIP, Lima, Peru.

2 Inst itute de ln ves tigac i6 n Nutr icional, Lima, Pe ru .

Th ere is a good po tenti al for improv in g

th e di et, and th ereby a child 's nutriti o nal
statu s, thro ugh th e in co rp o rati o n o f
swee tp o tato (/pom oea batatas) into th e di et.
Sweetp otato is pro du ce d in Peru , is eas il y
access ibl e, and lo w -c ost. Children like it.
Oran ge-fl es hed vari eti es are rich in hi ghly
bio ava il abl e ~-caro t e n e (pro-vitamin A ).
Th e use of sweetp otat o offe rs th e poss ibil ity
of ex p andin g th e use of the cro p w hil e
contributin g to th e preve ntion of v itamin A
defic iency prese nt in seve ral sectors o f th e
Peru v ian po pul ati o n (Pres idenci a d e la
Republi ca d el Peru , 1998) .
In thi s co ntex t, a co ll aborati ve proj ect
was condu c ted by CIP with the Nutriti o n
Res ea rc h In stitute (JIN), a non governm ental
orga ni za ti o n w ith ove r 30 yr experi ence
w orkin g w ith vuln erab le population s in
Peru . Th e obj ective of the proj ect was to
full y ex pl o it the sweetpotato's nutrition al
attributes and palatab i lity by deve lo pin g a
low-cos t in sta nt we anin g food as a co mpl ement to breast m ilk fo r children 6 mo to 3
yr old (Fi gure 1 ).

Materials and Methods

Weaning food ingredients
Fres h sweetpota to roots of comm erc i al
culti var Jonath an (o range) and the new
culti va r Imperi al-I NI A (YM89-05 2, w hite),
and co rn , ri ce, soybea n, pea, quinoa, and
w hea t fl ours we re u sed to prepare th e
weanin g food m ixes . In addition , po w dered
wh o le milk, freeze-dri ed egg, milk wh ey,
and soybea n pro tein i sol ate (PAS), (PR O
FAM 64 6) (Archer Dani els Midland Co mpan y) we re in c luded as protein sources.

CI PPwgram Report 1997-96

295

1993 ). Th e M ini stry's stan dard s are those

establi shed fo r in stant produ cts used in
th e governm ent' s suppl ementary feedin g
prog ram fo r presc hoo l-age children.
Development and preparation of the
instant weaning food
Th e wea nin g foo d was deve loped takin g
1nto cons iderati on th e:
optimum perce ntage of sweetpotato th at
shoul d be in c lud ed in th e formul ati on
and th e best fo rm of sweetpotato fl our
(raw, precoo ked, or extrud ed),
ava il ab ility of th e in gredi ents based on
th e ca lcul ation of different com bin ations
fo r th eir nu tri tio nal va lu e and cos t, and
laborato ry determination of th e o rga nolepti c characteri sti cs, si ze of rati on,
amount of wa ter needed, and energy
density.
Si x fo rmul as we re se lected th at d iffered
by type of sweetpotato (white or ora nge
fl es h) and fl our (w heat or soybea n) used.

f;gure 1. Ayoung lad enjoying his weaning food

during the acceptability trial.

W hi te sugar and vege tab le o il we re used
as energy so urces, and a premi xture of
v ita mi ns and min erals fo rmul ated for in stant
food fo r child re n (Pro du ct# 2760960
Roc he S.A.) was add ed. Ronoxa n A (Roc he
S.A.) was in c lu ded as an antiox ida nt fo r the
fo rm ula and c innam o n (Sd spray-dry,
Prod uct # 0803 5, M o nta na S.A.) to improve
fl avo r.
Nutritional criteria for an
instant weaning food
. Th e nutrition al criter ia used for deve lop111g th e wea ni ng food we re:
th e reco mm end ed intake of energy and
protein fo r children und er 3 yr of age
and other nutrients (FAO/W H O/ UNU ,
1985; FAO/ OM S, 199 1 ),
the nutritional defi c it th at ex ists between
th ese requirements and actu al foo d
co nsumption by in fa nts in Peru , and
the minimum nutriti o nal standa rd set by
the Pe ru vian M ini stry of Hea lth (M O H ,

2 96

Sweetpololo

Indu stri al prepa ration of the in fa nt food

was ca rri ed out in AS A Alim entos S.A. , a
pri vate co mp any loc ated south of Lim a in
Can ete, Peru. Two p1eparation meth ods
we re used.
1. Precoo ked sweetpotato flour mi xed w ith
th e oth er fl ours (ri ce, co rn , w heat, or
soybean), soybea n protein (PAS), o il , and
Ronoxa n A. Th ese were feel direct ly in to
th e drum-drye r to obtain precoo ked
fl akes.
2. Sweetpotato puree was prepared by
was hin g, pee lin g, and steam coo kin g
fres h roots . Th e puree w as th en co mbin ed w ith ri ce, co rn , w heat, or soybea n
fl ours and soybea n protein (PAS) that had
prev iously been mi xed w ith oil and
Ronoxa n A . Thi s mi xture was fe el direct ly
into th e d rum-drye r to obtain precoo ked
fl akes.
In both cases, the fl akes we re groun d
and then mi xed w ith th e pow dered mil k,
suga r, c inn amon, and th e premi x of v itamin s and min erals to form the w ea nin g
fo od .

Nutritional and microbiological analysis

The following analyses were performed:
Proximal ana lysis (total protein, fat, ash,
fiber, and carbohydrates), using the
standardized method of AOAC 1983.
~ - carotene (AOAC, 1995).
Percentage of gelatinization ana lysis
made by DEMSA, Lim a, Peru (us in g the
method of the Mppimbianti Company,
1987).
Digestibility in vivo ana lysis conducted
by the Nutrition Department, Ag rari an
University, La Molina, Lima, Peru, using
the balance of rat method.
Amino ac id conte nt (AOAC, 1994).
Microbiological ana lysis (total aerob ic
flora; total col iform s: Staphylococcus
aureus, Bacillus cerus, and Clostridium
perfringens; molds; yeast; and sa lm onella) using 3M Petrifilm plates (Beuchat
et al., 1991) and plate dilution (AOAC,
1995).

Acceptability trial
An acceptabi 1 ity tria 1 was conducted
w ith mothers and infants li ving in periurban areas in Huascar and Bayovar, Canto
Grande, San Ju an de Lurigan cho District,
Lima.

The tri al co nsisted of determining th e

acceptab ili ty of the six different form ul as by
infants and their mothers through short-term
feeding trials (4 d), exp lor in g the incorporation of th e weaning food into the daily diet
of the infants, and eva lu at in g mothers'
perceptions of ease of preparation and use.
A double-blind design was used in
cod in g the formulas. Tests were conducted
with 120 mother-child pairs, 97 of which
comp leted the study. Twenty-three dropped
out because of illness (11) or parents did
not w ish to continue (12) . Before the tests,
the mothers were provided w ith 900 g of
weaning food formula (sufficient for 10 d), a
bowl, and two measures (one for wate r and
the other for the formula). They were also
in structed on how to prepare the weaning
food to avoid the risk of contamination.

Results and Discussion

Weaning food formula
The composition of the six weaning
foods is shown in Tabl e 1 for the eq ui valent
of a daily ration of 90 g. Weaning foods of
two fat levels (30% and 33%) were prepared to optimize the industrial process.
The 30% leve l was prepared w ith ora nge
sweetpotato on ly (white sweetpotato was
not ava ilabl e at the time of thi s preparation)
(Espinola et al., 1998).

Sweetpotato root was the only fresh

ingredient used in th e formulation. It
represented 24% of the total composit ion
on a dry basis. All ingredients were compl ete ly cooked, resulting in a mixture which
needs only hot wa ter to prepare in the
home.
In acco rd ance with the FONCODES
(1996) requirement for 20% of total protein
to be prov id ed from an imal sources,
powdered whole milk was includ ed.
Soybean protein (PAS), rice, maize, soybean, and sweetpotato flours provided th e
remaining protein. Vegetable oil co ntributed add ition al fat and refined suga r was
added as a comp lem entary source of
energy (up to 11 % of carbo hydrates). A
mixture of vitamins and minerals was
added to meet the required levels. Ronoxan
A (which contains a-tocofero l and palmitate
of L-ascorbi lo) was added to prevent
oxidation. Cinnamon was added as a
natura l flavor. Cinnam on and Ronoxan A
are permitted addi ti ves fo r processed
products for ch ildren under 3 yr o ld,
according to Codex A lim entariu s, 1994
(FONCODES, 1996).
Nutritional and microbiological quality
Th e protein, carbohydrate, and energy
density levels of the six weani ng foods are
all within the range of requirements
stipulated by official standa rd s (Tab le 2).
The fa t and fiber contents vary by type of
mixture. Indeed, the 97% level of ge lat inization in all form ulas surpasses the official
requirement, indicating that the foods were

CIP Progrnm Report 1997-98

297

Table 1. Ingredients of six weaning food formulations, CIP, Limo , Peru , 1997.
Weaning foods' (g/90g rationb)
30% fat

33% fat

Dry ingredients
White sweetpotato'
Orange sweetpotato'
Wheat flour
Soybean flour

22.07
6.89

Rice flour
Maize flour
Whole powdered milk
Soybean protein (PAS)
Sugar
Vegetable oil
Ronoxan A
Mix of minerals/vitamins
Tricalcium phosphate
Magnesium sulfate
Natural cinnamon

8.36
6.89
27.54
3.93
5.90
7.38
0.01
0.09
0.55
0.25
0.15

21.79

21.67

22.07
6.92
4.92
9.84
9.84
27.54
l.97
5.90
6.89
0.01
0.09
0.55
0.25
0.15

22.07
7.38

22.07
4.92

4.92
8.40
6.92
27.69
3.96
5.93
7.42
0.01
0.09
0.55
0.25
0.07

10.33
9.83
27.53
1.97
5.90
6.88
0.01
0.09
0.55
0.25
0.07

8.85
7.87
27.54
3.93
5.90
5.41
0.01
0.09
0.55
0.25
0.15

11 .80
l 0.82
24.59
2.95
5.90
5.90
0.01
0.09
0.55
0.25
0.15

" Weaning foods: l) orange sweetpotato + wheat, 2} orange sweetpototo + soybean, 3) white sweetpototo + wheat, 4) white
sweetpototo + soybean, 5) orange sweetpotato + wheat, 6) orange sweetpotato + soybean.
b To be consumed in two feedings/day.
' Before drum drying: orange sweetpotato fresh wt 129 g, white sweetpototo 116 g.

compl etely cooked , easy to digest, and met

establish ed standards fo r insta nt food
(FONCODES, 1996).

WHO/ UNU , 1985) for a given amino ac id

(Tab le 3) . Th ere we re no limiting amino
acids and th e protein quali ty was adequate.

Th e reti nol content of formu lations 1, 2,

5, and 6 (Table 2) exceeded offic ial requirements for children 6 mo to 3 yr o ld. More
importantl y, the results indi ca te that root
process in g did not significant ly affect
~-ca rotene co ntent. Woolfe (1992) reported
a similar finding in her ea rli er rev iew of th e
literature . Hence, inc lu sion of orange
sweetpotato roots in th e wea nin g food
permits either a contribution of v itamin A
beyo nd th at official ly req uired, or a redu ction in th e leve l of retinol required in th e
fo rmul ations .

Throu gho ut preparation-mi xin g,

cook ing, pac kagin g, and transporting-ca re
was taken to prevent co ntamination of th e
product. Microb io logica l analyses show th e
wea nin g food to be we ll w ithin th e prescr ibed limits of ae rob ic flora, total
co l iforrns, m o lds, and yeast, and sa lm onell a, permitted for an infant food , in
acco rdan ce w ith th e FONCODES (1996)
req uirements.

The amino ac id content of eac h formula

exceeded th e FAO chemi ca l score (FAO/
298

Sweetpototo

Howeve r, w hen th e acceptabi lity tri als

we re started (6 wk after preparation ), th e
bags co ntainin g formula 1 were externa ll y
contam in ated by fun gi. As a res ult, thi s
formu lat ion was dropped from th e acceptability tri als.

Table 2. Peruvian Ministry of Health's requirements for an instant weaning food, and the results of nutritional
analyses of six weaning foods, CIP, Lima , Peru, 1997.

Weaning foodsb

Requirement'
Fat33%
2

Ration
Energy (kcal)
350-400
Moisture (%)
< 5%
Protein (12-15% total kcal) 48-60 kcal
l 00-120 kcal
Fat (25-30% total kcal)
Sugar (g)
Up to 11 % CHO
::;2.0
Fiber (g)
Energy density (kcal/g)
1 to 1.5
Gelatinization (%)
> 94
Retinal (g)
400
Chemical score(%)
> 85%
cx-tocoferol (mg/kg)
300 max.
0

375.0
2.7
55.0
83.7
11.0
1.7
1.1
97.8
510.0
> 100%
60

Fat30%

390.0
3.2
54.4
114.6
10.9
1.7
1.2
98.5
563.0
> 100%
60

396.0
2.7
59.2
129.6
11.6
2.8
1.2
98.3
75.3
> 100%
60

385.0
3.4
49.6
113.4
10.8
2.5
1.1

98.0
101.0
> 100%
60

373
377.0
2.6
3.2
53.2
47.8
73.4
88.0
9.7
9.9
1.5
1.8
1.1
1.1
98.0
97.6
518.0
500.6
> 100% > 100%
60
60

FONCODES (1996).

Wea ning foods: 1) orange sweetpototo + wheat, 2) orange sweetpotato + soybean, 3) white sweetpotato + wheat, 4) white
sweetpotota + soybean, 5) orange sweetpototo + wheat, 6) orange sweetpotato + soybean.

Table 3. Analysis of the amino acids of the weaning foods formula vs chemical score for preschoolers.

Sweetpotato weaning foods'

2

Chemical
scoreb

(mg/g protein}
lsoleucine
Leucine
Lysine
Methionine + cystine
Phenyalanine +
tyrosine
Threonine
Valine
0

35.5
81.1
67.3
30.6

49.6
87.7
66.9
35.4

49.1
92.1
70.0
37.8

44.4
84.6
67.2
29.8

48.1
89.9
65.0
42.8

28
66
58
25

86.0
41.0
38.1

83.7
45.9
52.5

94.3
45.9
52.l

89.7
43.2
48.l

92.8
44.4
52.4

63
34
35

Weaning foods: 2 = Orange sweetpotato + soybean (33% fat), 3 = white sweetpotato + wheat (33% fat), 4 = white
sweetpotato + soybean (33% fat), 5 = orange sweetpotato + wheat (30% fat), 6 = orange sweetpototo + soybean
(30% fat).
For children under 5 yr old (FAO/WHO/UNU, 1985).

CIP Program Reporl 1997-98

299

Acceptability by mothers and children

Acceptability of the wea ning foods was
evaluated using the method deve loped by
the llN, based on IS0-6658 (ISO, 1985) and
Watts et al. (1992). Each formulation was
evaluated for the number of times consumed each day, amount consumed, and
the mother's perception of her child's
acceptance of the formula. A score was
assigned to each of these criteria to develop
an acceptabi I ity rating. Seventy-five percent
was considered good acceptance, in
accordance w ith the requirements established by FONCODES (1996). On this
basis, formulations 2 (o range sweetpotato
+soybean flour), 4 (white sweetpotato +
soybean flour), and 5 (o range sweetpotato
+wheat flour) were rated acceptable.
Consumption of the weani ng foods
increased on the 2nd and 3rd days as
children became accustomed to them.
Consuming these foods between main
meal s changed the child's eating patterns,
principally through replacing between meal
snacks, e.g., banana, porridge, cookies, or
milk. In general, the wean ing food contributed to increasing the c hild's total food
intake.
Almost all mothers found the formulations beneficial. Eighty-seven percent said
that the instant preparation, using just
warm, boiled wa ter and mixing, helped
them in their daily activities, both in saving

time and in satisfying the child. These

mothers perceived their children to be quiet
and well fed. A few children disliked the
food, accounting for the 13% of mothers
who did not consider it beneficial.
With the results of the acceptability
trials, formulations 2, 4, 5 we re selected
(Table 4). Formulation 2 had lower fiber
and higher retinal that Formulation 4, the
only ration with a higher acceptabi I ity
rating. Formulation 5 closely followed 2 in
acceptabi I ity, but had a slightly lowe r
retinal and was made with only 30% fat.

Production cost
We prepared preliminary estimates of the
costs and returns to production of the
wea ning food. They were based in part on
actual trial runs at the Canete plant,
assuming a raw material price for
sweetpotato of US$0.03/ kg. We made one
set of calculations assuming the plant
running at 100% capacity and a second set
assuming a 70% use factor. The costs of
wa ter, gas, or electricity we re not quantified.
Assuming the plant running at 100%
capacity, the cost per ration ranged from
US$0.16 to US$0.17 for th e six formulations. Assuming operation at 70% capacity,
the cost per ration ranged from US$0.16 to
US$0.18 (Table 4). The estimated costs
using both assumptions were well below

Table 4. Final characteristics for the selection of weaning food, 90g ration, CIP, Lima, Peru, 1997.

Weaning
food'
4
2
5
6
3
0

300

Acceptability

Fiber

Fat

Retinal

Modified

(%)

(g)

(%)

(g)

costs (USS)

82
80
75
70
66

2.46

33
33
30
30
33

101
563
518
500
75

1.74

150
1.80
2.84

Weaning foods: 2) orange sweetpotato + soybean, 3) white sweetpotato

5) orange sweetpotato + wheat, 6) orange sweetpotato + soybean.

Sweetpototo

+ wheat, 4) white sweetpotato + soybean,

0.17
0.17
0.18
0.16
0.18

the MOH acceptable prices for a weaning

food of US$0.26 per 90 g ration. They were
also below the average price of US$0.19
per 90g ration for the llN-FONCODES
supplementary feeding program for preschool children.
Formulation 6 was the lowest cost
(US$.16), fol lowed by 2 and 4 (US$.1 7).
Although it is not the lowest cost, Formulation 2 was judged to be the best candidate
for success because it combined high
acceptability, low fiber, and high retinol .
Conclusions

Sweetpotato roots are an excel lent

ingredient for weaning foods, which can
be used to contribute to optimum
nutrition of children in this critical
growth period.
Using a system of precooking
sweetpotato before converting it into
flour, orange-fleshed sweetpotato
retained most of its Ei-carotene content
resulting in a weaning food with a higher
reti no I content.
Formulation 2 is nourishing (high retinol,
low fiber), easy to prepare, and readily
accepted by children and mothers. Its
low cost makes it competitive with other
foods offered in the government supplementary feeding programs for preschool
children.
References Cited

AOAC 1995. Official methods of analysis

of AOAC International. 161h edition. 2nd
Volume. AOAC, Arlington, VA, USA.
Beuchat, L.R., B.V. Nail, R.E. Brackett, and
T.L. Fox. 1991. Comparison of the
petrifilm yeast and mold culture film
method to conventional methods for
enumerating yeast and molds in Foods. J.
Food Prot. 54:6:443-447.
Creed-Kanashiro, H.M., K.H. Brown, G.
Lopez de Roman, T. Lopez, and R.E.
Black. 1990. Consumption of food and

nutrients by infants in Huascar (Lima),

Peru. Am. J. Clin. Nutr. 52:995-1004 .
ENDES (Encuesta nacional de demografia y
salud). 1996. National survey of demography and family health . National
Institute of Statistics and Information
Science (INEI), DHS, Lima, Peru.
Espinola, N., H. Creed-Kanashiro, M.E.
Ugaz, and M. van Hal . 1998. Development of a food supplement with
sweetpotato for children from 6 months
to 3 years. Department of Social Sciences Working Paper 1998-8. International Potato Center (CIP), Lima, Peru.
FAO/WHO/UNU (Food and Agricultural
Organization of the UN/World Health
Organization/United Nations University).
1985 . Energy and protein requirement.
Technical report series 724. WHO,
Geneva, Switzerland.
FAO/OMS (Organizacion Mundial de
Salud). 1991. Requirements for vitamin
A, iron, fol ate, and B 12. Estudios FAO,
Alimentacion y Nutricion No 23, FAO,
Rome. 124 p.
FONCODES (Fondo de Compensaci6n y
Desarrollo Social) . 1996. Bas es de la
invitaci6n para la selecci6n de productos
alimentarios. Programa social de
suplementaci6n alimentaria para ninos
menores de 3 anos en el departamento
de Ancash. Convenio No 601271. Lima,
Peru. 33 p.
ISO (International Organization for
Standarization). 1985. Sensory analysis,
methodology. IS0-6658. General
guidance. 1st Edition. ISO, Geneva,
Switzerland. p 1-14.
MOH (Ministry of Health). 1993. The
composition of food of greater consumption in the Peru. 6th Edition. National
Institute of Nutrition, Lima, Peru. 63 p.
Mppimbianti Company, 1987. Spectrophotometric method for degree of gelatinization. Italy. Unpublished document.
Presidencia de la Republica del Peru .
1998. Plan nacional de nutrici6n y
alimentaci6n. Nutrici6n al alcance de
todos. Comisi6n Tecnica de

CIP Ptogtam Re pott 1997-98

30 l

Coordinaci 6 n Multisectorial de Politica

Nutricional. Supported by UNICEF/
PROMUDEH /PRONAA . 70 p.
Watts B.M., G .L. Ylimaki, L.E. Jeffe ry, and
L.G. Elias 1992. Metodos sensoriales
basicos para la evaluaci6n de alim en tos.

302 Sweetpotato

In ternat ional Development Rese arch

Center (IDRC), Ottawa, Ontario, Canada.
167 p.
Woo lfe, J.A . 1992. Sweetpotato - An
untapped food resources. Cambridge
University Press, Cambrid ge, UK.

AFLP Assessment of Sweetpotato Genetic Diversity in

Four Tropical American Regions
D.P. Zhang, M. Ghislain, Z. Huaman, J.C. Cervantes1, and E.E. Carey2

Sweetpotato (Jpomoea batatas (L.) Lam.)

was originally domesticated in the New
World (Austin, 1988). The exact site of
origin and domestication of the sweetpotato
has not been well defined; neither has the
wild ancestor of this species been found.
Based on the numerical analysis of key
morphological characters of sweetpotato
and the wild lpomoea species, Austin
(1988) postulated that the center of origin of
I. batatas was somewhere between the
Yucatan Peninsula of Mexico and the
mouth of the Orinoco River in Venezuela.
Linguistic evidence suggests that
sweetpotato was dispersed to the rest of the
world in three lines. The kumara line is
prehistoric and based on lexical parallels
between the Quechua name and
Polynesian word kumara. That could
explain the transfer of sweetpotato by
Peruvian or Polynesian voyagers from
northern South America to eastern
Polynesia around AD 400. The batata line
dates from the first voyage of Columbus in
1492, which resulted in the introduction of
West Indian sweetpotatoes to Western
Europe. Portuguese explorers transferred
sweetpotatoes grown in western Mediterranean Europe to Africa, India, and the East
Indies in the 16th century. The plant was
recorded in South China by 1 594 and in
southern Japan by 1698. The camote (name
derived from the word camotli in the
Mayan language Nahuatl) line represents
the direct transfer of Mexican
sweetpotatoes by Spanish trading gal Ieons
between Acapulco, Mexico, and Manila,

1 CIP, Lirna, Peru.

2 CIP, Nairobi, Kenya.

Philippines, in the 16th century (O'Brien,

1972; Yen, 1982).
Today sweetpotato, with an annual
production of 124 million t, is the world's
seventh most important crop after wheat,
rice, maize, potato, barley, and cassava
(CIP, 1996). It is widely grown in tropical,
subtropical, and warm temperate regions.
Developing countries account for 98% of
the world's sweetpotato production. Its
wide adaptability on marginal land and rich
nutritional content provide an enormous
potential for preventing malnutrition and
enhancing food security in the developing
world.
CIP has carried out numerous
sweetpotato-collecting expeditions in Latin
America and the Caribbean since 1985.
Donations from other countries and the
transfer of sweetpotato collections previously maintained in other international
research centers further expanded the
collection. The sweetpotato gene bank
maintained at CIP now contains 5,526
cultivated accessions from 5 7 countries,
2,589 of them from Latin America. Peru
alone contributed 1,099 accessions
(Huaman and Zhang, 1997).
The success of any genetic conservation
and breeding program is dependent on
understanding the amount and distribution
of genetic diversity present in the gene
pool. For example, information on diversity
distribution and domestic history is crucial
for constructing core collections. They are a
limited subset of accessions, derived from a
larger germplasm collection, chosen to
represent the genetic spectrum in the whole
collection . Maximum genetic diversity is

CIP Prngrom Report 199798

303

the key to establishing a good core subset

(Brown, 1989).
In the present study, we app li ed amplified fragment length polymorphism (AF LP )
fingerprinting to a group of Latin Ame rican
sweetpotato cultivars maintain ed at CIP to
answer two basic questions rel ated to
genetic structure in this gene pool:
1. Is there any regional differentiation in the
sweetpotato diversity from tropical
America? For purposes of this study, we
defined four regions: (1) Central
America, (2) Colombia-Venezuela, (3)
Peru-Ecuador, and (4) Caribbean.
2. What is the geographical pattern of
genetic diversity in these reg ions?
Thi s survey is part of CIP's molecular
assessment of genetic di versity in the
sweetpotato collection. Th e infmm ation
obtained w ill enhance our und erstand ing of
the genet ic structure in the American
sweetpotato gene pool and help to rationalize its management.

Materials and Methods

Plant materials
Four to ten sweetpotato cu lti va rs were
randoml y selected from eac h country of the
four regions, w hich result ed in a total of
eighty cultivars with a geographical coverage ranging from Mexico to so uth ern Peru
(Table 1 ). The plant materials were obtained
from the sweetpotato gene bank at CIP.
H ea lthy young leaves we re collected from
accessions maintained in a screen house
and in vitro. The leaf tis sue was immediately imm ersed in liquid N and then
transferred to -80C, freeze-dried, and
stored at room temperature until use.

AFLP protocol
A modified DNA miniprep procedure of
Doyle and Doyle (1990) was used to extract
DNA. Th e AFLP techniqu e is described by
Vos et al. (1 995). DNA restriction uses the
enzyme combination fcoRl /Msel. Afte r
adapter li ga tion , DNA fragments are
amplified using polymeras e c hain reaction

304

Sweelpototo

(PCR). Primer annealing is targeted at the

adapter and restriction-site sequence.
Three- nu c leot ide extensions o n both fcoRI
and Msel primers cause selective amp li fica tion of fragments. Primer comb in atio ns
we re chosen to produce a high numb er of
unambi guous polymorphisms in a set of 10
sweetpotato genotypes te sted. The eig ht
primer combinations that we re used are
presented in Table 2.

Scoring and data analysis . Different

fragments produced wit h each primer were
treated as a unit character and numbered
sequentially. Genotypes were scored for the
presence (1) or absence (0) of each fragment. Only those fragments w ith medium
or high intensity we re taken into account.
Fra gments with the same mobility on the
gel, but w ith different intensities, were not
distin gu ished from each other w hen
cultivars were compared. Mo nomorph ic
fragments we re not scored.
From these data, a matri x of pairwise
Euclide an distances defined by Excoffier et
al., (1992) and a matrix of simil arity based
on simple matching coefficients (Sneath
and Sokal , 1973) we re calculated using the
package RAPDistance vl .03 (Armstrong et
al. , 199 5) .
Th e similarity-based relation ships
between the 80 cultivars we re then presented in a dendrogram following th e
un weig hted pair group method with
arithmetic mean algorithm (UPGMA) using
SAHN clustering analysis of NTSYSpc
(Rohlf, 1992).
Based o n the Euclidean distances, the
analysis of molecular variance (AMOVA)
pro ced ure (Exco ffier et al., 1992) was
applied to estimate variance components
for AFLP phenotypes. Indi vidual variation
was partitioned w ithin and between
re gion s. The va riance components of
interest were extracted and tested using
nonparametric permutational procedures.
Va riation between regions was then
partitioned into pairwise distance between

Table 1. Name, collection number, and origin ol 80 sweetpotato cultivars from the collection held at CIP.
No. Name

l
2
3
4
5
6
7
8
9
l0
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42

Cuba 9
Linea 96
Cuba 3
Linea 475
Unknown
Columbiaorado
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Camote Blanco
Santa Rosa (A)
Camote Morado
(ornate Morado
Unknown
Unknown
Unknown
Unknown
Gumbs
Lover's Nome
Six Weeks White
Ras ta
Papa Blanca
Jiquima
Nigua
Pata de Gallina
Unknown
3-c-n
Tunita
Verna de Huevo
Unknown
De Dulce
Blanco Papa
Camote de Dulce
Bunuelo
lrizo Blanco
Unknown
Niguilla
Unknown
Camote Dulce

Collection

Country

No.

of origin

CTX 34
CTX 9
CTX 5
CTX 16
DLP 3874
DLP 3892
DLP 3837
DLP 3834
DLP 4617
DLP 4678
DLP 4686
OLP 4675
GUA 494
GUA STRA
GUA 940
GUA 948
OLP 4545
DLP 4494
DLP 4521
DLP 4558
SVG 12
SVG 24
SVG 27
SVG 8
INVT 72
INVT 51
INVT 63
INVT 74
CSDA 22
3-C-N
DLP 1567
SOSA 33
DLP125 7
DLP1149
DLP 1435
DLP1231
DLP 1484
DLP 1405
DLP 1186
DLP 1397
DLP1157
DLP 1498

Mexico
Mexico
Mexico
Mexico
Panama
Panama
Panama
Panama
Nicaragua
Nicaragua
Nicaragua
Nicoragua
Guatemala
Guatemala
Guatemala
Guatemala
Honduras
Honduras
Honduras
Honduras
El Salvador
El Salvador
El Salvador
El Salvador
Cuba
Cuba
Cuba
Cuba
Dominican Rep
Dominican Rep
Dominican Rep
Dominican Rep
Ecuador
Ecuador
Ecuador
Ecuador
Ecuador
Ecuador
Ecuador
Ecuador
Ecuador
Ecuador

No.

43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76

77
78
79
80

Name

Catemaco
Unknown
Chaco Morado
Unknown
Unknown
Unknown
Unknown
Patatilla
Chaco Columbiaorado
Morado
Unknown
Unknown
Blanco
Morada
Unknown
Exquisita
Camota
Morado
Blanca
Unknown
Horse Money
Tis-2544
Clipperu
Tis-5093
Papota
Sunny
Mojave
Toquecita
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown
Unknown

Collection

Country

No.

origin

DLP 868
DLP 869
DLP 2869
DLP 2902
DLP 2884
DLP 842
DLP 824
DLP 806
DLP 2868
DLP 2896
DLP 1870
OLP 1685
OLP 1731
DLP 1793
OLP 1858
DLP 1736
OLP 1771
OLP l 011
DLP 1792
OLP 2104
OLP 3205
DLP 3247
DLP 3211
DLP 3232
SPV 44
SPV 43
SPV 65
SPV 55
DLP 3433
DLP 253
RCB IN-199
DLP 909
DLP 1921
RCB IN- 90
ARB 355
ARB 455
DLP 2298
ARB 97

Venezuela
Venezuela
Venezuela
Venezuela
Venezuela
Venezuela
Venezuela
Venezuela
Venezuela
Venezuela
Colombia
Colombia
Colombia
Colombia
Colombia
Colombia
Colombia
Colombia
Colombia
Colombia
Jamaica
Jamaica
Jamaica
Jamaica
Puerto Rico
Puerto Rico
Puerto Rico
Puerto Rico
Peru
Peru
Peru
Peru
Peru
Peru
Peru
Peru
Peru
Peru

CI P P1ogromRepo1t 1997-98

305

Table 2. Primer combinations used in AFLP analysis of sweetpotato diversity.

Number
E34/M33

E+ 3/M + 3 nucleotide extensions

5'-GACTGCGTACCAATTC AAT-3'
5'-GATGAGTCCTGAGTAA AAG-3'

E38/ M38

5'-GACTGCGTACCAATTC ACT-3'

E33/M38

5'-GACTGCGTACCAATTC AAG-3'

E38/ M32

5'-GACTGCGTACCAATTC ACT-3'

5'-GATGAGTCCTGAGT AA ACT-3'
5'-GATGAGTCCTGAGTAA ACT-3'
5'-GATGAGTCCTGAGTAA AAC-3'
E36/M37

5'-GACTGCGTACCAATTC ACC-3'
5'-GATGAGTCCTGAGT AA ACG-3'

E38/ M34

5'-GACTGCGTACCAATTC ACT-3'

E36/M33

5'-GACTGCGTACCAATTC ACC-3'
5'-GATGAGTCCTGAGTAA AAG-3'

5'-GATGAGTCCTGAGTAA AAT-3'

E36/M34

5' -GACTGCGTACCAATTC ACC-3'

5'-GATGAGTCCTGAGTAA AAT-3'

reg io ns to exa min e th e reg io nal co ntri buti o n to total mol ec ul ar d ive rsity (Excoffier et
al. 1 199 4).

Results
AFLP profile, similarity and cluster
analysis
Th e eight primer co mbin ati o ns generated
2 10 po lymo rphic and c lea rl y sco rab le
frag ments ac ross the 80 culti va rs. Th ere
w ere no reg io n-spec ific markers (prese nt in
o ne reg io n but abse nt in th e oth er). H oweve r, th e mea n gen eti c simil ariti es va ry
co nsid erab ly betwee n reg io ns, rangin g fro m
0.592 fo r Ce ntral Ameri ca to 0.660 for
Peru -Ec uado r (Tabl e 3).
Th e ove rall geographi c prox imi ty is low
as refl ected by th e U PG M A dendrogram
(Fi gure 1). No clu sters c lea rl y di stin gui shed
c ulti va rs fro m Central A meri ca, Ca ribbea n,
and Co lo mbi a-Venez uela. Within eac h
c lu ster, howeve r, culti va rs fro m th e sa me
geographi ca l ori gin s tend to gro up togeth er.
In sharp co ntrast w ith th e c lose r re lati o nship between culti va rs fro m Ce ntral

306

Sweetpololo

A meri ca , th e Ca ribbea n, and Co lom bi aVe nez uela, the Peru vian-Ecu ado ri an
culti va rs we re c lea rl y distin gui shab le from
th e oth ers. All of th e 40 culti va rs fro m Peru Ecu ado r we re gro uped into o ne sing le
c lu ster at a hi gher si mil arity leve l.
Th e mea n si mil arity and th e U PGMA
dendrog ram both suggest that Ce ntra I
A meri ca n c u lti va rs apparentl y have th e
hi ghest geneti c di versity, w hereas Peru Ecuado r c ulti va rs have th e lowest. Th e
Car ibbea n and Co lo mbia-Ve nez uela
c ulti va rs have an intermediate leve l. Th e
c lu ster analys is demo nstrated a substa nti al
reg io nal d iffe renti ati o n in th e Latin A mer ica n sweetpotato c u ltivars. Swee tpotatoes
fro m Peru-E cuado r co ntributed most to th e
reg io nal di ffe renti ati o n.

Analysis of molecular variance

The results of AMOVA (Tabl e 4) revea led
large w ithin-reg io n va ri ation s, w hi c h
accou nted fo r 9 1.3% of th e total mo lec ul ar
va ri ance. Thi s large w ithin-reg io n va ri ation
co ul d be th e co ntributi o ns fro m two
sources. Th e fi rs t is th e ex pected large

Table 3 . Number of AFLP-generated polymorphic markers and mean similarity in sweetpotato cultivars from four

Latin American regions.

Region

Accessions

Polymorphic fragments

Mean similarities

(no.)

within regions

within regions
(no.)

Central America
Colombia-Venezuela
Caribbea n
Peru-Ecuador

24
20
16
20

va ri atio n betwee n indi v id ual cul tiva rs,

beca use sweetpotato is an outcross in g
hexap lo id, therefo re the va ri ation due to
sex ual reprodu ction is large. Th e seco nd
source is the betwee n-co untry va ri at io n
w ithin a reg ion. In thi s study, eac h reg ion
in cl uded more than o ne country, and the
between-country va ri atio n was co nfounded
by the betwee n-indi v idu al va ri ati o n. The
betwee n-country compo nent was not
separated in the analys is beca use the
sa mple sizes in some co untri es, pa rticularl y those from Central A meri ca and the
Ca ri bbea n, w ere too sm al I to rep rese nt one
co untry.
The betwee n-reg ion va ri ati on, although
accounting for onl y 8.7% of tota l molecul ar
va ri ance, was stati sti ca ll y sign ifica nt using a
nonparametri c test (Excoffier, 1994). That
means there is m eas urabl e d ivergence
betwee n the fo ur reg io ns. Ind iv idua ls from
d iffe rent reg io ns we re mo re d ive rgent, on
ave rage, than individu als from the sa me
reg io ns.
Pa rt it io nin g between-reg ion va ri ab ility
into pa irw ise di stance gave a clea rer
picture of the extent to w hi ch each reg ion
con tributes to the total mo lecul ar d iversity
(Tab le 5). Pairw ise di stances between
reg io ns differ greatl y. Th e shortest d istance
was betwee n Ce ntral A meri ca and the
Car ibbea n (0.02 34), w hereas the longest
d ista nce w as betwee n Ce ntral Ame ri ca and
Peru-Ecuado r (0.1 269). Peru -Ecuador has
the greatest mea n di stance to the rest of th e

21 3
210
202
195

0.592
0.611
0.639
0.660

reg ions (0. 1007), in d icatin g th at sweetpotato from Pe ru -Ec uador has th e most
differenti at ion thu s co ntributed most to th e
betwee n-reg io n va ri ati on. Thi s res ult fully
agreed w ith the obse rved groupin g patte rn
of Peru- Ecuador culti va rs in th e clu ste r
analys is.
There is a signi fica nt geographi c pattern
of di versity d istri buti o n (Tab le 4). The
Central A meri ca reg ion co ntain s th e most
intern al d ive rsity (0. 1990), w hereas the
Peru- Ec uador reg ion co ntains the least
(0.1 677) . The inte rm ed iate within-reg ion
di ve rsities we re fo und in Colombi aVenez uela (0.1837) and th e Caribbean
(0.1795). Thi s pattern is th e sam e as th at
based on mea n simil arity and clu ster
analys is. It thu s co nfirms the hi gher leve l of
genetic di vers ity in Central A meri ca. Th e
di versity leve l decreased as sweetpotato
was di spersed from th e center of ori gin
toward th e so uth and the east.

Discussion
Th is study is the first case of ap pl y ing A FLP
fin gerpri nts fo r d ive rsity assessment in
sweetpotato. The hi gh po lymorphi sm of
A FLP makes it a powe rful tool for
genotypi ng a large numbe r of access ions. It
is suitab le for geneti c diversity assessment
in large sweetpotato gene banks.
Au stin (1988) postul ated that the ce nter
of ori gin of I. batatas was somewh ere
betwee n the Yucata n Penin sul a of M ex ico

CIPProgrom Reporl 1997-98

307

Simple match similarity coefficient

0.48

0.64

0 .80

0 .96

1.12
OLP 3874
OLP 3892
OLP 2869
OLP 2868
OLP 2884
OLP 4521
OLP 4558
GUA 494
GUA 948
OLP 1011
GUA STRA
OLP 4678
OLP 2902
CSDA 22
SOSA 33
OLP 2896
GUA 940
INVT 72
INVT 74
CTX 34
OLP 3232
INVT 63
OLP 1567
SPV 65
3-C-N
SVG 27
SVG 8
OLP 3247
OLP 4686
OLP 4675

Panama
Panama
Venezuela
Venezuela
Venezue la
Honduras
Honduras
Guatemala
Guatema la
Colombia
Guatemala
Nicaragua
Venezuela
Dominic Republic
Dominic Republ ic
Venezuela
Guatemala
Cuba
Cuba
Mexico
Jamaica
Cuba
Dom inic Republic
Puerto Ric o
Dominic Republic
El Salvad or
El Salvador
Jamaica
Nicaragua
Nicaragua

INVT 51
CTX 9
SPV 43
SVG 24
CTX 16
SPV 55
SVG 12
SPV 44
OLP 4494
OLP 1736
OLP 3211
OLP 3205
OLP 4545
OLP 4617
OLP 868
OLP 869
OL P 1793
OLP 1771
OLP 1792
OLP 1397
ARB 455
OLP 1149
OLP 1435
OLP 1405
OLP 1257
OLP 1498
OLP 253
OLP 3433
RCB IN-199
OLP 2298
ARB 355
RCB IN-90
OLP 909
OLP 1921
OLP 1231
OLP 1484
OLP 1186
OLP 1157
ARB 97
OLP 842
OLP 806
OLP 824
OLP 1870
OLP 1858
OLP 1685
OLP 1731
OLP 2104
OLP 3837
OLP 3834
CTX 5

Cuba
Mexico
Puerto Rico
El Salvador
Mexico
Puerto Rico
El Salvador
Puerto Rico
Honduras
Colombia
Jamaica
Jamaica
Honduras
Nicaragua
Venezuela
Venezue!a
Co lombia
Colomb ia
Colomb ia
Ecuador
Peru
Ecuador
Ecuador
Ecuador
Ecuador
Ecuador
Peru
Peru
Peru
Peru
Peru
Peru
Peru
Peru
Ecuador
Ecuador
Ecuador
Ecuador
Pe ru
Venezuela
Venezue la
Venezue la
Colombia
Colombia
Colombia
Colombia
Colombia
Panama
Pa nama
Mexico

Figure 1. UPGMA dendrogram of 80 sweetpotato varieties based on simple matching coefficients from AFLP
mar ke rs. J = Regional differentiation of Peru-Ecuador varieties and the lower internal diversities in
the region.
308

Sweetpotoio

Table 4. AMOVA format for the extraction of components of AFLP variation of sweetpotato clones between
regions, and between individuals within regions.

Variation source

Df

SSD

MSD 0

Variance component

Totalh {%)

P value'

Between reg ions

Individuals within regions

3
76

1.593
14.129

0.531
0.1859

0.0174
0.1859

8.55
91.45

< 0.001

Central America
Colombia-Venezuela
Peru-Ecuador
Caribbean
Total

23
19
19
15
79

4.5758
3.67 45
3.1868
2.6920
15.722

0.1990
0.1837
0.1677
0.1795

Peru-Ecuador

Caribbean

a. Mean squared deviations.

b. Percent of total molecular variance.
c. Probability of obtaining a larger component estimate. Number of permutations

1000

Table 5. Genetic distances between sweetpotato cultivars from four Latin American regions
(Distances = PhiST between pairs of regions) .

Regions

Central America

Colombia-Venezuela

Central America
Colombia-Venezuela
Peru-Ecuador
Caribbean

0.0000
0.0391
0.1269
0.0234

0.0391
0.0000
0.1130
0.0622

0.1269
0.1130
0.0000
0.0622

0.0234
0.0622
0.0622
0.0000

Mean distance to other regions

0.0631

0.0714

0.1007

0.0493

and the mouth of th e O rin oco Rive r in

Venezue la. That is where /. trifida mi ght
have been c rossed with another putative
ancestor, /. triloba, and mi ght have produced th e wild ancestor of/. batatas. By at
least 2,500 BC, the cu ltigen had most like ly
been spread by the loca l people to a lmost
the limits for cultivation in Ce ntra l and
So uth Ameri ca that ex isted when th e
Europeans arrived nea rl y 4,000 years later.
The grea ter molecul a r dive rsity in Ce ntral
Ameri ca provides strong evidence th at
Centra l Ameri ca is th e prim a ry center of
diversity. It shou ld a lso be th e center of
orig in as well , conside ring th e richness of
wild lpomoea spec ies that a re close ly
related to sweetpotato.

The mu c h lowe r mol ec ular diversity

found in Pe ru-Ecuador suggests that
cu lti va rs in this region constitute a subset
from the ce nter of prim a ry dive rsity. Thi s
res ult refl ects the traditi o na l patte rn of c rop
distribution from the primary cente r of
ori gin . As the c rop dispersed a lo ng mi gra tion and trade routes, and through exchanges between native commun iti es, the
di versity leve l decreased due to the limi ted
donor popul ation s. A similar pattern was
de monstrated in ou r prev ious wo rk on a
group of Papua New Guinea sweetpotatoes
(Zh ang et a l. , 1998).
Sweetpotato, howeve r, is an a ncient c rop
in Pe ru (Enge l, 1970; Ye n, 1974). The

CIP ProgmmReporl 1997-98

309

uniqu e anc ient agriculture and oth er

ethnographic factors ma y have led to a
greater morphologica l di vers ity based on
relatively narro w don o r di ve rsity. Th at
could exp lai n w hy Peru- Ecuador hosts a
large number of sweetpotato culti va rs but
has a lower leve l of diversity. Th e natura l
and hum an selection fo rces also ca used a
signi ficant regio nal d ifferentiati on , w hi ch
was reflected as th e greate r genetic di stance
betwee n Peru-Ecu ador and the other
regions. Th erefo re, Peru-Ecuador cou ld be
consid ered as a secondary ce nter of
sweetpotato di versity.
References Cited

Armstrong, J., A. Gibb s, R. Peakall , and G.

Weil ler. 1995. RAPOi stance, Package
Manual. Vers ion 1.03. A ustralian Na tion al Uni ve rsity, Canbe rra.
Austin , O.F. 1988. The taxo nom y, evo lutio n
and ge netic diversity of sweetpotatoes
and related w ild spec ies, In: Grego ry, P.
(ed. ). Exploration , maintenance, and
util iza ti on of sweetpotato genetic
resources. CIP, Lima , Peru. p. 27-60 .
Brown, A.H.O. 1989. Core collection: A
pract ica l approac h to ge net ic res ources
manageme nt. Genome 3 1 :8 18-824.
CIP, 1996 . Sweetpotato facts. In te rn ati o nal
Potato Center, Lim a, Peru. (Brochure.)
Doyle, J.J. and J.L . Doy le. 1990. Iso lati o n
of plant DNA from fres h tissue. Focus
12: 1 3-15.
Excoffie r, L. and P.E. Smo use. 1994. Using
allel fr equ encies and geog raph ic subdi v ision to reco nstruct ge ne trees w ithin a
sp ec ies: Molecular va riance pars im o ny.
Genetic s 136:343-3 5 9.
Excoffier, L., P.E. Smou se, and J.M . Quattro.
1992. Ana lysis of mol ec ul ar va rian ce
inferred from metr ic di sta nces among

3 10 Sweetpototo

ONA hap lotypes: App li catio ns to human

mitoch o ndrial ONA restriction data.
Genetics 1 31:479-491.
En ge l, E. 197 0. Exp lora tion of th e Chi lea
Canyon. Cur. Anthropol . 11 :55 -58.
H uaman, Z. and O.P. Z hang. 1997 .
Sweetpotato. In : Fu cc i I lo, D. , et al. (eds).
Biodi vers ity in tru st - Conservati o n an d
use of plant genetic reso urces in CG IAR
Cente rs. Cambr idge U ni ve rsity Press,
Cambridge, UK. p. 29-38.
O ' Bri en, P.J. 1972 . Th e sweet potato : Its
origi n and disp ersa l. Arn. Anthropo logist
7 4: 34 3-3 65.
Roh lf, F.J . 1992. N TSYS-PC numerica l
taxonomy and multi va riate analys is
syste m, Ve rsi on 1.70 . O w ner's manual.
Exte r Publications, Setauket, NY, USA .
Sneath, P.H.A. and R.R. Sokal. 197 3.
N um eri cal taxonomy. The prin ciples and
pract ice of nu merica l c lass ifi cat io n. W .H .
Freema n, San Franc isco, CA, USA.
Vos, P., R. Ho ge rs, M. Bleeke r, M . Reijans,
T. va n der Lee, M. Hornes, A . Fri j ters, J.
Pot, J. Peleman, M. Kuiper, and M .
Zabeau. 1995. A FLP : A new tec hnique
fo r ON A fin gerprinti ng. Nuc. Ac. Res .
23: 44 07 -4414 .
Yen , O.E. 1974. Th e sweetpotato and
Oceania. Bi shop M useum Bull. (Ho nolulu) 236 :1-3 89.
Yen, O. E. 1982. Sweet potato in hi sto rical
perspect ive. In : Villarea l, R.L. and T.O.
Griggs (eds. ). Sweet potato, Procee dings
of First Internati o nal Symposium. AV ROC
Publ. No. 82- 172. p . 17-30.
Zh ang, O.P. , M. G h is lain , Z. Hu aman, A.
Golmirza ie, and R.J. Hijman s. 1998.
RAPO va riation in sweetpotato [lporn oea
batatas (L. ) Lam] cul tivars from South
A meri ca and Papu a New Guin ea. Gen.
Res. Crop Evo lution 45:27 1-277 .

Introduction and Use of Exotic Germplasm in the

Chinese Sweetpotato Breeding Program
D.F. Ma1, H.M. Li1, and I.G. Mok2

Sweetpotato (/pom oea batatas) was intro du ced to Chin a ove r 400 yea rs ago. Its hi gh
yield, w id e adaptability, tol erance of abi otic
stresses, and ve rsatili ty ha ve co ntributed to
its tradition al and co ntinuin g imp ortan ce in
Chin ese ag ri culture. Sweetpotato rank s
fo urth amo ng crop s in pl anted area, about
5.6 million ha, with mea n yi eld s higher
than those alm ost anyw here else in th e
world. In th e 19 50s and 1960s, th e crop
was used mostl y fo r foo d and fee d, w ith
only 10% of total produ ction go ing to
sta rch extractio n and foo d pro cess in g. In
th e 1 990s, th at has ri se n to about 45 % of
tota l produ cti on. Gi ve n the ch anges in use
pattern , it is c lear th at hi gh dry matter (OM )
or hi gh starch co ntent is becomin g more
imp ortant. M ost Chin ese culti va rs have low
or mod erate DM co ntent of <3 0% (XSPRC,
1993) . O ve r th e la st 20 yea rs, ma ny
prov in cial breedin g prog ram s wo rked to
in crease DM co ntent and yi eld w ith li ttle
success. Gi ve n th e lac k of progress , th ere
appea rs to be a need fo r an infu sion of new
so urces of geneti c di ve rsity.
Chinese Sweetpotato Breeding:
Past and Present

Ove r th e past 400 yea rs, a num be r of loca l

culti vars emerged du e to sel ecti o n pressure
by fa rm ers and env ironm ental changes.
Forty-ei ght w ere li sted in th e Ca talog of
Sweetpo ta to Culti va rs in Chin a (X SPRC,
1993), w hi ch co ntain ed 184 im po rtant
culti vars. Th ose culti va rs pos sess tolerance
of bioti c stress and res istance to abioti c
stresses . Th ey are still gro w n in small areas
in So uth Chin a. H owever, most of th em had

1 Xu zhou Sweetpotato Research Center, Xu zhou, Chin a.

2 CIP, Bogar, Indon esia .

low y ield and poor adaptability w hen

pl anted away from their plac e of ori gin.
Th at ha s bee n a se ri ous co nstraint to th e
expansion of sweetp otato produ cti on. In
1938 , Nan cy H all w as introdu ce d from th e
USA. Prelim inary eva lu ations were condu cted in Na njin g and Sichuan. In 1941 ,
th e Japane se cultivar O kinawa 100 wa s
introd uced and soon beca me the most
w id ely grown in north ern Ch ina. Nancy
H all and Okin awa 100 becam e important
in both produ ction and in breedin g.
Modern breeding programs
M odern sweetpotato breedin g beg an in
th e late 1940s and ea rl y 1950s w hen th e
popul ation ex pl os ion and seve re fo od
shortages led the gove rnm ent to em ph as ize
sweetpotato producti on. Major efforts w ere
devoted to sweetpotato breedin g in provin ces w ith large sweetpota to pl antin g areas .
A number of new cultiv ars w ere deve lop ed
usin g Okin awa 100, N ancy Hall , and som e
loca l cu lti va rs as parents. M ost of th e new
culti va rs had medium to low DM co nte nt.
Xu shu 18 is th e culti va r most wid ely grown,
w ith a planted area o f 1.5 milli on ha. Its
DM co ntent is about 25% (s umm er croppin g in sand y loa m so il in Xu zhou) and
extractabl e starch content is 11-1 5%
(XS PRC, 1993). Variou s in stitutes released
many cu lti vars. On e of most success fu I after
Xu shu 18 w as Nanshu 88, whi ch w as
released in 1988 by th e Nan chong Agri cultural Resea rch In stitute. It is, howeve r,
grow n onl y in Sichu an Provin ce.

To in cre ase produ ct ion and effi ciency,

key breedin g units in Chin a we re orga ni zed
into a nati onal collaborative sweetpotato
breedin g pro j ect. Greatl y improved col laboration among prov in cial- and c ity-l eve l

CIP Pw gwm Repmt 1997-98

3 11

research uni ts resu lted. Co nve ntio nal

hyb rid ization is sti ll th e pr in c ipal approach
for sweet potato breeding. Since the late
1970s, Ji angsu Aca dem y of Agricultura l
Sciences (AAS) and Shandong AAS paid
more atte ntion to col lectio n, evaluat io n,
and use of wild -relati ve spec ies. Some
new ly deve loped culti vars, such as Sushu 2
and Pushu 83-91, were bred with th e
introgress ion of genes from w il d relatives .
New cu lti va rs , Yi 306 , Lushu 7, and Yushu
7, we re se lected from c rosses with th e
Japan ese cultivar M in amiyutaka , w hi ch has
12 .5% of wi ld speci es ge nes in its pedigree.
M utation breeding with 6 Co radiation used
to indu ce somaclona l var iation has been
used to improve sweetpotato. Th is work
was do ne at Beijing Agricu ltu ral University.
As a res ult, Nongda 601 , hi ghl y res ista nt to
b lack ro t ca used by Ceratosys tis fimbriata,
was selected fro m Xushu 18; and, No ngda
321 -10, w ith resistance to stem nematode
(or brow n rin g caused by Oitylenchus
destructor), was se lected from Ning 12 - 17.
Yantai Institu te of Agricultural Sci ences
se lected a number of culti va rs using fast
neutron radiation. Th ese se lected cu lti vars
are hi gh y ieldin g and ha ve hi gh v itamin C
co ntent. The Xuzho u Sweetpotato Research
Center (X SPRC) selfed elite c lo nes, and
se lec ted two superior clones, Xuzi 39 and
Pen gwe i S1- i 2-48 . Non e of th ese efforts,
howeve r, res ul te d in a cul ti var sup er ior to
Xushu 1 8.
In rece nt years, postharvest process in g of
sweetpotato has becom e an important way
for farme rs to increase th eir in co me. Th at
requires a hi gh- y ieldin g sweetpotato hi gh in
OM conte nt. Existing cultivars do not meet
thes e dema nds. Dise ase-res istant culti vars
with high starch content and h igh OM y ield
are urgently needed.

Inbreeding tendency in sweetpotato

breeding
Na ncy Hall and Okinawa 100 have been
w id ely used as parental mater ial si nce the
beginning of the mod ern breedin g prog ram .
According to Lu (1990), w ho analyzed the
ped igree of cultivars li sted in th e Cata log of
Sweetpotato Cult iva rs (198 1), 41 cu ltiva rs
312 Sweelpototo

have been released from crosses wi th

O kinawa 100 as a parent, and 33 from
Nancy H all . Th irty-one culti va rs were
deve loped from cross combi nati ons
betwee n the two, inclu din g important
cultivars suc h as Li z ixiang, Yiwohong ,
Fen gs houhuang, and Huab ei 52-45. Nancy
Hall and Ok in awa 100 we re domin ant
among parents in breedi ng prog rams in
Beiji ng and Nan j ing. In Beijing, Hu abei
11 7, H uab ei 166, Bei j ing 553, Beij ing 284,
Beijing 169, and othe rs were se lected fro m
crosses of these two parents . More than 70
cul tiva rs were se lected from the crosses of
Nancy H al l and Okinawa 100 after prov in cia l bree din g units we re establi shed in th e
early 1950s. Som e of them , suc h as
Yiwohong, Beij in g 553, and Lizixi ang a1e
still planted in some p laces.
In tu rn , Xushu 18, w hich has 75 % of its
genes from Na ncy H al l and Okina wa 100,
has been w idel y used as parental mate ri al.
Although the use of Nancy H al l and
Okinawa 100 led to the developm ent of
many su ccess fu l cultiva rs , the genet ic
bac kg round of sweetpotato has become
narro w and lim ited. From 1986 to 1990, 21
of 23 cul tivars release d in China carried a
substanti al num ber of genes fro m Na ncy
Hall and Okinawa 100 . Nans hu 88 and Ji
1 8-1 we re the on ly two excep tions. Of 21
new cu lt ivars registered from 1990 to 1995,
onl y Sushu 4 and Sushu 6 w ere not genetica l ly related to th ese two parents.
Th e General See d Company of Chi na
(1989) li sted 42 cult iva rs gro w n on more
than 667 ha. Of these, 26 we re consid ered
closely relatt: d to Nancy Hall o r Okina wa
100 o r both , in that they have more than
12 .5% of thei r genes in commo n.
Sweetpotato breedin g programs al I over
Ch ina ha ve ex tensi ve ly used Na ncy H all,
Okinawa 100, Xushu 18, and oth er c losely
related cultiva rs as paren ts. This practice
caused severe inbreed ing during sel ection,
and limited th e release of new, pro m ising
cultivars , es peciall y tho se w ith hi gh OM
content. Alt hough more th an 30 new

culti va rs have been relea sed sin ce 1990,

no ne ca n repl ace Xush u 18. These new
culti va rs have low to med ium DM co ntent
co mpared w ith many Japanese culti va rs.
In rece nt yea rs, mo re and more exotic
germpl as m ha s bee n introdu ced through
internati onal coope rat ion . Th e breeding
strategy of CIP (Mo k et al., 1998) strongly
in flu enced sweetpotato breed ing in Chin a.
Breeders paid mo re attenti o n to broaden in g
the genetic base by in trodu c in g exoti c
germ p las m . W ith the introdu cti on of
superior germpla sm from CIP and Japan, th e
in breedin g tend ency ca n be reversed to
ac hi eve hi gher y ield s and DM co ntent.
Breeding materi als from va ri ous sources,
including those from South A merica and
Afri ca, are ex pected to furth er broaden th e
geneti c ba se to in co rporate resistance to
viru ses and other diseases. The chall enge is
to introgress th ese new genetic reso urces
into hi ghly-a dapted, hi gh-y ield ing Chin ese
co mmercial culti va rs. Obvious ly, culti va r
introdu cti o n alo ne w ill no t ac hi eve these
goa ls. There is a need for further breedin g.

Material and Methods

Va ri o us sources o f seed fam ili es have bee n
eva lu ated in Chi na. Seed fa mili es from CIPLim a had hi gh DM co ntent w ith res istan ce
to root-k no t nem atode (Meloidogyne spp. ).
CIP-N airobi se nt seed fam ili es of hi gh DM
co ntent w ith sweetpotato v iru s di sease
res istan ce . Those from CIP-Bogor had hi gh
DM content w ith w id e ada ptabil ity. Seeds
we re also rece ived from the cooperative

shuttl e breed in g proj ect w ith Japa n (Tab le

1). Seve n adva nced cl ones w ere introd uced
from CIP-Bogo r. Th ey had bee n se lected at
three sites in Indones ia: Bogor (humi d
tropi cs, poor so il , 200 m), Lemba ng
(hi ghl and, coo ler cl im ate, 1600 m) and
Ma lang (hi ghl y ferti le so il , 700 m). Th ese
clon es were initia ll y se lected based on
yield, DM con tent, starc h co ntent, and
adapta bi li ty to poo r so il s. Two clones
(190070-1 and 194039- 1) selected at
Xuzhou from CI P-Lim a seeds were in cluded. Xush u 1 8 was used as a chec k.
Seed ling eva lu ation was done at Xuzhou
by pla ntin g durin g the first week of Jun e.
Ad va nced cl o nes w ere multi pli ed at XSPRC
and eva lu ated at Chengdu , Jin an, and
Xuzhou. Clones we re plan ted in random
co mplete bloc k des ign w ith three repl ica tion s. The size of a single plot was 9.4 m 2
Cultiva ti on was don e acco rdin g to th e
stan da rd meth od of eac h provin c ial AAS.
H arvest was 120 cl after p lantin g. DM
con tent of storage root was meas ured using
200 g of longitudin all y sli ced strips. Dry in g
oven temperature w as set at 70C until the
sa mples reac hed co nstant we ight.

Results and Discussion

Botanica l seed fa mili es from CIP have bee n
introdu ced into Chin a since 1990; more
th an 34,000 seeds have been eva lu ated.
Th e number of fa mili es and th eir or igin are
li sted in Tab le 1. Seeds introdu ced in the
ear ly 1990s had poor germin ation and low
adaptab ili ty, res ul ting in a low se lect io n

Table 1 . Evaluation of botanical seeds introduced to XS PRC from CIP.

Year

Crosses

Seeds received

Seedlings planted

Clones selected

1992
1994
1995
1996
1997
1998
1999

33
100
80
30
14
160
148

1,700
8, 100
6,700
2,650
1,483
12,000
20,520

600
3,571
4,000
1,010
920
8,600

12
34
70
22
30
400

Origin
CIP-Lima
CIP-Limo, CIP-Bogor
CIP-Limo, CIP-Bogor, (JP-Nairobi
CIP-Limo, CIP-Bogor
Shuttle-breeding fa milies
CIP-Limo, CIP-Bogor
CIP-Limo, CIP-Bogor

CIP Pmgrurn Repo111997-98

3 13

ratio. Se lected hi gh-y ieldin g c lon es such as

490074-2 (from CIP-Lima , 1992 ) and
B0080- l (fro m CIP-Bogo r, 1994) had low
DM co ntent. These c lones have alrea dy
bee n cli sca rcl ecl from th e CIP breed in g
program.

th e nurse ry we re disca rcl ecl , 8,600 seedlings

we re transp lanted to th e field. Fin all y, 400
good clones we re se lected from th e f irst
yea r seedling eva lu at ion. Sel ection was
based on good root shape and hi gh DM
co ntent.

In 1995, large qu antities of seeds we re

introclu cecl, most of th em from crosses of
hi gh OM-content parents. Two clo nes from
CIP-Lim a were sel ected, 194038-3 and
194039-1. Both had hi gh DM co ntent, 35%
for 194038-3 and 32% for 194039 . Th ey
had hi ghe1 dry matter co ntent th an th at of
Xushu 18 (25 -26%). After initi al se lecti on at
XS PRC, the 2 clon es were sent to more tha n
10 breeding units around the co untry fo r
furth er eva lu ation. Although th e c lones
we re not released as culti va rs, th ey we re
used as parenta l mater ial in many provin ces . In 1996, 22 good clon es we re
se lec ted from CIP seed fam ili es. Of these,
Kawa 090-1 had 30% DM content, 4%
hi gher th an Xushu 18. Further eva lu at io n
was clon e in 1998 w ith about 12,000 seeds,
9,600 of w hi ch came from CIP-Lima, 2,200
from CIP-Bogo r, and 52 4 from th e Nat io nal
Agr iculture Resea rch Ce nter (NARC), Japa n.
Th e Japanese seeds were produ ced as a
part of CIP-NARC's shuttl e-breed ing
acti v ity. With the adoption of improved
seed germin ation meth ods, about 90% of
seeds germinated. After inferi o r plants in

In 1998, seve n c lo nes we re introduced

from CIP-B ogor. Of th ese, enough plant in g
material was produ ced to test three c lones
(G roup I) in Xuz hou, Jinan , and Chengdu.
Four clones (Group II ) that did not produ ce
enough cuttin gs we re tested o nl y in
Xuzhou. Th e res ult of the eva lu at ion of
Group I is presented in Tab le 2, and th at of
Group II in Table 3. The results indi ca ted
that AB94001 .8 (code name CIP-2) had
hi gh DM co ntent and high yield. Fres h
storage root y ield of CIP-2 was alm ost the
same as Xushu 1 8 (Tabl e 2), but DM
co ntent was hi ghe1 th an Xushu 18 at all
three sites. This clon e w ill be tested in
demonstration tri als in 1999.
Clone AB9400 1. 8 was initi all y se lected
in Indonesia from seed fami li es obta in ed
through shu tt le breed in g. In eva lu ations of
va ri ous i ntrocluction s from m any countri es,
seed fam ili es from Japa n we re obse rved to
produce progen ies havi ng very hi gh DM
co ntent. Th e paren tal clones of these
fami li es also had hi gh DM content in Japa n
(Tarumoto, 1989; Yamakawa, 1995).

Table 2. Multilocation trial of introduced high OM-content clones {Group I), China , 1998.
Clone

Xuzhou
DM

FY

(%)

(t/ha)

Chengdu
DY

DM

(t/ha) (%)

FY

DM

DM

FY

DY

28.4a

39.4

4.lab
3.9 b 1.6 e
6.2 b 2.4 be

34.0

19.7

33.3
31.6

8.5
11.3

5.1 b 2.2 e
8.2ab 3.3abc
12.4a
4.3a

36.2
34.4

10.2

AB95002.3
AB97001 .1

30.4
8.5c
28.2 l 4.4abe

2.6 b
4.lab

190070-1

31.8

12.1be

3.8ab

33.3

194039-1

30.2

21.2ab

32.3

Xushu 18

24.4

24.2a

6.40
5.9a

13.5 e
17.1 b

30.3

30.7a

13.2 e
13.2 e

9.2a
3.7 (

41.6

3.7 e

38.5

4.5 be

43.6

5.5 b

40.6
35.1

9.3a

DM = dry matter content, FY = fresh storage root yield, DY = dry matter yield.
a. Mean separation within columns by Duncan's multiple range test at P = 0.05.

Sweetpotato

DY

(%) (t/ha) (t/ha)

5.9a

3 14

FY

{%) (t/ha) (t/ha)

30.2 19.5ob'

28.3

DY

Mean

(t/ha) (t/ha)

AB94001.8

32.3
27.8

Jinan

l 1.2a

29.9

15.5
22.4

6.4
2.6
3.4
3.5
5.1
6.5

Table 3. Single-location trial of introduced high OM-content clones (Group II), Xuzhou, China, 1998.

Clone
AB94002.7
AB95002.7
AB94078. l
AB95012 .4

Xushu 18
OM

dry molter content, FY

DM(%}

FY (t/ha)

29.7
32.8
33.7

4.9

1.46

10.8
13.8

25.0
24.0

8.4
20.6

3.54
4.65
2.10
4.94

fresh storage root yield, DY

O btaining botani cal seeds in qu antity

from Japan was diffi cult du e to th e cos t of
seed produ cti o n. Moreove r, all of th e
rece ntly-released culti va rs, w hi ch are
frequ ently used fo r hybridi zat ion, are
protected under th e breeders' ri ghts law in
Japan. To overcome th ese limitati o ns,
shuttl e breedin g was suggested (Takag i, H.,
Japanese Intern ati onal Resea rch Center fo r
Ag ri cultural Sc ience and Komaki , K.,
Na ti o nal Agri cultural Resea rch Ce nter, pers.
co mm .) . CIP sent a trainee to NARC to learn
hybridi zation tec hniqu es usin g hi gh OMco ntent cultivars such as Min amiyutaka,
Satsum ahikari, and Hi-starch. lpomoea nil,
a dwa rf mornin g glory, w hi ch had strong
fl owe rin g indu cti on ability, was used as
graftin g stock. Three train ees from deve lopin g countri es we re sent to Japan, o ne eac h
in 1996, 1997, and 1998. Hybrid seeds
produ ced w ere evalu ated in Chin a, Indonesia, and Vi etn am . OM content in th e
breedin g popul ati o n was rapi d ly in creased
usin g th ese seed fa mi Iies.
An oth er c lon e, AB9407 8.1 , was se lected
at CIP-B ogo r from seed fa mili es of CIPLim a. Thi s clone had the hi ghest OM
co ntent of all cl o nes tested at Xu zhou
(Tabl es 2 and 3), and also gave the hi ghest
OM co ntent (39%) at three di stin cti ve ly
di ffe rent env ironm ents in Indones ia (M ok et
al. , 1997a). Thi s cl o ne has already been
di stributed to oth er breeding units in Chin a.
Two main selection criteri a at CIP-Bogor
w ere hi gh OM co ntent and wid e adaptability (M ok et al. , 1997b). Th ese AB series
cl o nes, se lected for w id e adaptabi Iity in

DY (t/ha)

dry matter yield.

Indones ia, seemed to perform sa ti sfacto ril y

at th e eva lu ati o n sites in Chin a. Clo nes
sel ected in Indones ia for hi gh OM co ntent
almost always tested hi gh for OM co ntent
in Chin a. Th e heritability of OM content
was 61. 2% (Mo k et al ., 1997b). It is likely
that thi s trait could have a small genotype
by env ironm ent interaction .

Conclusion
In Chin a, popul ar sw eetpotato culti vars
mostly have low to medium OM content.
Two major reason s for th is are th e narrow
gen eti c base and th e high yi eld focus of
breedin g programs. Rece nt deve lopm ents in
starch process in g from sweetpotato in
Chin a require culti va rs w ith hi gh OM
content and hi gh sta rc h co ntent. Co l labo rative resea rch w ith CIP makes the exc hange
of geneti c resources possibl e w ith unprecedented speed. As a result, a number of
high OM clo nes have bee n se lec ted and
introdu ced. Th ese materi als are und er
evalu ati on in many prov in ces w here
sweetpotato is important. Th ese prelimin ary
result s indi ca te it is possibl e to se lect
culti va rs for hi gh OM content, hi gh yi eld,
and w id e adaptability in Chin a. Thi s effort
will have a major impact in many prov in ces
w here starch process ing is a major so urce
of fa rmers' in co me.

References Cited
Gen eral Seed Co mpany of Chin a (GSCA).
1989. Stati sti ca l table on major crop s in
Chin a. GSCA, Bejin g, Chin a. 93 p.

CIP P1og1om Rcpo1t 1997-98

31 5

Lu , G.Q. 1990. Th e ped igree of

sweetpotato var iet ies bred in Chi na.
Zh ong Guo Ganshu. Chi nese
Sweetpotato J. 4:26-28.
Mok, l.G ., L. Ningsih, and Tjintoko had i.
l 997a. Se lect ion of new sweetpotato
va ri et ies fo r hi gh dry matter co ntent in
Ind onesi a. In : CIP Pro gram Report 1995 96. p. 45-48.
Mok, l. G. , Tjintokoha di , L. Nin gs ih , and
T.D . Hoa ng. l 997b. Sweetpota to breeding strategy and germpla sm test in g in
Southeast As ia. In: CIP Program Repo rt
1995 -96. p. 104-1 09 .
Mok, l .G. , D. Zha ng, and E.E. Carey. 1998 .
Sweetpotato breeding strategy of CIP. In :
LaBo nte, D.R. , M . Yamashita , and H.
Moch id a (e ds.) . Proc eed in gs of Inte rn ation a I Workshop on Sweetpotato Production System towa rd th e 21st Century held
9- 10 Decemb er 1997 in M iyako noj o,
Mi yaza ki, Japan. p. 9-27 .
Tarumoto, I. 1989. Sweetpotato breeding in
Japan : Its past, prese nt and future. In : CIP

3l6

Sweelpololo

(In ternationa l Potato Center). Im provement of sweet potato (Jpomoea batatas)

in As ia. Report of th e Wo rkshop on
sweet potato im provement in Asia held
24-280ct 1988 at ICAR, Indi a. p. 137146.
Xuzhou Sweetpotato Researc h Center
(XSPRC). 1981 . Catalog of Swee tpotato
Culti va r. XSPRC , Xuzh ou, China .
__ . 1993. Ca talog of Sweetpotato Cul tivar
in Ch ina. XS PR C, Xuzhou, Chi na. 149 p.
Ya makawa, 0. 1995. Sweetpotato in Japa n,
produ ction , utili za tio n and breed ing . In :
Proceed ings of the 1st Chinese-Japanese
Sympo sium on Swee tpotato and Potato.
Beij ing, Chi na. p. 163- 167.
Yuan B. 1995. Swee tpotato production and
breedin g in Chin a. In : Procee dings of th e
1st Chin ese-Japanese Sym posium on
Swee tpotato and Potato held in 1995 in
Beijing, Ch in a. p. 156 -162.

Farmer Maintenance of Sweetpotato Diversity in Asia:

Dominant Cultivars and Implications for In Situ
Conservation
G. Prain and D . Campilan 1

In situ conservation of plant ge net ic

diversity has received growing attent ion in
the last 30-40 years, primarily in relation to
wi ld spec ies. In Asia, the number of
protected areas tripl ed betwee n the 1960s
and ear ly 1990s (MacK inn on, 1995). In situ
maintenanc e of c rop genetic divers ity,
howeve r, was not co nsidered a viab le
co nserva tion strategy by interna tional
organ izations or lea ding specia li sts unti l
rece ntl y (Wi lli ams, 1988; Franke l, 1970).
The ear li er assumpt ion was that land 1aces
we re used onl y in tradition al agr icultural
systems . Tec hn o logica l modernization th en,
espec iall y th e introduct ion of new cultiv ars,
wou ld in ev itab ly lead to the loss of
landraces in most cases . Since it wou ld be
both impracticable and immoral to preserve
an entire system in its traditio nal sta te, it
was argued, ex si tu rather than in situ
co nse rva ti on was th e only so lution.
In fact, th e cul tivat ion of diversity has
survived enormous historical changes in
these supposed un changin g soc ieties (Wo lf,
1982). There is c lea r ev id ence that
landraces continue to be cul tivated und er
modern conditions and w ith modern
tec hn o logy (Brush , 199 1, 1993). Neve rth eless, some of th ese find in gs give rise to
co ncern. Modern or se lected comme rcial
cul tivars seem to have contribu ted in some
cases to the dec lin e of area, fragm entation
into pockets or islands of product ion, and
eve n th e di sappea rance of indi ge nous
cu lti vars altogether (Brush, 1993).
Th at is an imp ortant issue for
sweetpotato (lpomoea batatas (L.) Lam. ) in

Asia, th e secondary center of diversity an d

th e prim ary cen ter of production and use of
th e crop. There is an urgent need to identify
adequate national conservation strateg ies to
ens ure the co ntinued avai lab ility of genetic
di ve rsity to fa rm ers and plant breeders. In
situ conserv at ion co uld provide greate r
access ibility to diversity and, by sp read in g
th e costs of co nservat ion, redu ce the
vul nerab ility of publicly funded ex situ
genebanks. It co uld maintain di ve rsity and
hab itat viability where co ntinu ed crop
evo luti on co u ld take place pro vided a few
commercia l culti va rs do not dominate and
displace loca l crop genetic diversity.

To ga uge the real potent ial of in situ

conservatio n, it is necessary to fully
und erstand culti var maintenance practices
of hous eholds und er different soc ioeconom ic and eco logica l cond itions, and the
dynam ics of va ri eta l dominan ce . Does
dominance mean displa cement? This paper
dra ws conc lu sions for th e future role of in
situ maintenance in crop conservat ion
strateg ies. It is based on case studies in
three areas where var ietal domi nance is
important.
Approach and Methods
To document and eva luate sweetpotato
diversity in Asia and to understa nd how
farme rs ma intain cu lti vars, CIP has supported germp las m col lection and soc ioeconom ic studi es in the Philippin es, Indon es ia,
and Vietnam sin ce th e early 1990s. Fi eld work methods ha ve varied considerably.
Som e focused on collect ion and characte rization w ith m inim um docum enta tion of

1 CIP, Los Banos, Ph i l ippines.

CIP Program Report 1997-98

3 17

indi geno us knowledge. Others are extremel y detai led , lon g-t erm docum entation
of local practices usi ng participatory rural
appraisa l, ethnob ota ni ca l tec hniqu es, and
mon ito rin g of loc al ge rmpl as m col lect ions .
Th e st udi es und ert ake n by UPWARD
(Users' Perspecti ve w ith Agricultural
Re searc h and D evelopme nt), a CIP network, have id entifi ed fo ur major
sweetpotato production system s: (1) sh ifti ng
cultivation in hillsid e and mountain areas,
(2) home garden s, (3) upland rotations , an d
(4) low land postrice system s (Prain , 1995) .
Th e last is th e mo st co mm erc ial ized
produ cti o n system in Southeast As ia and by
vol um e the most impo rt ant sys tem in
Vietn am, In do nes ia, and the Phil ippi nes.
Th e different systems in fl u- enc e farmer
mai nte nance of swee tpotato di ve rsity.

Case Study 1. Shifting Cultivation in

Mountain Environments
Th e case st ud y on sh ifting cultivatio n in
mou ntain env ironm ents is based o n
publi shed an d unpubli shed data co llec ted
in W agawaga, Central lrian Jaya, Indon es ia
(Schn eider and Yaku, 1 996; Schn eider et
al. , 1997; and Pr ain , W idias tuti , and Yaku ,
unpub l.) . Wagawaga is in the north ern part
of th e Baliem Vall ey, a large, flat ag ric ul tu ra l area. Th e va ll ey is surround ed by
steeply rising mountain wa lls, wh ich also
suppo rt shiftin g agriculture. Th ere are two
mai n sweetpotato produ ction systems. In
one , hipere we n, sweetpotato is p lanted in
in dividual mo und s on hi gh rid ges surround ed by d eep drain age ditc hes. Th e
other, yab u waganak, is similar to hipere
wen bu t wit hou t th e drainage d itc hes. Yab u
waganak is sometim es practiced in ga rd ens
nea r the co mpound s, particularly o n
hill sid es .
As in mos t part s of hi ghland lrian Jaya,
steam ed or baked sweetpota to is the stap le
food of Wagawaga fam ili es, ac co un tin g fo r
as mu c h as 90% of th ei r diet (H eid er,
1979). Produ ction sys tems are des igned to
suppl y fr es hl y harves ted sweetpotato yea rround. H ouse hold s maintain several beds at

318

Sweetpototo

di ffe rent stages of developme nt as wel l as

mi xtures of culti va rs in each bed to ensu re
supplies. M ixtures are impo rta nt . Lo cal
peo pl e co mp are th e mi xin g of man y
sweetpotato culti va rs in one ga rd en bed
w ith th e need to mix up the c lans for
marri age to ensure th e co ntinuity of th e
tribe. Short durat io n c ulti vars are impo rta nt
com pon ents of the mixtures . Th e number of
cu ltivars ma intain ed per bed ranged fro m 7
to 20, with an ave rage of 12. Wo men w ho
maintain multipl e sweetpotato beds do not
plant the same set in eac h bed, bu t the most
important cult ivar-s are like ly to be commo n
in all beds .
Wagawaga was des ignated as an in situ
site in 1994. A preliminary study on the
genet ic di versity (Yuli an tin ingsi h, 1995)
id entified 47 cul tivars in th e 30 be ds of th e
in situ stud y garden . A fo llow-up documentat ion of 30 beds in 1998 also docum ented
47 cul tivars, but o nl y 27 of th ese were the
same as the cult iva rs p lanted in 1994 (Tabl e
1). Th e d istrib ution of th e different sets of
culti vars in 1994 and in 19 98 are qui te
simi lar, al thou gh th e beds doc um ented in
1998 were not exact ly the same as th ose in
the 19 94 study si nce beds are periodica lly
rema de wi th in th e sa me garden area . We
ca n better und ersta nd the cultivar di st ributio n by co nsid erin g two ke y c haracter istic s
of fa rm ers' cultivar manageme nt: th e
number of pl ants of a c ulti va r an d the
freq uency wi th wh ich th e culti va r is planted
in different beds (Fi gures 1 and 2) . The
pl antin g density of cult iva rs was calcula ted
differentl y in 19 94 and 1998. In 199 4, the
to tal numb er of pl ants in th e garden was
cou nted and th e perce ntage of the to ta I
ca lcul ated for eac h cu lti v ar. In 1998, eac h
bed was mapped and the perce ntage eac h
cul tiva r occu p ied in th e bed was ca lculated.
Hel aleke asl i was the most w id ely
pl anted culti va r in both 1994 and in 1998,
w ith 34.9% of bed area in 1994 and 4 2% in
1998 (C lu ster 1, Fi gures 1 and 2). Thi s
cu ltivar is th e most important hum an foo d
in Wagawaga and , in particular, is the

Table 1. Cultivars planted in Wagawaga, Irion Joya in 1994 and 1998.

Cultivars

Cultivars

Cultivars

Cultivars

Cultivars

Cultivars

present 1994

present 1998

present 1994

present 1998

1994

present 1998

Heloleke osli
Hupuk
Hopoye
Muson
Nomukero
Musaneken
Abukul
Alukulek
Kilooke
Helokeke boru
lnin
Juaiken
Duok
Ebe osli
Fibisok
Ogopem
Nobokum
Hulok
Hupoleke
Kilu
Hupuk Sumunoh
Pilko
Pogo re ken
Pulugio

Heloleke osli
Hupuk
Hopoye
Mu son
Nomukero
Musaneken
Abukul
Arukulek
Kilooke
Helokeke boru

Ponoi
Tuke osli
Puluk
Mikmok
Musaneken baru
Saporeken
So lisike
Suweol
Sobolok
Tabogole
Tinto
Tamue
Welelom Sumunoh
Welelum Boru
Wenoboge Moloh
Wenoboge
Wereneh
Heloleke Hulek
Heloleke Wologin
Yibiloken
Wosilolo
Kumbuk Aidek
Moyugwe

Juoiken

Hibisok
Nobokum
Hulok

Pilko

ap propri ate food for women to present to

men. Th e seco nd most w idely pl anted
cultivar, in area and in number of beds
wh ere it w as grown, was Mu sa n. It occupi ed 16% of bed area in 1994 (Figu re 1)
and 24% in 1998 (Figure 2). Thi s cultivar
produ ces large roots and is an important
source of pig feed in the area . Accord in g to
women (who plant and tend the beds), the
proportion of beds dedicated to parti cul ar
culti va rs va ries, dependin g o n th e propo rtion of peo pl e and pi gs in house holds and
compound s.
Th e second clustering (C lu ster 2, Figures
and 2) includes cultivars pl anted in more
th an 10 beds and which occupy more than

Mikmok
Soporeken
So lisike
Suweol
Tabogole
Tinto
Tamue
Welelom Sumunoh
Welelum Boru
Wenoboge
Wereneh

Kofiar
Suweol boru
Legel pilodok
Kokum eko
Musan baru
Kentong
Wiyayuken
Tinto hitom
Juoiken boru
Wopem
Tinto putih
Kulomeke
Tinto meroh
Yokik
Ho book
Lisuge
Boruke
Alu age

Wosilolo

Wortel

1 % of bed area. Th ese clusters compr ise

cultivars th at are also important, but for
different reaso ns. O ne is conn ected with
the use of sw eetpotato for baby food, such
as cultivars Arukul ek, W elelom, and
Wel elum Baru. The rapid diffusion of the
new cultivar W ortel between 1994 and
1998 seems to be due to its orange co lo r
and less fl oury texture, preferred characterist ics for baby food. So me, such as Hupuk
are ancestra l cultivars. They must be
planted, often befo re other cu lti va rs and by
the fertility chi ef, for the sake of th e ancestors and to ensure a bountiful harvest. M any
of th ese cultivars are valu ed as adult food
and are also fed to pi gs, especially th e
small er roots.
CIP Program Reporl 199798

319

Variety area (%)of all plants

35
18

Clust er 1
Helaleke (34.9% )
Mus an (16%)

14
10

Cluster 2

6
Cluster 3

25

30

Nu mbe r of beds with variety

Figure 1. Distribution ol 47 sweetpotato culti vars in 30 beds, Wagawaga , Irion Jo ya, Indonesia , 1994.

Variety area (m ean % of all plants)

45
Cluster 1
Helalek e (42.9%)
Musan (24 %)

35

25
Clu ster 2

Clu ster 3

15

..

5
I

'

15

20

25

30

Number of beds with variety

Figure 2. Distrib ution of 47 sweetpotato varieties in 30 beds, Wagawaga , Irio n Jo ya, Indonesia , 1998.
Cl uster 3 in volves c ulti vars that appear

residua l. In 1994, 30 of th e 47 culti va rs

were grown in fewe r than 10 beds and
accounted for on ly 4% of area. In 1998, 33
cul ti va rs were grown in fewer than 1O beds
3 20

Sweetpototo

and accounted for 6% of area. Wo men say

they do not forget to pl ant any of the
cult ivars, eve n th ese w ith such tin y rep resenta tion. Howeve r, they ma y not repl ant
some cul tiva rs if th ey do not find cuttin gs,

w hi c h is es pec iall y poss ib le w ith thi s las t

c lu ste r. Th ese culti va rs acco unt fo r most of
the va ri ati o n betwee n 1994 and 1998 as
they sli p in and o ut of th e system. Of the 20
c ul tivars th at d isappea red, almost all
cove red less than 0 .5% of beds and we re
p lanted in fewe r th an 5 beds.
A simil ar situat io n appli es to the new ly
in troduced cu ltiva rs in 1998. A lm ost al l
were fo un d in fewe r th an 5 of the 30 beds,
w ith ve ry sma ll area p lanted. Except io ns are
Tinta (in 18 beds in 1998) and Wo rte l (14
beds). Tin ta is gro w n as p ig fee d. Worte l is
gro w n as baby fo od and fo r sa le to no nlria nese co nsum ers w ho li ke its ye ll ow
fles h. Bot h c ulti vars are new (baru),
mean in g they have rece ntl y been introdu ced or identifi ed as nove l types, probably
hav in g emerged as seed I in gs.

Case Study 2. Upland Semicommerc ial

Rotations
Th e up la nd ro tat io ns case study invol ves
a se m ico m me rc ial syste m practi ced in
Maambo ng Prov in ce o n Mindan ao, the
south ernm ost of th e large r Philippin e
islands (Prain and Pi niero, 199 9). It is one
of th e wettest areas of th e cou ntry w ith
we ll -d istr ibuted rainfa ll al lowi ng crops to
be p lan ted thro ugho ut the yea r. Th at favors
sweetpotato co nse rva tion, beca use vegetative p lantin g materi al is rea dil y ava il ab le fo r
repl antin g. Maa m bo ng is situated in the
re lative ly fla t, no rth ern part of a genera ll y
un dul at in g grass land p latea u, crosscut wi th
deep river va ll eys.
Farmers prac ti ce adapted plow ag ri cu 1ture and rota ti o ns o n th e pl ateau and in the
hill s, es pec iall y fo r ma ize, roo t crops, and,
mo re rece ntl y, vege tab les. A noth er commo n pro ductio n syste m is the ho me garden,
w hi c h no rm all y su ppo rts a large amoun t of
bi odi versity. Sweetpotato is a supplementary stap le in both system s. There has bee n
a steady inc rease in sweetpotato geneti c
divers ity w ith th e influ x of cult ivars bro ught
by m igrants fro m th e no rthern coas t and
from nea rb y small er island s. Ho weve r, the
grow th in di vers ity may have peaked

arou nd the ear ly 1980s. Th at was aro un d

th e tim e a multin ati o nal p in eap pl e pl an tati o n beca me an in creas in gly impo rtant
emp loyer in th e area; thi s in creased
oppo rtun ities fo r vegetab le prod ucti o n.
Th ro ugh an ag ree ment with an in fo rm al
group of wome n in Maa mbo ng, a conservation garden was establi shed . Th e ga rd en,
in itial ly co nsisti ng of 24 beds, was mode led
after ho use hold ga rd ens. Fo ll ow in g initial
ex per ime nta tio n, however, th e wo men
adopted a rota tion syste m w ith pea nu t for
repl anting their roo t crop co ll ect io ns. Eac h
wo man culti va ted her ow n pl ot and ini ti;:i lly
p lanted cultivars fro m her own fi eld s and
garden, lea din g to co nsiderab le d upl ica tion
of cul tivars across the plots.
There was a sharp rise in sweetpotato
di ve rsity in the seco nd p lantin g, fr om 11 to
19 cul tivars, w ith o ne lost (Tab le 2) . That
ca n be exp lain ed by seve ral facto rs, th e
most im port;:i nt of wh ich is the infl uence of
th e pro jec t itse lf. Th e co nsiderab le interest
and enthu sias m amo ng the wome n in
di ve rsity mai ntenance res ulted in th eir
scour in g the loca li ty fo r add iti onal cul tiva rs.
H oweve r, five of the added cul ti va rs had
already d isap peared from the ga rd en by the
thi rd pl antin g; all had d isa ppeared by the
fifth pl anting. It was fe lt th at th e project had
heighte ned th e no rma l practi ce of eva lu atin g any new cul tiva r fo r w hi ch pl antin g
materia l is avai lable.
That imp ress ion was sup ported by the
sim i lar patte rn that c haracter ized subsequ ent pl antin gs, with a few rare local o r
exot ic cul tiva rs added on ly to be abando ned later. A c lu ster of preferred cul tivars
prese nt fr om the beginn ing of the ga rde n
(Klari n, Fi ve Fingers, lgoro t Pula, lgorot
Pu ti, Amerikan o, Kamada, and Tapol ) we re
co nsistentl y pl anted by mos t wome n and
gradu all y became the mai n culti va rs (Tab le
2). Furthermo re, in later p lantin gs th ere was
a ten dency to reduce the numbers of pla nts
of other c ul tiva rs in favor of the prefe rr ed
culti va rs. That was parti cul arl y true of
Kl arin and Fi ve Fin ge rs, w hi c h are popul ar

CIP Program Report 1997-98

321

Table 2. Sweetpotato cultivars planted in the Maambang genebank over five seasons.
First planting

Second planting

Third planting

Fourth planting

Fifth planting

Klarin

Klarin

Klarin

Klarin

5-fingers

5-fingers

5-fingers

5-fingers

5-fingers

lgorot Pulo

lgorot Pula

lgorot Pulo

lgorot Pulo

lgorot Pula

Topal

Topal

Topal

Topal

Topal

Ameriko no

Amerikano

Amerikano

Amerikano

Am erikono

Biloko

Bilaka

Bilaka

Bilaka

lgorot Puti

lgorot Puti

lgorot Puti

lgorot Puti

lgorot Puti

Kamada

Kamada

Kamada

Komada

Kamada

Valencia

Valencia

Valencia

Valencia

Kinampay

Kinampay

Kinampay

Kinampoy

Kinampay

Kaligatos

Kaligatos

Kaligatos

Kaligatos

Kaligatos

Tinangkong

Tinangkong

Tinangkong

Tinangkong

Kapitlok

Kapitlok

Kabohol

Kabohol

Sil-ipon

Sil-ipon

Kaligatos
Kabohol
Sil-ipon

Magtuko

Magtuko
Maranding
Senorita
Imelda
Kitam-is

Ma randing
Senorita
Imelda
Ki tam-is

Lila
lnitlog

Total
Additions
Losses

11

Klari n

Lila
lntilog
PNGL

PNGL

Kawakwak

Kawakwak

Pl6

Pl 6

NPSP

PNSP

Sa layaw

Sala yaw

UPLSP

UPLSP

Kabato
Turay

Kabato
Turay

19

17

17

13

+ 9
-1

+ 3
-5

+8
-8

+ l
-5

Note: Cultivors in bold ore additions.

Cultivors in italics ore losses.

in th e fresh market. Characteristics of the

preferred culti vars are give n in Tabl e 3. The
wide genetic diversity suggests c lose
relat ion between diversity of use and
genetic diversity.

322

Sweet po to to

Case Study 3. Lowland Postrice Systems

Th e low land postri ce sweetpotato
production system case study (Anganon et
al. , 1998; D ata et al., 199 7) is based on the
key co ~1111erc i a l sweetpotato producin g

Table 3. Morphological characterization' of continuously planted, dominant cultivars in the Maambong,

Philippines, genebank, 1998.
Cultivar

Klarin

Plant Vine

Leaf

Leaf

Immature

type color

vein

shape

leaf col.

5332

32

Petiole

Root

Skin

Flesh

Extracts of farmers'

pigment shape

color

color

characterizations

830

100

Grows easily; roots

dry, sweet; tops far

vegetable, forage;
Five Fingers 7

6755

72

634

442

salable
Early maturing;
leaves far
vegetable;
roots sweet, watery;

Amerikano

3131

95

126

543

lgorot Puti

5332

62

230

200

salable
Profuse vines; few big
roots; roots sweet,
powdery
Profuse foliage, good
for weed control,
forage; roots sweet,
dry

lgorot Pula

5334

92

230

l 00

Topal

3151

95

930

196

Roots dry, sweet;

leaves for
vegetable;
salable
Easy to maintain;
Roots unsweet,
watery; leaves for
vegetable; good

Kamada

5332

23

520

100

market far this

color; small area
Early maturing; easily
harvested; root
sweet, watery; tops
suita ble for forage

Key: Plant type = l erect, 9 extremely spreading. Vine color = l green, 9 totally dark purple. Leaf vein color = l yellow, 2
green, 3 purple spots main rib, 4 purple spots several veins, 5 main rib partially purple, 6 main rib mostly purple, 7 all veins
partially purple, 8 all veins mostly purple, 9 lower surface and veins totally purple. Leaf shape (combination of 4 numbers): l"
Outline = l rounded, 7 almost divided; 2dType of lobe = l none, 9 very deep; 3'd Lobe number; 41h Shape of central lobe = 0
absent, 9 linear. Immature and mature leaf color (combination of 2 numbers) = l yellow-g reen, 9 totally purple. Petiole pigment
= l green, 2 green with purple near stem, 3 green with purple near leaf, 4 green with purple both ends, 5 green with purple spots
th roughout, 6 green with purple stripes, 7 purple with green near leaf, 8 some petioles purple, some green, 9 totally/mostly purple.
Root shape = l round, 9 long, irregu lar or curved. Skin color (combination of 3 numbers): main = l white, 9 dark purple;
intensity = l pale, 3 dark; secondary = 0 absent, 9 dark purple. Flesh color (combination of 3 numbers): main = l white, 9
strongly pigmented; second color = 0 absent, 9 dark purple; distribution of color = 0 absent, 9 covering all.

CIPPtogrnmReport 199798

323

reg ion in Centra l Lu zo n, the Philippin es.

Located about 50 km north of Ma nil a, it is
th e main supplier for the urban fres h root
market. Swee tpotato is comm erc iall y grow n
in ri cefie lds durin g the dry mo nths from
Decemb er to April, wi th peak harvest in
Marc h. Central Lu zo n' s six prov in ces had a
sweetpo tato production are a of app rox imate ly 9,000 ha in the early 1990s. Rece ntl y, th e area has ex pand ed to 10,000 ha
th ro ugh increased demand for sweetpotato
as a ra w mate ria l for large-sca le starch
processin g. Moreover, destru cti ve fl ows of
vo lca ni c deb ri s res ultin g from th e erupti o n
of Mount Pinatub o in 1991 forc ed th e shi ft
fro m ri ce and sugarca ne to sweetpotato .
A total of 25 sweetpotato culti va rs have
bee n grown in the reg ion sin ce th e ea rl y
20th century, acco rdin g to farm ers. On ly
four are now wide ly gro w n and a sin gle
c ul tivar Sup erburea u cove rs at least 80 % of
th e total product io n area. H igh root yiel d,
marketabilit y, and ea rly maturity are th e
mo st important traits co nsid ered by farn1 ers.
Hi sto ri ca ll y, fa rmers grew a m ix of cu ltivars
for both general and spec iali zed uses. For
in sta nce , two pop ul ar cult iva rs, Bu rea u and
Tai wa n, were prefe rr ed for th e hi gh qu ali ty
of their bo il ed roots. Girayan and
Lampan gog we re grow n beca use th ey
co mmand ed hi gh pr ices fo r their use in
tradi t ional d is hes. Fin al ly, th e es tabli shme nt
of starc h facto ries crea ted a demand fo r
Sup erburea u as a p rocess in g c ulti va r.
Avai lability of pl anting materials has also
bee n a maj or determinant of cu lt iva r. Most
fa rm ers main tai n th eir ow n mini- seedp lots
of preferred c u ltiva rs. Wi th th e hi gh demand fo r plantin g material s at th e start of a
cro ppin g seaso n, so me enterpri sing fa rm ers
have offered spec iali zed seed product ion
and marke tin g serv ices. The cul tiva rs
propagated by the se co mm erc ial suppli ers
determin e the va ri eta I com bi nati o ns for the
next seaso n. For in sta nce, Tar lac fa rmers,
faced wi th th e ac ute shortage of their
prefe rred Burea u, resorted to pl antin g
Supe rbu reau, th e on ly materi al ava il ab le
from seed suppli ers from th e neighborin g

32 4 Sweetpoto to

prov in ce . Th e seed suppli ers have in turn

bee n res po nd i ng to ag roecolog ica l and
socioeconom ic c hanges in th e reg ion.
Burea u was bei ng severely affected by
fe at hery mott le vi rus (SP FMV) and possib ly
other viru ses in th e ear ly 1990s and seed
growe rs bega n lo oking for an alte rnati ve
cult iva r. Superb ure au had many of th e
cha racte ristics of Bureau, but was mo re
resis tant to SPFM V and more appropr iate
fo r the develo ping process in g in dustry. In
19 98, Superbureau itse lf bega n ex hibi tin g a
simi lar suscept ibility to the viru s co mpl ex,
promptin g far mers to seek newe r, more
to lerant culti va rs. Mea nw hil e, wi th in creasin g att ention to the processi ng market,
cultivars hi gh in dry matter became imp ortant.
Despite th e relati ve uni fo rmit y of th e
postr ice syste m in Centra l Lu zo n,
agroecologica l var iation affecti ng culti var
cho ice stil l exists . Va riat ion may be stable
and lo ng-term such as in so ils, topograph y,
dra in age, etc. For exampl e, in so me areas
Su perbureau performs ve ry poor ly. Or
va ri ati o n may be produ ce d th roug h natura l
di saste rs suc h as th e acc umul atin g volca ni c
deb ri s in agricul tural land after the Mo unt
Pinatubo erupti o n. For insta nce, there has
bee n grea ter preference for Ube over
Bin ico l in so me affected areas.

Patterns of Diversity: Maintenance,

Dominant Cultivars, and In Situ
Conservation
In al l three case stu di es, one o r two
c ul tivars are preferred. Th ey are planted by
most fa rm ers an d ove r mu ch large r area s
than oth er culti vars . Th at reflec ts the fact
that th e do min ant loca l co nce rn s wi ll be
reflected in th e one o r two cultivars that
best add ress those con ce rn s. Even w hen
dom in ant co nce rn s are the sa me, such as
cult ivars fo r home co nsumpti on or for sal e
in th e mark et, they ca n res ult in different
dom in ant culti va rs. Th at is beca use th e
ag roeco logica l co nd itio ns th emse lves va ry
and differe nt cu lti vars adapt in d iffere nt
ways suc h as between provinces in Cent ral
Luzon or be tween lri an Jaya's hi 11-dw el I ing

an d va ll ey-botto m fa rm ers. Similarly, in

M ind anao com parati ve studi es of nei ghborin g co mmuniti es revealed th at different, bu t
iso morphi c, ge notypes esse nti all y fulfil the
sa me function (P rain an d Pini ero, 1999).
G iven th e va riabi I ity of the env ironm ent
that ex ists in th e upl and s, espec ially in th e
mountains, we can expect to find cons id erabl e genetic diversity, eve n w hen cultural
preferences are similar.
W hen we observe culti va r maintenance
along th e dimensions of frequency of
planting in different beds and the more
com mon meas ure of plant density per bed,
the regul ar and often substanti al occ urrence
of other cu lti vars indi cates other important
soc ieta l co nce rns. In M aa mbong, in
addition to th e two dominant culti va rs
grown primarily for the fresh market, other
culti va rs are maintain ed for ear ly maturi ty,
vegeta bl e use, pi g feed, etc. In Wagawaga,
w hil e Helaleke as li and Musan are th e main
sources of food and feed, re specti ve ly,
seve ral cultivars are also maintain ed for
in fant food or for sale. The most str ikin g
examp le of w hat we ca n ca ll cultural
sa liency, howeve r, is the rol e of particular
cul tiva rs in rituals of first plan tin g and
co nsecrati on of land .
In both W agawaga and M aambo ng, the
dynamis m of loca l maintenance appears to
occ ur prim aril y among culti va rs that are
pl anted in a few beds with few plants per
bed. Ritu al cultivars also have few pl ants
per bed, but the man y women who plant
them do not let them disappear. Th e old
cul tivars are planted first, not on ly to
comp ly w ith custom but because they are
sa id to need more fertilization than exotic
cu ltiva rs. Th e extreme case is lnin, which is
planted in a pile of ash from burnt branches
of clea red garden to ens ure production .
Th is ca refu I management contrasts w ith the
treatment of cultivars in few beds and small
areas. These res idu al or exper imen tal
culti va rs, w hi ch mi ght ce rtainl y include the
products of fo rtu itou s seed ge rmination ,
usua lly enter and exit individu al collections
quite rapidl y. Occasionally, a cu lti va r like

Wote l or Tinta w ill be take n up by increasin g numbers of fa rm ers.

The Central Luzon case indi cates that in
th e past di versity of use stimul ated diversity
of cu lti var. More recent ly there has been a
shift from spat ial to temporal diversity, a
process shared w ith other hi ghl y co mm ercial areas (P rain and Fano, 19 91). Th e one
or two commercial cu ltiva rs cove rin g a
wide area are quite rap idly rep laced as
their productivity declines du e to disease,
they are outpe rfo rm ed by new cu ltiva rs, or
market dem and changes.

Conclusions
Th ese findings have implication s for in situ
conservation strateg ies. A more co mprehensive analys is of dominance sugges ts th at
evolution to wa rd mo dern sweetpotato
agriculture dominated by a single cultivar is
not in ev itabl e. Even if sweetpotato were to
become a more co mm ercia l crop in
W agawaga, the maintenance practices are
res i I ient enough to in corpo ra te one or two
comme rci all y dominant culti va rs wi thout
overall dec lin e in variability. Maambong
also gives reason for optimism. There th e
same kind of cultural sal ience characterizes
several cul tiva rs besides those dominant in
pl ant density. Moreover, in the upl and,
particularly the hi I lside env ironm ents of
th ese regio ns, there is isomo rphi c va riability. That is, geneticall y diverse cul tiva rs
w ith simil ar cultural sa li ence are fou nd at
different sites in a reg ion. One co uld
therefore envision a netwo rk app roac h to in
situ co nservat ion. Sets of culturally sa li ent
and genetica ll y hi ghl y diverse cu ltiva rs
could be maintain ed in se lected sites w ithin
a region. The co ntributi on of the national
genetic resou rces conservation system
toward this type of fa rm er-led co nserva tion
shou ld focus on five thi ngs.
1. Pub li c recognition of the va lu e of lo ca l
germp lasm and loca l knowledge of it.
2. Public recognition of the particip ation of
site represe ntatives in a nati onal conse rvation effort.

CIP ProgramRepor t 199798

325

3 . Simple reward syste m in vo lv in g agr ic u 1tura I seeds and low- input tec hnologies.
4. Facilitating access to exoti c germp lasm
from other sites and from bre eding
progra ms.
5. Supporting continuing di ve rsifi catio n of
use and the cultural sa lien ce of di ffe re nt
c u Iti va rs.
Alt hough the kind s of lowland s represe nted by Central Luz on are unlik ely to
become repositor ies of ge net ic d iversi ty, in
situ conservatio n strateg ies at least sho uld
attempt to increase loca l diversity throu gh
div ersify ing uses (Amihan -Vega and
Bacus mo, 1999 ). Th at cou ld help to avo id
th e se ri o us co nse qu en ces tha t v iru s disease
is now hav in g on th e mono-culti va r cro p in
Central Luz o n .
References Cited

Aganon , C., P. Tan gonan , and C.

Bagala no n. 199 8. Document at ion of
fa rming practices of sweetpotato growers
in different depth s of lahar depos its in
Centra l Lu zo n, Philippines . U npublished
project termination report. UPWARD,
Lo s Bano s, Lagun a, Phi Ii ppines.
Amihan-Vega , B. and J.L. Bac usmo, 1999.
Community-based sweetpotato
ge nebanking and distribution system. In:
Prain, G. and C. Baga lan on (e ds.).
Conservat io n and c hange: Farmer
m ana ge ment of ag ro-biolo gica l di ve rsity
in th e Philippin es . UPWARD, Los Banos,
La gun a, Philippin es .
Bru sh, S.B . 1991 . A Farmer-based approach to conservin g crop germp las m.
Econ. Bot. 45 (2): 153 -16 5.
. 1993. In situ co nservat ion of
landraces in centers of crop di vers ity.
Paper presented at the Sympos ium o n
Global Implications of Germpla sm
Conservation and Util izati on at th e 85 'h
Annual Meetin gs of the Ameri ca n
Society of Agronomy at Cin c innat i, OH.
Organ ized b y A ppli ed Behav ioral
Sc iences, Uni ve rsity of Cali fo rnia , D av is,
CA, USA.

3 2 6 Sweelpololo

D ata, E., J. Roa, and P. Tangonan. 1997.

Sweetpotato food sys tems in Central
Lu zon , Philippin es. Working Paper No.
3, UPWA RD , Lo s Banos, Laguna,
Ph ilippin es .
Frankel, O .H. 1970. G eneti c conservation
of plants useful to m an. Biologica l
co nse rvat ion , Vo l. 2, No. 3. El sev ier
Pub li shing Company Ltd , Amsterdam,
Ne th erl ands .
H eid er, K.G . 1979. Grand Va lley Dani:
Peaceful warriors. H ol t, Ri nehart and
W in sto n, A ustin , TX, U SA.
MacKinnon, J. 1995. Ana lytical status
report of biodi ve rsi ty conservation in th e
As ia-Pac ific Region. In : Mc N eely, J. A.
(ed.). Biodiversity conservat ion in th e
As ia Pac ific Reg ion: Constraints and
op portuniti es. As ian Development Bank,
Ma nil a, Philippin es.
Pr ai n, G. 1995. Swee tpotato in Asian
prod ucti o n systems: A n overview of
U PWA RD' s fi rst phas e research. In:
Tak ing root: Proceedings of th e Third
UPWARD Rev iew and Plannin g Workshop. U PWARD , Los Bano s, Laguna,
Ph ilippin es.
Prain , G. and M. Pini ero. 1999. Commu nal conservation of roo t crop ge neti c
di versi ty in so uth ern Phi Ii ppi nes. In :
Prai n, G . and C. Bagalanon (eds.).
Conservat ion and change: Farm er
mana ge m ent of ag ro -b iological di ve rsity
in the Ph iii ppin es. UPWA RD , Los Banos,
Lagu na, Phil ippin es.
Prain , G. and H. Fano. 1991 . Swee tp otato
in the food system s of Lat in America and
th e Carrib ean. In : Sweetpotato cultures
of Asia and South Pac ific: Proceedin gs of
the Seco nd A nnu al UPWA RD Intern ational Conference. U PWA RD , Los Banos ,
Lagun a, Philippin es.
Prain, G. , C. A . Widyastuti, and A . Yaku.
n.d. Unpub lished f ield data . UPWARD,
Los Banos, Laguna, Philippines.
Schn eid er, J. and A. Yaku. 1996. Conservation fo r d eve lopment: Th e releva nce o f
indi ge nous rootcrop kno w ledge in lri an
Jaya. In: Into ac tio n resea rch : Partn ership in As ian root cro p resea rch and

developme nt. UPWARD, Los Banos,

Lagun a, Philippines.
Schneider, J., C.A. Widyastuti, and P.
Schmiediche. 1997. Preservation of
sweetpotato Biodiversity in Indon es ia.
Final Report. CIP-ESEAP, Bogor, Indonesia.
Williams, J.T. 1988. Identifying and
protect in g the origins of our food plants.
In: Wilson, E.O. (ed.). Biodiversity.
National Academy Press, Washin gton,
DC, USA.

Wolf, E.R. 1982. Europ e and the people

without history. University of California
Press, Berkeley, CA, USA.
Yuliantiningsih, T.C. 1995. Peranan tenaga
kerj a wanita dalam kegiatan usa hata ni
ubi jalar dan penyeba ran berbagai jenis
ubi j alar (lporno ea batatas (L) Lamb.) di
desa Waga-waga, Kecamatan Kurulu dan
Desa Woogi , Kecamatan Assologaima,
Kabupaten Ja yaw ij aya. Skripsi. Fakultas
Pertanian, Universitas Cendrawasih ,
Manokwari, Indon es ia.

CIP Program Report 1997 -98

327

Sweetpotato for the New Millennium: Trends in

Production and Utilization in Developing Countries
G.J.

Scott and L. Maldonado 1

Sweetpotato (/pomoea batatas L. Lam. ) is

among th e world's most important, versatil e, and und erexploited food c ro ps. With
more than 133 million tons (FAO STAT,
1998) in annua l production , sweetpotato
currently ranks as the fifth most impo rtant
food c rop on a fresh-weight ba sis in
develop in g countries after rice, wheat,
mai ze, and cassava. Sweetpotato is cultivated in over 100 deve lopin g co untri es and
rank s amon g the five most imp ortant food
cro ps in more than half of th em (FAOSTAT,
Jun e 1998). Onl y in the last deca de has the
crop been th e focus of an intense, coordinated, globa l effort to realiz e its fu ll poten ti al as a so urce of food, feed, pro cessed
produ cts, and income for million s of smal l
fa rm ers and low-incom e consumers in
Africa, A sia, and Latin Ameri ca .
Sweetpotato is also one of th e most
m isunderstood of the maj o r food cro ps. Not
al l sweetpota toes are parti c ul arl y sweetmany ha ve a neutral to dry fla vor. None are
lik e potato, where the tubers are harvested.
In stead sweetpotato roots are du g and eaten
like ca rrots or cassava. Potato foliage is not
us ed for human consumption. But
sweetpotato leaves and tips are an important so urce of vitamin A in seve ral deve lopin g cou ntri es in Asia and Afr ica. Th ey are a
trend y, lu xury vegetable in Japan . Sti ll ,
relat ive ly few sweetpotatoes are produced
and consumed in indu str iali zed co untries.
Over 98% of global sweetpotato output is
currently harvested and utilized in deve lopin g co untries.
Thi s paper analyzes recent trends in
sweetpotato production and use from a

1 CI P, Lima, Peru.

global , reg ional , and sub-region al perspecti ve. Deve lop ments in Asia, China in
parti cul ar and Sub-Saharan Afri ca me rit
parti cular attentio n as upw ard trends in
output are link ed to the chang in g rol e of
sweetpotato in local food systems. Th e
resul ts prese nted se rve as the basis for
recomm endat ions for policymakers and
res earch scientists and are aimed at building on the loca l mom entum to fully ex ploit
sweetpotato's untapped potential in th e
new mil lennium.

Materials and Methods

An alysi s don e for this paper relies heav ily,
although not exc lu sive ly, on FAQ produ ction and utili za tion data (FAOSTAT, Jun e
1998) . Th ese stat ist ics are supplem ented
with in formation from othe r sources
includin g official publications of natio nal
governments, survey res ults, and related
ana lyses found in scie ntific pap ers and th e
gray literature.
Estimated growth rates for production for
particular tim e periods and descriptive
statistic s on utiliz ation for selected yea rs
were ca lcu lated on an average annua l
basis. Th ese fi gures are presented for
different reg ions and countr ies to prov id e
the matrix for co mp arati ve analysis.

Results and Discussion

Four facts have generated growing interes t
in trend s in sweetpotato production and
utili zation. First, sweetpotato is typi ca lly a
small-farm er c rop and often grown on
marginal so il s with limited outputs. Furthermore, alth ough the crop is w id ely c ul tivated in Asia (3 1 countries), Afr ica (39), and

CI PPIOQIOITI Repo1t 1997-98

329

Latin A meri ca (3 1 ), prod uct ion te nds to be

con ce ntrated in th ose countri es wit h lower
per cap ita in comes (Fig ure 1) and with in
th ose co untries in reg io ns suc h as Sich uan
Prov ince in Ch ina or wes tern Ke nya w here
in co me leve ls are relative ly low. H ence,
in creas in g sweetpotato prod uct ion is ofte n
con sidered as a mea ns to im prove foo d
sec urity amo ng the po orer seg ments of the
rural and urba n pop ul ati o n. Simil ar ly,
ex pand ing utili za tion thro ugh improvements in processi ng and feed use is see n as
a way to raise in co mes and thereby redu ce
poverty.
Seco nd , average yield s in severa l
cou ntri es are we ll below th e ave rage of 15
t/ ha fo r develo pin g co un tri es as a w ho le
(Tab le 1). Thi s average in turn is we ll below
c urrent ave rage y ields obtai ned in coun tries
such as Chin a. Potential yie ld in creases
based o n ex per im ent station trials are much
hi gher still. So me rece nt deve lop men tssuch as the tend ency in some As ian
cou ntri es to push th e cro p o nto more
margina l land- do raise add iti o nal challenges to deve lo p and d iffuse yield-in creasin g tec hn o log ies for sweetpotato and hel p
ex pl ain th e preva ilin g yield gap . Neve rtheless, rap id im proveme nts in productivity are
co nsid ered mo re re adil y feas ible w ith
relatively less investmen t in resea rch and
ex tensio n for sweetpo tato than ot her crops,
such as ri ce. This is beca use fa rm ers' y ields

Cro p
Soybean
Lentil

USS Ran king

I
I

17,004

_J

Wheat

6, 130

2,014

Maize!

1,782

11

Potato

1,550

14

1,419

16

Cassava

Rice

Swee po tato

1, 112

17

881

18

Average incorre'capita =average, over all dev. counlries, of IJl"Op:Jrtion of overall

prcxfl.dion of the etop in each oountry rrultiplied by that country's per capita GNP.

Figure 1. Average income per capita and commodity

ranking of major food crop production in

deve ioping countries. (Source : Scott and
Maldonado, 1999.)
330

Sweetpototo

are far below w hat has bee n show n alrea dy

to be tec hni ca ll y feas ib le, and because by
us ing standa rd scie nt ific tec hniq ues ca n
raise experi menta l y ields eve n h ighe r.
Thi rd , the last decade has w itnessed a
return to a pos itive growth rate fo r
sweetpo tato product io n in Chi na-a
remarkab le reversa l of prev io us trends in a
co untry w here so me 85% of the wo rl d's
outp ut is harvested (Figure 2). Simi lar
up wa rd tre nds have eme rged i n a num ber
of ot her deve lop in g co untr ies. So me of
these trends are ev id en t fro m FAO statisti cs;
oth ers are not.
Fo urth, seve ral easte rn and so uthern
Afr ica n co untri es have w itn essed ma jo r
increases in sweetpotato production in
rece nt yea rs. With thi s ex pansio n in outp ut
have come sh ifts in both the orien tation of
prod uct io n and the uti lizat ion of the
harves ted cro p. G ive n th e heigh te ned
co nce rn abo ut foo d sec uri ty and pove rty
alleviation in ma ny of these loca tions as
we ll as for the reg io n as a w ho le, th ese
deve lo pm ents have ra ised th e prospects fo r
a more prom in ent and d ive rsified ro le fo r
sweetpota to in th ese cou ntri es in the
decades ahead.

Sweetpotato Utilization: Food, Cash,

Feed and Processed Products
Th ough co m mo nl y ca tego ri zed as stri ctl y
a " su bsistence," " foo d sec urity" o r "fa m in e
reli ef" crop, sweetp otato uses have d ive rsified cons ide rab ly in deve lo p ing co un tri es
over the last fo ur d ecades. H ence, w hi le
th ese longstand ing uses are stil l im portant
in so me cou ntri es or so me regio ns wit h in
these countries, other uses have c lea rl y
emerged, pa rticul ar ly in As ia (Tab le 2),
whe re th e share used as feed has increased
from 14 .5% in 196 1-63 to 44.6% in 1993 95. Ma ny of these c hanges are either not
qu antified o r blurred in th e pub li shed
agg regate utili zat io n figures . A mo ng th e
most notable examp les are (1) the use of
vi nes in add itio n to roo ts for anima l fee dth is is sim pl y not quan tifie d into the
estimates of fee d use as these are typical ly
based so lely on es ti mated ro ot p ro du ct io n,

Table 1. Average sweetpotato production, yield, and area in developing countries.

1995-97
Region/country

Average annual growth rate'

Production

Area

Yield

Production

Area

Yield

{000 t)

(000 ha)

(t/ha)

(%)

(%)

(%)

2
Asiab(n = 31)
Chino
Indonesia
Vietnam
Africa' (n = 39)
Uganda
Rwanda
Malawi
Kenya
Burundi
Latin America (n = 31)
Brazil
Argentina
Cuba
Peru
Developing countries

125,058
117,848
2,013
1,675
6,957d
1,888
967
962f

725
663
1,850
655
339
220
191
l 33,865d

7,178
6,160
212
292
l,519d
513
150
991
74
108
247
58
20
60
11
8,944d

17
19
10
6
5d
4
6
lOf
10
6
7
11

17
4
18
15d

l.l

1.2
-1.2
l.l
2.1
3.7
l. 9
n.a.
4.7
1.7
-1.2
-2.3
-0.3
0.4
0.7
l. l

0.8
l.O
-0.4
-1.7
l.5/2.7'
0.9
0.5
25.3
3.7
1.0
-1.7
-1.5
-1. l
-2.6
3.6
0.8d

2
l.4
-1.5
-2.3
0.6
2.6
3.5
2.1
n.a .
3.2
1.7
-1.0
-2.7
-1.8
0.9
-1.2
-1.0

-0.3
-0.l
-1.5
-1.l
l.9/2.3'
2.8
l.5
14.0
3.8
1.4
-2.2
-2.9
-3.9
0.2
-0.7
O.Od

2
2.5
l.l
2.7
l.l
l.l
l.l
0.5
-0.5
-0.4 -0.4/0.3
0.1
-1. 9
-0.2
-1.0
n.a.
9.7
-0.1
l.5
-0.1
-0.4
-0.2
0.5
0.4
l.5
l.5
2.9
-0.5
-2.8
l. 9
4.3
0.8d
2.1

Sou rce: Scott and Maldonado, {1999); (Source for Malawi data only, Ministry of Agriculture and Irrigation, Malawi).
al. 1961-63 to 1995-97, 2. 1985-87 to 1995-97.
bAsia (excluding Japan, Israel) + Oceania {excluding Australia, New Zealand).
' Excludes South Africa.
dfotals do not include data far Malawi.
' includes data for Malawi.
1
For 1995/96 to 1997/98.
and (2) the lac k of division in the share of
output devoted to food consumption
between fresh roots and processed products, e.g., noodles made from starch.
Consequent ly, a ny ana lysis of th e trends in
utilization for sweetpotato must be done
w ith cons ide rab le ca uti on.
Average ann ual per cap ita co nsumption
of fresh roots for 1994-96 is estimated at:
Africa, 9 kg; Asia, 18 kg; Oceania, 73 kg;
Latin America, 2 kg; Japan, 9 kg; and USA,
2 kg (FAOSTAT, Jun e l 998). In contrast to
potato, per cap ita sweetpotato cons umpti on
in Canada, Europe, and Australia is ex-

tremely limited and often confined to a n

immi grant popu lat ion .
The quantiti es of sweetpotato consumed
can vary treme ndously w ithin developin g
country reg ions. In Africa , for example,
annual per cap ita sweetpotato consumpt ion
in Rwanda is estimated at 160 kg; Burundi ,
102 kg; and Uganda 85 kg. Sweetpotato
consumption a lso varies w ithin countries by
reg ion s, by tim e of year, and by in come
group. In northeast Uganda, one of the
poorest parts of that co untry, sweetpotato
becomes a seasona l stap le during the dry
season whe n supplies of most other food

CIP Pmgmm Report 1997-98

33 l

Gr owt h rate (%)

6.0

-4. 0
-5.0

'-------'---~--~--~--~

to

20

15

25

1 0-years inter va ls*

*1 =Average annua l growth rate 1961-63 to 1971-73:
2 =Average annual growth rate 1962-64 to 1972-74:
24 =Ave rage annual growth rate 1985-87 to 199597.

Figure 2. Average annual growth rates(%) for

sweetpotato in China calculated in 10year intervals from 1961-97. (Source:
Scott et al., 1999.)
stuffs are ex hausted (H all et al., 1998). Eve n
under such circu mstances the importance
of the crop may be und erestim ated given
the in creasing ly ap pa re nt fl aws in the
ca lcul ation of such est im ates.
For developing countri es as a w hole, per
cap ita co nsu mption of sweetpotato,
particul arl y in fresh form , has dec lin ed over
tim e. Acco rdin g to ava il ab le statisti cs,

average per cap ita co nsumption fell from

37.5 kg/yr in 1961 -63 to 18.8 kg/yr in
1994-96 (FAOSTAT, June 1998). As in comes
increased, urbani zati o n acce lerated, and
the ava i labi Iity of cheaper, more prefe rred
substitutes (mea t, w heat breads, etc.)
became more ab undant in many parts of
Latin Amer ica and As ia, in parti cul ar,
co nsumers o n and off the farm red uced
th ei r intake of fresh sweetpotato. H owever,
the trend has been more vo lat il e in SubSaharan Afr ica w ith consumpti o n ri sin g and
fa ll in g and th en ri sin g aga in . In so me
countries of th e reg ion , e.g. , Ma law i, the
trend has been a rapid ri se in cons um pt ion.
Fo r th e most ly poo r co nsumers in eastern
and southern Afr ica, per capita sweetpotato
consumption has eith er stayed rou ghl y
co nsta nt or risen as rea l in co mes have
deteri orated, imports of traditi onal subst itutes have bee n cut bac k, and produ ction
of (other) loca l sta pl es stagnated or shrank
as a co nseq uence of wea knesses in th e
ove rall eco nomy and co ntinu ed stro ng
pop ul atio n growth. In some cases thi s was
exace rb ated by prob lems in spec ifi c
commod ity subsecto rs, e.g., maize in
Malawi, cassava in Uganda. In Asi a, rapid
expa nsion of co nsumption of processed
sweetpotato in the fo rm of suc h prod ucts as
nood les has helped offset the declin e in
fres h co nsumpti on.

Table 2. Distribution of the uses of sweetpotato in Africa, Asia , Latin America , and the industrialized countries,
1961-63 and 1993-95.

1993-95 (%)

1961 -63 (%)

Total (ODO t)
Food
Feed
Processing
Seed
Waste
Net export'

AF

AS

LA

IC

AF

AS

LA

IC

2,755
84.8
3.7
0.0
1.1
104
0.0

87,143
79.7
14.5
0.0
0.2
5.5

2,578
65.8
24.5
0.0
0.2
9.5
00

7,483
60.1
23.6
10.7
3.8
1.8

5,748
85.3
2.9
00
0.7
11 .0
0.1

120,771
49.9
44.6
0.2
0.0
5.2
0.1

1,628
71.1
18.5
0.0
0.3
9.6
0.6

2,284
65.5
19.3
9.1
3.5
2.6

Source: FAOSTAT (June 1997, accessed July 1997).

Note: AF = Africa; AS = Asia; LA = Latin America; IC = Industrialized countries. Regions as defined by FAOSTAT.
= None recorded.
0

332

Sweelpolo lo

Tastes and preferences for fresh

sweetpotato are also highly variable and
show some signs of evolving as well. Some
consumers prefer sweet varieties with a
floury taste; others prefer bland roots.
Yellow-to-orange fleshed cultivars with high
G-carotene content have been introduced
and diffused recently in East Africa as part
of an integrated effort to reduce vitamin A
deficiency (Low et al., 1997). Wide genetic
variability found in sweetpotato means that
these and a number of other desirable traits
can be found in existing germplasm.
Recent research has documented the
widespread use of sweetpotato by smal I
farmers in their efforts to sustain local
I ivestock production systems (Scott, 1992;
Woolfe, 1992). In fact, virtually wherever
sweetpotato is cultivated, from Brazil to
Madagascar to China, some part of the
plant in some form is used in some type of
animal production. The steady increase in
the use of sweetpotato roots and vines in
pig and other livestock systems in China
over the last 30 years now means that from
30 to 50 mi 11 ion tons or more are used
annually as feed. Vines also play an
underexploited, but important role in
animal production in other Asian countries
such as Indonesia, the Philippines, and
Vietnam.
Processed products made from
sweetpotato including starch, noodles,
candy, desserts, and flour have long been
made by farm households to extend the
availability, diversify the use, and increase
the value-added for the crop. In China, in
particular, production of sweetpotato starch
in recent years has evolved into a cottage
industry that utilizes millions of tons of
roots per year as raw material inputs . The
magnitude of these new uses is not easy to
quantify in a systematic way; partly for that
reason, available statistics on processing do
not always reflect their true level of importance. Recent estimates from China suggest
that 3 to 5 mi 11 ion tons of sweetpotatoes are
transformed into starch to make noodles for
both the domestic and export market (see
Fuglie et al., 1999).

Sweetpotato Production
Driven by strong demand for feed and
starch, growth rates for sweetpotato output
and area planted have turned upward in
China after years of decline (Figure 2).
Trends for other countries and regions have
been mixed (Table 1 ). Given weaknesses in
the data, these figures should be interpreted
with caution.
Asia. Sweetpotato production in Asia
has been characterized by four trends. (1)
The continued overwhelming dominance of
China with recent positive growth rates
reversing an earlier decline. (2) Shrinking
area planted in sweetpotatoes-a trend that
accelerated in much of the region during
the last ten years. (3) Leveling off of yields
as the rate of growth has slowed in many
countries, including China. As sweetpotato
cultivation has been pushed onto more
marginal land and average yields have
improved to 17 t/ha, it has become more
difficult to maintain the rate of growth of
improvement in yields. (4) The possible
shift in the future prospects for regional
sweetpotato production due to recent
changes in relative prices for sweetpotato
versus traditional substitutes such as
imported wheat flour, as a consequence of
the economic crises in Southeast Asia.
Latin America and the Caribbean. For
much of this region, production and area
planted to sweetpotato is most important in
smaller, poorer countries such as Cuba,
Haiti, and Paraguay. In Cuba, a recent
sharp decline reflects the pressure on
sweetpotato yields resulting from a shortage
of chemical pesticides in the current
transition to biological control of important
pests. In bigger and/or wealthier countries,
the decline during the last decade represents more of a long-term trend toward
highe1-value crops, the use of farmland for
alternative purposes, or the migration of
small farmers to other occupations outside
agriculture. In Peru, production and yields
rose spectacularly over the last decade as
agro-climatic conditions improved, the
general economy went through structural
adjustment, and many small growers turned
CIP Program Report 199798

333

to sweetpotato because of the shortage of

farm credit and the low costs of production
per hectare.
Africa. The continent produces near ly 7
million tons of sweetpotato yearly and
near ly all of it south of the Sahara; Egypt in
North Africa is the prominent exceptio n.
Grow th rates in sweetpotato production
and, in particular, area planted are the
hi ghest of any region though most of the
major producers saw growth rates decline
over the last decade. As area planted
conti nu ed to expa nd , th e annua l average
rate of improvement in yields turned
negative in some cases (e.g., Uganda: 1.9%), and offset what wo uld have otherwise been faster rates of growth in production . In othe r words, as pl anting took place
under more marginal conditions, and
perhaps by farmers less acq uai nted w ith the
most approp riate cultural practices , yields
suffered in the process.

An average yield of 5 t/ha for

sweetpotato in Africa (Table 1) is the lowest
of any developing re gion, and less than a
third of the average yi eld in Asia. This
suggests amp le room for improvement in
productivity in the years ahead even taking
into co nsid erat ion the co mm on criti cism
that FAO data underestimate average yiel ds
for sweetpotato in this regio n. Prospects are
brightest for the adoption of yield-inc rea sin g production techn ology for sweetpotato
in Ea ste rn and Southern Africa where cash
sa les of fresh roots ha ve risen in importance
in a number of countri es. Although only a
minor share of current cons umpti on , the
eme rgi ng demand for processed products
made from sweetpotato shows signs of
increasing. This should further bolster
prospects for additional production. Africa n
farmers will continue to increase th eir
production of sweetpotato for food secur ity
and in come generati on in response to
continued high population growth and the
econom ic hardships assoc iated w ith
political instability, natural disasters and
limited production infrastru ctu re .

3 34

Sweetpototo

A Word of Caution
Any review of the re cent tre nd s in
sweetpotato prod uction, consumption , and
us e in developing countries would be
incompl ete were it not to draw attention to
discrepancies in data for this commodity. It
is difficult to est im ate produ ct ion for a crop
produced by smal l farmers on non-contiguous p lots , harves ted several tim es a year,
and not sold through regulated domestic
marketing channels or trad ed abroad in
appreciable quantities. Therefore, FAO
statistic ians frequently resort to using the
availab le national statistics to estimate
production , area, and y ield . U nfortunate ly,
there are ofte n dis crepancies between FAO
figures and national data. For example,
FAO reports Malawi produces no
sweetpotato, while Ministry of Agriculture
figures show the coun tr y harves ted over
800,000 t during 1995-97 (Phiri, 1998).
Th is is not an isolated case, especially
w hen one goes beyo nd the national
statist ics and compares tho se figures with
data gathered in farm surveys as reported in
the gray literature or consults with com modity spec iali sts based in the countries
th emse lves. Similar problems apply in the
case of the figures for utilization. In conc lu sion, readers are adv ised to use these
" trends" wi th cautio n.
Conclusions

Sweetpotato is an importa nt so u rce of food,

cash, feed , and raw material fo r processing
in developing countr ies. Although production is hi gh ly co ncentra ted in China, recent
yea rs ha ve witnessed a surge in
sweetpotato output in a number of SubSaharan Afr ica n countries. With the
germp lasm for sweetpotato large ly
und erexploited, amp le prospects exist to
deve lop va rieti es with particular traits in
addit ion to yield to satisfy the emerg in g
demand for diversified encl uses. Development of the crop's full potential in the new
millennium wi ll require a fully integrated
approach in wh ich efforts aimed at
germplasm improve ment are c lo se ly ti ed to
th e following.

 Identification of market segments for

which there is the greatest demandcurrently this involves feed and starch in
East Asia, and flour and fresh uses in
Latin America and Africa.
Improvements in the technical efficiency
of processing (e.g., via new or improved
small equipment) as well as feed use and
preparation (e.g., vine fermentation).
Reductions in per unit production and
marketing costs through better cultural
practices and trading procedures.
Strengthening linkages of producers as
sources of raw material to processors,
managerial improvements in sweetpotato
processing at the enterprise level as well
as more informed utilization of fresh
and processed sweetpotato by final
consumers.
References Cited

FAOSTAT, 1997. Statistics Database

(Online). June. Available HTTP: http://
apps.fao.org
. 1998. Statistics Database (Online).
June. Available HTTP: http://apps.fao.org
Fuglie, K., L. Zhang, L. Salazar, and T.
Walker, 1999. Economic impact of
virus-free sweetpotato planting material
in Shandong Province, China. Program
Report 1997-98. CIP, Lima, Peru. p. 249254.
Hall, A., G . Bockett, and S. Nahdy. 1998.
Sweetpotato postharvest systems in
Uganda: Strategies, constraints and
potentials. Social Science Department
Working Paper 1998-7. International
Potato Center (CIP). Lima, Peru.
Low, J., P.Kinyae, M.A.Oyunga, E.Carey,
and J.Kabira. 1997. Combating vitamin A

deficiency through the use of

sweetpotato. Results from phase 1 of an
action research project in south
Nayanza, Kenya. A co-publication of the
Kenyan Agricultural Research Institute
(KARI) and CIP. CIP. Lima, Peru.
Phiri, M.A. 1998. Grain legumes: Issues and
options for research in Malawi. Paper
presented at the international training
course on Methods for Analyzing
Agricultural Markets held 9-15 November 1998 in Nairobi, Kenya. (Mimeo.)
Scott, G.J. 1992. Sweetpotatoes as animal
feed in developing countries: Present
patterns and future prospects. In: Machin
D. and S. Nyvold (eds.). Roots, tubers,
plantains and bananas in animal feeding.
FAO Animal Production and Health
Paper No. 95. FAQ, Rome, Italy. p: 183202.
Scott, G.J. and L. Maldonado. 1999.
Sweetpotato facts. A compendium of key
figures and analysis for 30 important
sweetpotato-producing countries. CIP.
Lima, Peru.
Scott, G.J., M.W. Rosegrant, and C. Ringler.
1999. Roots and Tubers for the 21st
Century: Trends, projections, and policy
for developing countries. IFPRI Food,
Agriculture and the Environment Discussion Paper. A co-publication of the
International Food Policy Research
Institute (IFPRI) and CIP. IFPRI, Washington, D.C., USA.
Woolfe, J. 1992. Sweetpotato. An untapped
food resource. History, chemical composition, postharvest procedures, livestock
feed and consumption and uti I ization
patterns and trends. Co-published by CIP
and Cambridge University Press. Cambridge, U.K.

CIP Program Report 199798

335

Sweetpotato in Ugandan Food Systems: Enhancing

Food Security and Alleviating Poverty
G. J. Scott\ J. Otieno1, S.B. Ferris 2, A.K. Muganga 2, and L. Maldonado 1

Uganda is one of the wor ld 's poorest and

least developed countries wi th a GN P per
cap ita of US$320 (Wor ld Bank, 1998/ 99).
Uganda's eco nomy and 90% of its population depend on ag riculture fo r food,
emp loyme nt, and income (UNDP, 1999).
Curre nt population stands at sli ghtly over
20 million and is growin g at over 2.8%/yr.
(UNDP, 1999). Some 85% of the people
reside in rural areas .
Given the fragile bal ance between
population, food produ cti on, and eco nomi c
gro wth, government pol icy seeks to ensu re
the co untry's con tinu ed ab ility to susta in
food se lf-suffic iency. In creases in agr icultural productivity and agroe nterpr ise
development are key elements in the
governm ent's strategy to in crease the
in comes of rural household s and to fac ilitate further expans ion of the overal l
eco nomy. These concerns are particularly
acute in the Ea stern Reg ion in Soroti and
Ku mi districts, and in th e North ern Region
in Lira and Apac districts. Th ese districts are
among th e poores t in th e country (UN DP,
1999).
Uganda is the biggest sweetpotato
(ip omoea batatas) produ ce r in Africa with
an annu al ave rage output of near ly 2
million t (Sco tt and Maldonado, 1999).
Output ha s stea dil y expanded in rece nt
yea rs. Co nsequ entl y researchers, development project personnel , and Ministry of
Agriculture officials have become in creasin gly interes ted in th e rol e of sweetpo tato in
Uga nd an food systems and its potential to
en hance food security and in crease rural

in comes (Bashaas ha et al., 199 5; Fow ler

and Stab rawa, 1992; Hall et al., 1998) . Th e
attention to sweetpotato also reflects a more
general interest in roots and tubers, and
matooke (cook in g banana) over th e last
decade (Bashaasha an d Mwanga, 1992;
Otim-N ape and Opio-Odongo, 199 2).
Thes e c rops accou nt for a major share of
annu al food production and consumption.
Th e focus on sweetpotato in Uganda has
also in te nsifi ed sin ce 1990 wi th th e spread
of th e more virul ent Ugandan form of the
cassava mosaic v iru s disease (CMD).
Th is pape r ana lyzes trends in
sweetpotato producti on, area planted,
prices, co nsu mpti on, and us e in Uganda
over th e last 1 0-15 yr. After a brief rev iew
of food produ ct ion and use nationwid e, th e
an alys is turns to particular regions and
di str icts. The analysis addresses the fo llow ing key questions regarding th e evo lv in g
role of sweetpotato in Uga ndan food
systems:
What has been the evo luti on of production and area planted nation al ly by
region and by district?
H as this evo luti on shifted the di str ibu tion
of sweetpotato production across th e
country and its relat ive importan ce in
loca l food systems, es peciall y in areas
w here it substitu tes for or complements
cassava?
W hat has bee n the effect of increased
produ ct ion on pri ces, relative pri ces,
consumption, and use?
How have these trends affected the ro le
of sweetpotato in Ugandan food systems,
parti cul ar ly in the poorer regions and
distri cts o f the co untry?

1 CIP, Lima, Peru.

2 llTA-ESARC, Kampa la, Uganda.

CIP Prngrnm Report 199798

33 7

Material and Methods

Information resources
This paper draw s upon a ra nge of
prim ary and secondary sources. Secondary
inform ation includes (1) official Ugandan
stati stic s on production , prices, and co nsurnption , (2) FAO estimates of th e calories
and protein pro v ided by th ese com mo dities
in relati o n to ot her major foods , and (3) a
selecti ve rev iew of th e rece nt literature on
sweetp otato and other root and tuber crops
in Uganda. Pr imary sources in c lud e
inform at ion gathered in inforrnal interviews
with key in fo rmants in 1994, 1996, and
1998, as wel l as with knowl edgea bl e
spec ialists in organizations elsewhe re, e.g.,
N atural Resources Institute, U I<, and FAORom e, and participant observat ion of rural
and urb an root crop process in g and
m arketin g.
Conceptual framework
Thi s stud y uses a food sys tem s co nce ptu al fr am ewo rk co nsisting of four cent ral
tra its. One is the focus on th e interrelation
and interaction between different subsystem s, e.g., farming (p rod uc ti o n), rnark eting (exc hange), and u se (co nsurnption )
sys tems .
A secon d feature is that eac h su bsystem
is co nsi d ered as a hori zo ntal conce pt. In
thi s framewo rk, farrnin g sys tems as a
co nce pt encompasses al I aspects of rural
act iv ity (c ropping, livesto ck, and fores try
system s) associated with agricu ltural
production. In contrast, the food syste m is
esse ntially a ve rtical concept. Th e foc us of
ana lys is then is not th e interrelati o n of
acti v iti es at a particular ho ri zo ntal leve l but
rath er the interre lati o n betwee n sub systems
of th e foo d syste m - from produ ction
through the marketin g chai n to th e consum er (bo th rural and urban), inc ludin g any
proces sing and vai ue- added acti v iti es (Hall
et al. , 1998).
A third key fe ature is that of a hi storical
pers pecti ve . Instead of analyz in g in detail
intera c tions at a gi ve n point in tim e, the

338

Sweelpololo

food syste rn s approach focuses mo re on

med iurn- to long-term trends in ke y va riabl es. In this way we ca pture th e evolv in g
nature of the interrelation betwee n sub sys tems.
Finall y, w ithin a co untry, the food
syste ms approac h freq uentl y examin es
nation al and interreg ional deve lopments as
opp osed to mme locat io n-specific trend s.
Thi s as pect of th e food systerns ap p roach is
intend ed as a macrocomplement to an alys is
of part ic u lar agr icultural acti v ities at the
m ic rol evel (Morris, 1995).

Results and Discussion

Role of sweetpotato in Ugandan
food systems
Th e ave ra ge Uganda n diet contains
staples such as ma too ke, cass ava,
sw eetpotato 0 1 cereals suc h as maize. Th e
typi ca l mea l co nsists of o ne or more of
these co mm o diti es ac companied by a
sauc e rnad e of beans, veg etables, fish, o r
meat (Bas haa sha et al., 1995). H al l et al.
(1 998) id entifi ed at least five distinct rol es
for sweetpotato in Uga ndan food systems,
four of th em pec uli ar to rural areas and one
associJted prim aril y w ith urban areas.
Th ey are :
1. A pred o m inant staple provi din g the
ma jority of ca lor ies for most o f th e year
(swee tp o tato product io n is seasona l and
processed roots rarely la st 4 mo).
Ex ampl es are found in Soroti, Kumi , and
parts of no rth ern districts (Fi gure 1 ).
2 . On e of a numb er of major compl em entary stapl es co nsumed throu ghout th e
yea r but w ith seaso nal peaks, for example , in Kabarole , Masindi , Kab ale, and
lga nga di stri cts.
3. A famine re se rve stap le ty picall y co nsumed on ly in sig nificant quantiti es
durin g shortages of the d om in ant stap le.
4. A source of cas h income either produced
str ictly for sa le as in Mpigi Distri ct nea r
Kamp ala (Bas haasha et al., 19 95 ), or
more commo nl y as a source of reve nu e
from petty tr ad in g. That is especially tru e
in th e Northern an d Eastern regions

Democratic
Republic of
the Congo

Lake Victoria

Ro ug hly defines the stud y area which incl udes high altitude, temperature zones in the
north west, pastoral dry to sem i-arid rangeland in the area around Lake Kyoga and grassland
in th e north and east.
District boundaries and nam es shown here reflect th e situati on at the tim e this study was done .

Figure 1. Ugandan administrative districts and agroecological zones.

where small farmers have relatively few
opportuniti es to ea rn cas h.
5. A low-p riced , complementary stap le for
the poor and lowe r-in co me urb an groups
(Mwesigwa, 1995).
Production trends
Sweetpotato produ ct ion in Uganda rose
from 1.7 to 1.9 million t, or slightly ove r
12%, between 1987-89 and 1995 -97 .
During the same period, maize outp ut
jumped 70% and matooke produ ct ion
in creased by an estimated 1. 9 million t. In
contrast, cassava produ ct ion fell from 3.3 to

2.2 milli on t, or roughly 32% (Tabl e 1).

Hence, sweetpotato assumed slightl y
greater impo rta nce in domesti c food
supplies. Th e in crease in sweetpotato
production, however, was less noteworthy
than the sharp decline in cassava output or
th e production in creases for maize and
matooke.
Fi ve of th e top 1 0 cassava-produ cing
distri cts in 1987 -89 were among th e top 10
sweetpotato-prod ucin g districts as well. By
1995-97, the number had ris en to 7 of 10.
Cas sava production declined in 9 of the top
CIPPmgrom Reporl 1997-98

339

Table 1. Uganda food crop production, 1987-89 vs. 1995-97.

1987-89
Crop

Production

1995-97
%

{0001)
Matooke (cooking banana)
Cassava
Sweetpotato
Maize
Millet
Sorghum
Beans
Potato
Groundnuts
Rice
Soybean
Sesame
Pigeon peas
Cowpeas

7,267
3,313
1,683
474
569
335
341
207
134
13
35
38
38
14,446

Production

{000 t)
50.3
22.9
11.6
3.3
3.9
2.3
2.4
1.4
0.9
0.0
0. 1
0.2
0.3
0.3

9, 153
2,2 53
1,888
804
525
330
282
360
120
BO
83
72
58
46
16,055

57.0
14.0
11 .B
5.0
3.3
2. 1
1.8
2.2
0.7
0.5
0.5
0.5
0.4
0.3

Source: Ministry of Agriculture.

Estimates are of economic production (ofter making allowance fo r postharvest losses) not of harvested product.

10 d istri cts betwee n 1987-89 and 1995 -97 .

Furthermo re, th e to p 10 producing d istricts
in 1995-97 acco unted fo r nea rl y 50% of th e
1. 1 m illi o n t decl ine in nati o nal prod ucti o n
sin ce 1987-89 .
The oppos ite occurred in the top 10
sweetpotato di stri cts (Tab le 2). The average
in crease in sweetpotato product io n fo r
these d istri cts fo r the per iod was ju st 9, 155
t. In ad diti o n to the adva nce of CMD, th e
period 1987-1 99 0 was ma rked by c iv il
unrest in ce rt ain eastern di stri cts that
tempora ril y redu ced prod uctio n of al I food
c rops. Betwee n 1987-89 and 1995-97,
o nl y Kitgum showed a sharp increase in
sweetpotato o utput (26,567 t) in re lati o n to
the decl in e in cassava product io n (- 17,536
t). In G ulu , cas sava o utp ut inc reased ove r
the pe ri od by 5, 1 76 t, but sweetpotato
prod uct io n jumped by 48%, o r 29,68 1 t.
Fo r the other to p 10 di stri cts, the in c rease in
sweetpotato o utput per se was modest and

340

Sweetpototo

also in relat io n to th e mu c h large r decl in e

in cassava . O nly th ree of th e cassavaprod ucing districts reco rd ed a comb in ation
of a sharp decl in e in cassava o utput and a
j ump in sweetpotato p rodu cti o n. In some
di stricts prod uctio n of both crops decl in ed ,
al though typ ica ll y th e dro p was mu c h
greate r for cassava than fo r sweetpotato.
Fo ll ow in g a po li cy of dece ntrali zatio n,
U gand a has c reated a number of new
di stri cts by subd iv id in g some large r di stri cts, e.g. , Katakw i o ut of So roti . Th e
c hange in the num ber of di stri cts and their
respecti ve size may affect o ur analys is of
di stri ct prod uctio n figures fo r 1987-89 and
1995-97. Th e acc uracy of di stri ct- leve l data
o n acreage and y ields may also suffe r.
H owever, there is li tt le reaso n to be li eve
th at th e data fo r sweetpotato are mo re
acc urate o r less acc urate than those fo r
oth er root c rops. The relati o n betwee n
maj o r product ion in c reases in sweetpotato

Table 2. Uganda sweetpotato and cassava production, variation(%) and ranking, 1987-89 versus 1995-97.'

Cassava

Sweetpotato
District

Production
1995-97

Change(%)
VS

1987-89

Ranking

District

1995-97

1987-89

127
105
98
98
94
93
92
91
80
72
1,888

Change(%)

Ranking

1987-89

1987-89

VS

(000 t)

(000 t)
Mbale
lganga
Hoima
Apac
Kitgum
Masindi
Ku mi
Gulu
Kamuli
Soroti
Total

Production

5.6
2.8
l.l

7.9
39.5
22.2
-2.8
48.3
1.2
3.4
12.2

2
3
7
11
9
5
13
8
10

Mb ale
lgonga
Apac
Arua
Kitgum
Gulu
Lira

Ku mi
Kamuli
Nebbi
Total

169
169
141
139
125
119
118
117

107
100
2,253

-37.4
-12.8
-27.4
-41.7
-12.3
4.6
-33.8
-38.7
-39.l
-26.2
-32.0

l
5
4
2
11
16
7
6
8
13

Source: Ministry of Agriculture.

'These district-level data ore for the top ten sweetpotato- and cassava-producing districts in 1995-97 only. The totals are for
all districts.

and declines in cassava at the district level

has been limited to a small number of
districts.
Sweetpotato rose in importance in both
the Eastern and Northern regions (Table 3).
Matooke accounted for the largest absolute
increase in food production in the Eastern
Region followed by maize and sweetpotato
(Table 3). These three crops then substituted
for the decline in cassava output. In
contrast, cassava fell sharply in absolute
output and in re lation to the other principal
food crops. In the Eastern Region farmers
increasingly see sweetpotato as a source of
cash income, particularly in the absence of
cassava. That is consistent with Uganda's
objective of raisin g rural incomes in the
region given th e greater incidence of
poverty and lower levels of human development (Table 4) .

high er than any other crop in absolute

tonnage (Table 3). Maize production rose
by 142,000 t. The increase in matooke
output has also been notewo rthy (84,000 t).
But not on th e scale of growth in the
Eastern Reg ion where growing co nditions
for matooke are more favorable. These
production changes are particularly
important because the Northern Region and
its districts are the poorest in all the country
(Table 4).
Statistics for cassava and sweetpotato
production in the Eastern and Northern
regions in 1997 indic ate that the slight
upturn in cassava output has been accompanied by a similar rebound for
sweetpotato (Figure 2). That suggests
farmers are producing more of both crops
rather than opting for one over the other.

Price trends
In the North ern Region, the increase in
sweetpotato production (151,000 t) was

Fresh cassava and sweetpotato, along

with matooke, serve as substitutes in

CIP Program Report 1997-98

341

Table 3. Food crop production in the Eastern and Northern regions of Uganda, 1987-89 vs. 1994-96. 1
1987-89
Crop

Matooke (cooking banana)

Sweetpotato
Cassava
Maize
Millet
Potato
Beans
Sorghum
Groundnuts
Total

Production (000 t)

1994-96
%

Production (000 t)

East

North

East

North

East

North

East

North

1,024
631
l, 198
186
259
9
122
85
49
3,563

138
349
854
105
167
2

28.7
17.7
33.6
5.2
7.3
0.2
3.4
2.4
1.4

7.3
18.5
45.5
5.6
8.9
0.1
4.1
7.6
2.4

1,401
794
782
345
261
84
84
8
46
3,882

222

36. l
20.5
20.l
8.9
6.7
2.2
2.2
2.2
1.2

9.9
22.4
29.4
l l. l
9.5
0.7
4.8
9.5
2.6

77

142
46
1,880

500
656
247
212
17
107
211
59
2,230

Source: Ministry of Agriculture.

Eastern Region includes the districts of Kapchorwa, Mbale, Katakwi, Kumi, Pallisa, Tororo, Busia, Bugiri, Soroti, Kamuli, lgango,
and Jinja; Northern Region includes the districts of Aruo, Moya, Adjumani, Kitgum, Kotido, Nebbi, Gulu, Apac, Lira, and Moroto.

Ugandan diets (Bashaasha et al., 1995).

Prices for these two crops have tended to
move together in recent years (F igure 2).
With changes in production, however,
relative prices for sweetpotato vs cassava
have shifted. In a number of urban markets, sweetpotato has become cheaper than
cassava, reflecting the drop in cassava
production and the increase in sweetpotato
output (Figure 2). That suggests urban
consumers may be inclined to eat more
sweetpotato, particularl y those w ith limited
food budgets.

Consumption
Roots and tubers, including matooke,
account for some 44% of the average daily
caloric intake of 2194 kcal in Uganda (FAQ
STAT, 1998). The decline in cassava
production and the concomitant rise in
matooke and sweetpotato output have
shifted the percentage that each commodity
contributes to the diet. Sweetpotato's share
declined from 12% in 1984-86 to 10% in
1 994-96. Cassava's share shrank from 19%
to 10%.

34 2

Sweetpolalo

Diets in Uganda differ by region and by

time of year. Matooke, for example, is
much more important in the area around
Kampala for agronomic and cultural
reasons (Mwesigwa, 1995). Sweetpotato
assumes greater relative importance in parts
of the Eastern and Northern regions and
generally in July, August, and September,
during and immediately after peak harvest,
and during seasonal shortages of other
crops (Bashaasha et al., 1995; Smit, 1997).
Recent shifts in production certainly suggest
sweetpotato's increasing importance in
regional diets, particularly in the Northern
Region.
The last decade has witnessed the
emergence of sharp differences between
rural and urban food expenditures (Table 5 ).
Rural food expenditures were two-thirds of
urban expenditures in 1989/ 90. They had
dropped to nearly half of urban expenditures by 1993/ 94 . Inasmuch as 85% of the
population resides in the countryside, the
national average nevertheless remained
closely tied to expenditure levels in rural
areas.

Table 4. Regional and selected district human development and poverty indices, 1996.

Region

Human development index'

Central

0.4970

Eastern

0.3620
0.3353
0.3230

Ka mu Ii
Soroti
Kumi
lgango
Northern
Kitgum
Gulu

Lira
Apoc
Western
Uganda

Human poverty indexb

31.2
40.0
52.9

0.3170
0.2644

52.0
51.0
48.4
45.7
59.5

0.3 179
0.3496

53.2
51.5

0.3670
0.3730

46. 9
39.3
39.3

0.2985
0.3661

0.4046

Source: UNDP, 1999.

The human development index {HDI) measures human prog ress on the basis of achievement in th ree broad indicators:
longevity, educational attainment, and standard of living. Longevity is measured by life expectancy at birth, educational at
tain me nt by a combination of adul t litera cy and school enro lment ratio, and standa rd of living by real GDP per capita {PPP$). The
HDI is an aggrega tion of these key indicators. The index ranges from0 to 1 where 0 indicates total absence of development and
l indicates the highest level of development.
b Poverty is interpreted to mea n total deprivation in a range of human capabilities. The three types of deprivation used in
computing the human poverty index ore only representative, but they capture a wide range of dimensions of life that
matter most. The first aspect of deprivation relates to survival - the vul nerability to death at a relatively early age. The
second relates to kn owledge - being excluded fromthe world of reading and comm un ication. The thi rd relates to a decent
standard of living.
0

Rece nt ho use ho ld ex penditure surveys

also prov id e in fo rm ati o n o n th e size and
evoluti on of mo netary o utl ays for
sweetpotato and cassava (Tabl e 5). Th ey
reco rd ed not o nl y cas h purc hases fo r food
but im puted expenditure equi va lents fo r
own -pro du ced foo d, es pec ially in rural
areas. M o reover, all fi gures are in no rnin al
US do ll ars, i.e., with no adju stmen t fo r
infl ati o n. Fo r sweetpo tato, rural o utlays
in creased both pe rce ntage-w ise and
no min all y. U rba n ex pend itures dec lin ed
percentage-w ise b ut rose sli ghtl y nomin all y.
Total expe nditures o n foo d rose mu ch mo re
rapidl y in the c iti es th an in the countrys ide
betwee n 1989/90 and 1993/94. Cassava
expenditures fo ll ow ed a simil ar pattern , but
the no min al in creases in rural areas we re
mo re modest. Co nsequ entl y, expenditures

on sw eetpotato in rural areas both pe rce ntage-w ise and nomin all y w ere hi gher than
for cassava in 1993/94, th ereby effecti ve ly
reve rsing th e situ ati o n th at ex isted in 1989/
90 . Fo r cassava, th e percentage dec lin e in
ex pend itures in urban a1eas was less
pron oun ced th an for sweetpotato, but not
eno ugh to co mpensate fo r th e mo re rn odest
ri se in ru ra l areas. As a res ult, th e ave rage
percen tage of total food ex penditures
devoted to cassava nati o nw id e decl in ed,
w hereas th at fo r sweetpotato remain ed
co nsta nt.
Conclusions and Recommendations
Sweetpotato produ ct io n has stea dil y
increased in U ga nd a ove r th e last 10- 15 y r.
Th e in c rease in sweetpotato o utput,

CIPProgram Repor t 1997-98

343

Ugandan Sh./kg

500

- - Fresh cassava
Sweetpotato

400

1.5

Fresh cassava/swe etpotato

300
200

1: ~"~. . . . . . .
1988

1989

1990

1991

1992

1993

1994

1995

1996

1997 1998

Years (beginning in December)

Figure 2. Trend of nominal and relative monthly prices for fresh cassa va and sweetpotato prices in Kampala ,

Uganda, December 1988 to Dece mber 1998. Source: Statistics Department, Ministry of Finance,
Planning and Econom ic Development, as presente d in Ferris et al., 1999.

(000 t)
1400
1200

--

-= -

1000

--

- --

800
-

400
200

.....

1988

1989

.-

1994

1995

--

1996

1997

,...

-~

Sweetpotato (East)

D Cassava (East)
D Sweetpotato (North)
D Cassava (North)
,_

600

1990

1991

1992

1993

Figure 3. Cassa va and sweetpotato production in Eastern and Northern Regions of Uganda , 1988-97. Source :

Ministry of Agriculture.

344

Sweet polo to

Table 5. Expenditures on roots and tubers as a percentage of total food expenditures, Uganda.'

1989/90
Urban
Food expenditure
(000 Uganda Sh.)
31.57
Total roots and tubers
8.0
Potatoes
0.8
Sweetpotato
3.7
Cassava
3.4
Yams
0.1
Yams/other tubers
Matooke (cooking banana)

1993/94

Rural

National

Urban

Rural

National

22.23
14.4
0.9
6.4
6.9
0.3

23.45
13.3
0.9
5.9
6.3
0.3

75.66
10.9
0.8
2.1
1. 9

33.26
24.l
0.7

39.59
19.7
0.7
5.9
5.5

0.0
8.4

00
7.6

6.0

7.7
7.3

0.0

Source: ROU, 1994; 1997.

All monetary figures are in nominal Uganda Sh.; includes cash and imputed expenditures on own-produced food. (1993/94 US$ l
= Uganda Sh. 1, 100)

however, represents only a minor percentage of the total decline in cassava production. The data further show that the
precipitous drop in cassava output in
certain districts was only w eakly related to
changes in sweetpotato production. But at
the regional level, particularly the poorer
Northern Region, the shift to sweetpotato in
response to the decline in cassava output
was more pronounced. The recent rebound
in cassava production has been accompanied by increases in sweetpotato output.
Average monthly retai I prices over the last
10 yr for Kampala, the capital, and four
regional urban markets indicate that
sweetpotato and cassava prices tend to
move together. More important,
sweetpotato is now less expensive than
cassava. Food expenditures for
sweetpotato in rural areas have also
increased.
The impacts of the various trends in the
role of sweetpotato in Ugandan food
systems then are as fol low:
Sweetpotato has achieved a higher
relative and absolute importance in
cropping systems, given the simultaneous sharp decline in cassava
production .

Sweetpotato has increased in importance

as a source of cash income, with lower
relative prices currently making
sweetpotato more competitive with
cassava.
The decline in cassava production has
seen sweetpotato play a greater rol e in
ensuring food security in the country,
especially in those areas where the crop
is a staple.
The bulk of the increase in sweetpotato
output has been achieved as a result of
increases in area planted (FAOSTAT, April
1999). Hence, the introduction and diffu sion of improved germplasm has the
potential to increase sweetpotato output
and yields, lower costs per kilo harvested,
and help to make sweetpotato even more
competitive in the years to come. Increased
availability and lower relative prices for
sweetpotato enhance the prospects for
economically viable processing activities.
Further work on product development
offers the prospect of providing growers/
processors with alternative markets and
therefore an added incentive to adopt yield i ncreasi ng technology. Production and sale
of processed sweetpotato products can

CIP Program Report 1997-98

345

contribute eve n mo re tha n in the 1e cent

past to generati ng value-add ed income fo r
rete ntio n in rural areas. That is es peciall y
tr ue for th e Ea stern and North ern reg io ns
w here fa rm ho use ho lds have few altern ative sources of incom e and the in c id ence of
rural po ve rty is most ac ute.
References Cited

Ba shaa sha, B. and R.0.M Mwanga. 1992.

Sw eetpotato: A so urce of in co me fo r lowincome rural fa mil ies in U ga nda. In:
Scott, G. , P.I. Ferguso n, and J.E. Herrera
(eds.). Produ ct deve lopment fo r Roo t and
Tu ber Crops. Vol. Il l - Africa. Procee din gs of th e w orksho p on process in g,
marketin g, and utili za tion of roo t and
tuber c rops in Africa, held Octobe r 26 November 2, 199 1 at the In ternationa l
In sti tute for Tropica l Ag1 icul ture (ll TA),
Iba d an, N ige ria. CIP (Intern atio nal
Potato Cente r), Lima , Peru .
Bas haasha, B., R. Mwanga, C. Oci tt i
p' Obo yoa, and P. Ew ell . 1995 .
Sweetpotato in th e fa rm in g and food
system of U ga nd a: A farm survey report.
CIP (In tern ational Potato Center), Lim a,
Peru .
FAOSTAT. 1998 . FAQ (Food and Agr iculture
Organiz ati o n of the U nited Nation s)
Stati sti cs Database (O nlin e) . Accesse d
Ju ne 1999 . Ava il ab le HTTP: http://
app s. fao .org.
___ . 1999 . FA Q (Food and Agricu lture,
Organiz at ion of the United Nat ion s)
Stati stics D atabase (O nlin e). Accesse d
Apri l 1999. Ava ila bl e HTTP: http://
app s.fao.org.
Ferris , S. , A. M uganga, R. M atov u, S. Kolijn ,
V. H agenima na, and E. Karu ri. 1999.
Marketi ng o pp ortunities fo 1 starc h and
hi gh qu ality flou1 produc tio n from
cassava and sweetpotato in U ganda.
Col laborati ve proj ect by Internat ional
In st itu te of Tropi ca l Ag ri cu lture (llTA),
U ga nd an National Postharvest
Programm e, Food and Ag ri culture
O rgani sat ion, Intern ation al Po tato
Ce nte r, and Mati lo ng Yo uth Mi xed
Farm ers 01ga ni sat io n (Mi meo.)

3 4 6 Sweetpototo

Fo w ler, M. H. and A.M . Stabrawa. 1992. A

rap id ap prai sal of sweet potato marketing
in U ganda. In : Scott, G., P. I. Fe1guson,
and J.E. H errera (eels.). Produc t deve lopmen t for Root and Tuber Crops . Vo l. Ill A frica . Proceedi ngs of th e wo rkshop on
processing, marketin g, and uti li zat ion of
root and tuber crops in Africa, held
October 26 -No ve mber 2, 199 1 at the
Intern atio nal In st itute fo r Trop ical
Ag ri cul ture (llTA), Ibadan, N ige ria. CIP,
Lim a, Peru .
H al l, A., G. Boc ke t, and S. Na hd y. 1998.
Swee tpotato po st harves t sy stems in
U ga nd a: Strategies , co nstrai nts , and
po tentia ls. Social Sci ence D epartment
Wo rkin g Paper No. 1998 -7. In ternati o nal Potato Cen ter (CIP ), Lim a, Peru.
Morr is, M . 199 5. Rap id reco nn aissa nce
methods fo r diagnos is of sub-sector
l imitati ons: Ma ize in Parag uay. In: G.
Scott (eel. ). Prices , produ cts and peop le:
A nalyz in g ag ricultur al marke ts in
deve lopi ng cou ntri es . CIP, Lima , Peru.
Mwes igwa, T.W. 1995. Swee t potato
co nsumption leve ls and pattern s in urban
ho use ho ld w ith pa rt icul ar refe rence to
Kampala c ity. Unpub li shed B.Sc. thes is.
Departme nt of Agr iculture, M akerere
U nive rsity, Kamp ala, U ganda.
Otim -Nape, G. W . and J.U .A. Opio Odon go . 1992. Cassava in U ga nda. In:
Scott, G ., P.I. Ferguso n, and J. E. Herre ra
(e ds.). Produ ct deve lopm ent fo r Root and
Tube r Crops . Vo l. Ill - Africa . Proc eedin gs of th e wo rk shop o n processi ng,
ma rketin g, and utili za ti o n of root an d
tub er cro ps in Africa, hel d October 26 Nove mber 2, 1991 at th e In ternational
In stitute for Tropic al Ag ricu lture (llTA),
Ibada n, N ige ri a. CIP, Lim a, Peru.
ROU (Repub li c of Uga nd a) . 199 4. H ouseho ld Bud get Survey 1989/ 90. M ini stry of
Finance, Pl annin g, and Econo m ic
Deve lopm ent. Kampal a.
____ . 1997 . H o usehold Bud ge t Survey
1993/9 4. M ini str y of Fi nance, Plannin g,
and Econo m ic Develo pment. Kampa la.
Scott, G .J. and L. Maldonado. 1999. CIP
sweetp otato facts . A compe ndium of key
fi gures and anal ys is fo r 33 important

sweetpotato-producing countries.
International Potato Center (CIP). Lima,
Peru. (brochure.)
Smit, N. 1997. The effects of indigenous
cultural practices of in ground storage
and piece meal harvesting of sweet
potato on yield and quality losses caused
by sweet potato weevil in Uganda. Ag.,
Ecosys. Env. 64:191-200.

UNDP (United Nations Development

Programme). 1999 . Uganda Human
Development Report 1998. UNDP,
Kampala, Uganda.
World Bank. 1998/99. World Development
Report. World Bank, Washington, D.C.

CIP Program Re port 199798

34 7

Research on
Potato and Sweetpotato
Impact on a Changing World

In Vitro Conservation of Potato and

Sweetpotato Germplasm
A. Golmirzaie and J. Toledo 1

In vitro conservation of genetic resources

has advanced considerably durin g this
decade. Since 1975, CIP has contributed to
developing tissue culture techniques for
conserving germplasm of potato (Solanum
tuberosum) (Roca, 1975), sweetpotato
Upomoea batatas) (Siguenas, 1987), and
Andean root and tuber crops (Toledo et al.,
1994). In vitro conservation is the most
useful and efficient way to distribute clonal
materials. It facilitates the availability of
planting materials at any time, avoids the
transfer of major pests and pathoge ns, and
makes possible virus eradication through
meristem culture (Roca et al., 1979;
George, 1993) . In addition, in vitro conservation is less expensive than cryopreservation of field-grown clonal materials
(Florkowski and Jarret, 1990).
Short- and Medium-Term Conservation
In vitro plantlets grow in a MurashigeSkoog (MS) culture medium containing
minerals, a carbon (C) source, vitamins, and
a low concentration of growth regulators
(Table 1). Plants exhaust the nutrients in this
medium in 2-3 mo; therefore, in vitro plants
have to be transferred frequently to fresh
medium. The culture rooms generally have
a temperature range similar to those needed
to grow a given crop in the field. That is
called short-term conservation. The
interval between subcultures, however, can
be extended through growth rate reduction
by modifications to the environment or
changes in some media components.
The addition of osmoticums or growth
retardants to the medium has proved

efficient for reducing growth rates of

different plant species. Osmoticums such as
mannitol or sorbitol reduce min eral uptake
by cells through differences in osmotic
pressures thereby retarding plant growth
(Dodds and Roberts, 1985; Thompson et
al., 1986). Growth retardants can produce
some physiological changes or generate
mutations, which can threaten the genetic
stability of the materials conserved in vitro
(Hughes, 1981; Lizarraga et al., 1989;
Wescott, 1981 ). Reducing growth temperature close to 0C for temperate plant species
or several degrees below normal for
tropical crops can also minimize the growth
rate in many crops (Dodds and Roberts,
1985; George and Sherrington, 1984). In
vitro plants growing in closed culture
vessels have low concentrations of C0 2 ; C
absorption is maintained by supplementing
the medium with sugar. Reducing light
intensity also affects growth rate by reducing photosynthetic requ i rem en ts and
therefo re metab olism (Hughes, 1981 ).
A combination of osmoticums, low
temperature, and low light intensity has
been the most effective in lengthening
periods between subcultures. At CIP, these
procedures are applied to potato and
sweetpotato with good resu It~; .
In vitro potato conservation
Th ere are well-established protocols for
micropropagation of potato (Espinoza et al.,
1992), meristem culture (Liza1 raga et al.,
1991 ), in vitro tuber inductior (Estrada et
al., 1986), medium-term storage
(Golmirzaie and Toledo, 1998), and
cryopreservation (Golmirzaie and Panta,
1997).

1 CIP, Lima, Peru.

CIP Program Report 1997-98

351

Table 1. Components of culture media al potato and sweetpotato.

Sweetpotato culture media'

Potato culture media

Ingredients

Propagation Conservation

Propagation I Propagation II Conservation

Tuber
induction

MSb(I L)
Stock l
Stock solution
25
Sucrose (g/L}
0.1
Gibberelic acid (mg/L}
Espermidine (mg.IL}
Naphthalene acetic acid
(mg.IL)
Sorbitol g/L
Benzylaminopurine
(mg.IL}
Chlorocholine chloride
(g/L}
Phytagel (g/L)
3.5
pH
5.6
18 - 20
Temperatu re range
0

Stock l
25

Stock 2
30

80

Stock 2
30

Stock l
30

20
0.5
20

40

0.5
3.5
5.6
6- 8

5.6
18 - 20

3.5
5.8
23 - 25

3.5
5.8
23 - 25

3.5
5.8
18 - 21

a. Liquid media is also used in sweetpotato culture.

b. Murashige-Skoog salts.
Stock solution l: 2 mg/L glycine, 0.5 mg/L nicotinicacid, 0.5 mg/L pyridoxine, 0.4 mg/L thyamine HCI.
Stock solution 2: 0.1 g/L argi nine, 0.02 g/L putrescine, 0.2 g/L ascorbicacid, 0.002 g/Lcalcium d-panthotenic.

Th e protocol for medi um -term in vitro

co nse rvation of the potato co ll ect ion at CIP
is as fo l lows. Accessions are co nse rved in
a co nservatio n medium containing 4%
so rbito l at a temperature of 6-8C, and li ght
intensi ty of 1,000 lu x. That extends th e in
vitro co nse rvat ion of th e potato co ll ect io n
fo r 2-4 yr without subcu Itu re (Tab le 2). The
use of so rbitol as an osmoticum is app lied
to many crops w ithout any physiological
changes such as cal I us formatio n or
vitri ficat io n. But th ese und es irab le reactions are produ ced whe n the media co ntains mannitol, w hi ch can affect potato
geneti c stability (H arding, 1994 ). After
seve ral yea rs in in vitro culture, plantlets
can recove r normal growth afte r one to two
subcultures in propagati on media. This
co nse rvation method is one of the most
effic ient for managin g a large number of
potato access ions and the time in terva l

3 52

Poto to ond Sweetpototo

between subcu ltures is lon ger th an for oth er

c rops.
The production of microtubers in in vitro
culture as an altern ati ve method fo r longterm co nse rvation of potato cu lti vars has
also been eva lu ated at CIP. Microtubers are
produ ced in 2 to 3 mo and can be sto red at
10C for up to 1 0 mo after harvest (Estrada,
et al., 1986). The dormancy of th ese
microtubers can be co ntroll ed by env ironmental changes (Estrada et al., 1986; Tovar
et al., 1985 ). A ltern ati ve ly, once th e
m icrotubers sprout, grow th ca n be reta rd ed ,
as wi th in vitro plants, for 2-4 yr by storin g
th em embedded in a conservation medium .

In vitro sweetpotato conservation

In v itro techniques for micropropagati on
of sweetpotato cu Iti va rs are wel I estab1ished. Pl ant lets grow from 1 to 2 mo in

Table 2. Number of accessions per Solanum species conserved in vitro and interval between subcultures.

Solonum species
andigena
stenotomum
phureia
tuberosum
chaucha
gonioca!yx
Juzepczukii
aianhuiri
Other
Total

Interval between subcultures

2 yr

3 yr

4 yr

5 yr

Total

569
57
37
38
23
15
4
6
184
933

666
54
37
24
23
6
8
3
125
946

1,555
129
66
68
61
31
14
8
280
2,212

70
5
9
8

2,860
262
149
138
108
53
27
17
607
4,204

culture med ia co ntaining minerals, a C

source, polyamin es, and growth regul ators
such as giberelli c ac id (propagation medium I, Tab le 1) in a culture room at 23250C and 3,000 lux (Li zarraga et al. , 1990).
Some plants show multipl e shoot form ation , pheno li zation, or shoot dormancy as a
respon se to hi gh stress. W e have overcome
this prob lem at CIP by usin g a culture
medium co ntainin g nap hth alene acet ic ac id
(propagat io n m edium II , Tab le 1).
A nu mber of problems prevent the longterm conse rva tion of sweetpotato. M any
attempts to establi sh an efficient slow
growth medium have fa il ed du e to a strong
genotyp ic response to the modifi ed culture
medi a, low surv iva l percentage under
restri ct ive growth co nditi ons, or th e forma tion of ca llu s and vitrifi ca tion durin g
storage.
Although plantlets cu ltured in growth
retard ants m ay have surv iva l ra tes ranging
from 70 to 90%, genotyp ic effects or to xic
effects are see n. For examp le, plants grown
in a medium co ntai ning absc icic ac id at 520 mg/L had a surv iva l rate of 70- 85% after
8 mo, but showed stro ng genotypi c effects
(Oesamero, 199 0; Jarret et al., 199 1). Plants
grown with the retard ant maleic hydraz ide
at 5 mg/L had a surv iva l rate of 70-90%
after 6 mo (Oesa mero, 1990). Plants grown

0
18
113

with cycoce l at 500 mg/L had the sa me

surviva l rate after 1 yr (Guo et al., 1995).
But tox ic effects we re observed w ith both
maleic hydraz ide and cycoce l. U sin g
kinetin e as a growth regul ator, Guo et al.
(1995) achi eved a 70% surv iva l rate in 20
access ions after 1 yr in storage, but with
some ca llu s presented. Plants grow in g in
MS medium diluted to 30% and 50%
showed low survival (36-48%) after 4 mo in
storage (Abreu et al. , 1992; Aguil ar and
Lopez, 1993) .
Sweetpotato pl antl ets ca n rema in in 1040 g/L so rbitol for 6-1 2 mo with out subcu 1ture (Desa mero, 1990; Acedo, 1993;
Cubillas, 199 7). They resume no rm al
growth w hen pl aced in a culture medium
w ith out osmoti cum . Sorbitol, howeve r, ca n
be meta bolized by th e pl antlets after so me
month s of sto rage and ex hibit an in cremental growth rate, effecti ve ly redu cing storage
time. M annitol at 20-40 g/L has also been
used in sweetpotato to extend th e interva l
betwee n subcultures to 1 yr (Aguil ar and
Lopez, 1993; Acedo, 1993; M and al and
Chandel, 1990; Guo et al., 1995;
Desa mero, 1990) but so me callus form ation
and vitrifi cation was reported.
Trop ica l crop spec ies usually lose their
v iability at temperatures lower than those
req uired for growth in the field .
CIPProgram Report 1997-98

353

Sweetpotato, however, can continue to

grow at temperatures several degrees below
normal field temperature. Sweetpotato
plants grown at temperatures below 15C
for long periods have shown detrimental
effects such as phenol ization and necrosis
(Desamero, 1990). CIP research has shown
that temperatures of 16-18C lengthen the
storage period up to 1 yr in cultures
containing 2% sorbitol. However, some
detrimental effects were produced in 25%
of accessions in the in vitro sweetpotato
collection. Temperatures of 18-21C are
being tested. A culture medium containing
2% sorbitol, storage at 23-25 C, and 3,000
lux (Lizarraga et al., 1990) successfully
maintained the sweetpotato collection at
CIP 4-6 mo without subculturing. More
recently, a new culture medium containing
2% sorbitol and 2% mannitol was tested
using 30 accessions. Their survival rate
was 82% after 16 mo of conservation (Table
3) (Cubillas, 1997). An evaluation of this
method using several hundred accessions
in the collection is under way.

Procedures for Optimum In Vitro

Germplasm Conservation
In vitro conservation entai Is the risk of
losing material due to cooling equipment
failure, contamination of cultures, or

mislabeling accessions. Therefore, the

following recommendations will help to
maintain an in vitro collection more
efficiently.
Frequently evaluate in vitro grow th
during the first month to detect plants
with growth problems.
Maintain aseptic conditions in the
culture growth room to avoid sources of
contamination (dust, dirt, mites, or
contaminated material ). Treat the room
as a restricted area to prevent contamination.
Equip the in vitro laboratory with an
electronic alarm connected to a control
panel that monitors en v ironmental
conditions.
Develop a database with unique codes
for each accession in the collection.
Each accession should have at least two
identification numbers for proper
identification. Labels for the cultures
should be generated directly from the
database to prevent human errors while
transcribing.
Closely monitor the cultures with some
frequency during in vitro storage. Isolate
contaminated cultures as soon as they
are detected to prevent spreading the
problem to clean plantlets. Include
specific antibiotics in culture media to

Table 3. Percentage of plant survival evaluated after 22 months of storage with three teatments of a modified
culture medium for medium-term storage of sweetpotato.

Mo

2% mannitol

1.5% mannitol

2% sorbitol

2% sorbitol

6
8
10
12
14
16
18
20

100
98
92
87
85
82

22

66

354

76

71

Potato ond Sweetpototo

100
96
88
82
77

70
63
57
51

2% sorbitol

100
94
84
79
70
61
53
43
37

eliminate endogenous bacteria in the

cultures.
References Cited

Abreu, l.N ., J.E.B.P. Pinto, R. lnne cco, C.

Deschamps, A.Y. Sato, B.M. Rodri guez,
L. Zica, and W.R. Maluf. 1992. Effect of
th e culture media co nce ntration on
gro wth of sweet potato in v itro.
Congresso Brasileiro de Olericultura
Araca ju (Brazil) 32:48.
Acedo, V.Z. 1993. Slow growth culture for
in vitro maintenance of Philippine
sweetpotato va ri eties. Int. Sweetpotato
Newsletter 6(1 ):5.
Aguilar, M. and M. Lopez . 1993. In v itro
conservation of sweetpotato clone N1437 w ith minimal growth rates. In :
Reunion anual de la Soc iedad del
Programa Cooperativo Ce ntro
Americano de Cultivos y Animales
(PCCM CA). Guatemala. p. 276-278.
Cubillas, L.A. 1997 . Uso del estres Como
metodo de conservac i6n in vitro de
germ op las ma d e lp omoea batatas (L.)
Lam. Thesis. San Marcos National
University, Lima, Peru.
Desamero, N.V. 1990. Minimal growth
storage for in v itro germ plasm conservati on of sweet potato (Jpomoea batatas L.,
Lam. ) Ph.D. Thesis. Clemso n Univers ity,
Clemson, SC, USA. 1 85 p.
Dodds, J.H . and L.W. Roberts. 1985.
Experiments in plant tissue culture,
Second Edition. Cambridge Uni ve rsity
Press, Cambridge, UK.
Espinoza, N., R. Li za rr aga, C. Siguefias, F.
Buitron, J. Bryan, and J.H. Dodd s. 1992.
Ti ssue cu lture: Micropropagation,
conservation, and export of potato
germpl as m. CIP Resea rch Guide 1. CIP,
Lim a, Peru.
Estrada, R., P. Tovar, and J.H. Dodds. 19 86 .
Induction of in vitro tubers in a broad
ran ge of potato genotypes. Plant Ce ll ,
Ti s. Organ Cult. 7:3-10.
Florkowski, W.J. and R.L. Jarret. 1990. Th e
cost of altern ative strategies for th e
maintena nce of sweet potato
Germpl asm. Quarterly J. Int. Agric.
29(1):79-87.

George, E.F. and P.O. Sherrington. 1984.

Plant propagation by tissue cu lture.
Exegetics Ltd ., Edington, Wetsbury, Wilts,
Eng land .
George, E.F. 1993. Shipment and distribution. In : Storing and distributing c lona l
material. In: Pl ant propagation by tissue
cu lture. Part 1. Th e tec hnol ogy. Exegetics
Ltd., Edington, Wetsbury, Wilts, England.
p.181.
Golmirzaie, A. and J. Toledo. 1998. Non cryoge nic, lon g-term ge rmpl asm storage.
In : H all, R. (ed.). Methods in molecular
biology, Vo l. 111 : Plant ce ll cu lture
protocols. Hum ana Press, Netherl ands.
p.95-101 .
Golmirzaie, A.M. and A. Panta. 1997.
Advances in potato cryo prese rvat ion by
vitr ifi cat ion. CIP Prog ram Report 199596. CIP, Lima, Peru . p. 71-76.
Guo, X., Y. Wu, and J. Sheng. 1995.
Maintenance and use of sweetpotato
germp las m in China. In: Root and tuber
crops. Ministry of Agr icul ture, Forestry
and Fisheries; National In stitute of
Agrob io log ical Reso urces; Research
Counci l Secretariat, Tsukuba, Japan. p.
12 3-133 .
H arding, K. 1994. Th e methylation status of
DNA derived from potato p lants recovered from slow growth. Pl ant Ce ll Ti s.
Orga n Cult. 37( 1):31-38.
Hu ghes, K. W. 19 81. In vitro eco logy :
Exogenous factors affect ing growth and
morphogenesis in plant culture system .
En viron. Exp. Bot. 21:28 1-288.
Jarret, R.L. and N . Gawel. 19 91. ABAinduced qui escence of sweetpotato in
vitro. Hortsci. 26(6):737.
Lizarraga, R., Z. Hu ama n, and J.H . Dodds.
1989 . In vitro conserva tion of potato
Germpla sm at the Internati onal Potato
Cente r. Am. Potato J. 66:253 -276 .
Liz arraga, R., A. Panta, N. Espinoza, and J.
Dodds. 1990. Tissue culture of lp omoea
batatas: Micropropagation and co nservati on. CIP Research Guide 32. Intern at ional Potato Center, Lima, Peru . 21 p.
Li za rraga, R., A. Panta, U. Jayasin ghe, and
J.H . Dodds. 1991 . Ti ss ue culture for
elimination of pathogens. CIP Research

CIPProgram Report 1997-98

355

Guide 3. International Potato Center,

Lima, Peru. 21 p.
Mandal , B.B. and K.P.S. Chandel. 1990.
Utilization of tissue culture technique in
preservation of sweet potato germplasm.
J. Root Crops (Special issue ):291-295 .
Roca , W.M. 1975. Tissue culture research
at CIP. Am. Potato J. 52(9):281.
Roca , W.M., J.E . Bryan , and M.R. Roca.
1979. Tissue culture for the international
transfer of potato genetic resources. A m.
PotatoJ. 56 (1):1-10.
Siguefias, C. 1987. Propagaci6n y
conservaci6n in vitro de dos cultivares
de camote (ipomoea batatas Lam).
Thesis. La Molina, National Agrarian
Uni versity. Lima, Peru , 108 p.

356

Poto to ond Sweetpototo

Thompson , M.R., T.J . Douglas, H. ObataSasamoto, and T.A. Thorpe. 1 986.

Mannitol metabolism in cultured plant
cells. Ph y siol. Plant. 67:365-369.
Toledo, J., C. Arbizu, and M. Hermann.
1994 . In v itro internat ional collection of
ulluco, oca, mashua and yacon. Abstracts of VIII International Congress of
A ndean agriculture systems. Valdi v ia,
Chile. p. 1. (In Spanish. )
To var, P. 1 R. Estrada , L. Sch i lde-Rentsch ler,
and J.H. Dodds. 1985. Induction of in
v itro potato tubers. CIP Circular 13 (4):
1-4.
W escott, R.J. 1981. Tissue culture storage of
potato germplasm . 2. Use of grow th
retardants. Potato Res. 24:343-352.

Global Projections for Potato and Sweetpotato to the

Year 2020
G.J. Scott1, M . Ro segrant, 2 C. Ringler,2 and L. Maldonado'

M any of th e deve loping world's poorest

producers and undernourish ed households
are hi ghl y dependent on roots and tubersincluding potato and sweetpotato-as a
major, if not principa l, so urce of food,
nu tr ition , an d cash in come (A lexa ndratos,
1995). Po tato and sweetpotato are currentl y
p lanted o n abo ut 27 million hectares
wo rld wide, producing 432 m illi on metr ic
tons (t), 56% of thi s is harves ted in deve lopin g cou ntri es (FAOS TAT, Jun e 1998). By the
mid-1990s, an estim ate d 158 mil li on tof
potato and sweetp otato we re consumed
ann uall y in deve lop in g co untries, represe ntin g part of the diet of severa l billion
co nsum ers in th ese reg ion s (FAOSTAT,
Janu ary 1999). An improved und erstandin g
of the importance of production, utili zatio n,
and future ro le of potato and sweetpotato in
th e globa I food system fo r the 21st cen tury
has potentia 1ly fa r-reaching imp licat io ns fo r
investments in agr icultural resea rch at both
the intern atio nal and , perhaps eve n more
imp orta ntly, nat ion al leve l.
This paper represents some of the key
findin gs of a rece nt co ll aborative stud y
ca rri ed out by C IP and IF PR I w ith in put
from sc ientists at CIAT and ll TA (Scott et al.,
1999) . It presents future projections for
potato and sweetpotato to th e yea r 2020 for
th e purpose of contr ibu tin g to a clea rer
v isio n of the ro le th ese c rops ca n pl ay in
the globa l foo d system in the deca des
ahead. Prev io us proj ec tion s fo r th ese
com moditi es, parti cul arly potato, have
proved to be hi ghly mi slead in g w hen
compared to FAO stat isti cs on actual output
trends . Those fo r sweetpotato and potato

1. CIP, Lima, Peru

2. IFPRI, Wash ington, D.C.

are also conside red suspect because th e

FAO production data frequently underieport production and/or yields for particular
developing co untries. Hence, the re is a
se nse among commod ity spe ciali sts and
so me globa l food analysts th at a new set of
proj ect ion s using a different methodo logy is
req uired in order to provide proj ec tions fo r
potato and sweetpotato that merit co nsideration in th e hi ghest internation al ag ri cul tural policym ak in g c ircles.
In the fi rst part of th is paper, the novel
elements of th is app roac h, whic h emphasizes presenting d isaggregate resu lts and
in co rp ora ting potato and sweetpotato in a
mul ti-co mm od ity mod el, are exp lain ed .
Subseq uen t sect ion s th en spell out findings
w ith rega rd to utilizati o n, produ ct io n, trade,
and the est im ated va lu e of output acco1d in g
to alternati ve sce narios. A concludin g
sect io n prese nts two sets of po licy reco mmendations . O ne focuses on reg ional
priorities based o n the projections th em se lves. Th e o th er concerns the process of
co ll ect in g and analyz in g potato and
sweetpotato data fo r the purpose of
generatin g such estim ates.

Material and Method s

Global food proj ecti ons in this paper were
calcu lated using IFP Rl's intern at iona l model
for pol icy analys is of agric ultural com moditi es and trade (IMPACT). Th e mod el,
covering 37 co untri es an d regions and 18
commod ities, (includin g all ce1ea ls,
soybea ns, roo ts and tubers, meals, and
dairy produ cts, accounting for th e vas t
majority of th e wm ld 's foo d production and
co nsumpt io n), is specified as a set of
country-level dema nd an d suppl y equations

CIP P109mm Repo1t 1997-98

357

linked to th e rest of th e wor ld through

trade. Demand is a function of pr ices,
in come, and population growth w ith to tal
demand equa l to the sum of food, feed, and
other (non-food, non-feed) demand.
Growth in crop producti o n in eac h co untry
is determined by crop prices and the future
rate of productivity growth, which in turn is
est im ated by its component sources
including advances in management and
plant breeding research. Other so urces of
growth considered in th e mod el include
private sector in vestm ents in agric ultura l
researc h and deve lop ment, ag ricul tural
exte nsion and education, markets, infrastru cture, and irrigation (see Rosegrant et
al. , 1995, for details of the methodology).
Th e resu lts presented here are ge nerated
from an updated version of IMPAC T that
takes as its base pe riod 1992-94. Th e
projections are for the year 2020. Projected
growth rates then are an nu al average rates
estim ated for the period 1993 -2020. All
other modeling methodology is co nsistent
with the IMPACT mod el reported in
Rosegrant et al. (199 5).
In orde r to put the following set of resu lts
into perspective, it should be pointed out
that these calculations attempt to go
beyond past projections mad e for these
c rops in a number of important respects
in c lud ing:
Previous attempts often focused only on
a sin gle root and tuber crop, e.g., potato
(FAO, 1995), thereby making less expl icit
the possible linkage s in production,
consumption, and trade with ot her food
commodities.
Other attempts to mod el future trends for
potato and sweetpotato were calcul ated
on an aggregate basis, for roots and
tubers in tota l, overlooking the comp lex ities that can be more effect ively captured
in a disaggregate approach.
Past efforts at estimating the future
relative va lue of major food commo d iti es
including potato and sweetpotato have
assu med fixed relati ve prices throughout
the planning period. IMPACT spec ifica ll y allows prices to va ry.

358

Pololo ond Sweetpoloto

Finally, previous attempts at mul ticommodity projections we re often

ca rri ed out w ith ou t the particip atio n of
commodity specia lists for roots and
tub ers (Roseg rant et al., 1995). Given
the re lative shortage of published
information on potato and sweetpotato
(th ere is no extens ive bod y of literature
o n projections for these crops, particularly for developing cou ntries), thi s
represents an important aspect of any
model in g exe rcise for these commodities. Th e results presented here represe nt
the output of co llaboration over the last
three and a half years between CIP and
IFPRI econom ists .
Results and Discussion

IMPACT posits two sce narios for futu re

growth in demand, product ion , trade, and
prices fo r potato and sweet potato in the
context of generatin g projection s to the year
2020 for the wo rld ' s major food commodities: (1) the baseline scenar io; and (2 ) the
high demand and production growth (HOP)
scenario. In the baseline scenario, projecti o ns for potato and sweetpotato are driven
by more conserva ti ve estimates of the effect
of income growth o n demand for th ese
commodities and the ir deri vatives. As a
corollary, the rate of technological change
is more modest tha n in the HOP scenario
contributing to less rapid increases in
production and yields.
Baseline scenario
IMPACT' s baseline scenario for projected
growth rates in food and animal fe ed
dem and from 1993 to 2020 are hi gh and
uneven across the globe . Th ey are highest
for potato wi th an ave rage annual in crease
in total demand of 2.02%, and 2.33 % for
average an nu al foo d dema nd (Tabl e 1). For
sweetpotato, total foo d demand is projected
to grow at 0.44% ove r the period reflecting
reduced demand for fr es h sweetpotato for
direct hum an co nsumption in China .

Regio nal growth rates for food dema nd

for potato are hi ghest for Sub-Sah aran
Africa (3.10%) and India (3.09%), and

Table 1. Trends and projections for potato and sweetpotato in developing countries, 1987-97 vs. 1993-2020
according to different scenarios.

Potato
Category

1987-97

l 993-2020Ab

Sweetpotato
1993-20208'

1987-97

0.84
l.85
2.71

1.3

Growth rates in production (average annual %)

- Area
2.4
0.51
- Yield
2.0
l.50
- Production
4.5
2.02
Growth rates in utilization (average annual%)
- Total food demand
2.33
- Total feed demand
0.37
- Total use
2.02'
Net trade (mil. mt)
-1.2
3.91
Value(%)

2.75
2.66
2.76'
-4.6
4.9f

0.9
2.2

l 993-2020Ab

1993-20208'

0.27d
0.97d
l .25d

0.35d
1.1 Od
l.4Sd

0.44d
l.81 d
l .25d
0.6d
l.51

0.SOrl
2.23d
l .46de
0.4d
1.91

Source: Scott et al. (1999) and FAOSTAT (June 1998, accessedJuly 1998).
Note: Developing countries as defined by FAOSTAT.
Calculated for the period 1985-87 to 1995-97, without central Asia in the case of potatoes.
b 1993/baseline scenario.
' 1993/ high demand and production growth scenario.
dThe growth rate projections are for sweetpototo and yam combined with a ratio 80/20 for sweetpotato to yam. China produces no
yam.
' Total food, feed, and industrial (non-food, non-feed) demand.
1See Table 2.

lowest fo r other East Asian co untri es (e.g.,

North Ko rea), and Latin Am erica (1.69%).
These stron g res ults reflect:
th e high status of potato as a preferred
food in these regions;
population growth, parti cularly stro ng in
Africa; and,
th e extent to w hi ch the crop fits into
loca l food systems in terms of, e.g., labor
requ irements, soi l, and cl imate conditions .
For sweetpotato, reg ion al growth rates
for food demand are high est in Sub-Saharan
Afri ca (2.74%) and lowest in Chin a
(-1 .02%) w ith most othe r reg ions showing
littl e or no growth. Feed demand presents
just the opposite picture w ith strong growth
in Chin a (1.81 %) . This di chotomy partly
reflects the relative availability of basic
foodstuffs in the two reg ions; Ch in a is
relatively well endowed and Sub-Saharan
Africa much less so. Differential population

and in come growth rates , slower demographic expansion and more rapid economic growth in creas in g demand for meat
in Ch in a, w ith an oppos in g scenario in SubSaharan Afr ica, also dri ve the di chotomy.
Projected production growth rates are
va riable and driven by yie ld increases.
Average ann ual growth rates in produ ction
from 1993-2 020 for potato (2 .02%) and
sweetpotato (1.25%) are domin ated by
growth rates in y ield (1. 50% and 0.97 %,
respectively) (Tab le 1). However, in
rel ation to rece nt actual trend s in produ ction in th e period 1987-97, th ese est im ated
production growth rates are less than half
th at for potato, and 25 % high er fo r
sweetpotato (Tab le 1). Whereas in abso lute
terms th e projected in creases from 1993 to
2020 in area planted in potato (1.0 million
ha, or 15%) and in syveetpotato (0.9 million
ha, or 8%) are by no mea ns insignific ant,
the projected average ann ual dec lin es in
(IPProgram Report 1997-98

359

the growth of area planted from the periods

1987-97 to 1993-2020 is more noteworth y
than th e slowing of production growth over
the period.
Trad e is not projected to increase to any
extraordinary exte nt, although the pubished FAO database for trade used for
generating these projections is particularl y
suspect because it often lacks statistics on
trade in processed products for particular
countries. Current estimates suggest that
processed potato exports, for example,
wou Id at least double the overal I trade
volume (FAO, 1995). For sweetpotato,
recent research in China suggests that
exports of starch-based products may be far
more important and are growin g more
rapidl y than prev iously rea lized. Notwith1

standing, some marked increases in trade

are projected to take place . Southeast Asia
and Sub-Saharan Africa will each import
300,000 t more of seed and tabl e potatoes
in 2020 than the estimated 100,000 t
imported in 1993. China's ex ports of
sweetpotato are projected to rise by
900,000 t to 1.4 million t.
The baseline scenario proj ects th e value
of potato and sweetpotato as a share of the
total value of the major food commodities
in the stud y to fall from 6.9 % in 199 3 to
5.4% in 2020 (Table 2). Much of this
decline is attributable to th e fall in price for
sweetpotato. Hi gher than previousl y
estimated prices for ce reals in 2020 also are
dri ving thi s result (Rosegrant at al. , 1995).

Table 2. Total value of selected commodities for developing countries, 1993 vs. 2020 according to different
scenarios.

2020Ah

1993

20208'

Price Production Value Total Price Production Value Total Price Production Value
(USS/t) {000 t) (USS)

(%) (USS/t) (000 t)

(millions)

(USS)

(%) (USS/t) (000 t)

(millions)

(USS)

Total

(%)

(millions)

160

94,336

15,094

4.1

137

161,994

22,193

3.9

145

194,006

28,131

4.9

Sweetpotato 80

124,703

9,976

2.8

56

154,722

8,606

1.5

68

163,369

11,186

1.9

All roots and

422,573

38,586

105

654,504

50,076

8.8

714,584

60,946 10.5

928, 115 17 6,622

48.0

Poto to

tubersd
All cereals'
Soybean

57,705

15,176

4.1

All meat1

88,326 137,752

37.4

Total
1993 share of R& Tin all
commodities
Share of R&Tin cereals +
R&T+ soybean

368,136

1,407,478 241,253 42.6

106,149

24,839

1,409,470 242, 195 41.9

4.4

106,203

182,749 249,862 44.l

182,831

566,030

360

PototoondSweelpototo

4.3

250,467 43.3
578,567

10.5

8.8

105

16.7

15.8

18.6

Source: Scott et al. (1999). Note: Roots and tubers= R&T.

Average for the three years; 1993 equivalent ta 1992-94.
b Baseline scenario.
' Highdemand and production growth scenario.
' Includes potato, sweetpatato, cassava, yam, and other raots and tubers.
' Includes wheat, maize, barley, sorghum, millet, and rice. 11ncludes beef, pork, poultry, sheep, and goat meat.
0

24, 958

High demand and production growth

scenario
Rece nt trends indi cate that the stru cture
of dema nd and supp ly for potato and
sweetpotato ma y be und ergoing fund amental shi fts in parts of th e deve loping world
(Table 1). The HOP scena rio is used to
exa min e th e imp act of faster suppl y and
demand growth in se lected region s.
Th e HOP scenario projects that growth
of both food and feed us es will account for
the faster growth in over a 11 demand for
potato and sweetpotato. In the HOP
proj ection , the jump in demand for pota to
is almo st entirely clu e to faster gro w th in
food use in Ch in a and Ind ia. In China,
average an nu al growth in food use for
potato from 1993 to 2020 is 2.74 % ve rsu s
2.20% for feed use. In India, ave ra ge
growth in food use during the period is
3.80% per yea r ve rsus 3.09 % for feed use .
Overa ll average annua l per capita demand
for potatoes in d eve loping countries w ill
ri se on ly mod est ly beyo nd the level projected in the baselin e sce nario- from 16.18
to 18.05kg. Th ese consumpt ion leve ls
remain a minor percentage of th e leve l
currentl y observed in indu striali zed co untrie s, 74 kg/ca pita/ y r.
For sweetpotato , th e maj or gro w th is in
total fe ed demand (2.33% average annu ally
from 19 93 to 2020). Thi s result large ly
refl ects th e strength of feed dem and in
Ch in a, w here sweetpotato w i 11 beco me a
more effic ient an im al fee d so urce to help
sa ti sfy the country's grow ing dem and for
meat. This reflects high er prices for mai ze
and improvements ass um ed by th e HOP
sce nar io.
Proj ected growth ra tes for produ ct io n of
potato (2.7 1 %) and sweetpota to (1.45%)
are high er-and fo r potato cons id erably
so-than those of the ba se line scenar io of
2.02 and 1 .2 5%, res pectively. For potato,
high er growth rat es res ult from faster
growth in China, 2.72% ve rsus 1 .4 9% in
the baselin e scenario, and in Indi a 3.67%
ve rsus 3.10%. In both cou ntri es, produ ction growth is a result of stro nger ex pan sion

in area pl anted. It should be noted here

that eve n these high er projected growth
rat es ac tu all y represen t a maj or dro p from
rece nt histor ica l leve ls, e.g., from 6.2 % per
year in Ch in a durin g 1987-97 (FAOS TAT,
Jun e 19 98) . For sweetpota to, stronger
growth in area planted in Sub-Saharan
Afri ca and moderately hi gher yields in
China (th oug h area pl anted is proj ected to
fa ll by 600,000 ha) acco unt for th e acce lerated growth rate for production of
sw eetpo tato (ro ughl y 2.7%).
In th e case of trad e, th e bi g change is
with potato-a lth ough th e lac k of data on
processed products st ill appli es. From 1993
to 2020, tota l impmts fm develo pin g
countries jump from 1.2 to 4.6 million t as
Chin a, Indi a, and the West As ian and North
Afri ca n reg io ns go from net ex porters to net
imp orters. Strong demand un der th e HOP
scena 1io is li ke ly to generate greate r
imp o rt s of processe d potato produ cts for
hum an co nsumption (e.g., fro ze n french
fr ies, chips) from indu stri alized co untri es .
Th e shares of potato and sweetpotato in
th e pro jected value of th e commodit ies is
proj ected to remain stable, if not ri se (Tabl e
2). Thi s result is th e combi ned effec t of the
faster growt h rat es in produ ction and a
mod erately slower rat e of decl in e in pri ces
for these co mmoditi es with th e bulk of th e
in crease clu e to fast er production . Furthe1more, all of the more rapid proj ected
in creases are we ll w ith in the ran ge of
rece nt trends in produ cti on and utili zatio n
for roots and tub ers in deve lopin g co untri es
as deriv ed from FAOSTAT statistics. Th e
additi o nal ex pansion in area planted in
roo ts and tubers is a small fra ction of th e
total Janel area devote e/ to crop produ ction
in 2020.

Conclusions
Th e impli ca tions of th ese proj ection s for
th e locat ion of potato and sweetpotato
produ cti o n in th e year 202 0 and the val ue
of produ ctio n for these cro ps can be
summ arized bri efl y.

CIP Progrorn Report 1997-98

36 1

Potato and sweetpotato are, as fa r as

developing countries are concerned ,
overwhelmingly Asian commoditi es (F igure
1 ). Based on these projections, particularly
the HOP scenario, the y will becom e
increasingly so in the ye ars ahead, albeit for
different reasons. Rising income s w ill
stimulate added consumption of potatoes as
consumers look to diversify their large ly
cereal-based diets. This increas ed consumption of potatoes will include high
percentage increases, but modest in
absolute qu antities on a per capita basis ,
in the consumption of fast foods and snacks
as Asia's largely rural-based popul ation
becomes more u1banized and incom es
improve.
For sweetpotato, rising incomes and
urbanization will catalyze increased use of
the roots and vi nes for animal feed and
processed food products such as noodles
made from starch.
From a global perspective, th e relatively
minor importance of potato and
sweetpotato in Africa is noteworthy.
Howeve r, potato an d sweetpotato will
maintain , if not ri se, in relati ve importa nce

1993

= 219

in particular sub-regions within Africa,

particularl y in East Africa.
Finall y, the estimates of the value of
production permit us to calculate the
geographic distribution of the future value
of production for potato and sweetpotato.
Thi s comparison highli ghts the importance
of China and India for potato and China
and Sub-Saharan Africa for sweetpotato
(F igure 2). Policymakers and research
scientists should keep th ese consumption
and production trends in mind when
formulating plans for future research and
related development efforts.
Recommendations

Experi ence in working with th e available

data to generate these proj ections suggests
a series of measures are needed to improve
the pro cess.
First, an international effort is needed to
examine the databases for potato and
sweetpotato in a select number of developing countries with the specific intent of
reducin g th e margin of di ffe rence between
FAQ statistics and national data, as we ll as

million t

2020 = 359 million t

Other
5%

Saharan
Africa 3%

Sub-Saharan Africa
3%

China

19%

China

24 %
India

7%
America

China
38%
India

China

12 %

49%
Others
5%

Latin America

63

6%
Other

6%

C==1

C==1

Sweetpotato

Potato

WANA

6%

Figure 1. Potato and sweetpotato production (million t) in developing countries 1993 vs. 2020, according to the
HOP scenario. (Source: Scott et al., 1999.)

362

Pototo ond Sweetpototo

2020

1993 = US$26.4 billion

Sub-Sa haran
Africa 2%

Sub-Saharan
Afr ica 2%
Chi na
26%

China
37%

Chin a
27%

In d ia
9%

Others
3%

30%

Ind ia
15%

La tin
Ame rica 8%
8%

= US$41 .7 billion

WANA
8%

[===:J

[===:J

Sweetpotato

Po tato

America 7%

Figure 2. Value of potato and sweetpotato production (USS billion) in developing countries 1993 vs. 2020,

according to the HOP scenario. (Source : Scott et al ., 1999.)

between agg regate co untry figu res and the
find ings of spec ifi c case studi es and fa rm
surveys.

wo uld make, as we ll as co ncrete suggestio ns on how the imp rove ments mi ght be
impl emented .

Seco nd, a systemati c atte mp t to improve

th e FAO Foo d Balance Sh eet data should be
undert aken. This should in c lud e reco nsiderat ion of how food and non-food ca tegories are reported. Sp ec ifica ll y, in cl usion of
sweetpota to leaves in nutritio n est imates
wo uld see m to make good se nse. Feed use
figures sho uld includ e so me estim ate for
sweetpotato vin e use; if not in th e Food
Balance Sheet fi gures th emse lves, then in
some other internation all y acce pted
sta nd ard of fee d availability.

Impact
Thi s joint effort of CIP and IFPRI to
an alyze hi stori ca l trends and generate
comm odity pro jecti ons has res ulted in two
major impacts . Fi rs t, CIP, in co njun ct io n
wit h CIAT as wel l as llTA, and FAO are
cur re ntly pre parin g th e next set of proj ection s fo r potato, sweetpotato, cassava, and
yam fo r TAC. Thi s more pro- acti ve approac h w i 11 redu ce th e need for pos tpubl ica ti on debates about th e acc uracy o f
parti cul ar fi gures .

Third , c lose r stati sti ca l mo nitorin g and

pu bli shin g o f trade data on potato and
sweetpota to should be a goa l, so as to
regul arly in c lud e inform ati on on at least th e
major processed produ cts, e.g., froze n
french fri es, chips, and dehydrated produ cts
fo r potato as we l I as starch and noodl es for
sweetpota to. Aga in , perh aps an ex perim ental pil ot effort in a fe w key co un tri es co uld
help identi fy the types of low-cost impro vements in reporting procedures, esti ma tes of
diffe rence in results th ese procedures

Seco nd , the interaction betwee n CIP and

IFPRI has strength ened IFPRl 's IM PACT
model fo r th e global fo od econ omy and
helped CIP bette r apprec iate th e re lati onship betwee n potato and sweetpotato and
other major food co mmodities . As a res ult,
th e CG IA R's interaction w ith intern ati onal
and nati onal poli cy makers w ill p rov id e a
more co nsistently acc urate and optimi st ic
ou tloo k for th ese co mm oditi es in th e yea rs
ah ea d.

CIP Progrom Reporl 1997-98

363

References Cited

Alexandratos, N. 1995 . World ag riculture:

Towards 201 0. An FAO Study. FAO and
John Wiley and Sons, NY, USA.
FAO (Food and Agriculture Organization of
the United Nations). 1995 . Potatoes in
the 1 990s. Situation and prospects of the
world potato economy. International
Potato Center and FAO, Rome, Italy.
. 1998. FAOSTAT Statistics Database
(On line). June. Available HTTP : http://
apps.fao.o rg.
. 1999. FAOSTAT Statistics Database
(O nlin e). January. Available HTTP: http://
apps. fao.org.

36 4

Polo lo ond Sweelpolalo

Rosegrant, M.W., M. Agcao i li-Sombr il la,

and N.D. Perez. 1995. Global food
projections to 2020: Imp licatio ns fo r
investment. Fo od, Agriculture, and the
Environment Discussion Paper 5.
Internationa l Food Polic y Research
Institute (IFPRI ), Washington, D.C., USA.
Scott, G.J., M. Rosegrant, and C. Ringler.
1999 . Roots and tubers for the 21 "
century: Trends, projections and policy
for deve loping cou ntri es. Food, Agr icultu re, and the En v ironment Discussion
Paper. In ternat iona l Food Policy Research Institute (IFPRI), Was hin gton,
D.C., USA.

The Role of Potato and Sweetpotato in Disaster Relief:

The Case of Rwandan Refugees in South Kivu,
Democratic Republic of the Congo (ex-Zaire), 1994-96
M. Tanganik 1 , P. Phezo1, P.T. Ewell2 , N.B. Lutaladio 3 , and G. Scott4

After the c iv il war and assoc iated genoc id e

in Rwanda in June and July, 1994, over one
million refugees streamed over the border
into neighboring areas of Zaire, a country
whose new government has since changed
its name to the Democratic Republic of the
Congo. This area around the Great Lakes is
one of the poorest and most densely
populated in all of Sub-Saharan Africa.
Prior to the outbreak of vio lence, Rwanda
and th en Za ire had estim ated per capita
in comes of US$265 and US$327, respectively (Wor ld Bank, 1992). In res ponse to
the hum an tragedy in 1994, refugee camps
were set up under the auspices of seve ral
branches of the United Nations and a
number of non-governmental o rgan izations
(NGOs). The displaced population lived in
these camps for a little over two yea 1s, until
they we 1e dispersed in October 1996. Some
went back to Rwanda, their home country.
Others fled to lo ca tions deeper into the
D.R. of the Congo. Five camps were set up
near the town of Bukavu, at the southe rn tip
of Lake Kivu (Figure 1 ). The agroecological
conditions, farming systems, and dietary
habits of the local population are very
simi lar to those across the border in
Rwanda. The refugees' presence generated
a huge increase in the loca l demand for
food. They also constituted a massive pool
of hired labor for local farmers. Their
presence led to a surge in the production of

1 ln stitul Nation<JI d'Etudes et de Recherch es Agr ico les

(INERA), Mulungu, South Kivu, Democrat ic Repub lic of
th e Congo.
2 CIP, Nairob i, Ken ya .
3 Programme Re gional de I' Amelioration de la Culture de la
Pomme de Terre et de la Palate D o uce en Afrique Centrale
et de l 'Es t (PRAPACE), Kilmpala, Uganda.
4 CIP, Lima, Peru.

International boundary
National capi ta l

Uganda
Democratic
Republic
of the
Congo

@Kigali

RWANDA

c~wda

Democrat ic
Repub lic
or the
Congo

un di

Figure 1. D.R. Congo, Rwanda, and Burundi showing

Bukavu, near the five refugee camps.

both potato and sweetpotato, important
crops histo1ically on both sides of the
border (Scott, 1988; Tardiff-Douglin, 1991 ).

Materials and Methods

The national potato and sweetpotato programs of IN ERA, the national agricultural
research in st itute of D.R. of the Congo, are
based at the Mulungu Research Station just
outside Bukavu. The programs are part
of the PRAPACE network, and have rou-

CIPProgromReportl997-98

365

t in ely received advanced ge rmp lasm and

training from CIP for the past 20 years
(Ewel l, 1992; Rueda et al. , 1996). Improve d
cul ti vars and other tested technologies were
locally available. One of the refugee camps
was located right on the edge of the station;
others we re in the imm ed iate vicin ity. The
station became in vo lved in the multiplication of planting material for local farmers,
who so ld their harvest directly to the
refugees as wel I as to the relie f agencies.
Given these dramatic developments,
researchers at M ulun gu were intereste d in
analyzing the response of local growers to
these eve nts. Among th e key considerations
was a better appreciation for local farmers'
abi lity to respond quickly to upheavals in
supply and demand for food, to begin to
quantify the extent of this response, an d to
exa min e the ro le of new technology in
these adj ustments. To that end, a total of
128 farmers in the imm ed iate su rround in gs
of five refugee camps we re interv iewed
about the impa ct of the cr isis on loca l
agriculture and food supplies during Apr il
and May, 1998 (Tanganik and Phezo,
1998).

Results and Discussion

The local farme rs are predominantly
members of t he M ushi tr ibe. Th e hi ghlands
of South Kivu are mountainous, and
agricultural land is located between 1,400
and 2,500 m. Th e predominant crop is
banana, consumed primarily as local beer.
Th e major staple crops are sweetpotato,
beans, maize, cassava, and potato (Table 1).
Sweetpotatoes are kno wn as cilera abana,
or "protector of the children", and are
universally grow n on a sma ll scale for food
security. Potatoes are wide ly grown as a
cas h crop, mostly by richer members of the
community, in drained swamps alon g with
other vegetab les . Farm ers were ab le to
rapidly intensify production of both crops to
sel I to the refu gees .

Potato
Most farmers increased their cropp in g
intensity from one to two or even three
366

Pololo ond Sweelpololo

crops a year (Table 2). Over half of the

farmers interv iewed decre ased the area
planted to other crops to inc rease potato
production , at the expe nse of beans, maize,
cassava, and sorghum. Two-thirds of th em
adopted new cu li vars. Cruza 148 (720118)
was in troduced into the region from Mexico
in the late l 970's. It is rel ative ly hi gh
yie ldin g, resistant to late blight, and tolerant
to bacterial w il t. Adoption has been limited
by its relati ve ly poor culi n ary quality that
makes it difficult to market (Rueda et al. ,
1996). Th is was obvious ly not a major issue
for the refugee market, and it quickly
became the most w id ely grown. Other
established cu lti vars Montsama (720049)
and the Rwandan se lection Mabondo were
also adopted. Seed was obtained from the
Mul un gu research station, from agricultural
exte nsion agents as well as from NGOs. In
some camps, seed was obtained fr om the
refugees themsel ves, w ho had carri ed it
from their homes in Rwanda .
Th e availab ility of refugees as laborers
encouraged local farme rs to adopt relatively
labor-i ntensive practices. These included
heavy mulching at planting, deep til lage,
the preparat ion and appl ication of organ ic
compost, draining of swampy plots, and
higher hilling . Ove r hal f of the farmers
increased the use of fungicides to control
late bl ight, and 42 pe rc ent rogued out
diseased or other susp icious plants to
improve the quality of their seed (Ta ble 2).

Sweetpotato
The presence of a ready market encouraged a noteworthy increase in the number
of plantings of sweetpotato by near ly all of
the farmers interviewe d, even more so than
potato (Table 2). Eighty six percent of them
went from one to two plantings per yea r,
and reduced the areas dedicated to maize,
beans, peanuts and vegetables. O ve r half of
them plan ted new cu lti va rs avai lable on the
Mulungu station- aga in even sli ghtl y more
so than in the case of potato. Th e new
cultivars had been se lected primarily for
earliness and hi gh y ield. Karebe II and
Mugande are regional farmers' cultivars

Table 1. Sampled households reporting major annual crops in areas surrounding refugee camps in the
Democratic Republic of the Congo before arrival of Rwandan refugees in July 1994.

Site
Mulungu"

Bugobe

Kalehe

Ludaha

Nyangezi

Total

Altitude (m)

1,850

2,000

< l,800

l,965

l,500

Sample size

32

22

24

28

22

128

Households reporting production (%)

Crop
Sweetpotato

l 00

JOO

100

100

100

JOO

Beans

100

82

92

79

91

89

Maize

JOO

82

50

JOO

82

83

Cassava

l 00

64

100

21

100

77

Potato

65

100

JOO

25

100

Vegetables

19

82

50

71

19

48

Soybean

50

37

67

21

37

42

Sorghum

31

72

37

21

45

38

Peanut

19

83

36

27

Mulungu is the research station of the lnstitut Notional d'Etudes et de Recherches Agricoles.

initially ide ntified as superior by th e

Rwand a n research program . Yanshu 1 is a
very hi g h yielding and ea rly Chinese
c ultiva r, which is norm a lly sco red as
marg inal by ta ste pan e ls in Africa. Mulungu
I is a loca l Congolese farmers' cultivar
recently id e ntified as superior by resea rchers. Benikomachi is a Japa nese cultivar with
relative ly low yield but exce llent taste. The
low emphasis by farmers o n the bette r
tastin g Benikomachi and th e acceptance of
the Yanshu 1 are indi cative of the willingness of th e market to accept cu ltivars with
relatively low culinary quality (Yanshu 1)
during the refugee crisis.
Plantin g material w as obtained from the
research stat ion, extension agencies, and
NGOs and again in some cases from the
refu gees themselves . Again, agricultural
farmers used refugee labor to inten sify
production through heavy mulching at
planting, the preparation a nd use of organic
compost, better ridging, d eep tillage, a nd
drainin g swampy areas. Overall the movement towa rd improved ag ronomic practices

was more pronoun ced than potato (Table

2). In sect icides were the only purchased

input mentioned, but o nly by eight percent
of th e fa rm e rs.

Conclusions
The in c reased incom e from the sale of
potato and sweetpotato was used to
increase the wealth and status of th e
farm e rs. Dowries, livestoc k, home improvements, and consumer ite ms such as bicycl es and radios were the most com monly
mention ed. School fees and improved diet
and hea lth for their families were rated as
somewhat less importa nt, which m ay imply
that the most basi c n eeds of these farm
house holds were already met.
Th e departure of th e refuge es m ea nt that
the ex traordinary surge market demand
return ed to pre-crisis levels, and labor was
no longe r so easily available. About half of
the farmers said that they lost money in the
season when the Rwa ndans left , but this
may also have been partially due to the

CIP Program Report 1997-98

36 7

Table 2. Impact of the presence of Rwandan refugees on potato and sweetpotato production in areas surrounding
refugee camps in the Democratic Republic of the Congo, 1994-1996.
0

Mulungub
Altitude (m)
Sample size

1,850

32

Farmers responding ( %)
l. Increased number of plantings per year?
No
0
6
94
98
From l to 2
6
13
From l to 3
2. Reduced other crops to grow more potato?
No
25
19
69
Yes'
50
Less beans
38
Less maize
25
25
Less cassava 1
6
Less sorghum 1
6
Less peonur
0
Less vegetables9
6

3. Adopted new cultivors?

No
Yes
Cruza {P) I Ko rebe 11 {S)
Montsama {P) I Muga nde (S)
Mabondo {P) I Yanshu I (S)
Murhula (P) I Mulungu I (S)
Enfula (P) I Benikomachi {S)
4. Where seed obtained?
Research station
Extension/NGO
Refugees

69

Bugobe
2,000
22

9
91
0

0
100
0

17
65
18

0
86

0
86
14

6
84
10

86

14

18
82
0

0
82
28
0

18

0
17
0
17
0
0

21

18

21
21

28
19

6
51
20
14

28
19

27
19

64
36

82
9
9

9
9
0
0

100

94

35
0

0
0

72
45

36

100
100
100
45

l 00
100
l 00

55

25
13
13
25

90

100

25

81
82

13

75

63
63
6

82

91

45

36
27

55

a. Doto collected April and Moy 1998.

b. Research station of the lnstitut National d'Etudes et de Recherches
Agricoles, South Kivu, Democratic Republic of the Congo.
c. Additional area studied for sweetpototo production only.

Potato and Sweetpototo

83
8

36
17
67

14
14
0
7

75
25

l 00

64
36

l 00

0
0

l 00

85

21
14
14

0
42
100

17

0
17
92
100
l 00
50
92
0

17

29
43

8
8
14
7

17
8

19

27

50
50
31
38
35
19
6

75
19

Nyangezi'
Total
1,500
22
106 128

Ludaha
1,965
28

24

13
13
13
0
13
0

5. More intensive practices adopted?

Mulching at planting
94
Deep tillage
88
Organic compost
75
Drain swamps
19
More hilling/plant on ridges 56
Fungicides/insecticides
44
Negative selection 1
56

368

Kale he
< 1,800

27

15

12

14

73

68

57
36

27

28

18

74

22

29
30

27
26

18
0

47
l0
9
6

20
20
7

56
34

100
64
0

7
43
40

72

68

45

27

47
5

21

l 00
100
l 00

l 00
l 00
l 00
79
l 00
0

l 00
l 00
l 00
55
100
9

80
75
72
65

97
83

64

91

54
42

86
86
79
14

13

95
60

d. P = potato, S = sweetpotato.
e. 'Yes' response not documented for sweetpototo production.
f. Applicable for potato production only.
g. Applicable for sweetpotato production only.

effects of the c ivil war in D.R. of the Congo

that broke out aga inst then Presid ent
Mobutu at the same time. Nearly 100
percent of the respondents said that they
immedia te ly reduced the area planted to
both crops. Nevertheless, a majority
reported that they have continued to benefit
from the improvements made in their farms
during the uph eava l, and that they co ntinue
to use improved practices to m aintain
higher yields than before. Th e major
probl em s are shrunken market and lower
prices. Lo ca l labor is available, but requires
payment in cash, w hereas Rwandan
refugees would accept payment in the form
of food.
This case study illustrates that new
cultivars and labo r-based technologies to
improve the productivity of the farming
systems are ava ilable in the reg ion, and that
they can be adopted rapidl y by large
numbers of farme rs. The wider and more
rapid adoption of yield-enhancing technologies is constrained by th e size of the
local market and limited effective demand
and to some exte nt by the supply of
affordable labo r. Furthermore, in many
areas of Sub-Saharan Africa, wars and
natural disasters unfortunately seem likel y
to continue in the foreseeable future. Both
potato and sweetpotato will ha ve important
roles to play as short-season, nutritious,
locally avai labl e foods. CIP and PRAPACE
will continue to work with partn ers and
donors to make them available as quickly
and efficiently as possible emergency
situations, while simultaneously pursuing a
longer-term strategy to expand and diversify

mme permanent commercial oppo rtunities

for th ese commodities.

References Cited
Ewe ll, P.T. 1992. Wo rkin g with NARS in the
PRAPACE network to develop an information system for monitoring and
evaluation. Discussion paper for
Meeting of CGIAR Social Scientists held
17-20 August 1992 in The H ag ue,
Netherlands.
Rueda, J.L., P.T. Ewell, T.S. Walker, M. Soto,
M. Bicamumpaka, and D. Berrios. 1996.
Economic impact of high-yielding, lateblight resistant cultivars in the eastern
and central African highland s. In:
Walker, T.S. and C.C. Crissman. Case
studies of th e economic impact of CIPrelat ed technology. Internation al Potato
Center (C IP), Lima, Peru.
Scott, G. 1988. Potatoes in Central Africa: A
study of Burundi, Rwanda, and Zai1e.
CIP, Lima, Peru.
Tanganik, M. and P. Phezo. 1998. Impact
socio-economfque de la presence des
refu gies Rwandais a l'est de la
Republique Democratique du Congo sur
la production des pomme de terre et
palate douce. INERA-Mulun gu, Bukavu,
Ki vu Sud, D.R. du Congo. (M imeo. )
Tardiff-D o uglin , D. 1991. Th e marketing of
sweetpotatoes in Rwanda:
Commercialising a perishabl e crop under
adverse co nditions. Unpubli shed PhD
dissertation. Department of Agric.
Economics. Cornell, Ithaca, NY, USA.
World Bank, 1992. African Development
Indi cators . 1992. Washington, D.C.

CIP Program Report 1991 98

369

Research on Natural
Resource Management
in the Andes
Impact on a Changing World

Estimating Frost Risk in Potato Production on the

Altiplano Using Interpolated Climate Data
R.

J.

Hijmans 1

The Altiplano is a high plateau in the And es

of Peru and Bolivia. Thi s paper deals only
with the part of the A lt iplan o ca ll ed the
TOPS System for Lake Titicaca,
.Qesc.guad ero River, Lake .Eoop6, and .S.alt
Lake of Coipasa (OEA, 1996), an area of
149,000 km 2 (Figure 1 ). That excludes th e
southernmost part of the Altiplano, which is
rath er arid , sparsely populated, and less
important for ag ri cu lture. Seventy-five
percent of the TOPS system, hereinafter
ca ll ed the Altiplano, is between 3,600 and
4,300 m; the other 25% is higher. Lake
Titicaca, coveri ng 8,400 km 2 is a conspi cuous part of th e topography that greatly
influences loca l precipitation and temperature.
In 199 3, the Altiplano had about 2.2
million inhabitants. About 5 1 % of the
population li ves in rural settlements, down
from about 59% in 1980 (OEA, 1996).
Rural population density is hi ghest along
the shores of Lake Titicaca and
Desaguad ero River, the are as most suited
for ag ricultur e. Due to emigration to urban
areas, population growth rates on the
Alt iplano are below national leve ls (1.4%
vs 2.1 %) (OEA, 19 96). The Altipl ano is one
of the poorest areas of the Americas, and
poorer than most other parts of Bolivia and
Peru. About 76% of th e population li ves in
poverty, with 33% of those considered to
live in extreme poverty (OEA, 1996).

Altiplano cannot be used for crop production and have a land cover of natural grass
and shrub th at is used for graz ing. Where
crops ca n be grown, co ld and dry condition s allow for on ly a small numb er of
crops. Potato is by far th e most important
crop, accountin g for 63% of th e gross va lu e
of crop production (OEA, 1996). The potato
area is about 63,000 ha (G-DRU, 1996;
INEI , 1996). Potato y ields are low at 5.2
t/ ha in the Peru vian part and th e northern
Bolivian part of the Altiplano, and 3.6 t/ha
in the southern part (OEA, 19 96; G-DRU ,

1996).
Th e growing season in the Alt ipl ano is
betwee n O ctobe r and March, when the
warmest tim e of the year co in cides w ith th e
rain y season. In the agr icultural zones,
ma ximum temperature is around 18( and
minimum temperature aro und 4( durin g
th e grow in g seaso n (INTECSA, 1993; Frere
et al., 1975). There is an ave rage frost-free
period of about 140 d for the north ern
Altiplano and 110 d for the south ern
Altiplano (L e Tacon, 1989). Precipitation is
hi ghest, around 800 mm/year, in th e
north east and near Lake Titicaca and
lowest, about 200 mm, in the southwest.
Production risk for potato is hi gh, es peci ally
beca use of drought, hail , and fro st. Drou ght
is a rec urrent prob lem th at can be especially dama ging in El Nino years, such as
happened in 1942-43 and 1982-8 3 (OEA,

1996).
The severe poverty is partly due to the
harsh climate that limits option s for agr icultural production . About 65% of the economically active population is engaged in
ag riculture (OEA, 1996). Large parts of th e

Frost is th e cond iti on that ex ists when

th e air temperature nea r the earth's surfa ce
drop s below 0( (Kalma et al. , 1992).
However, the temperature at which frost
dam age to crops occurs ca n be lower,
depending on the crop species and cu lti var.

1 CIP, Lima, Peru.

CIPProgrnm Report 1997-98 373

69.,.

7P

le

Bolivia

16 '

Peru

".

. ... x .
. -A

. . .' ...

.,

... .: : .

OiUro.

18 '

nv ers
19 ""

100 ha of pot ato

x vve ather stati on

maj o r town

g.uu~h

Chile

oi
C<>i~~

kil ometers
50

71

10

69 ""

or

Figure 1. The Altiplano (TOPS system) , distribution of potato production and sites of the weather stations used
for the construction of the extreme temperature maps.
For 5. tuberosum subsp. andigena, frost
dam age is likel y to occur w hen the temperature drops to -2C or lowe r (Ca rrasco et
al., 19 97). Th e refo re, we wi ll focus on the
occ urrence of -2C eve nts. 5. tuberosum
subsp. tuberosum, the potato culti vated
3 74 Natural Resource Management in the Andes

outside the Andes is sli ght ly more susceptib le a nd frost dam age may occ ur at -1 C.
Othe r culti vated potato spec ies suc h as 5.
ajanhuiri and 5. curtilobum suffer da mage
betwee n -3 a nd -5( (Huanco, 1992 ; Tapia
and Saravia, 1997), w he reas 5. juzepczukii

generall y tol era tes -5C and eve n lowe r

te mp eratures (Huan co, 1992 ; Ca nahu a an d
Aguil ar, 1992 ; Tapi a and Saravi a, 1997).
With th e exception of 5. ajanhuiri, th e
tub ers of these frost-tolerant spec ies are
generall y bitter du e to a hi gh leve l of
glycoalkaloids and the need to be processed before con sumpti o n.
Fro st ca n have a direct and indi rect
nega ti ve imp act on potato produ cti o n. Th e
direct effec t is the partial o r co mpl ete loss
of lea f area th at lead s to a redu cti o n in
ph oto sy nthesis and henc e yield s. In additi o n, c rop failure may lead to a dec rease in
th e nex t se ason ' s area pl anted du e to tuberseed shortag e (Morion, 198 9). Hi gh produ cti o n ri sk also lea ds to low inves tm ents in
agr iculture res ultin g in lowe r produ cti o n in
yea rs w ith relati ve ly good weather.
Farm ers can avoid or redu ce fr ost
damage by se lectin g appropri ate p lant
materi al, by careful site se lection , by usin g
appropri ate cultural and management
practi ces , and by modifyin g th e physica l
environm ent of the crop (Kai ma et al. ,
19 92). Altiplano farmers try to dimini sh
fros t ri sk by pl antin g potatoes on "wa rm "
so il s (w ith hi gh thermal co ndu cti v ity), and
o n slo pes w here frost in c id ence is lowe r
th an on the va ll ey fl oor (D e Bo uet du
Portal, 1993). Frost-re lated ma nage ment
practices include th e use of smo ke, ru stic
gree nh ouses (Aguirre et al ., 19 99), and
rai sed bed s (Smith et al., 1968; Eri ckson,
n.d.; Sanc hez de Loz ada et al., 1998) . In
th e mo st frost- affected areas, farm ers grow
fr ost-tol erant potato spec ies su ch as 5.
ju zepczukii in stead of 5. tuberos um subsp.

andige na.
Fro st ri sk can be as sessed w ith data from
c l imate station s, as wa s do ne fo r th e
Al tipl ano in Peru by Morion (1 989) and in
Bo li via by Le Tacon (198 9) . H oweve r, gi ve n
th e hi gh spati al vari ability of temp erature
and frost risk , c lim ate stati o ns are too fa r
apart to draw ge neral con c lusi ons for the
potato growing areas in th e Altipl ano. For
th e prese nt stud y, m aps o f the probabi I ity of
ex treme minimum temp erature eve nts w ere

ge nerated. Th ese maps we re sub sequ entl y

analyzed to assess th e exte nt of fro st r isk in
potato produ cti o n o n th e Altipl ano.
Th e prese nt study wa s carri ed out in
collaboration w ith th e Fundaci 6 n PROINPA
(Prom oc i6n e lnvesti gaci6n de Produ ctos
Andin os), Bo li via, to support th eir proj ect
on th e integ rated mana ge ment of abi ot ic
stress in potato produ ct io n. PROINPA's
acti v iti es in c lud e breedin g for fros t-tol erant
potato culti vars (Carr asc o et al. , 199 7) .
M aps of fr ost ri sk wo uld help target areas
wh ere new fros t-tol erant culti va rs mi ght be
adopted and wo u Id perm it an as sessment o f
the potenti al impact of th es e new culti vars
or of oth er tec hnol og ies.

Materials and Methods

W hen frost damage is repo rted in the
literatu re, it is not always c lear w heth er
reference is made to temperatu re at sc ree n
height (1.5 to 2 m) or at ca nop y hei ght. In
thi s stud y, temperature refe rs to screen
height. At the ca nopy lev el, temperatur e
may be abo ut 1C low er (De Bo uet du
Portal, 1993) .
Th ere is a st ron g relati o n betwee n
temperature and altitud e. Hen ce altitud e is
often used as a proxy fo r te mpera ture and
al so fo r fros t ri sk. H oweve r, th e relati o n
betwee n temp era ture and altitud e ch anges
over sp ace and tim e and ca nn ot be extra polated simpl y. Hutchin son (1995 , 1997)
developed a method that ca n be used to
interpol ate c lim ate data. It fits Lapla c ian
smoothin g splin e fun ctions of two or more
ind epend ent va ri abl es (lon gitud e, latitude,
and usu ally altitud e) to th e clim ate v ari abl e
to be mapped. Th e method takes ad va ntage
of th e st ro ng depend enc e of clim ate
(es peci all y temp erature) o n alti tud e, but
all ows th e size of thi s depend ence to va ry
ove r tim e and space (Corbett, 1998).
Monthly ext reme minimum (i. e., low est,
ab solute) temperature data from 44 wea th er
station s (Figure 1) fro m th e database
desc rib ed by IN TECSA (1993 ) we re used.
Dependin g o n th e wea th er station , data for
CIP Prngrnm Report 1997-98

375

15-33 yr (avg. 25.2 yr) were obta in ed. For

each station , th e probabi I ity of a 0, -1 , -2 ,
-3 , -4 , and -5 ( ab so lute minimu m temperature event was calculated fo r all
months of th e yea r. It should be noted th at
the probabi lity of, for example , a 0( event
does not ex c lude eve n lower absolute
minimum temperatures. These probabil ities
were interp olated on a 30-sec (a pproximate ly 1 km 2) grid usi ng the method of
Hutchinson (19 97 ), w ith altitud e as the
third independent var iab le. Altit ud e data
were taken from th e GTOP0 30 database, a
globa l digital elevation model provid ed by
the US Geo logi cal Survey.
In the Altip lano, most potato es are
pl anted in October or November and
harvested in Ap ril or May. Frost risk is not
important befo re emergence. When there is
frost damage just afte r emerg ence, yield
loss may be Ii m ited because of re- emergence. A killin g fro st in Apr il is not as
harmfu I because there wi 11 be so me crop to
harvest. Therefo re, in th is stud y, the average
frost risk between December and March is
co nsidered.
Th e prob abi liti es of an extreme minimum tem perature of 0, -1 , -2 , -3, -4 and
-5C we re group ed in four cl asses: a fro st
eve nt in less than 3.3% of years, betwee n
3.4 and 10%, between 10 and 33.3 %, and
in more than 33.3% of yea rs . The limits of
th es e c lasses represent a frost eve nt once in
every 3, 10, and 30 yr.
A map of potato distribution in the
Altiplano was made. For Peru , it was bas ed
on district leve l census data (INEI, 1993).
For Boli v ia, departmental level census data
(G-DRU, 1996) and maps of c rop distribu tion (ZONISIG , 1998) were used. Overlays
(simu ltaneous qu eries) of the potato map
and the frost risk maps we re made to
est im ate frost risk in potato producing
areas.

Results
Frost risk is lowest in th e w armest parts of
the Altip lano around Lake Ti ticaca and in
the relatively lo wer parts in the southeast. It
is hi ghest in the eastern Alt ipl ano an d in th e
hi gh parts of the ca tchm ent (Fi gure 2). Tabl e
1 prese nts the prob ability of extreme
tempera tu re events for the potato area in
the Alt iplano.
Of the 63,000 ha area plan ted to potato,
on ly 4% ha s a neg li gible risk of a -2C
eve nt, and 1 8% has a -2( eve nt less than
once eve ry 10 yr. Twenty-five percent of the
potato area has an extremel y hi gh fros t ri sk,
w ith a prob ability of a -2 ( eve nt once
eve ry 3 yr (>3 3.3 %) or more. All the area in
this latte r class is most like ly planted w ith
potato species oth er th an 5. tuberosum.
This assumpt ion is supported by Canahua
and Aguilar (1992) and Huanco (1992) who
est imate th at about one-third of the total
potato area on th e Peruvian Altiplano, is
planted w ith bitter potatoes, of w hi ch 60 %
are 5. juzepczukii, and 33% are 5.
curtilob um (Ca nahua and Agui l lar, 1992). It
is also supported by Rea 's (199 2) estimate
that 15% of the potato area in Boli via is
planted to bitter pota toes . On the A ltiplano
there are relati ve ly more bitter potatoes
than in most other zones in Bolivia.
Table 1 indicates that the introdu ction of
potato cult iva rs w ith increased fro st
tole ran ce co uld stro ngl y redu ce frost
damage. If frost tolerance in potato increases from -1 to -2(, and damage onl y
occ urs at -3( and lower, the percentage of
th e current potato area w ith a frost eve nt
less th an once eve ry 1 0 y r nea rly doubles,
increa sing fro m 18 to 32%. W ith a furth er
in cre ase of to lerance to -3 C, there w ould
be anoth er 15% in creas e of the current
potato area with a frost event less than once
eve ry 10 yr.

Discussion and Conclusion

Current frost risk in potato producti on is
rather hi gh, on ly 18% of the potato area has

3 7 6 Natura l Resource Management in the Andes

70"

os

69 *

61

10

71*

67'

-wl

IW

-14

-w 1 11s

-u

Peru

Peru
11 -

11

J8

-ii::

Prnbabilty [%] of a
-3C event in

Pro babilty [%] of a

-2C event in
December to !'.;l arch

Decem ber t o Mar ch

19

19"

CJ0- 3 2
CJ3.3 - 99
CJ 10-33 .2
CJ 333- 100

co

-w

16 '

'"

CJ
CJ
CJ
CJ

Chile

50

20

JOO

71) *

69 *

50

20

c
11

0-3 .2
3.3- 9.9
l O 332
333- 100

Chile

ki lo1Cie te1s

b lo m e l~r:;

os

61 '

1 1

;;;:

63 .

'--J
'--J

69 *

Figure 2. Probability of a -2( (A) and a -3( (B) event on the Altiplano between December and March.

100

70

Table 1. Distribution (%)of the potato area (63,000 ha) in the Altiplano over the extreme temperature
probability classes (for the period December to March).
0

Temperature (

Probability

C)

-1

-2

-3

-4

-5

2
4

4
14

8
24
52
16

15
42
34
9

34
38
25
3

(%of yr)
< 3.3
3.3-10
10-33.3
>33.3

2
31
66

57

57

36

25

a -2( event once every 10 yr or less. That

confirms the importan ce of the frost
probl em in potato produ ction on th e
Altiplano. The introdu ct ion of potato
cu ltivars with frost tolerance of -2 ( co u Id
lead to a sharp decrease in frost risk and
damage, and hence increased produ ct ion.
A furth er increase of to lerance to -3 or -4C
would have a simil ar effect. Thi s assumes,
however, having potatoes that not on ly
resist these low extreme temperatures, but
also grow we ll in the co ld weather associated with these low abso lute minimum
temperatures .
Because farmers select favorab le microclimates to p lant potatoes, frost risk may
have been overestimated . Neverthe less, the
estimates may approach reality. Beca use of
the prese nce of potato spec ies more
tolerant than 5. tuberosum, and because
other reports indicate similarl y hi gh frost
risks, Gade (1975) distin guishes between
the effective and the absolute I imits of
cu lti vat ion in the And es. Beyond th e
effect ive limit, y ields are unsatisfactory and
crops are not important. Beyond the
abso lu te limi t, the crop w ill not grow or
y ield at all. In the Vil canota Val ley of Peru,
Gade (1975) defin ed the effective limi t,
where hardy crops y ielded satisfactori ly in
at least 60% of the years, as 3,910 m. Up to
the abso lute limit of 4 ,3 20 m, th ere we re
scattered fie lds with "patheticall y lowy ield ing" crops, wh ich were owned by
pastorali sts that do not prima ril y depend on

378

tfoturol Resource Management in the Andes

crop harvests . Gade's 60% good years is

close to the 33% bad years used in this
paper. The resu Its have even greater
corresponde nce tak in g into account that
this study co nsid ered only frost risk,
ignoring other produ ction risks such as
drought or hail.
Th e introduction of new frost-tol erant
potatoes cou ld lead to an expans ion of the
potato-growing area. That cou ld in vo lve
p lant ing more potatoes in the current
product ion zones or shift in g toward new
zones that could not be used before. With
th e exception of the shores of Lake Titicaca ,
however, the A lti plano is not dense ly
popu lated and agricu lture is not very
intens ive.
Land ava il ability does not seem to be
limi ting in potato production. Instead of
increasing area or shiftin g into new and
co ld areas, farmers wo uld probably choose
to in crease produ ct ion in the current
produ ction zones . That cou ld make potato
production more efficie nt and profitab le.
On the other hand, potential new production areas at high altitudes might be
assoc iated w ith better (not eroded) so ils and
hi gher prec ipitat ion. Hence, some ex pansion of potato area into the higher parts of
the catc hments mi ght be expected. More
in -depth knowledge of farmer strategies in
relati on to frost ri sk on the A lt iplano is
needed to predict their response to the
introduct ion of new cultivars.

To ana ly ze frost dama ge, we used the

average frost risk from D ecember to March .
However, a frost event in January ha s a
different effect on yield than one in March.
Thi s could be taken into account by using a
computer model that simulates th e effect of
frost on potato yield using, for exa mple, a
daily tim e scale. Moxey (1991) followed
this approach to estimate the benefit of
non-ice nucleation active bacte1ia for frost
protection of potato. A model th at simulate s
growth , development, and response to frost
stress of Bolivian potato cultivars is curren tl y under development, and w ill be
linked to the climate surfaces . That will
allow a more refined zonation, ri sk estimation, and ex ante impact assessment of new
frost-to lerant cultivars.
Thi s study has focused on the potential
impact of breeding for po tatoes w ith
increased frost tolerance, but ignored the
fro st tolerant species such as 5. juzepczukii
and 5. ajan huiri. With a few except ions
(Rea and Vacher, 1992; Tapia and Saravia,
1997), these species have not been the
subject of scientific research . Wh ereas they
are often said to be low yielding, Tapia and
Saravia (1997) report yields of 37 t/ha for 5.
juzepczukii and 22 t/ha for 5. ajanhuiri.
Th ese native species m eri t more bas ic and
adaptive research, and need to be consid ered wh en designing strategies to diminish
frost ri sk in potato production on the
Altiplano.

References Cited
Aguirre, G., J. Calderon, D. Buitrago, V.
Iri arte, J. Ramos, J. Blajos, G. Thiele, and
A . Devaux. 1999. Ru st ic seed bed, a
bridge between formal and traditional
potato seed systems in Boliva. Impact on
a Changing World. CIP Program Report
1997-98. Lima, Peru. p. 195-203.
Canahua, A. and P.C. A guilar. 1992.
Agroecologica de las pap as ama rgas en
Puno. In: Rea, J. and Vacher, J.J. La papa
ama rga . I m esa redonda: Peru - Bolivia,
La Paz 7-8 Mayo 1991. ORSTOM, La
Paz, Bolivia. p.57-61.

Carrasco, E., A. Devaux, W. Garcia, and R.

Espre lla. 1997. Fro st tolerant potato
var iet ies fm the A ndean Hi gh l ands. In:
Program Report 199 5-1996. Internationa l
Potato Center, Lima, Peru. p.227-232.
Corb ett, J.D. 1998 . Classifying m aize
produ ct ion zo nes in Kenya through
multivariat e cluster analysis. In: H assa n,
R.M. (ed.). Maize technology development and tra nsfer. A GIS application fo 1
research and planning in Kenya.
p. 15-25 .
De Bouet du Portal, D. 1993. Etud e des
gelees sur I' Al tiplano Bolivien. ORSTOM
and ENSAM, Montpellier, France. 41 p.
Erickson, C. L., n .d. lnvestigaci6n
arqueol6gica del sistema agricola de los
camellon es en la Cuenca d e l Lago
Titicaca del Peru. Centro de lnform ac i6 n
para e l Desarrollo, La Paz , Bolivia.
336 p.
Frere, M., J. Q . Rijks, and J. Re a. 1975 .
Estudio agroc lim atol6g ico de la zo na
A ndin a. FAO, Rome, Italy. 375 p.
Gade, D.W. 1975. Plants, man and the
Land in the Vilcanota Valley of Peru.
Biogeographica 6. Junk, Th e Hagu e,
Netherlands. 240 p.
G-DRU (G ru po lnterin stitucional d e
Desa1rollo Rural). 1996. Anuario
Estad fstico clel Sector rural 1995-1996 de
Bolivia. Centro de lnforma c i6n para e l
Desarrollo, La Paz, Bolivia. 297 p.
Huanco, S. 1992. Potencial de las papas
amargas en el Altiplano de Puno. In: Rea
and Vacher, Rea, J. and J.J. Vacher. 1992 .
La papa ama 1ga. I mesa redonda: Pe1u Bolivia, La Paz 7-8 Mayo 1991 .
ORSTOM, La Paz, Bolivia. p. 25-26.
Hutchin so n, M.F. 199 5 . Interpol ating mean
rai nfal l using thin plate smoothing
splines. Int . J. Geo. Info. Sys. 9(4):385403.
Hutchin so n, M.F. 1997. ANUSPLIN version
3.2. User G uide. Centre for Reso urce and
Environm enta l Studies, Australian
N at ion al University, Canbet"ra, Australia.
39p.
INE I (ln stitut o Nac ion a l de Estadistica e
In formatica) . 1993. Ill Censo Nac iona l

CIP Program Reporl 1997-98

379

Agropecurio. Database . INEI, Lima,

Peru.
INTECSA (Internacional De lngen ieria y
Estudios Tecnicos S.A.). 1993. Estudio de
Climatologia. Plan director globa l
binacional de protecci6n - prevenci6n de
inundaciones y aprovechamiento de los
recursos del Iago Titicaca, rio
Desaguadero, Iago Poop6 y Iago Salar de
Coipasa (S istema T.D.P.S). INTECSA, AIC,
CNR, La Paz, Boli via. 85 p.
Kalma , J.D., G .P. Lau ghlin , J.M . Capr io and
P.J.C. Hamer. 199 2. Th e b ioclimato logy
of frost. Advances in Bioclimatology 2.
Springer-Verlag, Berlin, Germany. 144 p.
Le Tacon, Ph. 1989. Manifestation des
risques climatiques a l'eche ll e de
l'exp loitation agricole, consequenes sur
les pratiques paysannes. Cas de
I' Altiplano Bolivien. Memoire de stage.
ORSTOM, ENSSAA, CNEARC, ENSAM,
Montpellier, France.130 p.
Morion, P. 1989. Du climat a la
commercialisation: !'example de
I' Altiplano peruvien. In: Eldin , M., and P.
Milleville. Le risqu e en agriculture.
ORSTOM, Paris, France. p. 187 -224 .
Moxey, A., 1991. Estimation of bid prices
for biotechnological frost protection of
maincrop potatoes. Ag. Sys. 37:399-413 .
OEA (0 rganizaci6n de los Estados
Americanos). 1996. Diagnostico
ambiental del sistema Titi cacaDesaguadero-Poopo-Salar de Coipasa
(S istema TOPS), Bolivia - Peru. OEA,
Washington, D.C. USA. 19 2 p.

3 80

Natural Resource Monogement in the Andes

Rea , J. 1992. Vigenc ia de las papas nativas

en Boli via. In : Rea, J. and J.J. Vacher. La
papa amarga. I mesa redonda: Peru Bolivia, La Paz 7-8 Mayo 1991. ORSTOM ,
La Paz, Bolivia. p.15-23 .
Rea, J. and J.J. Vacher. 1992. La papa
amarga. I mesa redonda: Peru - Bolivia,
La Paz 7-8 Mayo 1991 . ORSTOM, La
Paz, Bolivia. 111 p.
Sanchez de Lozada , D. , P. Baveye and S.
Riha. 1998. Heat and moisture dynamics
in raised field systems of the lake Titica ca
region (Bolivia). Ag . Fo rest Meteorol.
92(4):251-265.
Smith, C.T., W.M . Denevan, and P.
Hamilton. 1968. Ancient ridged fie ld s in
th e region of Lake Titicaca. Geo. J.
134:353-367.
Tapi a, N. and G. Saravia.1997.
Biodiversidad en papas amargas. Prov.
Tapacari, Dpto. de Cochabamba.
AGRUCO, Cochabamba, Bolivia. 53 p.
ZONISIG (Proyeto Zonificaci6n Agroeco l6gico y Establecim iento de una Base
de Datos y Red de Sistema e lnformaci6n
Geografica en Bolivia). 1998.
Zonificaci6n agroeco l6gica y
socioecon6mica de la Cuenca del
Altiplano del Departamento de La Paz.
Ministerio de Desarrollo Sostenible y
Planificaci6n, Prefectura del
Departamento de La Paz, DHV-ITC, La
Paz, Bolivia. 213 p.

Simulating the Response of Potato to Applied Nitrogen

W. Bowen1, H. Cabrera2 , V. Barrera3 , and G. Baigorria4

In potato (So/anum tuberosum) production,

nitrogen (N) is applied more frequently and
in grea ter amounts than any other nutri ent.
It is also th e nutrient th at most ofte n limits
y ield . W ithout added N, gro w in g plants
often show N defici ency characte ri zed by
ye ll ow leaves, stunted growth, and lower
yie ld s. Because it is an imp ortant input, N
and factors affecting its availability have
been th e subject of much invest igation
(Harris, 1992) . A common obj ect ive of
many of those studies ha s been to develop
N recommendation systems to ass ist
growe rs in determining th e amount of N
needed and the best tim e to app ly it.
Knowl edge of when and how mu ch N to
app ly is esse nti al, not only because N
inputs have an economic cost but Nin
excess of that used by the cro p may have
an environmental cost. The amo unt of N
needed for crop growth and its ultim ate fate
are important issues re ga rdl ess of w hether
the N so urce is organic matter or mineral
ferti li ze r.
A n app ro ac h often used in field expe riments for estimating th e amount of N to
app ly to potato in vo lves meas uring y iel d
response to in c rea sing rates of N fertili ze r.
Yi eld response is quantifi ed by fitting a
mathematical equation to the data. This
eq uat ion is then used together with economic information to inves ti gate retu rn s to
in vestments in fe rtilizer N. Although this
app roac h has prove n usefu I for demonstratin g th e concept of dimini shing return s, it is
nothin g more than a black-bo x approac h,
w hi ch offe rs limited inform ation for derivin g N recommendations in anoth er year or
at another site.

1 IFDC/C IP, Lima, Peru.

2 INIA, Ca jamarca, Peru .
3 INIAP, Quito, Ecuador.
4 CIP, Lima , Peru .

Nitrogen is a dynamic and mobile

nutri ent, hence its effec t on crop production
is rarely the same from yea r to yea r. At best,
equations from va ri ab le N-rate ex perim ents
descr ibe only hi sto ric al relations in th e
data. As such they offer little insi ght into th e
pro cesses that must be understood to
man age N inpu ts appropriately.
More inform at ive ap proaches for mak in g
N reco mm endation s have follow ed from a
better und ersta ndin g of the processes that
affect both crop N demand and soil N
supply. That has led to N reco mm end ation s
based on th e simpl e prin c iple that th e
amount of additi onal N to be applied ca n
be est im ated by kn ow in g the crop N
requireme nt and the pote ntial soil N supp ly.
The intern al N needed to obtain a specific
yield sets the crop N requirement. Soil N
supply is set by th e properties of th e so il ,
parti cul ar ly the so il organic matter (SOM)
con te nt and ot her properties that affect the
extent and rate of microbial decomposit ion.
Both also depend on the vagari es of
weather. For hi gh-yieldin g crops in most
soils, th e crop N requirement w ill exceed
the so il N supp ly. It is the differe nce
betwee n these two that approximates the
amount of supp lemental N need ed by a
growin g crop.
Generall y, the amou nt of N fertili ze r to
apply is calc ul ated by estimatin g the soi l N
supply, a target y ield leve l, and an expected
fertiliz er efficiency; thi s last beca use not all
applied N is normally recove red by th e
crop (Dahnke an d John so n, 1990). To
estim ate eac h of these components,
measurements of one or more of the
following vari ab les may be need ed: whole
plant N content required to reach a target
yield, min eral N released during soil
incub at ion , total N co ntent of the soil,
CI PPrngrnm Report 1997-98 38 1

rni neral N present in th e so i I profil e at

plantin g, and leaf N co nce ntration or
chlorophyll co ntent at a specifi c growth
stage. Th e relation betwee n an y of these
va riabl es and the benefit of ad ded N is
usuall y defined in fi eld ca libration studies.
Becau se of site-specifi c effects, such studies
need to be co nducted ac ross a broad range
of so il-c rop-c lirnate co nditi o ns. Th ere is,
howeve r, a practical limit to the number of
soil-crop-c lirnate conditions that ca n be
included in field studies.
An add iti o nal tool now ava ilabl e fo r
evaluatin g N rnan ageme nt practices and for
makin g N reco111111end ations is o ne based
on cornputer simul at io n. Using dynamic
sirnulation tec hniqu es, sc ientists have bee n
abl e to co nstru ct computer mod els ca pable
of simul at in g man y of the major processe s
that control N demand and supply. When
assembled w ithin th e framework of co mprehen sive crop-growth model s, th eir
dynarni c nature makes th em valuabl e fm
explorin g N m anagement options ac ross
a theoreti ca ll y unlimited number of cro pping prac ti ces, so il types, and weat her
conditions.
Th e purpose of thi s stud y is to illu strate
how we are using a potato growth model to
better und erstand N dynamics in potatobased cropping systems. We stress, how-

eve r, that simul ation model s are not a

substi tute for standard testin g and monitorin g of th e N status in soils and plants.
Rath er, the y are a tool for exte ndin g the
va lu e of such meas urements by fac ilitatin g
a systematic anal ys is of how plant, so il ,
wea th er, and manage ment facto rs intera ct
to affect N dynami cs . In thi s stud y, we use
th e SUBSTOR potato mod el (Ritc hi e et al. ,
1995) to systemati ca ll y anal yze so me of
th ese fac tors to demonstrate how the mode l
ca n be used to ga in insight in to N dynami cs
o n a site-spec ifi c basis. We w ill also show
res ults from mode l testing in Andean
co ndition s. Th e ve rsion of SUBSTOR used
in this study is th at released w ith the
Dec ision Support System for Agrotechno logy Tran sfe r (DSSAT v3) (Jon es et al.,
1998) .

A Comprehensive Crop Growth Model

Th e SU BS TOR mod el prov ides a usefu I too l
fo r anal yzin g th e quantitati ve effect that
co ntro ll ed factors (e.g., man age ment),
un co ntrolled fa cto rs (e .g. , wea ther), and
site-specific soil properties have on principal co mpon ents of N balance. Th e mod el is
co rnprehensi ve in that it simul ates the
majo r processes assoc iated w ith c rop N
demand and so il N suppl y (Tab le 1). Th e
mode l prov ides a balanced approac h in th e
leve l of detail used to descr ibe so il and

Table 1. Major processes that are simulated and environmental factors that affect those processes in the N
sub model of SUBSTOR-Potato {DSSAT v3).

Process simulated
Crop Ndema nd
Growth
Development
Soil Nsupply
Mineralization/ immobilization
Nitrification
Denitrification
N0 3 leaching
Urea hydrolysis
Uptake

382

Notuial Resource Management in the Andes

Main factors influencing process

Solar radiation, temperature
Photoperiod, temperature
Soil temperature, soil water, C:N ratio
Soil temperature, soil water, soil pH, NH 4 + concentration
Soil temperature, soil water, soil pH, soil C
Drainage
Soil temperature, soil water, soi l pH, sail C
Soil water, inorganic N, crop demand, root length density

plant processes. All of th e processes li sted

in Table 1 are simulated using a daily tim e
step. Beca use it is comprehe nsi ve and
dyn am ic, th e mod el ca n be used to eva luate, on a site-spec ifi c basis, many alte rnative man age ment pra cti ces aga in st th e
un ce rtainti es of weath er. Simulation res ults,
when ba sed on reli ab le input data , ca n then
provide critical information for definin g the
best N management practi ces from both an
eco nomi c and env ironmental perspec ti ve.
W e have used experimenta l data from
Ecuador and Peru to first test the performance of SUBSTOR und er Andean condition s (Fi gure 1). In Ecuador, Clav ijo (1999)
showed th at the mod el acc urately simulated th e res ponse of two cultivar s to N
fertili zer at two si tes in Carch i. In Peru ,
Yauri (1 997) showed that SUBSTOR
rea li sti ca ll y simul ated th e growth of two
cultivars w ith and without irrigation at a site
in Hu ancayo . A comparison of simul ated
and observed tuber yields from th e Ecuado r
and Peru stud ies is shown in Figure 1. Thi s
limited testing shows that the model
realistically simulated tuber yields that
ranged from 16 to 56 t/ ha due to differences
in weather, so ils, culti va rs, and man age ment. Furth er test in g is und erway as we
co ntinu e to cr iti c all y eva lu ate the performance of SUBSTOR and searc h for ways to
imp rove it as both a resea rch and management tool.
Simulated tub er yield (Vha)

60

~~~~~~~~~~~~~

50 -

......

40 30 -

20 -

.. "

.......

......'"

..1"

1
1
1
10 L ~-.L.._
~~1 ~~-~
~-~
~~

10

20

30

40

50

60

Observed tub er yie ld (Vha)

Figure 1. Relation between simulated and observed

fresh weight of potato tubers for sites in
Ecuador and Peru (dotted line is 1:1 line).

Components of a Systematic Analysis

A genera l approach for estim ating the
amount of N fertili zer (Nr) to appl y to a crop
may be based on th e fo ll owi ng calc ul ati on:

Nr= (Nv - N,)/Er

where N y is crop N demand (the internal
plant N required to attain th e expected
yie ld), N is the N suppli ed by the soi l, and
Er is th e ~x pected efficie ncy or frac ti on of
appli ed N th e crop is ex pected to recover.
This appro ac h provid es a useful fram ework
for exa mining th e main factors to be
consid ered w hen manag in g N inputs.
Within this framework, th e N bal ance in
agri cultural field s can be character ized by
definin g three major co mpon ents: Nv' N,,
and Nr To opt imi ze N manage ment for a
specifi c situat ion, quantitative estim ates of
each of these compon ents are needed
along with eco nomi c information .

Crop N demand
Crop N demand is th e product of th e
expec ted yie ld and internal N requirement,
w hi ch ca n be thought o f as th e minimum
amount of plant N assoc iated with max imum yield (Stanford and Legg, 1984).
A lthough a growin g crop may take up more
than th e minimu m N needed , extra N
(l uxury co nsumption) does not usu all y
resu lt in any yield benefit. Th erefor e, to
optimi ze N manageme nt and avoid its
ineffic ient use, it is important to know th e
expected max imum yie ld and its assoc iated
internal N requirem ent. Maximum y ield,
however, is not a co nstant. For a singl e
cultivar, maximum yi eld will vary from site
to site and year to year du e to the intera ction of genetic traits w ith photoper iod,
temp erature, so lar radiati on, wa ter, nutrient
avail ability, and mana ge ment. Many of
these interactions can be ca ptured in th e
SUBSTOR model .

To illu strate, Fi gure 2 shows how tuber

yields respo nded differently to app li ed N
wh en th e SUBSTOR mod el was run using
different weather years for a site in
Hu ancayo, Peru. Th e cultivar simu lated
was _Yungay, which was planted in midCIP Prng1om Repo111997-98

383

in simu lations designed to vary the amount

and tim ing of N applicat io ns.

Tuber fresh weight (t/ha )

60
50

40
30
20
10

50

100

150

200

250

300

N app lied (kg/ha )

Figure 2. Simulated response of potato tuber fresh

weight to applied Nusing daily weather

data from 1982-83 or 1990-91 for a site in
Huancayo, Peru.
November and harvested in ear ly April.
Whereas fresh y ields reached a maximum
of 55 t/ha using daily weather from 198283, the same amount of N in 1990-9 1
produced a maximum yield of on ly 26 t/ha.
The difference in N response betwee n
seasons was attributed mostly to a less
favorable distribution of rainfall during
1990-91. The simulation showed the plants
suffered a significant water deficit during
tuber bulking. A lthough yield leve ls varied
for the two seasons, the internal N requirement remained constant at about 16 g N/ kg
of total dry matter (OM ) (tops plus tubers),
whic h is within the range reported by Vos
(1995) for seve ral field experiments. The
simulated removal of N in fresh tubers was
about 2.8 kg N/t fresh weight in each
season, also within the range reported for
real data by Harris (1992).
Simulation can be used as well to
estimate the time course or N uptake
pattern of a growing crop. With output
provided on a daily basis, the model makes
it possible to define the period of greatest N
demand. Such information can then be
used to examine how management might
improve the synchrony between crop N
demand and N supply. For examp le, the
economic and yield impact of spl it applications of N fertilizer could be easily studied

384

Natural Resource l.\onogement in the Andes

Soil N supply
Nitrogen supplied by the soil comes
mostly from two sources: 1) mineralization
of soil organic N during the growing
season, and 2) mineral N initially present in
the soil profile at plant in g. Both sources
should be considered when estimating the
amount of supplementa l N needed by a
growing crop. Howeve r, the importance of
initia l min eral N tends to diminish in high
rainfall environments where significant
leaching can occur.
Since the mineralization of soil organic
N is a biological process, the amount of N
made available depends primarily on the
leve l of microbial activity and the amount
of carbon (C) substrate. Fo r most minera l
soils, th e C:N ratio in SOM is fairl y constant
at 10. That ratio favors a net release of N to
availab le mineral forms un less OM is added
wi th C:N ratios greater than 25. At a C:N
>25, there is usually a loss of plant ava ilable N as it becomes tied up through net
immobili zation.
The amount of mineral N present in the
soil profile at planting (ini tial mineral N)
often has a substantia l impact on the need
for supplementa l N, particularly in less
humid environments. In itial mineral N
usually va ri es across sites and years, with
the amount largel y determined by management and growth of the previous crop and
the residual N left from earlier applications .
If rainfall is not excessive, much of the
initial mineral N can remain available to a
crop throughout the growing season.
The process desc riptions in SUBSTOR
enab le captur in g the interaction of these
sources of available plant N w ith crop N
demand for innumerable combinations of
so i I type, weather, cu Iti va r, and ma.nagement. For example, simulation could be
used to examine how soil N supply might
vary across years for different quantities of
SOM and mineral N present at planting.
Moreover, a simu lation study like the one

used to deri ve Fi gure 2 wo uld obta in

so mewhat different respo nses if SOM or
ini tial min eral N inputs we re var ied.

Sources of supplemental N
In SUBSTOR, suppl emental N ca n be
add ed as a min eral fertili ze r or as a pl ant
res idu e such as gree n manure (G M ).
A lgori thm s d ea lin g w ith th e transfo rma ti on
of anim al manure have not ye t bee n
in clud ed, althou gh efforts to do so are
underw ay. Nitrogen management practi ces
th at ca n be exa min ed with th e model
in c lud e th e effect of va ryin g N rates, tim e of
appli ca ti on, pl acement depth , and fert ili ze r
so urce. Th ese pra cti ces ca n be studi ed fm
th eir effect not only on tub er yi eld , but also
0 11 nitrate leachin g or economi c return s.
Such studi es can be simulated for a sin gle
growin g season , or ac ro ss many seaso ns, to
qu antify th e imp act of wea th er var iability
and its associ ated risk .
A si mul ation exa mpl e that illu strates how
the model mi ght be used to es tim ate th e
y ield benefit of alternativ e N manage ment
opti ons is shown in Fi gure 3. In thi s exampl e, SUBSTOR w as run for 19 season s
usin g dail y wea th er data from Hu ancayo .

Cumul ati ve pro bability

1.0

~-~---~----~-~

0.8

On ce aga in, the culti va r si mul ated eac h

sea so n was Yun gay, p la nted in m id November and harvested in ea r ly A pr i I. Th e
N manage ment opti ons we re no extern al
source of appli ed N (0 N), a legum e gree n
manure (G M ) (4 t OM/ ha, N co ntent 2. 5 % )
incorporated just before pl antin g, and 250
kg N/ha app li ed as urea in tw o, equ al split
appli ca ti ons. Th e simul ated tube r y ield s fo r
th e 19 seaso ns are pl otted in Fi gure 3 as
cumul ati ve probability di stribution s. Th ey
cl ea rl y show th e superiority of th e urea
treatm ent in in creas in g y ield, but th ey also
dem onstra te th e ri sk o f low y ield s des pite
heavy N ap pli ca ti on. For exa mp le, beca use
of poor rain fa ll di stributi on and subsequ ent
drou ght stress, th e pro bability of obtainin g
essenti all y no res pon se to appli ed N w ould
be about 15%, or about 1 yr in 7. But the
prob ability o f obta inin g a si gnifi ca nt
resp onse to N is mu ch grea ter, w ith yield s
of 30-50 t/ ha ex pected abo ut 50% of th e
t i me.
W ith thi s type of in fo rm ati on prov id ed
on a site-spec ifi c bas is, dec isions rega rdin g
N manage ment ca n be made based on an
impro ved aw arenes s o f both th e potenti al
and limi ta ti ons of a give n env ironm ent. To
mak e a simu lat io n eve n more use ful , mod el
inputs and yie ld s co uld be co mbin ed w ith
actu al pri ce and cost data to obtain simil ar
di stributi ons fo r net return s. Th at would
better defin e th e eco nomi c ri sk, w hi ch is o f
more interes t to a farm er.

0 .6
0 .4

0 .2
0.0 '-----'-------'---~---'--'-----~
30
10
20
0
40
60
50
Tube r fresh weig ht (Vha)

Figure 3. The cumulative probability distributions

for potato tuber fresh weight resulting
from simulations made across 19 years of
weather from Huancayo, Peru. The N
management treatments were no applied
N (0 N}, green manure incorporated (GM),
or urea applied at 250 kg N/ha (250 N) .

Conclusions
Nitrogen manage ment ca n be imp rov ed
throu gh the in sight pr ovid ed by comprehen sive c rop simul ati on mod els such as
SUBSTO R. As a continu all y evo lv in g too l,
su ch models have th e potential to help
re sea rchers and fa rme rs better understand
how soi I, cro p, wea th er, and manage ment
factors in te ract to affect crop N demand,
soil N suppl y, and fertili ze r use effi c iency
on a site-spec ifi c basi s. Any number of
man age ment sce narios can be examin ed
and co mpared for th eir impact, not onl y on
econ omi c return s but also on the potenti al

CIP P1ogrom Repoil 1997-98

385

for excessive le aching of nitrates. Economic

and environmental (nitrate leaching) risks
due to uncertain weather can also be
quantified.
References Cited
Cla v ijo, N. 1999. Validaci6n del modelo de
simu lac i6n del sistema DSS AT en el
culti vo de papa (Solanum tuberosum L.)
en las cond iciones del Ca nton Montufar,
pro v incia del Carchi. Eng. Th esis.
Escu ela Superior Politecnica de
Chimborazo , Riobamba , Ecuador. 85 p.
Dahnk e, WC. and G.V. Johnson. 1990.
Testing so ils for avai lable nitrogen. In:
Westerman, R.L. (ed.). Soi l test in g and
plant ana lysis, Third Edition. SSSA,
Madison, W I, USA. p. 127-139 .
Harri s, P.M. 1992. Mineral nutrition. In:
Harris, P. (ed.). The potato crop.
Chapman & H all , London, UK. p. 162213.
Jones, J.W., G. Tsuji , G . Hoogenboom, L.A.
Hunt, P.K. Thornton, P.W. W il kens, D.T.
Im amura, W.T. Bowen, and U. Singh .
1998. Decision support system for
agrotechnology transfer : DSSAT v3. In :

3 86

Natural Resource Management in the Andes

Tsuji, G.Y., G. Hoogenboom, and

P.K.Thornton (eds.). Understanding
options for agricu ltu ra I production.
Kluwer Academic Publishers, Dordrec ht,
Netherlands.
Ritchi e, J.T., T.S. Griffin, and B.S. Johnson.
1995. SUBSTOR: functional model of
potato growth, deve lopment and y ield.
In: Kabat, P., et al. (eds .). Model l ing and
pa tameterization of the soi l-pl antatmosphere system. Wageningen Pers,
Wageningen, Netherlands. p. 401-435 .
Stanford, G. and J.O. Legg. 1984 . Nitrogen
and y ield potential. In : H auck, R.D . (ed .).
N itrogen in crop production. ASA-CSSASSSA, Mad ison , W I, USA. p . 263-272.
Vos, J. 1995 . Nitroge n and the growth of
potato crops. In : Haverkort, A.J. and
D.K.L . MacKerron (eds .). Potato eco logy
and mode llin g of crops under conditio ns
limiting growth. Kluwer Academ ic
Publishers, Wageningen, Nether lands . p.
115-128.
Yauri, H. 1997. Validaci6n de un mode lo
para simu lar el crec im iento del cultivo
de papa. En g. Thesis. Universidad
Naciona l Agraria, La Molina . Peru. 98 p.

Selecting Optimum Ranges of Technological

Alternatives by Using Response Surface Designs in
Systems Analysis
C. Le6n-Velarde 1 and R.A. Quiroz 2

The search for technologies that increase

the productivity of a farming system or a
component is one of the main objectives of
agricultural researchers. The researcher is
usually interested in finding maximums,
i.e., highest yield, highest net income, etc.
Usually, the bioeconomic outcome is
dependent on severa I i ntri nsical ly- i nterrelated factors. The fact that the explanatory
or independent factors are related limits the
usefulness of a one-factor-at-a-time strategy
to find the levels at which the dependent
variable attains the maximum (Box et al.,
1978). A joint, functional dependency of
the dependent variable on the independent
factors is thus desirable.
If all the independent factors involved
represent quantitative variables, one can
describe the behavior of the dependent
variable as a function of the levels of the
independent variables. This function is
called the response surface (Cochran and
Cox, 1957). Wh en the mathematical form
of the function is unknown, it may be
satisfactorily approximated, within the
experimental region, by a polynomial in the
independent variables (Brown and Rothery,
1993). First- and second-order designs can
be used to find the level at which each of
the independent variables should be set to
maximize the response surface (Box et al.,
1978; Cochran and Cox, 1957; LeonVelarde and Quiroz, 1994; Montgomery,
1984). In spite of its usefulness, this approach is seldom used in agriculture
because it is costly and complex to implement (Wilton et al., 1974).

Systems analysis applied to agricultural

field research constitutes a complementary
tool to deal with complex issues. The use of
simulation models enables the comparison
of hypothetical situations in agricultural
production systems, which are otherwise
difficult to assess in the real world. A wide
range of alternatives can be evaluated with
simulation models designed, constructed,
and validated to simulate a particular
system. Despite this flexibility, the common
trend in the modeling literature is to use a
one-factor-at-a-time strategy (Kropff et al.,
1997; Teng et al., 1997). By combining
simulation models and factorial designs,
response surfaces can be generated and
analyzed. Critical factors and levels can be
identified and ex-ante appraisal of technologies may be used to screen for most
appropriate recommendations to farm ers.
This paper describes how central
composite, rotatable designs can be used to
plan simulation studies to generate response surfaces and to determine optimum
levels of the variables studied. In addition
to the description of the method, an
example to illustrate how to generate and
analyze a simulated response surface is
given.

Materials and Methods

Response surface method
The response surface method (RSM) can
be represented mathematically as follows.
Let Y represent the response variable and \
the continuous independent variables that
influence Y. Suppose the researcher wishes

1 ILRI, Nairobi, Kenya, and CIP, Lima, Peru.

2 CIP, Lima, Peru.

CIP ProgramReport 1997-98

387

to find the levels of eac h var iabl e at w hich

Y is ma x imi zed . We may w rite the observed
res pon se as a function of\ as shown:

Y = f(X 11 X2, ... , X)+E

The expected res pon se E( Y) = f(X 1,X2, ... , X)
is called th e respo ns e surface . A two dimensional grap h of the response su rface
cou ld be draw n o n two axes, lead in g to
co ntours of consta nt va Iues of E( Y) .
In most RSM problems, the form of t he
relati o n between th e response and the
ind ependent va riable s is unknow n. Thu s,
the first step in RS M is to f ind a suitab le
approx im at io n for the tru e fun ct ion al
relation betwee n Y and Xi. U suall y, a low o rder polynomial can be used to desc ribe
this relation, particularl y in the re g ion
described by the leve ls of the ind ependent
variables stud ied (Montgome ry, 1984;
Brown and Rothery, 1993) .
Th e effic iency of th e res ponse surface f it
might be imp roved by the use of appropr iate experimental designs. Suited des igns
minimi ze the v ariance of the regress ion
coeffici ents. First-ord er orthogona l des igns
are th e simplest ones . A f irst-orde r des ign is
orthogonal if all the off-d iago nal eleme nts
of the X1X matrix= 0. Th e class of orthogonal first-ord er designs in clu des the 2'
fac tor ial fractions of th e 2k ser ies in wh ich
main effects are not related to each other
(Box et al. , 19 78; Co chran and Cox, 19 57;
Montgomery, 1984). Th e 2k desig n does not
afford an estim ation of the expe rimenta l
error unless some run s are repeat ed . A
common method of including replication is
to augm ent w ith sever a I obse rva tion s at the
center (w hen xi= 0) .
The most commonly used second -o rd er
respons e surface desi gn is the class of
rotatable des igns . Th e ana lysis of th e fitted
second-order response surface is often
called th e cano nical ana lys is. Th e term
rotatabl e indi cates th at the variance of the
predicted respo nse at some po int x is a
function on ly of the distance of the point
from th e design center, and not a fun ct ion

3 88 tlolural Resource Monogemenl in the Andes

of the direction (Montgomery, 1984; John

and Q uenov ill e, 1977). Anoth er pro perty of
this design is th at th e va rian ce of the
est ima tes is unch anged w hen the design is
rot ated about th e cente r, henc e the nam e
rotatab le designs.
A mong the rotata ble desi gns, the most
w ide ly u sed is the central composite
design . Thi s de sign co n sists of 2' facto rial ,
augme nted by 2k ax ial points and n0 center
points (o r 2k + 2k + n0 , where n0 is the
numb er of times the ce ntral point is repeated) . A centra l compos ite des ign is
made rotatable by the choice of a value "a" .
Th e va lu e of "a" depends on th e number of
po ints in the factorial porti o n of the design
(P = 2'), i.e. , a= P114 (For in stance, if k = 2
then P = 2' = 4 and a = 1 .414). The effi ciency of the desig n, estimated as the ratio
betwee n th e num ber of parameters to be
estimated and th e number of treatments,
max imizes w hen k = 3, i. e., 0 .67 (10
parameters and 1 5 treatments).

Alpaca production in the

altiplano: An example
In th e dry puna of the A nd ea n A ltipla no,
a sem iarid , trop ica l highland , nat ive
pa stu res are th e o nl y feed resource fo r
alp aca (Lama pacus). The c lim at ic va riati ons, mainl y prec ipita tion and low temper-atures, substantially affect the growt h
and digestibi I ity of the gra ss lands. Due to
he avy stock in g rates, the pastures te nd to
be deg raded, which adve rse ly affects fiber
and meat produ ction . A ph ys ica l ex periment to assess th e chara cteristics of all
po ss ib le combinat ions of the factors in an
opt imum grazin g management system is
imposs ible. Th e cost and the time required
to co ndu ct it wo uld be unreaso nab le.
A systems researc h approa ch, based on
sim ul at ion, was used to address the problem (1989-1991 ). After a good character ization of the system, an alpaca simul atio n
mode l was desi gned, co nstructed, and
vali dated (Quiroz et al ., 1990; Arce et al. ,
1994). This computerized mod el w as then
us ed to assess th e levels at w hi c h pasture

growth rate (GR), pasture digestibility (D)

and the stoc king rate (SR) would max imi ze
gross in come. The stocking rate direct ly
affects th e pasture growth rate, and the two
togeth er modify digest ibility.
A composite central, rotatable des ign
was used to simulate an ex periment to
estimate the expected outcomes of chan ges
in the leve ls of the ind epe ndent va riabl es
deem ed most relevant. The factors (X)
includ ed were stocking rate, pasture growth
rate, and pasture digest ibility. The central
point was se lected to mimi c the management leve ls of the variables used by farmers
(Quiroz et al. , 1991) Th e res ponse va ri ab les
(Y) were yearly fiber and meat produ ction
and th e annual gross in come of the farm
from alpaca production alon e. Th e simulation ti me for eac h scenario was 1 0 yr; the
averag e respo nse over the period was used.
The seco nd step consisted of simulating
the tim e-co urse of the impl ementation of
the optimal co nditions enco untered and th e
ex pected changes in gross income over
time. A logisti c model was fitted to the data
to estimate the param ete rs describing the
time changes in gross in come. Thi s exerc ise
was of particular importance, first to assess
the sustainability of the proposed chan ges,
and second to define th e milestones needed
to monitor progress.
Th e most sa lient features of th e graz in g
system simulated are summarized below
(Quiroz et al., 1991).
Type of pasture: bofeda l (spring fed yearround pasture) and rainfed native pasture
in th e dry puna of the Altiplano.
Surfa ce: 80 ha of nati ve pasture di vided
into 70 ha of rainfed pasture and 10 ha
of bofeda l.
Initial pasture avail ab ility: rainfed pasture
is 1 t dry matter (DM)/ha; bofedal is 1.5 t
OM/ ha.
Grazi ng pattern: 3 mo on rainfed pasture
(January, February, M arch); 9 mo in
bofed al (April to D ece mber).

Res idu al pasture availability: rainfed is

0.5 t OM/ha per mo; bofedal is 0.7 t
OM/ha.
Th e management ope ration s were:
matin g in Janu ary, birth of offspring in
Decembe r, weaning in September of the
second year, sh ea rin g every year. Sale of
products was assumed to be once a yea r.
Th e reso urces of land, number of
di visions (paddocks), herd composition,
matin g, weaning, and mortality and birth
rates were held constant.
Table 1 shows the va ri ab les and th e
leve ls for eac h variable assigned to eac h
treatment. It also identifies the leve ls
assi gned to the factori al, axial, and ce ntral
parts of the composite, rotatable design
desc ri bed above. Th e 15 trea tm ents were
coded for th e resp ecti ve analysis of production va ri ab les (fib er and meat). Five leve ls
of each fac tor were tested in the composite
central, rotatable des ign (2 3 + 2(3) + 1 (6)) .
Th e centra l point was repeated six times.
Th e biological respo nse variables, meat
produ ct ion and fiber production, were
multipli ed by their farmgate prices to
estim ate gross incom e.
Th e analys is of the bioeconomic response was programm ed in SAS (SAS,
1988) to ge nerate the res pecti ve response
surfa ces .

Results and Discussion

The si mul at ion process provided the
biologi ca l information for each sce nario
analyzed. All the factors studied significantly influ enced the response var iab les (P
< 0.01) and showed a significant qu ad rati c
effect (P < 0.01 ). Th e coefficients of the
mod el for fiber, meat, and gross in come are
shown in Table 2.
Th e co mbination of the leve ls of the
va ri ab les studied that maximized gross
incom e (refe rred to as optimum) used:
stockin g rate is 0.54 hea d/ha, pasture

CIPProgrnm Report 1997-98

389

Table 1. Variables, coded values, and treatments used for the application of the response surface method in a
simulated study of an alpaca production system in the dry puna of the Altiplano, Peru , 1991.
Code

Variables
-1.662

-1.00

0.00

1.00

1.662

X1: stocking rate (SR)

X1: growth rate' (GR)
X3: digestibility (D)'

0.56
6.233
62.4

0.70
6.913
63.25

0.90
7.913
65.2

1.10

8.913
67.25

1.24
9.533
68.6

Treatments:
1- 8, factorial part; 9- 14, axial
part; 15, central part repeated
sixtimes

Code

SR

GR

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15

(head/ha) (kg DM/ha

{%)

per d)

-1
-1

-1
-1

-1

-1
-1
-1
-1

-1.682
1.682
0
0
0
0
0

-1

1
0
0
-1 .682
1.682
0
0
0

-1
1
-1
0
0
0
0
-1.682
1.682
0

0.7
0.7
0.7
0.7
1.1

1.1
1.1
1.1
0.56
1.24
0.9
0.9
0.9
0.9
0.9

6.9
6.9
8.9
8.9
6.9
6.9
8.9
8.9
7.9
7.9
6.2
9.5
7.9
7.9
7.9

63.2
67.2
63.2
67.2
63.2
66.2
63.2
67.2
65.2
65.2
65.2
65.2
62.4
68.6
65.2

' Weighted values correspond to posture use, 3 mo for roinfed posture and for 9 mo bafedal. The ranges of growth rotes for rainfed
pasture isfrom 0.57 kg to 3.93 kg DM/ha day, and for bofedal from 8.12 to 11.40 kg DM/ha day. The digestibi lity of rainfed
ranged from 53.6 to 60.4% and of bafedal, 67.4 to 71.4%.

growth rate is 8.14 kg DM/ ha per day, and

digestib ility is 65.9 % . Si nce th e m ax imum
is ju st a point, we est im ated a 95 % co nfiden ce interva l of th e ma x imum and ca ll ed
it optimum ran ge. If th e leve ls of the
va ri ab les are maintain ed w ithin th e co nfi den ce interva l, th e expected optimum
ran ge fo r gross in co me va ried fro111
US $2 ,700 to $3,000/yr.
Far111e rs tend to 111 aintain stockin g rates
betwee n 0.6 and 0.9 hea d/ ha. That is

3 90

Nolural Resource Monogemenl in !he Andes

und ersta ndabl e because th ey mainl y

mark et the fiber and 111anage the system to
p rod uce as mu c h fiber as poss ibl e w hile
augmenting th eir animal stock. If we we re
to 111aximize the obj ective function for fiber
prod uct io n, we co uld say far111 ers are
already maximi z in g their produ ct io n
fun ction . As long as the perception is that
fiber is th e most important produ ct maintain ed, th ere is littl e hope for th e adoption
of tec hnologies to i111prove pasture
111anagement.

Table 2. Regression coefficients to generate the response surface and variable levels that maximize the surface'.

Regression coefficients
Poro meters

Fiber (kg)

Meat (kg)

Intercept (b 0}
Stocking rate (SR)
Growth rate (GR}
Digestibility (D)
SR x SR
SR xGR
GR xGR
D x SR
D x GR
D x D
Ri

167.0
-51.6
15.3
3.8
-20.3
-4.4
-8.8
-0.5
0.8
-7.2
0.68

2934.7
-1178.2
175.3
59.4
-324.3
17.0
-392.2
-33.3
-13.2
-152.9
0.76

Income (USS}
2123.30
-818.90
138.40
43.90
-238.60
0.60
-254.50
-21.00
-6. l 0
-107.20
0.76

Variable levels that optimize the response

and maximum response obtained.
Variable

Fiber (kg)

Meat (kg)

Income (USS}

SR
GR

0.62
8.98
65.93
213.70

0.52
8.05
65.96
4,040.80

0.54
8.14
65.93
2,858.90

Maximum obtained
0

Regressioncoefficientsaccordingta:

Y = /3

i=]

i=l

+ LtJ;X;+ L/3;;Xi + LLf3iiXiXj +

The system must be seen as a multipurpose one, instead of just a fiber-producing

system. When the stock, the fiber, and the
meat sold are valued, the figure changes.
The findings suggest that there are different
technological opportunities for increasing
farmers' income that must be explored. It is
clear that the fiber and meat markets should
be encouraged. Recommendations in this
regard are already being tested in Bolivia
by the Regional Project on South American
Camelids (International Fund for Agricultural Development and lnstituto
lnteramericano de Cooperacion para la
Agricultural and in Peru by Centro de
lnvestigac ion en Recursos Naturales y el

Medio Ambiente, a local nongovernmental

organization .
The levels of digestibility and pasture
growth rate that maximized the gross
income were a little higher than the existing
conditions reflected in the central treatment. Therefore, the main factor requiring
change to obtain the expected maximum
income is stocking rate. This needs to be
worked out in relation to birth rate and
mortalities.
An example of a 3-D response surface is
presented in Figure 1. Shown are changes
in gross income when digestibility is

CIP Program Report 1997-98

391

Gross income (US$)

3000
2 50 0 1---+--+-,.:IZ
20001--~~

1500~~~~~~~~~~~

1000U-+-~~~~~~~~
5oof-t-f---t--rll~~~~~~~~

oL-L----+----+-+-1--r~~~~~~~~~
-500 LJ--1---++-HI~

-1000
-1500

-2000~~~~~~1~-~~~~~
di
~\of\..
..tj-m~a a c:::; Cl?
di ~ O? O"i o o
;::

0)

,.._

...!.

ao,__.cy"'.>""1-dil.{)m"'

.,._ ........

Stocking rate
Head/ha

.
.,._

0 0 r--.:

L{)

rorom

f'....

.....-.....-

Growth rate
\<g oM/ha per daV

Figure 1. Response surface describing the simulated gross income of an alpaca production system as a function of
changes in stocking rate and pasture growth rate at 65% pasture digestibility.
maintained constant at 65% and levels of
stocking rate and pasture growth rate vary.
Changes in both stocking rate and pasture
growth rate are extremely sensitive. It is
important to show that when the stocking
rate is greater than 0.9 head/ha, income
decreases sharply. As a matter of fact,
stocking rates greater than 1 head/ha make
the system collapse in 5-7 yr (Arce et al .,
1994).
Figure 2 depicts an attempt to quantify
the path of implementation of the recommended management options to maximize
gross income. Three phases are identified:
1) existing, 2) adaptation, and 3) stable
(Leon-Velarde and Quiroz, 1994). The
existing phase has low producti v ity and an
income level that appears to have been
stable for man y years. The adaptation phase
is one in which technical changes take
place. The stable phase (resilient) is one of
higher production level quantified by the
asymptotic value in the logistic model.

3 92

Natural Resource Management in the Andes

According to these new sets of simulations

in time, with optimum management
practices the gross income of the system
can reach as high as US$3 ,90 0/farm per
year. It wi 11 take from 8 to 1 0 yr to reach the
asy mptotic income.

Conclusions
The paper shows how systems anal ys is and
composite central , rotatable designs can be
jointly used in agricultural systems to
search for optimum ranges of technological
alternatives. Our illustration was an alpaca
production system. But the principle applies
to any condition in agricultural research
w here a stochastic systems analysis tool
such as a validated simulation model exists.
Regarding the alpaca system, the results
showed that existing levels of income can
be substantially increased. Implementing
such changes, however, could take up to a
decade. The simulation also showed that in

Gross income (US$ x 10

6
5

.---~~~~~~~~~~~~~--,--,~......,.....~~,--,---,-,

Phase IInitial
sustainability

Phase 11technical
increments

3
2
1

10 11

12 13 14

Year
Figure 2. Simulated yearly gross income increase when the practices that maximize incomes ore implemented in

the alpaca system.

the medium- to long-term, the additional
income generated with the changes at the
fa rm leve l might be sustained.

References Cited
Arce, B. , C. Aguilar, R. Canas, and R.
Quiroz. 1994. A simulation model of an
alpaca system in the dry puna of the
Andes. Agri. Sys. 46:205-225.
Box, G.E.P., W.G. Hunter, and J.S. Hunter,
1978. Statistics for exper im enters. John
Wiley & Sons, Inc. NY, USA. 653 p.
Brown, D . and P. Rothery. 1993. Models in
biology: Mathematics, statistic s, and
computing. John Wil ey & Sons, In c., NY,
USA. 688 p.
Cochran, W.G. and G.M . Cox. 1957.
Experimental designs. John Wiley &
Sons, Inc. NY, USA. 616 p.
John, J. and M.H. Quenoville. 1977.
Experiments design and ana lysis. 2nd Ed.
Garden Griffin & Company Ltd. , London,
UK. 296 p.

Kropff, M.J., Teng, P.S., Aggarwal, P.K .,

Bouma, J., Bouman, B.A.M.,Jones, J.W.,
and Van Laar, H.H . (eds.) 1997. App li cation of syste ms approaches at th e field
leve l. Kluwer Academic Publishers/
Internat ional Rice Research Institute/
International Consortium for Agricultural
Systems Applications . Dordrecht,
Netherlands. 465 p.
Leon Ve larde, C.U . and R.A. Quiroz . 1994.
Analisis de sistemas ag ropecu arios: Uso
de metodos bio-matematicos. CIRNMA.
Editoria l EFl-GRAF, La Paz, Bolivia.
238 p.
Montgomery, D.C. 1984. Design and
analys is of experiments. 2nd ed. John
Wiley & Sons, NY, USA. 538 p.
Quiroz, R.A., B. Arce, R. Canas, and C.
Agui lar. 1990. Desarrollo y uso de
modelos de simulaci6n en la
investigaci6n de sistemas de producci6n
an im al. In Ruiz, M., and A. Ruiz (eds.).
Memorias de la reunion general de

CIP Prog ram Repoct 199798

393

RISPAL, Zacatecas, Mex ico. llC A-R ISPAL,

San Jose, Costa Rica. 43 p.
Quiroz , R., G., Mamani, R. , Revilla , C.,
Guerra, J. Sanchez, M. Gonzales, and G.
Pari. 1991 . Perspect ivas de ln vestigaci6n
Pec uaria para el D esar ro llo de la s
Comunidades de Puno. In: Aeguelles , L.
and R.D. Estrada (eds .). Perspectivas de
la ln vest igaci6n Pec uaria para el Altipl ano. Centro Internaciona l de
ln vest igaciones para el Desarrollo/
Proyecto de ln vestigacion de Sistemas
Agropecuarios Andinos, Lima , Pe ru . p

357-406.
Quiroz, R. , C., Leon-Ve larde, and W .
Bowen . 1999. Farming systems research
from a modeling perspective: Experiences in Lati n Amer ica. In : Collinson , M.
(ed.) . The history of farm in g systems.
CAB International , NY, USA.

3 9 4 Nolurol Resource /,\onogemenl in lhe Andes

SAS (Statistic Ana lys is System In stitute In c).

1988. SAS/STAT; User's guide, release
6.03 Edition. Gary, NC, USA. 1,028 p.
Teng, P.S ., M.J. Kropff, H .F. M. ten Berge, J. B.
Dent, F.P. Lan siga n, and H.H . Van Laar.
(e ds.) 1997. Application of systems
approaches at the farm and regional
leve ls. Klu wer Academic Pub li shers/
International Rice Research In stitute/
International Consortium for Agricu ltural
Systems App lications, Dordrec ht,
Netherland s. 468 p.
Wilton, J.W., C.A. Morris, E. A. Jenson, A.O.
Lei gh, and W.C. Pfeiffer. 1974. A l in er
programmin g mod e l for beef ca ttle
prod uction. Ca n. J. A ni mal Sc i.

54:693-707.

Mapping Aquatic and Semiarid Vegetation in the

Altiplano Using Multichannel Radar Imagery
R. Quiroz' and S. Saatchi2

Remo te sensing (RS) prov ides our first

opportunity to inventory th e surface
resources of the earth in a systematic,
repet itive manner. Remote sens ing methods
are beco ming increasin g ly important for
mapp ing land use and land cove r. So me of
the reaso ns for the in c reased use of RS
method s are (Sabi ns, 1997):
Large areas ca n be imaged quickly and
repetitively.
Images ca n be acqu ired wit h a spa tial
re sol ution that match es the degree of
detail required for the survey.
RS im ages e limin ate the probl ems of
surface access th at often hamper grou nd
surveys.
Images prov ide a perspective that is
lacking in grou nd surveys .
lm cige interpretation is fas ter and less
ex pensive than conduct ing gro und
su rveys.
Images prov ide an obj ec tive, permane nt
data set th at ma y be interpreted for a
wid e range of spec ific land uses and land
covers suc h as forestry, ag ri culture, and
urban growth.
In th e last 3 yea rs, C IP has bee n us ing RS
to map lci nd use/land cover in se lected
resea rch sites in the Andes. One of th ese is
the hi gh p lateau or Alt ip lano at an elevat io n
of mo re than 3,600 m. It is made up of a
ser ies of plains, hi ghland s, and isolated hi lls
ex tending from about 14 S latitud e in Peru
to 22 S lat itude in Bol ivia.
Fine- and coa rse -reso lu tion optica l
im agery, suc h as Land sat-Th emati c Mapper
ci nd th e Advanced Very Hi gh Resolution
Radiom eter (AV HR R), is be in g used to

1 CIP, Lima, Pe ru.

2 Je t Prop uls ion Laboral ory (NASA), Pasadena , CA , USA.

elaborate land us e maps and to monitor

land use/land cove r cha nges in the Alt iplano. Methods used are described elsewhere
(Townshend et al. , 1991; Defri es and
To w nshend, 19 94; Smith et al. , 1990).
Optical remote sens in g in cropping areas of
th e Altip lano has produced var ious results,
parti cul arly du rin g the cro pping season.
Seve ral reaso ns contrib ute to th is lack of
accurac y, the mo st important be ing: (1) th e
area is continuo usly covered w ith c louds
during th e cropp in g perio d, wh ich coi ncid es with the rai ny seaso n; (2) the smal l
size of the crop plots co mbined with the
rotation pattern of c rops and pasture
generates large va riability over small areas;
and, (3) the sim il ar ity in spectral attributes
(e.g., ca nopy architec ture) of the land cove r
types. Some of the li mitatio ns for the
succes sful use of optical remote sensi ng in
the A ltip lano ca n be c ircumvented by
spaceborne radar data. Radar is less
affected by clouds and ca n penetrate the
vegeta tion laye r (Saatch i, 1996; Saatchi et
al. , 1997).
Th e objective of th is paper is to demonstrate th e potentia l of space born e polarimetr ic radar in mapp in g land cover and
land use in th e Altip lano. Th e paper
desc rib es th e mo st important land cover
types in th e cropping area of th e Altipla no,
the aq uatic vegeta tion in Lake Ti ticaca, and
th e app li catio n of a maxim um l ikelih ood
Bayes ian prncedure to c lassify land cove r
types with ca nopy sim ilarities.

Material and Method s

Land cover types
Th e three cci tegories of land cover types
studi ed were aquatic vege tation , land

CIP P1ogram Repo1t 1997-9 8

395

vegetat io n, and v ill ages . The multil eve l

c lass ificat ion system based on remote
sens in g (A nderso n et al. , 1976; Sabins,
1997) was used to se lec t the clas ses. A
more detailed description of aquatic and
land vegetation fo ll ows.

Aquatic vegetation. Aq uatic vege tation

is composed of plankton and macroph ytes.
Within the rnacrop hytes , there are about 15
spec ies. The most importan t ones are the
I lach u (Elodea potamogeton, Myriophyllum
elatin oides) and the totora (Schoenoplectus
talora). Macrophytes, particularly ll achu
and totora , occupy th e sha llow sector of
Lak e Titicaca up to 15 rn deep (Co llot,
1982 ). Ll ac hu and totora constitute th e best
quality fo rage available to fee d milk/beef
cattle in the zo ne, wi th digestibility above
65% and protein co n te nts above 10%
(Quiroz et al., 1997). Totora is also used for
roofing and for makin g ree d boats. Th e
association of llachu an d toto ra is also of
c ru cia l importance for the subsistence of
nati ve f ishes and for nesti ng birds. In the
la st 50 y r, the area of l lach u and totora has
bee n reduced by more than 50 % (PELT,

1993).
Land vegetation. The agr icu ltu ral zo ne
incl ud es potato , barl ey, oat, alfalfa, and
quinoa, al l cu ltivated in sma ll plots. Th e
crop rotat ion begins with potato, fo ll owed
by any of the cerea l crops for th e nex t 2 or

3 yr.
Practicall y the entire A ltiplan o is a
natural grassland, varying accordi ng to
c lim ate and soil. It is richest in forage in
marshy areas, w hi ch yie ld more than 2.5 t
d ry ma tter (D M)/ ha per yea r. Th e product ivity of meadows, where grasses predorn in ate, ranges from 1 to 1 .6 t OM/ ha. O n
some types of scr ubl and , however, production drops to between 130 and 210 kg OM/
ha. The productivity of scrub or a grassancl-shrub meadow is also low, 150-2 10 kg
OM/ ha . Consequentl y, animal -carryi ng
capacity var ies great ly.
Grasslands are composed of differe nt
associations of pl ant spec ies. Th ey vary in
396

Natural Resau JCe Management in the Andes

size, density, growth habit, and ca nopy

arc hitectu re according to precipitation, so il
type, and altitud e, among other factors.
Mixed rangeland is presented in patches
all over the study area . It is an interrnixture
of grass land with shrub and brush land, and
in some in sta nces scattered trees. Th e most
abu nd ant tree is eucalyptus. Eucal yptu s
trees, w ith their thin trunks and sparse
canopy, resemble shrubs more th an forest.
Other types of vegetat ion in this c lass
includ e Polylepis sp. (quei'ioa), Budleia
coriacea (kisuara), Parasthrep hya sp. , and
Baccharis sp. (tola ). In zo nes near th e lake,
ye llow or inland totora is also considered as
mi xed range land vegetation .
Bofecla l de scribes a spring fe el , yea rrou ncl pasture, wh ich is characte ri zed by a
rich flora of good feed qua lity v ita l for the
production of camel ids. Since bofeclal es are
th e only source of yea r-round pasture, they
are usuall y heavi ly grazed. Another peculiarity of bofedales is the roughn ess of the
terrain. It res ults from so il eros io n if the
water is not man aged correct ly, trampl in g
by the animals, and the occ urrence of
patch es of grasses, particu Iarly in season a I
bo fe cla les .

Classification method
The quantitative ana lys is of the im age
was clone through a supe rv ised c lassification. Th e sequent ial steps are summar ized
below.
1. D efinition of the eight land covers into
w hi ch the ima ge was to be segmented:
wate r, ll ac hu , totora, crops/pasture,
grass lan ds, mi xed rangeland , bofeda l,
and v i I Iages.
2. Selection of 25 wi ndows of 3 x 3 pixels
(or training sites) for each of the 8 land
cove r types li sted above. Th e training
si tes we re chosen o n th e basis of v isual
in terpretation of the im age, aid ed by a
good know ledge of the area, plus field
data.
3. Use of the trainin g data to est im ate the
parameters of the class ifie r algorithm,
and to assess th e accuracy of the class ifi-

cation of the automated process over th e

entire im age.
4. Selection of another set of 25 w ind ows,
eac h 3 x 3 pi xe ls, for eac h land cover to
va lid ate the est imated param eters
(validation sites).
5. Label or c lassify every pi xe l in the image
into one of the eight land cover types.
6. Produce a th ematic map that summ ari zes
the res ults of the c la ss ifi ca tion.
A Bayes ian ma x imum likelih ood procedure was used for th e autom ated cla ss ification pro cedure (Rich ard s, 1993). Equations
1 through 4 (Table 1) descr ib e the dec ision
rul e appli ed for the classifi ca tion (Rao,
1973, Rignot and Ch ell appa , 199 3). With
equation 1, the pairwise squared distances
betwee n land cover types (c lasses) we re
est imated . A di sc riminan t lin ear functio n
was then calc ul ated fo r eac h c lass (equation 2). Eve ry pixel was th en refe renced
w ith respect to th e ce ntro id of eac h class
(equat ion 3), and th en cla ss ifi ed using th e
probab ility th at a pi xe l X belon gs to a
partic ul ar c lass (equation 4).
Th e c lass ification acc uracy w as ca lculated as the proportion of pixels correct ly
c lass ifi ed in the validation sites w ith the
discrimina nt function deve loped fo r the
tra inin g sites. The automated p rocedure
was program med in IDL (Intera cti ve Data
Language Resea rch Systems, Inc.) w here the
pr ior prob ability was ca lcul ated as the
number of pixels in each c lass of th e
tra inin g c lass over th e tota l numb er of
p ixe ls in the sample .
Remote se nsing scientists have access to
three co mm erc ial rad ar sate llites JER S-1 (LHH ), ERS-1 & 2 (C-VV), o r RADARSAT (CHH). (JERS , Japa nese Earth Resource
Sate llite; ERS, Ea rth Reso urce Sate llite,
European Space Agency; RAD ARSAT,
Synthetic Aperture RADAR o n boa rd a
Ca nad ian sa tellite.) Th e ex pected c lass ifi cations w ith JERS-1 or RADARSAT we re
simul ated with the corres pondin g c hannel s
and polarizations obtain ed from SIR-C data
(L-HH , and C-HH).

Remote sensing data

Durin g Apr il and Octobe r 1994, a
po lar im etri c shuttl e im ag in g radar- L-, C-,
and X-bands-s ynthe ti c ape rtur e rad ar
system (S IR-C/X-SAR) was flown on th e
space shutt le Endeavo r during two, 10-d
mission s. During both missions, shuttl e
rad ar co ll ected data ove r various reg ions in
South Am eri ca fo r defo restation , land use,
biodi vers ity, and flooding. Data acqui red in
April 1994 over Lake Titicaca betwee n
Huarina and Jes us de Machaca, La Pa z,
Boli via, was used fo r thi s stud y.
Th e im ages us ed in this paper are taken
from orbit 11 0.5, wi th 48.1 degrees in c idence angle, in multiple wavelengths (Lband , 24.0 cm; C-ba nd, 6.0 cm) in a du al
pol ari zat ion mod e: lik e- polariz ed (H H ) and
cross-polari ze d (HV) . The image size is 32
km in range (swath) an d 107 km in az imuth . The no min al re so lution of th e d ata at
th e 20 MH z bandw idth is approx im ate ly
12 .5 m.
Th e im ages we re furth er resa mpl ed to
25 m resolution to red uce speckle no ise.
Th e im age ri es were not corrected for
topograp hi c effect because a hi gh-reso lution , di gital eleva tion model (D EM) required
was not availab le.

Results and Discussion

Tabl e 2 shows the backsca ttering coefficients in the four channe ls for all land cove r
types (F igure 1). Table 3 summ arizes th e
clas sifi cat io n. Th e main diagonal shows the
perc enta ge of pixels correctly classified and
the miscla ss ifica tions are shown in th e off
dia gona l. Th e ove rall acc uracy for the
validation sites, as a sa mpl e of the ent ire
image, was 88%. Thi s is a low error rate
(12 %) cons id er ing the similariti es of th e
land cover types being discriminated. Th e
accuracy was furth er co nfir med in th e field
by vis uall y in specting a large portion of the
area mapped.
Aquatic plants w ere well discrimin ated
from eac h other (Table 3). Twe lve perce nt of

CIPProgram Repo1t 1997-98

39 7

Table 1. Maximum likelihood decision rule for classification .

Equation 1.
0 2 ri!iJ

=(x;-xd *f'- 1 *(x;-xi ) - 2* LnPriori

Equation 2.
1

Constant = -0.5 * x / * f' - *xi + LnPriori

Coefficient vector = f' - 1 * x i

Equation 3.
0 2 ; (x) = (x-x i ) 1 * f' - 1 * (x-x i ) - 2* LnPriori

Equation 4.
(-05*Df

Prij/x) =exp

(x)4

f.,exp (-o.sof (x))

Where,
2

0 1;/i! = squareddistancebetween land- covers iand j

x ; = mean value of class i

x i = mean value of class j

f' = variance covariance matrix

Priori =prior probability for class j

Ln = natural logorithm
Pr O/x) = probability that a pixel at a location X belongs to a class j

of (x) =distance between the pixel at location X and the centroid of the class j
exp= exponentia function

k =a class where the observatim in the pixel at location X is not included in the calculatim

toto ra p ixe ls we re m isclass ified as c rops.

Totora p lants are ex pected to have bac kscatter ing attri butes sim il ar to ann ual c rops
as they mature. A m ul t itempo ral spectral
analys is is then required to c lea rl y d iffere nti ate them from c rops.
The backscatter in g of the crop/ pasture
cl ass was lowe r th an ex pected beca use the
im age used was taken in mi d-Apr il , w hi c h
co in c ides w ith harvest. Thi s impli es that th e
39 8

Natural Resource Monogemenl in the Andes

bac kscatterin g was m ainl y the response to

the c rop res idu e remaining in the fields and
th e alfa lfa f ields, w hi c h we re still gree n.
Crop res id ues ca n co ntain 20-30% wa ter,
depend ing on rains. The ma in
m iscl ass ificati o n of the c rop/ pasture cla ss
was as toto ra.
Both the train in g and the testing areas fo r
grass lands we re se lected in the pasto ral
systems of Jes us de M ac haca . Thi s zo ne is

Figure 1. Thematic mop produced by maximum likelihood classification .

classified as dry, arid puna (GS/OAS, 1996).
Low-d ensity vegetation with patches of bare
soil is typical of this zone. The LHH
backscattering (Table 2) correspo nds to
patches of bunch grasses, locally known as
pajonales, and roughness va ri ation. The
low water co ntent of both so il and plants in
the semiarid zone was refl ected in low
responses in the C-band. Th e
mi sclass ifi cation of 8% of the pixels in this
c lass as mixed rangeland is not surprising.
There is a fuzzy boundary between the
spectral characteristics of th ese two classes.
The stro ng double-bounce scatterin g
res ulting from the mixed rangeland differen-

tiates this land cover from the other three

classes within the land vegetation. Th e
presence of taller trunks in vegetation such
as brush, sc rub, or trees was evidenced by
the magnitude of the stron g LHH signa l
(Tabl e 2). On the othe r hand, the vo lum e of
LHV scatte rin g was typical of shrubs, sc rub,
or trees containing a low -density ca nopy,
e.g., euca lyptus trees growing in the
Altiplano .
Bofedales showed th e highest error in
the classifi cation , w ith 16% of pixels
classifi ed as grass land. There are at least
two reasons for this mi sclass ifi cat ion. First,
bofed ales are small and assoc iated with
CI PProgromReporl 1997-98

399

Table 2. Backscatter statistics (dB) of training sites for land-cover type classes at L-Band and C-Band
frequencies.

Class

sx

cr0, HV

sx

-33.00
-2 5.37
-11.98
-16.58
-20.49
-9.97
-23.09
-3.35

l.00
0.40
0.39
0.58
0.38
0.86
0.58
0.43

-38.61
-36.49
-21.67
-2 2.55
-28.94
-22. 40
-32.99
-14.40

0.38
0.37
0.32
0.46
0.37
0.47
0.40
0.36

-39.92
-11.18
-10.52
-18.46
-2 4.40
-20.00
-17.53
-6.85

0.08
0.55
0.87
0.70
0.78
0.98
0.58
0.57

-25.93
-18.69
-14.23
-17.23
-22.22
-20.37
-19.04
-12.43

0.22
0.37
0.31
0.33
0.27
0.18
0.26
0.34

cr0, HH

L-Band
Water
Llachu
Totaro
Crops
Grasslands
Mixed rangelands
Bofedal
Village

C-Band
Water
Llachu
Totaro
Crops
Grasslands
Mixed rangelands
Bofedal
Village

S, =standard error of the mean; cr0 = Backscattering, dB.

patches of dry land pastures o r range land of

good quality, due to th eir close ness to water
and natural fertilization by graz in g animals.
Second, grasslands in th e lower land s
present higher soil moi sture than nei ghboring lands. These high -produ c in g grasslands
are ca lled seasonal bofedales. A more
acc urate differentiation betwee n permanent
and seaso nal bofedal es req uires a
multitemporal analysi s.

No sin gle chann el or pairs of channels

separated all land vegetation classes
accurately. Therefore, the information
contained in all four channels had to be
combined. All four channels combin ed
resulted in a good (89% accuracy) classifi cation of the land cover types. The resultin g
map containing a level II land cover
classification (Sabins, 1997) of the study
area is shown in Figure 1.

Villages were a sepa rate c lass. The effect

of co rner reflectors and meta l roofin g, in
both urban areas and widely dispersed
co mmunities, were cl ea rly di stin gui shable
in all four channels used (LHH, LHV, CHH,
and CHV). Areas with pronounced topography showed similar backscatte rin g to the
vill age class. This probl em co uld not be
corrected due to the la ck of a high -resolu tion DEM for the area.

The simulation of JERS- 1 and RADARSAT

classifications res ulted in less than 4 5%
accuracy. Neith er of the two sources of
commercial rada r data alone produced an
adequate class ifi cat ion of the most important land co ver types in the Altiplano. By
combining the LHH and CHH bandsmimi cking the combination of JERS- 1 and
RADARSAT data- the accuracy of classifi cation increased to around 78%.

4 00

Naturnl Resource Management in the Andes

Table 3. Confusion matrix of land-cover types derived from the Bayesian classifier.

Mixed
From class

Water

Water
100
Llachu
0
Totaro
4
Crops/pasture 0
Grasslands
0
Mixed
0
rangelands
Bofedol
0
Village
0
Error Rate

Grasslands rangelands

Llachu

Totora

Crops

0
92
0
0
0
0

0
0
80
12
0
0

0
0
12
84
0
0

0
0
0
0
92
8

0
0
0
0
8
88

4
0

0
0

0
0

16
0

0
0

20

16

Conclusions
Space borne polarimetri c rad ar data were
used to map aq uatic and semiarid vegetati on in the A ltipl ano. L- and C-bands and
two po lari zation mod es (like-po lari zed
bands HH and cro ss-po larized bands HV)
were used to c lass ify eight land cove r
cl asses. A thematic map was produ ced w ith
a cl ass ification acc uracy of 89%. Thi s
accuracy mi ght be improved if the image is
first corrected for the topog raphi c effect,
wh ich requires a hi gh-reso luti on DEM.
Some of the classes require a multitemporal
analysis, including at least two seaso ns, to
be co mpl etely spectrall y sepa rated. The HV
band s co ntributed more to the di sc rimin ati o n than th e HH bands.
Th e results provide new possibi li ties for
monito rin g land use changes in se mi arid
co nditi ons. Despite the fact th at most
vegetation presented simil ar spectral
attributes, particularly for optica l RS, radar
im age ry proved a good alternati ve for
separatin g eve n very simil ar cl asses such as
grass land s and mixed ran geland . The
ca pability of monitorin g croppin g systems
durin g th e rainy season gives rada r RS a
cl ea r advantage over opti ca l sensing.
Furth erm ore, different c lasses of aq uatic

12

Bofed al

Village

0
8

0
0
0
0
0
0

80
0

0
100

4
0

20

Total Error
11
0

vegetation, spec ifica ll y ll ac hu and totora,

ca n be eas il y differentiated.
An additi onal findin g was th at im agery
from th e two co ntemporary commerc ial
rada r sate II ites produ ced low expected
accuracy for the cl ass ifi ca tion of land cove r
types when used alo ne. When comb in ed,
howeve r, data from JERS-1 and RADARSAT
satellites in crease the c lass ification to an
acceptabl e 78% accuracy.

References Cited
Anderson, J.R., E.T. H ard y, J.T. Roach , and
R.E. Witm er. 1976. A land use and land
cover classification system for use with
remote se nso r data: U.S. Geologi ca l
Survey Profess iona l Paper 964. USGS,
Washin gton, D.C., USA.
Col lot, D.D. 1982. Mapa de vegetac i6n de
la bahia de Pu no. Revista Ecologia en
Bo li via 2:241-261.
Defri es, R.S. and J.D.G. Townshend. 1994.
NOVI deri ved land cover class ifi ca tion at
globa l sca les. Int. J. Remote Sens.
15( 17):3567-3586.
GS/OAS (Genera l Secretari at of the Organization of Am eri ca n States). 1996.
Diagn6sti co Ambi ental del Sistema
Titi caca-Desagud ero-Poop6-Salar de

CIP P1ogrom Repmt 1997-98

40 l

Co ip asa (S istema TOPS) Bo li v ia-Peru .

OAS, Was hin gton, D.C., USA . 192 p.
PELT (Proyecto Especial Lago Titicaca).
1993. Los Toto rales de l Lago Titicaca.
Estado, Tecnologfa y Potenc ial. lnforme
de trabajo. PE LT, Puno , Pe ru .
Quiroz , R., D. Pezo, D. Rearte, and F. San
Martin . 1997. Dynamics of feed resources in mixed farming systems in
Lat in America. In: Renard, C. (ed. ). Crop
residues in mixed crop/li ves tock farming
systems . CAB In ternatio nal , Wa lli ngton,

use in Amazon rai nfore st by usin g SIR-C

imagery, Remote Sens. En v. 59:19 1-202 .
Saatchi, S.S. 1996 . App li cat ion of SAR
remote se nsing in land su rface processes
over trop ical region, Proceedings of V II I
Simpos io Brasi leiro de Sensoriemento
Remoto held 14-19 Apr. 1996, Sa lvado r,
Bah ia, Brazil. p. 213-222 .
Sabins, F.F. 1997. Remote sens in g: Principles and interpretat ion. Third Edition.
W.H. Freeman an d Company, NY, USA.

UK . p. 149-180 .
Rao , C.R. 1973. Lin ear stat ist ical inference

Smith, O.S., L. Ustin, J.B. Adams, and A.R.

Gi ll espie. 1990. Vegetation i n deserts: A
regional measure of abundance from
multispectral images. Remote Sens.
Env iron . 31:1-26.
Townshend, J.R.G., C.O. Ju stice, W. Li , C.
Gurney, and J. McManus. 199 1. Glob al
land cover c lassific atio n by remote
se nsi ng: Prese nt capabi lit ies and future
poss ibilities. Remote Sens. Environ.

and its app lication. Second Ed ition. John

W i ley & Sons, In c., NY, USA.
Richards, J. A. 1993 . Remote sensing di gital
im age ana lys is: An introd uction.
Springe r-Ve rlag, Berlin, Germany.
Rignot, E. and R. Ch el lap pa. 1993. Maximum a poste ri ori class ifica tion of
multifrequency, multilook, synthetic
ape rture radar intensity data, J. Opt. Soc.
Am. 10:5 73 -582 .
Saatchi , S.S ., J.V. Soares, and D. A lves.
1997. Mappi ng defo restation and land

402

Noturol Resource /,\onogement in the Andes

495 p.

35:2 43-256.

A Process-Based Model (WEPP) for Simulating Soil

Erosion in the Andes
W. Bowen1, G. Baigorria 2 , V. Barrera3,

J.

Cordova\ P. Muck4, and R. Pastor 5

Soil erosion by wakr represents a major

threa t to the lon g-term productivity of
hill sid e ag ri culture in th e Andes. To help
co mbat thi s threat, research ers need a
quantitative understandin g of th e hydrologi c and physical proce sses th at drive so il
loss. Th ey also need to und ersta nd ho w
these pro cesses interact with land use and
mana gement to either diminish or increase
soil loss. This und erstanding, howeve r, is
not eas ily obtain ed in a cultural an d
ph ys ical env iro nm ent as di ve rs e and
complex as the A nd es, particul a1ly give n
th e pauc ity of ex periments and data needed
to quantify so il eros ion for the my1iad landuse systems in the reg ion.

To better und erstand soil erosion and its

impact in the Andes, we have begun a
multi - in stitutional and multidisc ip lin ary
effort based o n a systems resea rc h approach. This ap proac h is designed to ta ke
advanta ge of ex isting simulation mode ls,
using ex istin g ex perimental data to the
extent possible for model calibration and
model test in g. N ew experiments and other
data coll ection activities will be pursued as
this approach helps us to identify gaps in
ava il ab le data se ts and poss ibl y in acc uracies in proc ess descriptions incorporated in
present mode ls. W ith time, we expect the
iterative process of experimentation and
mod el eva lu ation to increase our understanding of the major processes th at cause
soil eros ion . W e also expect the syste ms
research ap proac h to result in improved
simulation mod e ls that can be used to help
pred ict how co nse rva tion meas ures-or the

1 IFDC/CIP, Limo, Peru .

2 CIP, Lim a, Peru.
3 INIAP, Qu it o, Ecuado r.
4 UNC, Caja marca, Peru.
5 UNALM, Lima, Peru .

lack of th em- might affect soil productivity

on a site-s pec ific basis.
A process-based si mulat ion model we
have beg un to evaluate und er Andea n
conditions is one rec entl y deve loped by the
Water Erosion Prediction Proj ect (WE PP ) of
the U.S. Depa rtment of Agriculture
(Flan ange n and Nea ring, 199 5). Th e WEPP
mod el mathematically describ es the
processes of so il particle deta c hm ent,
tran spo rt, and d eposition du e to hydro log ic
and mec hani ca l forces acting on a hill slope
pro fi le. It is co nside red to possess state-ofthe- art kn owledge of erosion sci ence, and
has become an important analytical tool for
glob al change studies (Favis-Mortlock et al.,
1996).
Th e purpose of this study is to provide a
first app rox im at ion of the capability of the
W EPP model fo r simulatin g so il loss in an
And ea n env iron ment. We w ill do thi s by
comparing model predictions of runoff and
soil loss to data meas ured from fallow
runoff pl ots during a 154-d study in the
eastern Andes of Peru.

Materials and Methods

Model description
Th e WEPP eros io n mod el is a con tinuous
simul ation co mputer program that ca lculates runoff and erosion on a daily bas is .
Erosion processes may be simulated at the
level of a hill slo pe profile or at the leve l of
a small wate 1shed. Our evaluation, however, is rest ri cted to th e hill slope profile
ve rsion (v 98.4) . The maj o r inputs fm
runnin g th e hill slope profile ve rsio n need to
be specified in four data file s: weather,
slope, so il, and management.

CI P Progrorn Repor t 1997-98

403

The weather file requires daily values for

precipitation, maximum and minimum
temperature, and solar radiation. In addition to rainfall amount, the model requires
three variables related to rainfall intensity
that are used to compute rainfall excess
rates and thus runoff. These are (1) rainfall
duration, (2) the ratio of time to peak
rainfall intensity to rainfall duration, and (3)
the ratio of maximum peak intensity to
average intensity.
With the WEPP model, a user can
simulate differences in soil types or management along the length of a hillslope
through the use of overland flo w elements
(OFE ). An OFE is a specified section of the
hillslope with a region of homogeneous
soils and management. Up to 10 OFEs can
be defined for one hillslope of a given
width. The number of OFEs to include in a
simulation is specified in the slope file ,
along with the length and slope characteristics of each OFE.
The soil file contains information on the
physical characteristics of the surface soil
and subsoil for each OFE . The surface soil
parameters required are the effective
hydraulic conductivity (Green and Am pt
effecti ve conductivity), i nterri 11 erodi bi I ity
(sheet erosion mostly caused by raindrop
impact), rill erodibility (small eroded
channels), critical shear stress (ri 11 detachment threshold parameter), and albedo
(fraction of solar radiation reflected back to
the atmosphere). For each soil layer, the
inputs required are cation exchange
capacity (CEC) and percentages of sand ,
clay, organic matter (OM), and rock
fragments. Up to eight soil layers to a
maximum profile depth of 1. 8 m may be
defined in the soil file.
The management file contains information needed to define initial conditions ,
tillage practices, plant growth parameters,
and residue management and crop management. It is the file with the largest number of
parameters.

404

Natural Resource Management in the Andes

Runoff plot study

Measurements of runoff and soil loss
w ere obtained from a runoff plot study
conducted during the first 6 mo of 1991 in
San Ramon , Peru (Pastor, 1992). In this
study, 30 runoff plots , each 10 m long and
4 m wide, were set up on slopes of 30, 35,
40 , 45, 50 , and 60 % . At each of the six
slope sites, five plots with the follo w ing
management systems were installed:
(1) clean fallow, (2) natural grass vegetation ,
(3) sweetpotato (ipomoea batatas) planted
in rows along the contour w ith hilling,
(4) sweetpotato planted in rows along the
contour w ithout hilling, and (5) sweetpotato
planted up and do w n the slope with hilling.
Treatments were not rep I icated, meaning
there was only one set of observations for
each unique combination of slope and
managem ent system . There were also fairly
large differences in the infiltration properties of the soils (Entisols) across slope sites.
lnfiltrometer measurements in the field
showed infiltration w as 21 .0 mm/ hr on the
30-35% slopes, 12 .4 mm/ hr on the 40-45%
slopes, 215.0 mm/ hr on the 50% slope, and
104.7 mm/ hr on the 60% slope.
Runoff and soi I losses w ere measured
after each rainfall from 21 Jan to 24 Jun
1991. There were 62 rainfall events during
this 154-d period, each producing measurable runoff and soil loss. The total rainfall
for the period was 1,053 mm. Total runoff
and soil loss measured for all treatments
and slopes are shown in Table 1 (note the
sweetpotato contour treatment represents
the mean of plots w ith and without hilling).
The greatest runoff and soil loss occurred
from the clean fallow plots. Any discernible
impact of slope on runoff and soil loss
apparentl y was masked by the differences
in soil infiltration properties across slopes.
The 50% slope, because of more rapid
infiltration, actually produced less runoff
and soil loss than the 30-45% slopes.

Model analysis
Our analysis of the WEPP model in this
initial study focuses on measured and
simulated data obtained from the clean
fallow plots only. Soil is often left bare

Table 1. Runoff and soil loss measured for four different management systems at six locations with different

slopes in San Ramon, Peru, from 21 January to 24 June 1991 (Pastor, 1992).
Slope(%)
Treatment
Treatments

30

35

40

45

50

60

mean

25.77
21.83
23.39
48.38
29.84

27.77
29.67
34.58
41.30
33.33

38.06
27.39
38.97
51.16
38.90

0.45
0.65
1.82
3.12
1.51

0.79
l.06
3.96
8.39
3.55

0.69
0.92
3.13
9.76
3.62

Runoff (mm)
Natural vegetation
Sweetpotato/contour"
Sweetpotato/up-downb
Clean fallow
Slope mean

42.18
30.95
59.10
62.85
48.77

41.57
26.24
44.69
50.99
40.87

38.25
26.29
38.52
43.40
36.62

52.84
29.36
33.53
60.05
43.95
Soil loss (mg ha1)

Natural vegetation
0.44
Sweetpototo/contour"
0.6 l
Sweetpotato/up-downb 2.42
Cleon fallow
14.88
Slape mean
4.58

0.48
0.80
3.33
13.03
4.41

110
0.98
3.21
11.38
4.17

0.86
1.42
4.01
7.74
3.51

" Sweetpototo planted in rows along the contour.

b Sweetpotato planted in rows up and down the slope.

between crops, thus it is important to

quantify the vulnerability of exposed soil to
the erosive forces of natural rainfall. Also,
we thought it better to first parameterize
and evaluate the performance of the model
for bare soil where there would be no
expected interaction between residue cover
and plant growth. If the model cannot be
shown to simulate soil loss realistically from
bare soil, then it would not be expected to
do so for other management systems.
Our analysis of WEPP in this study is
also restricted to results averaged across
slopes. There was insufficient information
about the on -site soi I properties to parameterize the WEPP model for each slope site.
Therefore, we evaluate the performance of
the model by averaging both measured and
simulated data across slopes. That is, WEPP
model input files were set up to simulate by
slope, but the simulated output values were
averaged ac ross slopes befo re comparing
them to meas ured data, which were also
averaged across slopes.

For evaluating model performance, we

followed the suggestion of Nearing et al.
(1994) to compare the frequency distribu tions of the mea sured and simulated events
rather than make event-by-event comparisons. Although the WEPP model should
eventually be evaluated by comparing
outcomes on an event-by-event basis, such
a comparison is not valid in this study. The
reason is that critical input parameters for
soil properti es and rainfall intensity were
not mea sured on site and could only be
approximated.

Model inputs
The weather file for simulating erosion
from the clean fallow plots was assembled
using daily rainfall recorded in
pluviometers at the experimental site. Daily
temperature and solar radiation values were
obtained from the CIP-San Ramon weather
station located about 1 km from the
experimental site. Rainfall intensity was not
recorded at the site, so we estimated values
for the duration and intensity parameters

CIP Program Report 1997-98

405

using statistical relations derived from

recorded values at the CIP-San Ramon
weather station.
Slope input f iles were assembled for
each of the six slopes by specifying one
OFE w ith a slope length of 10 m and a plot
width of 4 m. Soil input files we re made
identical for each of the slope sites. Soil
profile parameters for soi I depth , texture,
OM, CEC, and rock fragments were based
on measurements taken from a nearby site
(La Torre, 1985). Surface soil parameters
were estimated using either empirical
relations described in the WEPP User
Summary (Flanagan and Livingston, 1995)
or by calibrating simulated runoff and soil
loss to measured data from only the 30%
slope. Input values for surface soil parameters were 7.0 mm/ h for effective hydraulic
conductivity, 2.0 x 10 6 kg s/m 4 interrill
erodibility, 0.002 s/m for rill erodibility, and
4.0 N/m 2 for critical shear stress. The
management input files were also made
identical for each slope with parameters set
to simulate bare soil.

sensitivity analysis that the model wo uld

reproduce more of the measured runoff and
soil loss events by adjusting the rainfall
intensity parameters of specific rainfall
events. Thus , if the on-site rainfall duration
and intensity parameters had been known
w ith more certainty, the model output may
have more accurately reflected the observed number of events.
The cumulative frequency distributions
for runoff and soil loss measured during all
62 events are provi ded in Figure 2, expressed as the probability that any one
event will exceed a given runoff or soil loss

Runoff (mm)
4

(A)
3
2

-....

.....:.;:-:

-~
10

20

30

40

50

60

70

Rainfall (mm)
Soil loss (t/ha)

Results and Discussion

2.0

(B)

1.6

One of the most notable discrepancies

between WEPP model simulations and
measured data was in the number of rainfall
events that resulted in runoff and soil loss .
Whereas field measurements showed that
each of the 62 rainfall events produced
some runoff and soil loss, although often in
only small amounts (Figures 1Aand1 B), the
WEPP model produced runoff and soil loss
values for onl y 16 of the events. Closer
analysis revealed that the model failed to
pick up the less intense or smaller rainfall
events that produced measured runoff
values of less than 1 mm or soil loss values
of < .1 t/ ha.
The failure of the model to simulate the
smaller events was judged not significant as
the measured data clearly showed that any
substantial soil loss onl y occurred when
rainfall exceeded 10 mm (Figure 1 B) or
runoff exceeded 1 mm (F igure 1 C). Furthermore, we were able to demonstrate in a

406 Natuial Resource Management in the Andes

1.2
0.8

0.4
0.0

....

u.-.---._.,L.!:l!'----,___:_
__.::..
_____
0

10

20

30

40

50

60

__J

70

Rainfall (mm)
Soil loss (t/ha)

:l
2.0

.-----------~---~

(C}

0.8

..

0.4

0.0 ............

I
4

Runoff (mm)

Figure 1. Measured data from fallow plots showing

the relations between (A) rainfall and
runoff, (8) rainfall and soil loss , and (C)
runoff and soil loss. Runoff and soil loss
values represent the mean across slopes
for each event.

amount. These distributi o ns repres ent the

same data presented in Figure 1. H ere,
however, the hi ghl y skewed nature of the
data is more apparent; relatively few events
produced most of the run off and so il loss.
For examp le, only 4 of the 62 eve nts
resulted in a combined soil loss of 6.0 t/ha,
whe reas the cumulative so il loss from all
events amou nted to 9.8 t/ha.
Th e cumulati ve frequ ency distributions
for simu lated runoff and so il loss are also
shown in Figure 2. Th ese we re calcu lated
using the simul ated output for the 16 eve nts
reproduced by the mod el together w ith
zero val ues for the other 46 events not
reproduced by the mod el. For that reason
the simul ated distributions indicate that
most of the events would not have produced run off exceeding 1 mm or soi l loss
exceed in g 0.1 t/ ha, values simil ar to those
indi cated by th e measured distributions.
When th ese values were exceeded, the

Probability of exceeding(%)

simulated distribution of events simil ar ly

reflected the meas ured di str ibution .
For runoff, the simu lated distribution was
shifted to the ri ght of the measured distribution , indi cating the model tended to
overpred ict runoff. For soil loss, th e model
tend ed to overpredict small values (distribution shi fted to the right), but und erp red ict
large va lu es (distr ibuti on shifted to the left).
Thi s same trend in overp redicti ng small
va lu es and underpredicting large va lu es is
common w hen eva lu ating man y eros io n
models with measured data. Nearing (1998)
attributes th e trend to limitations in capturin g the random compone nt of measured
erosion data w ith a deterministic mode l.
Although measured and simulated distribution s we re somewhat similar, th e model
underestimated the 154-d totals for runoff
and so il loss. We attr ibute th at to the fa ilure
of the model to reproduce all events. The
total measured run off was 51. 2 mm compared w ith 40.4 mm for the simulation. Th e
total measured so i I loss was 9 .8 t/ha
compared w ith 7.6 t/ha for the simul ation.

100

Measu red
Simulated

80
60
40
20
2

Runoff (mm)
Probability of exceeding(%)
100

80

--- ~ ...

Measured
Simu lated

60
40
20

0.4

0.8

1.2

1.6

Soil loss (I/ha)

Figure 2. Measured and simulated cumulative

frequency distributions showing the
probability of exceeding a given runoff or
soil loss from fallow plots.

In add ress in g some of the uncertainty

regardin g input parameters, we found both
run off and so il loss to be quite se nsitive to
rain fa ll and surface soi l parameters. To
illu strate, Figure 3 shows how runoff and
so il loss wou ld vary w ith the duration or
intensi ty of a 61-m m rai nfall event for two
soils with different effective hydrauli c
conductivity values. Th e 61-mm rainfall
was th e largest recorded during the run off
plot study. In genera l terms, th e model
indi cates that the difference in soil loss
between a 4- or 6-h event cou Id be as
mu ch as 20 t/ha. Likew ise, an 8-h eve nt
mi ght result in no so il loss on a so il with
good infiltration properties, but erode more
than 10 t/ ha if infiltration is poor. Th e
meas ured soi l loss fo r this eve nt was 1.3 t/
ha. Clearly, it is important to have reliable
estimates of parameters like rain fa ll in te nsity and effective hydrau lic conductivity if
the model is to provide realistic estim ates of
runoff and soil loss .

CIPProgrnm Repo1t 1997-98

407

Runoff (mm)

Soil loss (Uha)

40 ,-----------------~-----~----~ so

-e- Runoff (Ke = 5)

- - Runoff (Ke = 11)
-a So il loss (K: = 5)
So il loss (Ke= 11 )

30

201 \.

40

10 L.. .. : .,.. .

20

_ __ _,_. -~~..........--=
- -W- - - ' 0

10

Figure 3. Sensitivity of simulated runoff and soil loss

due to variation in the duration of a 61
mm rainfall event for two soils with
different effective hydraulic conductivity
values (KJ

Conclusions
This initial analysis of the WEPP erosion
model has shown that it can be an effecti ve
tool for studying the hydrologic and erosion
processes that drive soil loss in the Andes.
The model should be particularl y va lu ab le
for focusing on the quantitative relations
and interactions between soil , weather,
topography, slope, and management factors
that determine runoff, soil loss, and deposition on a site-spec ific basis. Nevertheless,
we must caution that this study should not
be see n as a va lidation of the WEPP model
for Andean conditions . More rigorous
testing of mod el assumptions is needed.
That w ill requir e more complete experi mental data sets than the one used in this
study.
The results presented in this study
represen t our first attempt at wo rkin g w ith
th e WEPP model and collating the type of
measured data needed to param eteriz e and
evaluate the model. The runoff plot study
used in this analysis was certainly not ideal
for model testing as show n by the lack of
some on-site measurements needed to
parameterize the model. It did at least start
us on the road toward a better understandin g of model assumptions and data inpu t
Na turn l Resource Management in the Andes

References Cited

12

Duration of 61 mm rainfall event

408

needs. As part of an ongoing regional effort,

we are gathering expe rim ental data from
other hi storica l run off plot studies in th e
A ndes. W e w il l continue to ev aluate the
W EPP model with th ese data . As th e need
and opportun ity arise, we w ill aiso conduct
new exper iments using the WE PP mod el to
help us focus on improved understanding of
processes.

Favis -Mortlock, D.T., J.N. Quinton, and

W.T. Dickinson . 1996. The GCTE
va lida tion of soil erosion models for
global change stud ies. J. Soil Water
Conserv ation 51 :397-403.
Flanagan , D.C. and M.A. Near in g (eds.) .
1995. USDA-Water Erosion Prediction
Project: Technical documentation.
NSERL Report No. 10. National Soil
Erosion Research Laboratory, West
Lafayette, IN , USA.
Flanagan , D.C. and S.J . Li v ingston (eel. ).
1995. USDA-Water Erosion Prediction
Project: WEPP user summary. NSERL
Report No . 11. National Soil Erosion
Research Laboratory, West Lafa ye tte, IN ,
USA.
La Torre, B. 1985 . Efec to de sistemas de
cult ivos sobre la esco rrentia y perdida de
nutrientes en un enti so l de la Selva Alta CAP Jose Santo s Atahualpa - San Ramon
- Chanchamayo. MSc Thesis.
Un ivers idad Nacional Agra ria , La
Mo lina. Peru. 150 p.
Neari ng, M .A. 1998. Why soil erosion
models over-predict small soil losses and
under-predict large so il losses. Catena

32:15 -22.
Near in g, M.A., L.J. Lane, and V.L. Lopes.
1994. Mode lin g soi l eros ion. In : Lal, R.
(ed.). Soil erosion research methods.
Second edition. St. Luci e Press and Soil
and Water Conservation Society, St
Lu c ie, FL, USA. p 127-156.
Pa stor, R.P. , 1992. Evaluacion de la eros ion
hidr ica en la zona de Chanchamayo Junfn , Uti li zando como cobertura vegeta l
el cultivo de camote (lpomoea batatas
L.). Eng. Thesis. Universidad Nacional
Agraria, La Molina. Peru. 144 p.

Ecoregional Research: A Vision from the Andes

In response to growin g interest amo ng scientists and donors in natural resource
manageme nt and sustainable development, the CGIAR initiated ecoreg ional
research programs beginning in 1992 . Th e CGIAR defines the ecoregional approach
as:
"The main role of the ecoregional approach is to contribute to the goal of
increasing sustainability of agricultural production by providing: first, a
process that identifies the right research content due to its holistic and
forward looking approaches to research; second, a mechanism for partnership, among relevant actors with comp lementary fun ct ions, th at contributes to
achieving their common and individual institutional goals through app lied and
strategic research on the foundations of sustainab le production systems; and,
third, a mechanism that develops, tests, and supports effective research
paradigms for the sustainable improvement of product ivity." (Th e eco-regional
approach to research in the CIGAR. TAC, CGIAR, Rome Italy. 1993, p. 4 .)
Ecoregional Research Agenda in the Andes

CON DESAN, the Consortium for the Sustainable Development of the Andean
Ecoregion, ca rries out eco region al research in the Andes in the context of the
CGIAR approach. Operating under the umbrella of CIP, CON DESAN is a consortium of more than 75 assoc iated groups, including universities, N GOs, com muni ties, and loca l government agencies. CIP's natural resource man agement program
develops methodological tools for use in meeting the resea rch and development
objectives of CON DESAN whose ecoreg ional approach contains three objectives.
Sustainable natural resource management.
Improved rural in comes and emp loym ent.
Increased community input and co ntro l of the landscape.
To meet th ese objectives, energy is invested in both research programs and
ben chmark integrated research/development programs. The former are developing
mu ch needed tools for improved natural resource management and monitoring,
participatory methodologies, and productive technologies, whereas the latter are
working at the frontier of increasing agricultural productivity, resource use, and
community control. In al most all cases, the activities are inter-disc iplinary, multiinstitutional, and strive to develop a participatory agend a.
The resea rch agenda for the Andes is a work in progress, initiated with a se ri es of
Andean-wide project planning by objective (PPO) exerc ises. In the broadest terms,
the Consortium is focusing on four "cross-Andean " research components.
1. Soil and water management. Methodologies for the eva luation and rehabilitation
of natural resource -G IS, remote sensin g, ex-ante and ex -post watershed model ing, and promoting com munity management of natural resou rces.
2. Agrobiodiversity in Andean roots and tubers and pasture species. Identifying

micro-centers of genetic diversity; character izing clones; improvin g agronomic and

postharvest technologies; and understanding marketing pathways.

CIP Program Report 1997-98

409

3. Improved farming systems for the Andes: From producer to consumer

- Production systems research. M odeling (crop and anim al production, chemical leaching, soil erosion, and water run -off); resource management (soil biology,
drip/sprinkle irrigation, pesticides/ health); commodity research.
- Agroindustries and Marketing research. Deve lopment of new products;
identifying bottlenecks and new opportunities.
4. Policy research. Modeling trade-offs betw een producti v ity, profit, and en v ironmental impact; using government policy to promote Andean development; and
promot ing pr ivate investment in the rural Andes.

These research activities primaril y take place in the field at " research sites, "
where both the energy and funding originates from the research agenda. Several
examples are:
San Gabriel (Ecuador): Tradeoffs project.
Lake Titicaca fringe (Peru/ Bo li via): Remote sensing project.
San Antonio/Rio Recio/ Rio Guadalajara (Colombia): Ex-ante anal ysis of watershed management proj ect.
Pensil vania (Co lombia): Joining entrepreneurs and local and national governm ent
authorities in in vesting in the hillsides.
San Jose de Min as (Ecu ador), Sucse (Peru ), and San Juan de Miel (Bo livia):
Develop ing the local arracacha industry: methodologies for working w ith cottage
industries.
Two compan ion themes are part of the cross-Andean portfolio: developin g
human resources and comrriunications. In addition to conducting workshops, in the
case of human resources, CONDESAN is working to im prove postgraduate educat ion in agricultural production and natural resource management. With respect to
comm uni cation, INFOANDINA is the Consortium ' s information netw ork w ith nearly
500 participants. It is focusing on improving communications between Consortium
members, providing access to the "gray literature," and conducti ng forums on key
development issues.
Research and Development Agenda: Work at the Benchmark Sites

A crucial component of the ecoregional approach is "effective research parad igms"

and the concept of the be nchmark site. For us, a benchmark site is neither a pi lot,
" integrated rural development program," no r a poi nt along some heroic A ndean
transect, but rather, a place where susta inabl e development becomes operational. It
is here whe re the three goals of th e Andean Program come together: (1) sustainable
natural resource management; (2) rural poverty alleviatio n; and (3) community
empowerme nt. Each benchmark site represents a sma ll number of communities
that share a common resource base and have an interrelated set of management and
development challenges (usua ll y 10,000-100,000 ha). These benchmark sites have
been selected based on three criteria .

4l0

Naturnl Resamce Management in the Andes

1. The site selected represe nts a "typical" location within an important Andean
ecology (e.g., the green Andes, semi-arid inter-Andean valleys, the high plains). A
good database already exists at the site.
2. Local organizations are willing to collaborate to address an important productivity/natural resource management issue. This normally requires building a multidisciplinary and multi-institutional team, with a shared resea rch/development
agenda.
3. The benchmark sites are in different stages of development. Some have solidified
teams with a second round of funding; others are still forming and debating their
research agenda. Ideally, the team develops an integrated funding proposal,
funding from one or several sources are secured, and the research/development
activities begin. Over time, the "round table" of team members is strengthened
and new research and development activities are designed. Descriptive titles and
themes for the benchmark site projects are shown below.

Benchmark site projects.

Location

Themes

Comments

Pueblo Llano and Govindio,

Merida, Venezuela

Intensifying potato production

Motor for and menace to sustainable

development in Alto Merida

Lo Selva de Florencio, Monizoles,

Colombia

Managing the Florencia

bioreserve and buffer zone

Developing profitable alternatives

in the humid cloud forest

Rio El Angel, Carchi, Ecuador

Phase 1: Characterizing the watershed

with special emphasis on the
potato-pasture zone
Phase 2: Building a sense of
commonwealth in the watershed
and promoting rural prosperity

Managing a new irrigation

canal in Rio El Angel

Lo Encanada, Cajarmoca, Peru

Phase 1: Using GIS tools to help

characterize the watershed
Phase 2: Building a micro-watershed
management pion

From landscape interventions to

increasing incomes

Manozo, Puna, Peru

Intensifying production in a new

irrigation district

From water management to

marketing Andean products

Aroma, La Paz, Bolivia

Intensifying livestock production

in the semi-arid Andes

Landscape and risk management

Cochabamba, Bolivia

- Andean roots and tubers in

Candelario
- Cochabombina water forum

From prospecting to profit

Clarifying the debate

CIP Program Report 1997-98

4 11

The Future

Although not yet fully realized, the expectation is that the cross-Andean methodologies, research findings, and discussions will increase the effectiveness of the
benchmark teams, and that the benchmark successes will serve to focus the crossAndean themes. The future has a threefold thrust.
Research activities will result in new and user-fri endly tools for natural resource
management, better parti cipatory methodologies, and improved technologies.
The Consortium will grow stronger as the model of sh ared-risk, shared-ben efit,
multi-institutional teams has greater success.
The ability to have a real impact on sustainable rural development in the Andes
will increase as the synergism of the cross-Andean methodology and in-depth
benchmark site research activities take off.
Joshua Posner
CONDESAN Coordin ator

Acultivated mountainside near Cusco, Peru.

4l2

Natural Resource Management in the Andes

Research on
Andean Roots and Tubers
Impact on a Changing Wo rld

Reappraisal of Edible Canna as a High-Value Starch

Crop in Vietnam
M . Hermann\ N.K. Quynh 2, D. Peters 3

Ed ibl e ca nn a (Canna edulis Ker- Gaw ler) or

Qu ee nsland arrowroot is a vege tati ve ly
propagated root crop from th e And es.
Abund ant archaeologi ca l ca nn a re mai ns
found in th e Pac ific littora l in Ecuador and
Peru predate most oth er foo d cro ps in thi s
area (Piperno and Pears all, 199 8) . Th at
suggests cann a was a fairly common stapl e
in prehi stori c tim es, but was succe ssive ly
dimini shed by the introdu cti on of oth er
crop s, notab ly by maize fro m mesoAm eri ca and by Old World crops in th e
w ake of th e Spanish conqu est. In the
And ea n hi ghl and s, cann a co ntinu es to be
gro w n as a subsistenc e c rop or, to a lesse r
ex tent, as a cas h crop fo r occas ional ritu al
use. However, it is of no eco nomi c si gnifica nce and is largely unkn own in its native
range .
Produ ction advantages of ca nn a in c lud e
high N-u se effi c iency, hi gh y ield s and
harves t indi ces (Hermann et al. , 1997),
shading to lerance, outstandin g drought
to lera nce and w ater-use efficiency (Jurei t,
199 7), and th e absence of pest and di sease
and repl ant problems. A seve re co nstraint
to its wid er use is its lon g duration to full
produ ctivity.
Simil ar in tex ture and tas te to
sweetp otato (ipomoea b atatas), th e starchy
ca nn a rhi zo me is actu all y a palatab le fo od,
but it has an un attracti ve brow n co lor and
is rath er fib ro us. More import ant, ca nn a
must be boil ed or steam ed for several hou rs
to softe n ro o t ti ssue suffi ciently for it to be
co nsum ed. Th at constrain s its use not onl y

I CIP, Li ma, Peru.

2 Vie tn am Ag ri cultural Sci ence In stitut e, H anoi, Vie tna m.
3 CIP, H anoi, Vietn am .

in urban areas but also in poor, rural

highl and communiti es w here fu el wood is
scarce . Thu s, th ere is I ittle scope for ca nn a
to regai n so me of its p rev ious importance as
a food for direct co nsumption.
Yet th e ea se with which cann a starch ca n
be extra cted usin g makeshift equipm ent has
not esca ped th e attention of rural peo pl e in
th e sea rch fo r w hea t fl our sub stitutes.
Cann a has th e largest starch granul es
kno w n. Th ey qui ckl y settle out of a suspension of grated rhi zome tissue. In th e
north ern A nd es, co ttage industri es have
sprun g up to extract ca nn a starch, albeit o f
qu est ionabl e hyg iene and low produ cti v ity.
Al though th e starch is highly va lu ed loca lly,
it is limited to ce rtain bakery produ cts to
whi ch it co nfers a spon gy and li ght texture
not ac hi eved w ith alte rnati ve raw material s
(Herm ann , 199 4).
Cann a produ ction in the And es is
dwa rfe d by the crop' s use in tropi ca l Asia,
particul ar ly in Vietn am and Chin a. It sprea d
there in past ce nturi es presum abl y beca use
of th e orn amenta l va lu e of its flow ers.
Beca use of its hi ghl y localized importance ,
and generall y restri cted role as pi g fee d and
famin e fo od, its prese nce in bac kya rd
gard ens has rarely bee n noted. In rece nt
decades, howeve r, ca nna starch produ ction
in Vie tn am, alm os t exc lu sivel y for tran sparent noodl es, has beco me a maj or ag ri c u 1tural operati on. Thi s p aper exa min es the
prin c ipal fea tures o f ca nna use in Vietnam
ba sed on ra pid apprai sa ls in 199 2 and
1995, and resea rch fr om 1996 to 199 8. Th e
in fo rm at ion prov id ed fill s a gap in th e
agri cultural and foo d sc ience literature, and
identifi es pri oriti es for further resea rch and
bre edin g efforts.

CIP Program Reporl 1997-98

4 15

Canna production
Elderly rural in forma nts in var iably state
th at th e use of cann a as pi g feed has a long
trad iti on in Vietn am. In th e 1950s and
1960s, w hen ri ce harvests we re as low as 2
t/ha, cann a rh iz om es also ente red the
human di et to a co nsid erabl e deg ree. In
so me areas, ca nn a was prod uced t o a much
larger extent tha n toda y. At the tim e, so me
sta rch was ex tra cted and used rath er
un spec ifica ll y as a thi ckener. Beg innin g in
th e late 1960s and ear ly 1970s, the prin cipa l use of ca nna starc h fo r th e manu facture
of ce lloph ane noodl es emerge d and has
si nce fu eled the expans ion of ca nn a
produ cti on.
Cann a is gro w n in all mounta in ous
prov in ces and al so in som e low-ly in g area s.
Acco rdin g to offi cial estim ates, so me
20, 000 ha of cann a are cropp ed in northern
Vietnam. In the fi ve leadin g ca nna-prod ucin g di stri cts of Tanh H oa Prov in ce (Red
Ri ve r D elta) alone, 2,93 0 ha we re reg iste red in 1998 . Rece nt ev id ence suggests
th at from 4,000 to 5,000 ha are grow n in
the South, princip all y in the prov in ces of
Dong N ai, So ng Be, Tay Ninh , Gia Lai, and
Lam D ong. Other roots and tube rs used fo r
starch in Vietnam in cl ude cassa v a (Ma n i ho t
escu lenta) fo r a range of seco ndary processed pro du cts suc h as maltose and
ferm entati on produ cts, and ku dz u (Pu erari a
/obata), a tube rou s legu me y iel ding a hi ghly
pri zed sta rch us ed in a traditi onal soft drink.
In north ern Vietn am, cann a is cropped
pr in c ip all y in th e hi ghl and s (300-1 ,200 m)
of Hoa Binh and Son La prov in ce s, whi ch
bord er Laos and Ch in a, and in the Red
Ri ve r Delta at sea leve l. Croppi ng is str ictly
sea son al, beg innin g at the on set of rain s
and last in g fo r 10-1 2 mo. Propagation is by
api cal rhi zo me secti o ns th at are eith er
immed iately re pl anted or stored fo r 2- 3 mo
to allow fo r staggered harve stin g.
In th e rice-d om inated c rop pin g sys tems
of th e Red Ri ver D elta, cann a is mostl y
relegated to th e least fertil e and unirri gated
upl and s such as ri ve r dams, backy ard s, and

4l6

Andean Roots and Tubers

other areas that do not com pete with foo d

pro duction. Occas ionally, ca nn a is grown
on more fe rtil e land after a w inter vegetab le
crop. U nder high N fe rt il ization, y ield s ca n
exceed 80 t/ha in 8-9 mo. Usuall y, how ever, yields are und er 40 t/ ha and reflect
ma rginal in puts.
In the highl ands, most can na co mes from
ag roforestry systems or shi ftin g culti v ati on.
In ag ro fo restry systems, w hi ch may in vo lve
steep slopes, shadi ng ca n exceed 50 % and
the cro p is not fert il ized. No pl owi ng is
practiced. Produc t ion is th erefo re extensive
w ith no extern al in puts and low yie ld s.
In Vietnam, as in other parts of A sia,
edib le can na shows little varia bility an d
belo ngs to a broa d- leave d morph otype.
Full y ex panded leaves assu me a nea rhor izontal o ri entati on as opp osed to th e
erect leaf hab it of most A nd ea n cultiva rs.
A noth er spec ial feature of As ian starch
ca nn a is its hypogae ic rhi zo me co mp ared
w ith the more co mmo n rhiz ome pos iti on at
or above the so i l surface in A nd ea n materi al. Vietn amese ca n na is reported by
fa rmers to be see d- ster il e, an d ind eed we
neve r fo und sex ual seed in mature f ield s.
A ll six access ions exa min ed were fo und to
be triploid (2 n = 27), w hi ch suggests th ey
orig in ated in the northern A nd es, w here
suc h materia l is co mm on. Th at is corroborated by rand oml y ampli fied po lymo rphi c
DNA an alys is, w h ich pl ace s Viet namese
ca nn a next to tripl oi d Co lomb ian cul t iva rs
(He rmann , 1999).
Starch extraction
Some can na starc h is extracted in the
h ighl and s, but th e bu lk of th e harves t is
tru cked to proces sin g villages in the Red
Rive r Delta about 100 -200 km away.
Processin g of th e rhi zo mes into starch and
th en into nood les takes p lace in a nu mbe r
of hi ghl y spec ializ ed v illages . M any of th em
ad d domesti c rhi zomes to th ose imported
from th e high land s. So me v i IIages spec ialize in starch extra cti on, oth ers in no odl e
making, and st ill ot hers prov id e process in g
equ ipm ent such as dry in g rac ks, baskets,

graters, and noodle cutters. Several of th ese

functions may even coex ist in a given
vil lage.

soi I and sma l I sto nes they trap can damage

grater surfaces. Up to 20 person hr are
need ed to remove the roots from 1 t of
canna rhizomes.

The econom ics of can na production and

process in g are I ittl e understood, but
production in the low lands, i.e., near the
processing sites, and noodle-making seem
to be particularly profitable. Farmgate
prices for canna betwee n 1990 and 1995
were sl ightly higher than for cassava and
doubl e that of sweetpotato, alth ough
cassava and sweetpotato have much high er
dry matter (DM) contents. Canna rhizome,
starch, and noodle pric es fluctuate seasonally suggesting correspond in g profit
variation.

Vietnam ese sta rch processors use mostly

engin e-d riven drum graters. They are
locally made and consist of a hardwood
cylind er, 20-25 cm diam and 30-40 cm
long, with longitudina l rows of nails
providing an abrasive surface. Only
wea lthier households own such graters and
rent them out to other processors, mov in g
grater and engine around on wheelbarrows.
Figure 1 provides a schematic of a typical ,
fami ly-owned rural starch factory. Th e pulp
generated by the grater is co ll ected in
buckets and poured onto a cloth suspended
(imm ersed) in stat ion ary wa ter tanks . Th e
pulp is then moved across the cloth and
kneaded to release the starch (F igure 2),
which sett les to the bottom of th e tank. The

Starc h processors req uire the rhizom es to

be fr ee from the adventit ious roots that
firmly anc hor the canna rhizome in the soi l
and remain attac hed to it after harvest. Th e

Starch sto re
and
selling room

Farm
house

l _ _ - s helter

c:?
'

f - - - + -- i '

Washing
tank

Road

""l
Basin for
washing
canna
rhizomes

Worker

Grater
5m

Figure 1. Schematic of canna starch factory in Moc Chau, Vietnam.

CIP P1ogrom RepOll 199798

41

Figure 2. Rural women workers washing canna pulp in Moc Chau starch factory, Son La province, northern
Vietnam, 1995.
remaining fiber may be passed on to
workers at another tank (Figure 1) who
repeat this process to extract add ition al
starc h and the fiber is finally discharged.
The starch sediment accumulating in the
recta ngul ar tanks is shove ll ed into circular
was hing tanks and is given a number of
was hes by sti rrin g it in clean wate r. After
eac h was h, the starch is al lowed to settle
for an hour or so and a lighter, brown sl urry
on top of the firm starc h cake is scraped off.
Once supernatant water of th e final washin g is siphoned off, any free wate r rem ai nin g on th e sta rch sedi ment is soaked up
w ith hyg rosco pic mate rials . Bags filled w ith
as h, or bricks that have been dr ied over a
heat source and then coo led, are co mmon ly used . The final wa ter co ntent of the
sta rch as it co mes out of th e sed imentati on
tank is 46.5-49.0% (wet wt basis).
The farmer-extracted starch i 11 Vietn am
ranges from between 10 and 15% of fresh
rhi zo me we ight (Tabl e 1), w ith three out of
four observations being between 10 and
4l8

Andean Roats andTubers

11% . That is roughly eq ui va lent to half of

rhi zo me DM . Actual sta rch y ield s are about
three-fourths of total extractab le sta rch ,
w hi ch in c ludes starch released from
process resid ue by treatment in a food
blender. Extraction efficiency across
production zo nes is thus mu ch better than
in other tuber crop s (hav in g small er
granules w ith lowe r sed imentat ion rates)
but there is room for improvement.
Noteworthy are th e hi gh co ncen trati ons
of so lubl e so lids (shown to be sugars by
Ripperton , 1927), as revea led by the hi gh
refractom etri c index. This indi ca tes that
wastewate r must have high va lues of
biological oxyge n demand and illu strates its
pol lutin g potential. Table 1 does not suggest
differences between Vietnam and th e Andes
in ca nna DM co ntent, extraction effici ency,
or tota l extractable starc h, but no firm
co nclusion s ca n be drawn from th ese
preliminary data.
Trad iti o nall y, all sta rch has been dri ed in
the open. Otherw ise it quickly deteriorates

Table 1. Parameters of starch extraction from canna rhizomes in Vietnam and Andes (in % of rhizome fresh
matter unless otherwise specified).

Locality

Cultivar

Soluble

Dry matter

solids
0

Farmer
starchh

Laboratory

Extractable

Extraction

starch'

starchd

efficiency

Brix

(%)'
% rhizome fresh matter

Vietnam
Hai Hau, Nam Ha,
sea level, 2014' N
Hai Hau, Nam Ha,
sea level, 2014' N
Tu Ly, Hoa Binh, 300 m,
2053'N
Tu Ly, Hoa Binh, 300 m,
2053' N
Moc Chau, Son La,
1000 m, 2050' N
Moc Chou, Son La,
l 000 m, 2050' N

MH1400

5.0

MH1401
MH1402

22.5

10.l

3.8

13.9

73

10.8

4.3

15.1

72

21.0
6.2

MH1403

26.7
27.1

MH1170

4.5

24.7

10.9

6.3

17.2

63

MHll 71

4.9

22.5

14.5

4.4

18.9

77

MHIF1335

6.4 1

23.8

11.3

4.7

16.0

71

MHIF1344

9.1 1

22.l

9.0

2.0

11.0

82

MHll73

5.5

23.8

13.4

6.0

19.4

75

Andes
Briceno, Colombia,
2050 m, l 0 37' N
Las Delicios, Colombia,
1800 m, 157' N
Palate, Ecuador,
2350 m, 119' S

All starch contents are expressed as starch with 20% humidity (wet weight basis)
Starch extracted by former
' Starch extracted from process residue
d Extractable starch is the sum of former and laboratory starch and approximates the potential starch yield achievable with
better extraction equipment.
Ratio former/total starch
1 calculated by difference
0

due to fermentatio n. Since the 1990s,

however, most sta rch has been stored as
wet starch until sa le or furth er processi ng
(wh ich can be months ahead) by keeping it
in anaerob ic cond iti ons. To that end, the
sta rch cake is firmly pushed into polyethylene bags, containing an inner laye r of foil
and an outer layer of woven plastic. This
practice obviates the need for large areas

for sta rch drying and prevents ai rborn e

contam in ation. Table 2 gives retail prices
for ca rboh ydrate commod iti es on the Hanoi
markets and shows the hi gh relative value
of canna starc h.
In vill ages where hundreds of fami li es
engage in starch extraction, the disposal of
res idu e and wastewater poses serio us
CI PProgram Report 1997-98

4 19

Table 2. Commodity retail prices for carbohydrate products, Hanoi, December 1995.

Commodity

Comments

Conno noodles
Canna starch
Rice noodles
Cassava storch
Cassava flour
Wheat flour
lndico rice
Japonico rice
White sugar
Brown sugar
Poto toes

From Huu Hoo

From Duong Lieu
From Huu Hoo
From Duong Lieu
'From the mountains'
From Australia
Sticky rice

Price (dong/kg)
7500-8000
7500
6000
3500
2500
6000
4000
8000
6500-6800
5500
2500-4000

US$ l = Dong l 0,800.

problems. Wastewater, w hich is rich in

o rganic mate rial , especially suga rs, is
discharged wi thout treatment into streams
and bad odo rs emanate from fermenting
extraction res idue.
Although extraction residu e is used in
var ious ways, its abundance makes it more
of a liability than an asset. Some of th e
fiber-rich res idue is returned to fields as
mulch or is fed to pigs, either plain or
mixed w ith rice bran. During peak processing, more residue is produced than can be
co nsum ed as feed and needs to be dri ed . It
is spread out on all ava ilabl e surfaces in the
vi ll ages, including wa lls and sid ewa lks.
Larger processing units have also been
known to ferment res idue, a process th at
mi ght add to its nutritional value. The li ght,
brown slurry scraped off settled sta rch
makes a good feed , espec iall y if cooked
with other co mpon ents.

Noodle making
A small proportion of canna sta rch is
destined for unspecifi c househ o ld uses or
processed into min or products such as
ca ndies, cakes, and rice papers . Overw helmin gly, howeve r, canna starc h is
processed into noodl es.

420

Andean Raats and Tubers

Of th e two techniques for making canna

nood les in Vietnam , traditional extrusi on is
the older, but is practiced on a small scale.
It involves forcing a freshly ge lat ini zed and
hi ghly viscous starch paste throu gh the
perforations of a meta l plate fo rming the
bottom of an open-topped box. Th e
noodles are directly released o nto a dryin g
rack. As w ith other extrusion nood les, th ey
are round , but are unacceptably thick (1.82.0 mm di am) to most co nsum ers . That is
because of the large di e diameter chosen to
keep the extru sio n force suffi cient ly low for
manual operation.
Steam-sheeting, th e technique used for
wel l over 95% of ca nn a noodl e production ,
is appa rentl y derived from rice nood le
mak ing for which a similar process is used.
First, wet starc h is slurri ed into a so lution of
full y gelatinized starch, which acts as a
matrix keep in g th e starch granules in
suspension and accounts for 5-10% of the
sta rch in th e resultant batter. Thi s batter is
then applied to a steam-heated textile
membrane (F igure 3) an d gelatinizes into
tra nslucent ge l sheets. These are then
transferred o nto concave bamboo racks and
are stretch ed by a factor of 2. 5 into a
recta ngul ar shape (F igure 4). After preliminary dryin g to 55 -58% moi sture, th e sheets
assume a rubbery, nonsticky texture and

Figure 3. Canna noodle production in Huu Hoa village, Hanoi, Thantri District, 1995. Man applying starch batter

to steam-heated textile membrane.

ca n be eas il y handl ed fo r subseq uent
cuttin g into noodl es. Two perso ns ca n
produce 1 50 sheets/day or th e equ iva lent of
160 kg of dry nood les (retail va lu e
US $1 10.00).

Figure 4. Woman stretching hot gel to the limits of

Typ ica ll y, a guill ot in e is used to cut th e

nood les (Fi gure 5) . It moves a stac k of
fo lded gel sheets in crementall y towa rd a
verti ca ll y mov in g kn ife that cuts off the
nood les . This mac hin e, ori gin al ly deve loped fo r manua l operat ion, has been
se mi automated sin ce 1993. A wa tchlike
mec hani sm attac hed to th e cutter and
powe red by an electr ic moto r moves th e gel
stack and th e cuttin g knife in a coo rdin ated
fash ion. Automated cuttin g has grea tly
in creased productiv ity but requires relati ve ly hi gh in ves tm ent and thu s has not
comp letely replaced manu al operati on. Th e
ma ch ine's cap ac ity is 400 kg of nood les/
cla y per ope rator co mp ared w ith 100 kg/clay
per ope1a tor w ith manua l ope ration. Drying
of the nood les to 18-21 % fin al mo isture
takes place in a co nfined space (F igure 6).

a bamboo rack.
(IP Progrom Report 1997-98

42 1

Canna noodl es be long to the class of

cel loph ane or glass noodles w ith a glossy,
transparent appea rance. They are rectangu lar and are 0.8 mm thick. Bland and
slippery, they add texture to a var iety of
Vietn amese di shes, es pecially to soups and
stir-fri ed di shes. They are regarded as a
lu xury item and rese rved for cel eb rati ons.
Gl ass nood les are traditionally produ ced
from mun gbean (Vigna radiata ) starch,
w hi ch co ntinues to be a superior raw
material (Lii and Chang, 1981 ). H owever,
Vietn amese ca nn a noodles exh ibit sufficient ly hi gh tensi le strength to sat isfy eve n
dem andin g co nsum ers . Th ey stay firm and
lose a minimum of so lids during cooking.
Being less wasteful and less cost ly to
extract, ca nna starch has totally repl aced
mun gbea n sta rch as the raw materi al for
glass nood le production.

Figure 5. Semiautomated noodle cutter.

Figure 6. Final drying of noodles on roofs.

4 22 Andean Roots and Tubers

Attempts to substitute canna starch w ith

less costl y sweetpotato sta rc h have had
mi xed success beca use noodles made from
pure sweetpotato starch are unacce ptabl e
to consumers. A proportion of 70% ca nn a

starch mixed with sweetpotato sta rch is

dee med appropriate to retain the as pect
and consistency of canna noodl es (Lan and
Huy, 1999).

increas in g extra ct ion efficiency,

increa sin g rhizo111e starch content an d
starch yield, and
increasing labor productivity of starch
extraction.

Discussion and Conclusions

Declining direct consumption of canna in
Vietna111 and in the Andes is contrary to the
occasionally stated view that ca nna is a
promising staple for the poor in 111ontane
tropics. Future research and development
should focus on canna as a source for
starch with functional properties that make
it a possible substitute for mungbea n starch
in glass noodle production, not only in
Vi etnam but also in other Asian countries.
Th ere is mounting evidence for canna
starch also being functionally si111ilar to
potato starch in shared end products, high
peak viscosity, high amylase content, large
granule size, crystallinity, and a111ount of
covalently bound phosphorus (Hermann,
1994; Hulleman, 1995; lnatsu et al., 1983;
Tu and Tscheuschner, 1981 ).
Canna and commO'n commercial
starches, however, are currently un equal
competitors. For example, potato starch is
being developed aggressively for a myriad
of uses and novel products. In contrast,
canna is scientifically neglec ted. Advanced
technologies for its production and processing are unknown and the cultivars used
have not bee n bred for specific uses .
The price competition from other
starches, mainly from cassava and
sweetpotato, will restrict canna starch use
in the medium term to very specific applications. At current price s, it seems unlikely
that new 111arkets will open up for canna
starch. Efforts should therefo re be undertaken to reduce production costs. The
Vietnamese retail price of dry canna starch
was the equivalent of US$0.70/kg in
D ece mber 1995. At that pric e, Eu rope an
high-grade potato starch 111ay well be a
threat to the Vietnamese canna starch
industry. Opportunities for redu c in g
production cost include:

Current ext ract ion efficiency ranges from

63 to 77% of total extractable starch.
I 111prove d grat in g surfaces are Ii kely to
succeed in releas in g 111ore starch fro111 th e
fibrous rhizom e tissue.
The incre ase of starch content will
reduce shipping costs per unit OM and
potentially increase starch yields. That will
require vari etal selection or breeding using
Andean material, which is sexually fertil e
and has greater variability than Asian
cultivars (Her111ann et al., 1997). Breeding
would also ha ve to address the high
residual sugar co ntent, which exists at the
expense of starc h yields and incurs environmental costs as incre.ased organic load in
wastewater. Further research is also needed
to identify germplas111 with improved
processing characteristics (extractable
starch, oxidative browning).
Increased labor productivity, and
probably hi gher starch quality, could be
achieved by using rotary strainers to
separate starch milk fro111 fibrous residu e.
That wo uld rep lac e the onerous task of
manual batchwise separation. Such intermedi ate technology could be produced
dom estically using blueprints from existing
equipment. Stationary paddle washers,
which were used to clean canna rhizom es
earlier this century in Australia, could also
help to red uce processing costs. Alternative
technology includes rotating dru111 washers
currently used in s111all cassava operations
in South A111erica. Such equip111ent would
obviate the need for the very labor-inten sive
re111o va l of adventitious roots .
Other than in wheat-derived pasta with
its nonstarch components (especially
gluten), the qu ality of starch noodles
depends solely on the functional propert ies

CIP Progrom Report 1997-98

423

of the starch itself (Pagani, 1986). Hot

canna gels have minimal adhesiveness and
outstandingly high elasticity. That, in
combination with the producti vity of steamsheeting, explains the wide adoption of this
method in Vietnam and its apparent
superiority over manual extrusion methods,
which are so w idely used in the manufacture of sweetpotato noodles in China. Also,
hot canna gels are translucent. They do not
need to be immersed in water or subjected
to other cooling treatments following
extrusion to produce the vitreous aspect
and desired consistency required for other
raw materials.
An assessment of the potential of edible
canna , if conducted a few decades ago,
would have concluded it was no more than
an ethnobotanical curiosity, hardly worthy
of promotion and conservation. Howeve r,
canna use in Vi etnam shows how product
development can prov ide novel perspectives for the use of a seemingly obsolete
crop. It demonstrates the unpredictability of
future ge rmplasm use and illustrates the
need to conserve germplasm wi th no
apparent value. Current use patterns of
minor crops should not be taken for
granted; they may well change over time.
References Cited

Hermann, M. 1994. Achira and arracacha:

Processing and product development.
CIP Circular 20(3):10-12.
Hermann, M. 1999. Characterization and
classification of Andean root and tuber
crop variability. Subproject Annual
Progress Report. CIP, Lima , Peru . 14 p.
Hermann, M., R. Uptmoor, I. Freire, and J.L.
Montalvo. 1997. Crop growth and starch
productivity of edible canna. In: CIP
Program Report 1995-1996. CIP, Lima,
Peru , p. 295-301.
Hulleman, S.H.D. 1995. X-Ray diffraction
on va rious Andean root and tuber

424

Andean Roats and Tubers

starches. ATO-DLO, Wageningen,

Netherlands. Mimeograph. 10 p.
lnatsu, 0., I. Maeda, N. Jimi, K. Takahashi ,
H. Taniguchi, M. Kawabata, and M.
Nakamura. 1983. Edible canna starch.
1: Some properties of edible Canna
starch produced in Tai wa n. J. Japanese
Soc. of Starch Sci. 30(1 ):38-47.
Jureit, C. 1997. Effekte von W asserangebot
und N-Dungung auf Ertrag und
Wassernutzungseffizienz von Canna
edulis. Diplomarbeit, Universitat Kiel,
German y. 84 p.
Lan, D.T. and P. Huy. 1999. Improved
production of canna and enhanced
utilisatio n of canna and sweetpotato
starch for noodle making in Vietnam in
Than Hoa Pro v ince, 1996-1998. Final
Project Report, CIP-UPWARD-CIATPHTI , CIP, Lima , Peru. 36 p.
Lii , C.Y. and S.M. Chang. 1981. Characterization of red bean (Phaseolus radiatus,
var. Aurea) starch and its noodle quality.
J. Food Sci. 46:78-81.
Pagani, M.A. 1986. Pasta products from
non conventional raw materials. In:
Mercier, Ch. and C. Cantarelli (eds .).
Pasta and extrusion cooked foods: Some
technological and nutritional aspects.
Elsevier Applied Science Publishers,
Barking, England. 199 p.
Piperno, D.R. and D.M. Pearsall. 1998. The
origins of agriculture in the lowland
neotropics. Academic Press, San Diego,
CA, USA. p. 244-256, 267-280.
Ripperton , J.C. 1927. Carbohydrate metabolism and its relation to growth in the
edible canna. Hawaii Agricultural
Experiment Statio n Bulletin No. 56. U.S.
Government Printing Office, Washington, D .C., USA. 35 p.
Tu , L.N. and H.D. Tscheuschner. 1981 .
Untersuchung von wichtigen
Eigenschaften der Dong-Rieng-Starke.
Lebensm ittel i ndustrie (Leipzig, V EB
Fachbuchverlag) 28(11 ):515-516.

Compositional Diversity of the Yacon Storage Root

M. Hermann 1, I. Freire 2 , C. Pazos 3

Yacon is a little-known , nonstarchy Andean

root crop, which is eaten raw and functions
as fruit in traditional food systems. It
belongs to the Asteraceae (sunflower
family) and is vegetatively propagated.
Botanically, th e crop has been referred to
until recently as Polymnia sonchifolia
Poepp. & Endl., but the binomial
Smallanthus sonchifolius (Poepp. & End I.)
H. Robinson is gaining acceptance among
taxonomi sts (G rau and Rea, 1997).
Yacon is a very produ ct ive crop with
root dry matte r (OM) yields in soils of
moderate fertility exceeding 10 t/ha in 6-8
mo. The dark-skinned roots vary from
spherical to oblong and weigh from 1 00 g
to 1 kg . Th e concentration of a yellow
orange pigm ent responsible for yacon's
light flesh color varies by ge notype. Defining organoleptic attributes of the yacon root
are a succulent, tend er cr unchiness, w hich
approaches that of a watery radish or apple,
and a mildly resinous but pleasantly sweet
taste.
Smallholders in th e Andes cultivate
yacon fairly commonly for subsistence.
Rarely, however, do th e roots reach rural
fairs, and reliable produ ct ion estimates for
the crop are not available. Typicall y, only a
few plants or rows are c ulti va ted in field
corners or backyard gardens and, through
piecemeal harvest, provide a continuous
supply year-round. Th e lack of urban
demand for this root is poorly understood,
but constraints may includ e the root's short
shelf life of a few days and a lack of
consumer fam i I iarity.

Th e bulk of yacon OM has previously

been show n to consist of free sugars and
fructans of low polym erization, i. e., fructooligosaccharides (FOS) (Ohyama et al.,
1990, Wei et al., 1991 ). FOS consist of
short chains of fructos e units linked by
(2~ 1) G-glucosidic bonds. They carry a
single D-glucosyl unit at the non -reducing
end of the chain (1 ~2)-cx as in sucrose.
Th e nutritional significance of the sweettastin g FOS is that the human small intestin e has no enzyme to hydrolyze the
glucosidic bonds. Therefore, FOS are
considered indigestible and serve as
dietetic sweeteners. H ea lth benefits are also
claimed for fructan s. Th ey ha ve been
shown to stimulate the growth of
bifidobacteria in the human colon, to
suppress putrefactive pathogens, and to
reduc e serum cholesterol concentrations.
The y are thus increas ingly added to pastry,
confect ionery, and dairy products
(Campb ell et al., 1997) .
Th ere is limited information on the
chemical composition of yacon and it
mostly co mes from one cultivar, presumably Ecuadorian, grown under tem perate
conditions in Japan. The object of this study
was to determine the chemical composition
of th e yaco n root and its variation in a
germplasm sample taken from throughout
the crop's geographic range in a tropical
highland environment. Yacon chemical
composition in relation to product development and nutrient remova l by harves ted
roots is examined. This study intends to
contribute to the in c rea sed use of the
biodiversity of this neglected Andean crop
in its nat ive range .

1 CI P, Lima, Peru.
2 Universidad Ca t61 ica, Quito, Ecuador.
3 Nestle Resea rch and Development Center, Quito, Ecuador.

CIP Prngram Report 199798

425

Materials and Methods

hilled up. Plots we re weeded at 2 and 4 mo

after planting. Furrow irrigation was
provided as needed.

Field cultivation
Ten yacon accessions, for wh ich attributes are gi ven in Table 1, we re grown in
natural soil in an open-sided, insect-proof,
quarantine greenhouse in the Tumbaco
Valley near Quito, Ecuador, under thermic
and light conditions close to the surrounding equatorial environment at 2,500 m
altitude. Monthly mean temperatures varied
from 14.6C to 16.0C, with average daily
minimums ranging from 4 to 8C and
maximums from 23 to 29C. Monthly
sunshine was 110-230 h. Before planting,
samples taken from the sandy soi I had a pH
of 7.2-7.5. Nitrogen content of the soil
ranged from 22 to 30 g/ml soil (low), P
content 80-200 g/ml (high), and K content
0.29-0.68 meq/100 ml soil (medium to
high ). Three to four plots per accession
were used in a completely randomized
design .

When all accessions had become

se nescent (8 mo after planting), 2 plants/
plot were harvested, yielding 20-50 roots
w ith root fresh w t of 6-1 0 kg. Harvested
roots were speedily delivered to the
laboratory, where they were peeled and
chopped into 1-1.5 cm cubes. The cubes
we re thoroughly mixed and a sample of 1-2
kg was taken for subsequent analysis. Onethird of the sample was used to extract juice
w ith a kitchen extractor for determining the
refractometric index (0 Brix). The remaining
root cubes were immediately frozen and
subsequently freeze-dried. Dried samples
w ere ground to the consistency of flour and
stored at -80C until anal ysis. All va lues
obtained for freeze-dried material we re
corrected to a fre sh wt basis.

Pre-rooted apical cuttings we re planted

on 12 Aug 1995, at a spacing of 1 m x 0.7 m.
Mineral fertilizers at the rate of 50 kg N/ ha
and 100 kg K/ ha were applied 4 wk after
planting, at w hich time the plants were

Moisture content (MC) was determined

for fresh and freeze-dried root material by
vacuum oven drying for 24 h at 70C, fat by
Soxhlet extraction using petroleum ethe r as
solvent, protein as Kjeldahl N x 6.25, and

Chemical composition determination

Table 1. Geographical attributes and mitotic chromosome number of yacon clones used in this study.

Accession'

Year

Country

Province

Locality

Altitude (m)

Latitude

Longitude Mitotic

collected

chromasames

ASL 136
AW 5075
ARB5027
ARB5073
ARB5074
ECUl 243
HN1013
MHG919
MHG923
MHG927
0

1992
1991
1991
1991
1991
1984
1992
1991
1991
1991

Peru
Peru
Peru
Peru
Peru
Ecuador
Argentina
Bolivia
Bolivia
Bolivia

Cajamorco
Cajamorco
Lima
Cajamorca
Cajamorca
Az.uay
Jujuy
Cochabamba
Cochabamba
Cochabamba

Chota
Cancan
Tintin
Sucre
Sucre
Cum be
Barcena
Pairumani
Parocti
Locotal

2,900
2,600
3,200
2,600
2,600
2,560
l,900
2,600
2,100
l,800

0633'S
071l'S
l220'S
0656'S
0656'S
03lO'S
2358'5
l 72 5'S
l 7lO'S
l 7lO'S

7838'W
7819'W
7547'W
7808'W
7808'W
7809'W
6526'W
6620'W
65SS'W
6545'W

87
58
58
58
58
58
58
58
58
58

Accession ECU 12 43 is maintained by lnstituto Nacional de lnvestigaciones Agropecuarias, Quito, Ecuador; all other accessions are
kept in the CIP genebank.

4 2 6 Andean Roots and Tubers

as h (min eral s) by incin era t io n at 550C for

4 h. Potassi um and Ca we re determ in ed by
fl ame spectrophotometr y, P spectrop hotometri ca ll y (mo lybdovanadate method).
Cru de fibe r was determin ed o n defatted
sa 111pl es u sing the Fibertec sys tem (Teca tor
A.B., H bge nas, Sweden). W e ca lcul ated
tota l carbohydrates as total OM less prote in ,
fat, an d as h.
Free sucrose, glu cose, and fru ctose were
ex trac ted wit h water at 70C fo r 30 min and
dete rmin ed w ith a hi gh performa nce ani o n
exc hange c hrom atograp h (HPAEC).
Fructa ns were determin ed indirect ly by a
nove l method adapted from H oebregs
(1997). Thi s method reli es o n th e ac id
hydro lys is o f the sampl e, fo ll owed by
HPAEC determination of the re leased suga rs
as was done for free sugars.
H ydro lysis depo lyme rizes fructans and
sucrose into its compon ent suga rs , glu cose
and fr uctose. By determinin g th e glu cose
and fructose co ntent before hydro lysis
(Gfree' Frr 0 ) and after hyd ro lys is (G 101.11 , F10 " ) 1
and cons id ering the lo ss of w ater in the
hydro lys is of sucrose, w e ca n ca lc ul ate
glu cose and fructo se released from fr uctans
as

err - S/ l 9 and

Gr,ur1ans = G,01a1-

w here G = glucose, F = fru ctose, and S =

suc ro se. The average deg ree of po lyme ri za tion (DP) of fructan s is then

0p

= Ffr uctans /

Gfruclans

c (F

Fructa ns

Results
Chemi ca l co 111pos ition of yacon relati ve to
root fresh matter (FM) is shown in Tabl e 2.
Compo site factors relative to root OM are
shown in Tab le 3. For all 10 access ions low
val ues and nar row ranges of OM (98- 136 g/
kg) and carbo hydrate co ntent (89 -127 g/ kg)
we re found, es pec iall y w hen accession
AW 5075 was exc lud ed. Carbohydrates
accounted fo r 9 1-9 4% of OM. Discount in g
the outlyin g access io n AW5075, th e
access ions also had narrow ran ges for
fructan s (50-89 g/kg FM, 52 -6 6% OM ) and
total free suga rs (18 -3 1 g/ kg FM, 14- 29 %
OM). By con tras t, AW5075 had a mu c h
redu ced fructan co nte nt (3 1 g/ kg FM , 32 %
OM) and co rrespo ndin gly high free suga r
co ntent (42 g/ kg FM, 43 % OM). Thi s
accession m atured precoc iousl y and, whe n
harvested, had started develo pin g new
sprouts.

+l

and the quantity of fru cta ns original ly

present in the sa mpl e beco111es

Fru eta n s =

Th e ca lor ic val ue per 100 g ed ibl e

po rtio n was calc u lated usin g a fo rmul a
adapted fro 111 Merrill and Watt (1973),
nam ely: kca l = (3.36 x % protein ) + (3.60 x
% [total ca rb o hyd rates - fructansJ) + (8.37 x
% fa t). Thi s formu la ass umes that no
fructans a1e m etaboliz ed in human s. A
handh eld Atago refrac tometer was us ed to
deter111in e 0 Brix at l 8C. Duncan 's 111ultipl e
ran ge test and Pea rso n cor re latio n coeffi c ie nts were calc ul ated usin g SAS pro cedure s G LM and CORR, res pective ly.
Prin cipa l co111po nent analys is was don e
w ith NTSYS-pc, v. 1.80 (SAS, 19 89; Rohlf,
199 2).

G fruc la ns )

where C is a constant correct in g fo r

water ga in ed for each glu cos id ic bond
during polymer ization:
C = [180 + 1 62 (DP - 1)] / 180 DP.

Access io n ASL 136 had the hi ghest

va lues for OM, to tal carbo hydrates,
fru ctans, and "Br ix. This access io n from
Peru is also di st inguished by its
dodecaploid nature (9A + 3B = 2n = 87; A
= 7, B = 8) from th e other m ateri al (Tab le 1 ),
w hi ch has bee n shown by Sal gado (1996)
to be octop loid (6A + 2B = 2n = 58). In a
rec ent fi eld tria l of 24 Peruvian access ions,
ASL 136 rank ed among the high est in OM
content and y ield (C. Arbizu, CIP, Lim a,
Peru, 1999, pers. comm.) thus und erscoring

CI PProgram Repori 1997-98

42 7

v;

or

O:J

14

3.6-4.3

18-42

2.3-5.9

3.9-21.1

l 0-19

DP

Total free sugars (g)

Free glucose (g)

Free fructose (g)

Free sucrose (g)

3.1-4.1

112-464

148-224

Fiber (g)

Fat (mg)

Energy (kcal)

1,843-2,946

Potassium (mg)

2,282

240

87

5,027

174

244

3.6

15

17

25

10

12

43

19

12

18

58

32

27

23

(%)

C.V.

2,859 ab

291 abc

68 def

5,630 ab

148 c

191 cde

3.6 abc

3.3 de

2.64 b

12.6 a

12 cd

1,969 c

197 d

84 bede

4,881 abc

224 a

289 be

3.5 bed

3.5 cde

2,267 be

240 abed

94 be

5,071 abc

170 be

171 de

3.3 cd

4.7 ab

2.36 c

10.6 abc

9.0 c
2.84 ab

15 be

19 a

4.2 a
24 bed

3.6 de
31 b

103 b
204 d

92 bed
245 abed

1,843 c

4,357 be

4,931 abe

1, 999 c

197 ab

177 be

2,361 abc

302 ab

131 a

5,369 abc

168 be

2,327 abc

232 bed

76 edef

5,015 abc

165 be

224 cd

309 a
2, 946 a

182 d

2,065 c

2,382 abc

4,944 abc
56 f

156 e .

112 e

101 b

194 cde

118 de

331 b

234 bed
464 a

4.0 a

61 ef

3.6 abe

3.1 d
3.4 bed

4.1 a

3.8 ab

3.7 cde

6,014 a

2.7 e

2.7 e

3.8 bed

4.3 abc

3.7 cde

2.37 c

4,275 c

2.34 e

2.35 c
2.86 a

1. 95 d

11.1 abc

155 c

11.8 ab

10.8 abc
10.2 be

l 0.7 abc

9.9 be

11 d

4.3e

163 c

12 bed

13 bed
13 bed
10 d

16 ab

2.14 cd

3.9c

5.6c

9.3 be

6.6 be

2.8 d

3.3 ed
3.0 d

11.4 b

18d

3.9 be
20cd

4.0 b

3.9 be

72 b

112 be

120 be

MHG927

22cd

74 b

68 b

4.0 be

24 bed

3.7 cde

58 be

114 b

123 b

118 be
111 be

MHG923

MHG919

2.4 d

5.9 a

59 be

100 cd

109 cd

111 be
102 bed

HN1013

ECUl 243

50 c

100 cd

109 ed

ARB5074

Accessions'

311 b

3.7 abc

4.9 a

2.07 d

10.6 abe

14 bed

7.5 be

9.4 be

21.l a

4.5 b

2.8 d

2.3 d

2.8 d
4.6 c

26 bed

27 be

42 a

3.6 e

61 be

19 cd

62 be

31 d

105 be

115 be

ARB5073

3.8 bed

104 be

114 be

ARB5027

89 d

98 d

AW5075

4.3 a

4.2 a

89 a

127 a

136 a

ASL 136

Means followed by a common letter are not significantly different at P < 0.05 by DMRT.

182-309

Phosphorus (mg)

56-131

Calcium (mg)

4,275-6,014

2.7-4.9

Protein (g)

Ash (mg)

2.38

l. 95-2.86

F,/Gro, ratio

3.7

10.7

9.0-12.6

3.4

26

3.9

62

8rix

8.5

31 -89

Fructons (g)

106

89-127

Toto I co rbohyd rates (g)

115

Mean

98-136

Range

Dry matter (g)

Variable

Table 2. Chemical composition of 10 yocon accessions per 1 kg of root fresh matter.

Table 3 . Carbohydrate composition of 10 yacon accessions relative to root dry matter(%).

Accession

Total
Carbohydrates

Fructans

93
91

66
32

43

ARB5027

91

54

24

ARB5073
ARB5074

91

53
46

22
29

52
53

23
22
19

ASLl 36
AW5075

ECUl 243
HNlOl 3

92
91

MHG919
MHG923

92
94
93

58
60

MHG927

93

60

Total free
sugars

Free
glucose

Free
fructose

Free
sucrose

14

2.0
2.3

3.4
21.6

8.5
19.5

2.5
3.9

8.3
6.4

5.4

16

2.1
3.8
2.6
2.7

8.7
11.l
6.1
4.9
3.2

13.3
12.0
14.7
9.4
12.l
11.3
10.l

15

2.3

3.6

9.1

its production potential. At a root produ ction of 50 t/ha, which underestimates yi eld
potential of yacon under reasonab le soil
fertility, ASL 136 wou ld have yielded 4.5 t
fru ctans/ ha. From th e sa me root producti on,
fru ctose and sucrose tota lin g 5.8 t/ha could
have been obta in ed throug h hydro lysis of
fru ctans and from free sugars (calculat ions
not shown).
Free glu cose and fru ctose were among
th e most variable root parameters; wi th C.V.
32% for fre e glucose and C.V. 58% for free
fru ctose. Th ere was a sign ifi cant and hi ghly
negative correlation between fru ctans and
free fructose (-0.88, Tabl e 4) indi ca ting th e
interrelati on of these metabolites in depolymerization. Interestin gly, free fructose
was positively correlated to DP (0.42),
whi ch suggests th at polymer elongation
in creased w ith the size of the fructose pool.
Conve rse ly, high glu cose conce ntrations
are associated with lowe r DPs (r = -0.68),
perh aps beca use th ey in crease the number
of fructan mol ecules competing for free
fru ctose. Although th ere were signific ant
differences of DP betwee n access ions, th e
range was narrow 3.6-4.3 (C.V. = 6%). The
ratio of total fructose to glu cose (F 1j G 101 )
indi cates th at in a syrup obtained after ac id
hydrolysis of yacon root ca rbohydrates,
fru ctose would be present at a concentra-

tion 2-3 times greater than glu cose (Tabl e

2).
As shown in Table 4, 0 Brix was highly
and positively co rrelated w ith fru ctan
content (r = 0.84) and DM (r = 0.86), which
su ggests that refractom etric meas urements
provide convenient and quick assessments
of these important variab les. Free fructose,
su cro se, and total free sugars are noted for
th eir inverse co rrelation with th e refractometri c index as wel I as fructan s. Th at
impli es access ions hi gh in fructans are low
in free sugars. Noncarbohydrate co mpounds are om itted from Tabl e 4, sin ce
they were on ly weak ly corre lated amo ng
th emse lves and with ca rbohydrate
vari ables.
As Table 2 shows, yacon is a poor source
of protein (2.7-4.9 g/kg FM). It is also low in
lipi ds (1 12-464 mg/kg FM), but has mod erate leve ls of fiber (3. 1 -4 .1 g/kg FM). It is a
good so urce o f K (1. 8-2.9 g/kg FM), a littl e
less than half of its total mineral content.
Food energy ranged from 148 to 224 kcal/
kg FM and is several times lower than for
comparable foods.
U sin g the va riabl es shown in Tab le 2, a
prin cipal component ana lys is (PCA) was
computed. Th e first compo nent ex plained
54% of total va riation ; the second accounted for 20%. PCA did not revea l any

CIPProgramReport 1997-98

4 29

Table 4 . Pearson correlation coefficients' of chemical variables of yacon root fresh matter.
Dry

Free

Free

Free

Total

molter

glucose

fructose

sucrose

sugars

Variables

Free glucose

-.13

Free frucose

-.77 a

-.16

Free sucrose

-.58 a

.13

Total sugars

-.76 a

.08

.94 a

.91 a

.95 a

-.14

-.88 a

-.75 a

-.90 a

.86 a

-.20

-.69 a

-.57 a

-.71 a

.Fructans
Brix

-.04

DP

-.68 a

.42 a

.03

Principal component 2
1.0

D ARB5074
D ARB5073

1.0

MHG919 ..to. HN 1013

1.5

0 MH927

0.0

OMHG923

I ARB5027
D

ASL136

-1 .6

-1.6

-.07

l
.01

DAW5 075

Discussion and Conclusions

Th e yaco n accessio ns used in thi s study
we re co ll ected betw een 1984 and 1992
durin g multicrop missi o ns or as c h ance
co ll ectio n s (Tabl e 1 ). Thus they were not
th e res ult of a systematic col lection effort
aim ed at capturin g m ax imum yacon
di ve rsity. Northern Peru , w ith four accessions, and Cochabamba Departm ent,
Boli v ia, w ith three access ions, are likely to
b e over-represented in th e sa mpl e. Th e
co ll ection sites, howeve r, desc rib e an area
th at is co ngruent w ith the distributi o n of th e
species, w hich stretches from Ec u ado r to
north ern A rge ntin a. Al so, the accessions
come from w ide ly differin g eco logies and a
co nside rab le altitudin al range. Th erefore,
sa mpl es do encompa ss a reason ab ly large
part of yaco n diversity, permittin g co nclusio ns ab o ut compositional diversity from
th e d ata obta in ed .

* ECU12 43
-1.6

-1 .6

Principal component 1
.A. Arg entina O Bolivia * Ecu ad or 0 Northen Peru I Central Peru

Figure 1. Principal component analysis on

compositional data of l 0 yacon accessions.

Andean Roats and Tubers

.20

.84 a

P< 0.01.

geographica ll y defin ed clusters of accession s (Fi gure 1 ). That suggests yacon

chemi ca l di ve rsity is not p artition ed in to
geographi ca l subgroups, which wou ld
allow the searc h for specia l c haracteristics
in germplasn, from certa in area s. Carbohydrate va ri abl es were hi ghly correlated w ith
th e first and second prin c ipal components
(d ata not show n). Thus, th ese va ri abl es
contribute more th an others in differentiatin g yaco n c lones by c hemi ca l ch aracter istics.

430

DPb

.75 a

' Computed across repetitions: n = 36; a= significant at

h Degree of polymerization.

-1 .0
-1 .6

Brix

Dry matter

Fructons

Althou gh claim s for hi gh inulin co ntent

in yaco n co ntinu e to be made, m ost
promin ently in th e often cited Lost Crops of
th e In cas (N RC, 1989), the co n siste ntly low
DP in 10 y acon accessio ns (3.6-4 .3) shows
that yacon fructan s are of low mol ecular
we ight. This is in agreement with Ohyama
et al. (1990) . Yacon fru ctans are th erefore
different from th e inulins in chicory
(Cichorium intybus) w ith a DP range of

7-12 and individu al fru cta n chain s up to 80

fructo se units lon g (V;rn Wa es et al., 1998).
Hi gh mo lec ular we ight inulin also accounts
for most of th e carbohydrates in Jeru sa lem
artic hoke (Helianthus tuberosus) (Pra znik
and Bec k, 1987; W ei et al. , 1991 ). Yaco n
also differs from Jeru sa lem artichok e in that
it has significant fru ctose and glucose
co ntents. Fructose accounts for 3-22% of
roo t OM, glucose fo r 2-5% (Tabl e 3).
In th e ca lc ul ation of food energy (14 822 4 kcal/kg FM), we have ass um ed th at
fructa ns be have as di eta1y fiber in th e
intestin al tra ct (Q uemener et al. , 1994) and
mak e no ca lo ri c co ntribu tio n during
di gestion. Fructans are unlik ely to be
brok en down to a signifi ca nt exten t by
stomach acid ity, but so me deg radation
occ urs in the co lo n du e to bacterial ferm entati o n (Silva , 1996). Th erefore, our food
energy va lu es somew hat underestimate tru e
ca lori c va lu es. In any case , foo d energy of
yaco n is very low and full y ju stifies the
root's rep utation as a low-ca lor ic d iet food.
Th e d ata do not suppo rt th e noti on of
chem ica l diffe1entiation of yacon along
geog raphi c gradi ents (Figure 1 ). O vera ll,
th ere was littl e compos itio nal d ive rsi ty in
th e 1 0 yaco n accessio ns. Ca rbohydrate
va ri ation due to physio log ica l factors (e.g.,
plant age) and po stharv est co nditi ons (e.g.,
stora ge duration) appear to have much
more practi ca l relevan ce th an genetic
differences. Fo r exam pl e, O hyama et al.
(1990) observed fruct ans to be reduced to
20% of OM after storage for 3 mo 11 und er
co ld cond iti o ns. 11 Also, Wei et al. (1991)
report dec reas in g fru cta ns and increas in g
fructo se aft er sto rage. Co nsid erab le co mposi tion al changes also occ ur durin g th e
tradition al soleado trea tm ent after harves t. It
in vo lves sprea din g th e harv es ted root s in an
unsh aded place for a w ee k or so. Th at
res ults in in c reased root sweetness and a
co nco mitant ri se in Bri x (H erm ann,
unpublish ed data ), w hi c h, apart from
res pirati ve wa ter loss, is most likely du e to
th e in creme ntal de-polym erization of
fructans.

Hi ghes t OM and fru c tan yields (accession ASL 136) were assoc iated with
dodecap lo id y compared w ith octoploidy in
the ot her access ions. Thi s pro vides a
pointer fo r id entify in g superio r ge rmpla sm
and for breeding yacon.
Ba sed o n Table 2, nutri ent remova l from
th e fi eld pert root FM ca n be ca lcul ated as
0.4-0.8 kg N, 0.2 -0.3 kg P, and 1.8-2.9 kg
K. Th ese data allow an approximation of
minim al fe 1tilizer requirements for yaco n.
The ranges fm nutri ent removal per 100 kg
solubl e ca rbohydrate produced (fru ctans
and free suga rs) are 0. 5-0.9 kg N , 0.2 -0.4
kg P, and 2.3-3.2 kg K. These figures
identi fy yaco n as not on ly high in N-use
effici ency but also demanding in its K
requirements.
D es pite its hi gh fru ctan productivity,
yacon is unlikely to beco me a so urce of
purified di eteti c sweeteners or fr uctose
products in the nea r future. Th at is because
of seve ral factors in c ludin g 1) lack of
suitab le extraction tec hn o logy and indu strial sca le production , 2) co mpetition from
very low pric ed, hi gh-fructose sy rup s fr om
corn starch, and 3) protectionism of suga r
markets .
It is more likely th at processed yaco n
produ cts requiring little or no refinin g cou ld
be targeted as a natural o r low-ca lori e food
to a hea lth-co nsc iou s c li entele. Entrepreneuri al far mers in Brazi l and Japan have
al ready seized this o ppo rtunity and are
produ c in g a number of pro cesse d yacon
produ cts fo r nich e mark ets (Grau and Rea ,
1997; Kak ih ara et al. , 1997). One suc h
produ ct co nsists of air-dri ed tuber sli ces,
w hi ch resem ble dri ed ap ples. Current
res ea rch at C IP examines proc ess in g
par ameters bearing o n fin al product qu ality.
Anoth er pote ntiall y interesting produ ct
is unrefin ed ya con syrup. It co uld be
mark eted as a di eteti c sweetener th e
con sistency of hon ey and possibly pri ce d
at th e sa me level. Yaco n sy rup-makin g in
rur al And ean settin gs co uld find mu ch

CIPPrngrain Report 1997 -98

431

inspi1atio n from makeshift technology

developed in the ea1ly days of maple sy1up
production. The challenge for yacon
product development is to design research
interve ntions that will allow Andean
farmers to retain some of the value added
throu gh processing.
References Cited
Campbell, J.M. , L.L. Bauer, G.C. Fah ey,
A. J.C.L. Hogarth , B.W. Wolf, and D.E.
Hun ter. 1997. Sel ected fructooli gosac c haride (1 -Kesto se, Nystose, and 1 Fbeta-fru cto-fu ran osy Inystose) com pos iti on of foods and feeds. J. Ag. Food
Chem. 45:3076-3082.
Grau , A. and J. Re a. 1997. Yacon.
Smallanthus sonchifolius (Poepp. &
Endl.) H. Robinson . In: Hermann, M. and
J. H eller (eds.). Andean roots and tubers:
Ah i pa, arracacha, mac a and ya con.
Promoting the conservation and use of
und erutilized crops, IPK, Gatersleben/
IPGRI, Rome. p. 199-256.
Hoebregs, H. 1997. Fructans in food and
food products, ion exc hange chromatography method: Collaborative study. J.
AOAC Int. 80 (5):1 019-1037.
Kakih ara , T.S. , F.L.A. Omara, and S.M.C.
V ilhena. 1997. Cultivo e industri alizar,:ao
de yac on: uma exper iencia brasil eira
[Cultivation and processing of yacon: A
Braz i Ii an experience]. 1'' Ya con Workshop held 31October1997 in Botucatu
(SP), Brazil. (Portu guese)
M errill , A.L. and B.K. Watt. 1973. Energy
value of foods: Bas is and derivation.
Agricultural Handbook No. 74. United
States Department of Agriculture,
Was hington, D.C. , USA.
NRC (National Resea 1ch Council). 1989.
Lost crops of the Incas: Little-known
plants of the Andes with promis e for
wor ldwide cultivation. National Academ y Press , Wash in gto n, D.C. , 415 p.

432

Andean Roots and Tubers

Ohyama , T., 0. Ito, S. Yasu yoshi , T.

lk ara shi, K. Minamisawa, M. Kubota, T.
Tsu kihashi, and T. Asami. 1990. Composition of storage carbohydrate in tubers
of yacon (Polymnia sonchifolia). Soil Sci.
Plant N utr. 36(1 ):167-171.
Praznik , W. and R.H.F. Beck. 1987. lnulin
co mposition during growth of tubers of
Helianthus tubero sus. Agric. Biol. Chem.
51 (6):1593-1599.
Qu emener, B., J.F. Thibault, and P.
Couseme nt. 1994 . Determination of
inulin and oligofructose in food products
and integration in the AOAC method for
measurement of total dietary fibre.
Lebensmi ttel- Wissenschaft und
Tec hnologie 27:125-132.
Van Waes, C., J. Ba ert, L. Carlier, and E. Va n
Bockstaele, 1998. A rapid determi nation
of the total sugar content and the average
inulin chain len gth in roots of c hicory
(Cichorium intybus L. ). J. Sci. Food Ag.
76:107 -110.
Rohlf, F.J. 1992. NTSYS-pc, num er ical
taxonomy and multivariate analysis
system. Exeter Publishing Ltd., NY, USA.
Salgado, V.X. 1996. Evaluac i6n y
caracterizaci6n citogenetica de 16
entr adas de jicam a (Poly mnia
sonchifolia) de Sudamerica. Tes is, Ing.
Agr. Universidad Central del Ecuador,
Quito, 106 p.
SAS Institute Inc. 1989. SAS/ STAT User's
Guide, Ve rsion 6, Fourth edition , Vo lumes 1 & 2, Cary, NC. USA.
Sil va, R.F. 1996. Use of inulin as a natural
texture modifier. Cereal Food World 41
(10): 792-794.
We i, B. , M. Hara, R. Yamauchi, Y. Ueno,
and K. Kato. 1991. Fructo-oligosaccharides in the tubers of Jerusalem artichoke
an d yacon. Resea 1ch Bu ll etin of th e
Faculty of Agriculture, Gifu University
56:133-138 .

Training
Impact on a Changing World

Trainin g acti v iti es at CIP ove r 1997-98

foc used on res pondin g to prog ramm ati c
changes as w e/I as to o ur va ri ous stakeho ld ers' interests. A t the reg io nal leve l, spec ial
effo rt s we re m ade to stre ngth en know ledge
fo r acce lerat in g enh ancement in improved
va ri eti es of res istance aga in st late b li ght,
bacteri al w ilt, v iru ses, and pests. A noth er
area th at rece ived spec ial trainin g attenti o n
was th e use of tec hno log ies aimed towa rd
sustain abl e produ cti o n, such as im proved
potato propaga ti o n and c rop management
practi ces .
CI P uses a dynami c, pa rti cipa tory
training meth od . Thi s combin es expos ure
to recentl y developed adva nced techni ques
w ith hand s-o n ex peri ence emph as izin g
procedures th at are directl y appli cab le to
workin g co nd itio ns in deve lop ing co untri es.
The meth od enh ances acti ve partic ipati o n
and in formatio n exc hange and aim s to
ex pand the ca pa b i Iiti es of partn er o rga n izati ons and strength en co ll abo rat ion w ith

th em. Such a meth odo logy all ows fas ter

resea rch progress, id entifies th ose elements
th at need further work at the nat io na I leve l,
and gives the op portuni ty to so lve co nstrain ts of co mm on interest. We present
below an ove rv iew of group and in d ividu al
trainin g, and close w ith our vision fo r
trainin g in th e future.
G roup trainin g acti v iti es includ ed
courses, wo rk shops, co nferences, and
sy mpos ia co nd ucted both at headq uarters
and in -co untry. St ro ng emph as is was also
give n to in d ividuali zed trainin g, an effecti ve
way to co ntinu e deve lo pin g th e research
skill s of co ll abo rators and partn ers throughout the deve lop ing wo rld . O ve r 1,000
resea rch sc ienti sts and deve lopment agents
ac ross th e globe w ere train ed durin g 199798 in activiti es orga ni zed and co ndu cted
by CIP. A noth er 9 11 perso ns partic ipated in
27 ( IP-fac ili tated trainin g eve nts in Peru
(Tabl e 1).

Table 1. All group training activities and participants during 1997-98.

Course focus

Activities

Participants
372

Potato
Sweetpotato
Andean roots and tubers
Natural resources management

18
18
3
6

Total

45

123
1,019

23

25
27

888

474
50

(IP-facilitated activities
CONDESAN
Other Peru-based
Total

911

(IPPrngrom Reporl 1997-98

433

Potato
Trainin g courses and works hops in potato
addressed six subject matter areas. Courses
in th ese areas va ri ed in content according
to spec ific needs. Th e cou rse theme,
numbers of participants, acti v ities, and the
countries rep rese nted are show n in Tab le 2.
Courses o n improved potato propagation techniques and sanitary control are in
continuous demand because there is an
urgent and co nstant need to produce higher
quality seed of new ly developed genet ic
mater ials that have improved produ ct ivity
and res ista nce to pests and di seases . The
program co ntent of these acti v iti es in c ludes
agronomic aspects, formal and inform al
seed syste ms, sanitary co ntrol procedures,
seed sto rage and mark et in g, and organization al aspects of seed p rograms. All topics
are from the perspective of an integrated
seed potato industry.

The major focus in potato germplasm

management co urses is to train sc ientists
and help them develop research act iv ities
that use biotechnological tec hniqu es to
acce lerate breed in g progress for improved
res ista nce to pests an d di seases . Thi s
in vo lves equipping sc ientists from developin g co untries to use adva nced labo ratory
tec hniqu es in th eir own co untri es.
Trainin g materi als, such as a spec ialized
manu al, v ideos, and other publi cat io ns,
we re developed for supportin g training
activities in those locations having the
highest potential for ado pting true potato
seed as an alternative to traditional potato
propagation. M ainl y, the tec hniqu es and
method s used in TPS production and use
are add ressed.
Trainin g activities in integrated pest
management prov ide informati o n o n the
latest developments in co ntrol and m anagement of the prin c ipal potato pests using IPM

Table 2 . Group training activities on potato-related subjects.

Course
Potato propagation systems and techniques

Activities

Participants

43

Countries represented
Argentina, Bolivia, Colombia, Ecuador,
Ethiopia, Guatemala, Kenya,
Nicaragua, Peru, Venezuela,

True potato seed (TPS)

2
3

34
63

Bangladesh, Egypt, Indio, Nepal, Peru,

Integrated pest management (IPM)

112

Argentina, Bolivia, Colombia, Ecuador,

Integrated management of late blight

120

Argentina, Bolivia, Chile, Colombia,

Potato germplosm conservation

Kenya
Sri Lanka, Thaila nd
Guatemala, Peru
Panama, Peru, Venezuela, USA,

18

372

88

Seed production

5
3
6
11

Total

25

Total
CIP-facilitated group training in Peru
Natural resources management
Posthorvest and marketing
IPM

434 Training

63
310

472
888

tec hniqu es. Trainin g is b ased on sh arin g

expe ri e nces from pi lot locat ions in key
potato-pro du c in g areas end emi ca ll y
affected by th e m os t se ri o u s pests.
Th e

posth arvest and m arketin g. Th e number of

activiti es and p arti c ipants' co untri es are
shown in Table 3 .

integrated management of late

blight co urses co mbin e a w ide range of

components to redu ce th e effects of th is
devastatin g disease. Trainin g emph as izes
c urrent bi o tec hn o log ica l tool s for th e u se
and enh ance m ent of potato ge rmplasm and
co nve nti o n al breedin g n eed s as a m ea n s to
sea rc h for res istan ce that could improve
c rop p erformance in areas regularl y
threa tened b y th e di sease. In 1997-98, pil o t
test in g b ega n of th e farmer fie ld sc hoo l
(FFS) approac h, n ew to Latin Am eri ca but
used wid ely in Asia. Amon g other ben efits
id entifi ed by fa rm er parti c ipan ts was an
improved und ersta ndin g of th e prin c ipl es
in v ol ved in di sease prog ress io n and
m an age m ent. Thi s su ccess has led to
resea rc h into how fa rm ers' new kn ow led ge
he lp s th em to red u ce losses and increase
profits.
Sweetpotato
Subj ec ts cove red by th e tra inin g program
o n sweetpotato includ ed germplasm
m an agement, breeding and production ,
integ rated p est m an age m ent, and

Germplasm management training

focu sed o n taxo nomy, biodiversity, propagati o n, and co n se rva tion of sweetp ota to
ge rm p lasm. Spec ia l attenti o n was g ive n to
the m o rph o log ica l id entifi ca tion of ge notypes, m aintenan ce o f ge neb anks (in situ
and in v itro), and rapid multipli ca tion of
clon a l m aterial.
Th e o bj ective of breeding and production co urses was to train se lected fa rm ers
on sweetp o tato seed pro du c ti o n, w ith an
emph as is o n th e prac ti ca l knowl ed ge and
skill s needed in seed and parti c ipato ry
re search procedures to se lec t and eva lu a te
new, improved va ri eti es . Th e co urses
co ve red growth ph ys io logy, clim ate, so il
and nutri ent requirem e nts, geneti c resources, morphol og ica l c harac teri za tion,
breedin g and se lec ti o n approac hes,
eva lu ati o n and va ri ety re lease procedures,
man agem ent of c roppin g system s, p es ts and
di seases of sweetpo tato, tis sue c ulture and
mi cropropaga ti on tec hniques, produ c tion
m ethods, and posth arves t m anagem ent.
A program of training of trainers (TOT)
was initi ated in Indo nes ia at th e requ est of

Table 3. Group training activities on sweetpotato-related subjects.

Course
Germplasm management

Activities

Participants

93

Countries represented
Cameroon, Dem. Rep. of the Congo, Ethiopia,
Ghana, Indonesia, Kenya, Madagascar, Malawi,
Malaysia, Mozambique, Myanmar, Nigeria,
Philippines, Rwa nda, Senegal, Tanzania,
Thai land, Zambia

Breeding and production

12

338

Bangladesh, Burundi, China, Ethiopia, Indonesia,

Kenya, Malawi, Malaysia, Mozam bique,
Phi lippines, Rwanda, Tanzania, Uganda, Zambia

Integra ted pest management

11

Zambia
China, Ethiopia, Kenya, Malawi, Uganda, Zambia,

Postharvest and marketing

32

Total

l8

474

Zimbabwe

(IP Program Report 1997-98

435

several NGOs and in collaboration w ith

several deve lopm ent agencies operating
locall y. Th e parti c ipants planned the
impl ementati o n of sweetpotato fa rmer field
schools in six distri cts of Java, w ith a goal of
introd ucing sweetpotato production
technology into an FFS program .
The integrated pest management
courses emphasized managem ent of
sweetpotato weev i I, fresh and dry storage,
and maintenan ce and multiplication of
planting material , as we ll as planning,
impl ementin g, mo nitorin g, and preparing
follow- up an d evaluation of !PM program
acti v ities at the fi eld leve l. In turn , CIP
gained information from particip ants 011 the
importance and constraints in sweetpotato
produ ction in their co untry, and obtained
feedback from parti c ipants 011 the practi ca lity of suggested managem ent strategies.
Su ccessful IPM experiences in some
countries are used to illu strate th e potential
of a particular control strategy.
Th e general aim of postharvest and
marketing training was to deve lop the sk ills
of NARS researc hers in agricultural marketin g research. Thi s in vo lved familiarizing
parti c ipants w ith standard meth od s for
analyzing agr icultural marketin g in deve loping countries, providing hand s-011
instru ction in these method s in both
classroo m and field and offering feedback
about on-going and proposed marketin g
res earc h by spec iali sts and instituti ons,
including the organizing CG centers and
NARS in th e reg ion . Bri efings were give n
on funding op portunities for marketori ented researc h. Participants pract iced

developing concept notes for resea rc h

projects.
Andean Root and Tuber Crops
Fihy sc ientists from Andean countri es
rece ived training in ge rmplasm characterization of Andean root and tuber c rops
(A RTC), production, and posth arvest and
processing (Tabl e 4).
Parti c ipants of Boli via and Ecuador
atte nd ed a course 011 germplasm characterization, w hich was conducted in Peru. Th e
goa l was to train scientists in the tec hniqu es
needed to identify duplicates within ARTC
germpl as m collections. Twe nty-fi ve
nation al sc ientists from Peru , Bolivia,
Ecuador, Brazil, and Colombia parti cipated
in a course on the production of Andean
root and tuber c rops. A new training
manual espec iall y deve loped for production of ARTC was used in this co urse.
Twe nty-o ne participants from Boli via,
Colombia, Ecu ador, and Peru attend ed an
international wo rkshop o n Andean food
c ro ps held in Lima, Peru. The participants
rep resented different projects operating in
th e Andes. Workshop discussion res ults
we re used for coo rdin at ing future resea rch
and development efforts in Andean food
products, w ithin the framework of
CO N DESAN , the Consortium for th e
Sustainable Development of the Andean
Region.
Natural Resources Management
A first co urse on wate rshed management
was organized by CIP in Peru , in collabora-

Table 4. Training activities on Andean root and tuber crop-related subjects.

Course

Activities

Germplasm conservation

Participants

Bolivi a, Ecuador, Peru

Bolivia, Brazil, Colombia, Ecuador

Andean root and tubers production

Posthorvest ond processing

25
21

Total

50

436

Training

Countries represented

Bolivia, Colombia Ecuador, Peru

ti on w ith seve ral loca l orga ni za ti ons. Th ese

loca l organi za ti o ns acti ve ly parti cipate in
projects related to the co nse rva ti o n and
sustain abl e improvement o f agri culture in
Andea n wa tersheds. The course program
emphas ized relation ships between bi o log ica l and soc ial sc ien ces in th e co ntext of
wa tershed management. O th er co urse
topi cs included soil and wa ter management, co nse rva ti on and use of biodi ve rsity,
eco log ica l zo nin g, clim ato logy, GIS, and
indu stri ali zati o n potenti al of agri cultural
produ cts. Fi ve short co urses were o rgani zed fo r Latin Ameri ca n co untri es on
simul ati o n modelin g and o ne in Uga nd a fo r
suppo rtin g resea rch acti viti es (Tab le 5).

Individualized Training
N ati o nal sc ienti sts fro m Afri ca, As ia, and
Latin Am eri ca rece ived trainin g at CIP
laborato ri es and fa c iliti es at Lim a headqu arters and CIP's reg ion al office in Na iro bi ,
Kenya (Ta bl e 6) . Trainin g subj ects fea tured
w ere potato seed produ ction ,
mi cropropagati o n and ti ssue culture
tec hniques, ad va nced methods fo r viru s
characteri zati o n, anti sera produ cti o n and
detect ion tec hniques, as we ll as biotec hn olog ica l trainin g co nce ntratin g on meth ods
such as RFLP, A FLP, RA PD, and geneti c
mani pul ati on in potato.

International and lntercenter Collaboration

in Training
There are seve ral exce ll ent exa mpl es of
intern ati onal and interce nter coope rati on in
our trainin g over 199 7-98. A reg io nal

wo rks hop fo r th e co nse rva tion and utili zati on of sweetpotato, cassava, and ya m in
Sub-Sa haran Afri ca (SSA) took pl ace in
Kenya. Th e system-wid e geneti c reso urces
program compri sing CIP, llTA, IPGRI , and
the N ati onal Network of PRA PACE and
SAR RN ET sponso red th is workshop. Fortyeight nati onal sc ienti sts particip ated. Two
courses on posth arvest and marketin g of
sweetpotato we re co ndu cted in Kenya and
Ugand a in co ll aborati o n w ith ICRAF, ILRI ,
ICRISAT, and IFP RI. Thi s was a short
trainin g co urse on meth ods for analyz in g
agri culture mark ets in SSA and was attend ed by 23 pa rti c ipants from thi s part of
Afri ca. Parti c ipants w ere fa mili ari zed with
stand ard mark etin g analys is procedures and
re ce ived hand s-o n in stru ction on mark et
ori ented resea rch. In 1997, CIP staff gave
lectures on the potato and sweetpotato
related issues at an I ITA-o rga ni zed reg io n a I
course on prod uct and market deve lopm ent
for roots and tu be r crops.
In co njun cti on with national co ll abo rators, a work shop on root process ing
techno logy took pl ace in Chengdu, in th e
Peop les Repu bli c of Chin a. Forty-seven
parti cipa nts rece ived a so lid found ati o n in
the kn ow ledge of starch stru cture, properti es, and appli ca ti o n. Th e program in clud ed v isits to starch and noodl e fac tori es
in Sich uan to assess ado pti on of Sichu an
Aca demy of Agri cultural Sc iences-CIP
project res ults and fo r pl anning future
resea rch needs.
Co ll aborati on w ith reg ional netwo rk s in
Sub-Sa haran Afri ca res ulted in a course o n

Table 5 . Group training activities on natural resources management.

Course

Activities

Participants

Watershed management
Simulation modeling

14
109

Total

123

Countries represented
Argentina, Bolivia, Chile, Ecuador, Peru
Bolivia, Brazil, Chile, Colombia,
Ecuador, Guatemala, Mexico,
Nicaragua, Peru, Uganda

CIP Program Report 1997-98

437

Table 6. Individual training.

Subject

Participants

Countries represented

Venue

IPM

Colombia, Ethiopia, Indonesia, Peru

HQ(4), SSA(l)

Vi rology

Argentine, Senegal, Thailand

HQ(2), SSA{l)

Angola, Argentina, Bolivia, Brazil, Chile, Korea,

HQ(l 8)

18

Germplasm management

Peru, Zaire

and molecular techniques

Pathology

13

Argentina, Bolivia, Chile, Ecuador, Peru, Syria

HQ(l 2), SSA(l)

Breeding

5
15

Ethiopia, Iron, Korea, Peru, Uganda

HQ(4), SSA(l)

Angola, Brazil, Iran, Kenya, Peru, Senegal,

HQ(4), SSA(l l)

Potato propagation

Syria, Vietnam
2

Natural resources
Communications

HQ(2)
HQ(2)

63

Total

HQ

Argentina, Peru
Argentina, Peru

Headquarters in Peru; SSA

Sub-Saharan Africa (CIP Regional Office and KARI).

food dehydrati o n by sol ar dryin g and

sweetpotato fl our processing. N in e parti cipants from Kenya, Tanzani a, and Ugand a
attend ed thi s co urse held in Ugand a.
CIP coll aborated in th e organi zation of
two nation al courses in Ecuador and on e in
Peru on soil m anagement w ith 8 1 parti cipants from va riou s devel opment age nci es.
Al so, CIP staff coll aborated o n two courses
on management of natural resources
organi zed in Ugand a and Vietn am D.R. in
w hi ch 120 nation al personn el from different research and extensio n organiz ation s
parti c ipated.

The Future: Developing a Vision for

Training
In 1998 , a hi ghl y particip atory process set
the bas is for CIP' s trainin g prog ram strategy.
In thi s Trainin g Visi on, CIP's program is
seen as a mec hani sm to efficientl y ove rcom e current con straints and effect ivel y
ident ify emerging oppo rtunities. Fulfillin g
thi s v ision w ill (1) enabl e CIP to strea mlin e
resea rch pri orities to meet cli ents' need s, (2)

438 Training

enhan ce capab i Iiti es fo r faster and better

co llabo rati ve tec hnol ogy deve lopment, (3)
fac ilitate co mmunity-l evel ad aptati o n of
improved techn o log ies, and (4 ) create a
bridge between current and future partn ers
thro ugh out th e developin g wo rld .
CIP 's trainin g program w ill be a ve hi cle
fo r interaction , in put, and coll aborati on
w ith a w id e ran ge of partners enablin g the
Ce nter to ac hi eve its mandate fo r potato,
sweetpotato, And ean root and tuber crops,
and mo un tai n natural reso urces . It w ill be
in extri cab ly linked w ith the resea rch
age nda and w ill respond to partn ers' needs
fo r enh anced resea rch skills and meth ods. It
w i 11 prov ide effecti ve mec hani sms fo r the
in trodu cti on of tec hn o log ies to ac hi eve
sustain abl e improve ments in th e produ ctivity and utili zatio n of CIP's mand ate crops,
and in th e man agement of natural reso urces
in the deve lopin g w orld . Th e prog ram w ill
create an intern ati onal netwo rk of hi ghl y
ca pabl e research sc ienti sts able to co nduct
independent studi es, to offer skill s trainin g
to others, and to co ll aborate effecti vely in
th e CIP global co mmunity of interest.

Selected Publications
from

1997-1998
Impact on a Changing World

Ames, T., N .E. J.M . Smit, A.R . Brau n, J.N.

O'S u llivan, an d L.G . Skog lund . 1997.
Sweetpo tato: Major pests, diseases, and
nutritional disorders. CIP, Lima, Peru.
153 p.
Ames de lcoc hea, T. 199 8. Enfei-medades
fungosas y bacterianas de raices y
tuberculos and inos. CIP, Lima , Peru . 172

p.
Antl e, J.M., S.M. Capalbo , and C.C.
Crissman. 1998. Econometric and
si mul atio n modeling of the Carchi potato
productio n syste m. In: Crissman, C.C.,
J.M. Ant le, and S.M. Cap albo (eds.).
H ea lth and sustainabl e agr icu lture:
Pes ticide use in th e And es. Kluwer
Academic Press, Boston , USA. p. 145180.
Ant le, J.M., S.M. Capalbo , and C.C.
Cr issman. 1998. Tradeoffs in policy
Ana lysis: Co ncep tu al foundations for
disciplinary integration. In : Crissman,
C.C., J.M. Antle, and S.M. Capa lbo (eds. ).
H ea lth and sustai nabl e ag ri culture:
Pesticid e use in the And es. Kluwer
Academic Press, Boston, USA. p. 21-4 0.
Antle, J.M., S.M. Capalbo, D.C. Co le, C.C.
Crissman, and R.J. Wa ge net. 1998.
Integ rated simul ation mod el and ana lys is
of economic, env ironm ental and hea lth
tradeoffs in the Carchi potato-pasture
production system. In : Crissman, C.C.,
J.M. Ant le, and S.M. Capalbo (eds .).
Health and sustainabl e agr icu lture:
Pesticide use in the Ande s. Kluwer
Academic Press, Boston, USA . p . 243268.

Antle, J.M ., D. Co le, and C.C. Cri ssman.

1998. The role of pesticides in farm
producti vity and farmer hea lth, in
quantifying trad eoffs in the env ironment.
In : Cr iss man, C.C., J.M. Antl e, and S.M.
Capalbo (eds .). Health and sustainable
ag ri cultu 1e: Pesticide use in the Andes.
Kluwer Acade m ic Press, Bo ston, USA. p.
231-242.
Antle, J. M., D.C. Co le, and C.C. Cri ssman.
1998. Furth er evide nce on pesticides,
product ivi ty, and far mer hea lth: Potato
production in Ecuador. Ag . Econ . An. Int.
J. 18(2):199-208.
Arbizu, C., Z. Huaman, and A. Go lmirzaie.
1997. Other Andean roots and tubers. In :
Fuccillo, D. , L. Sears, and P. Stap leton
(eds.). Biod ive rsity in trust: Conservatio n
and use of plant genetic resources in
CG IAR centres. Cambr id ge U ni ve rsity
Press, Cambridge, UK. p. 39-56 .
Ar itua, V., E. Ad ip ala, E.E. Carey, and R.W.
Gib so n. 1998. Th e in c id ence of sweet
potato vi ru s disease and v iru s resistanc e
of sweet potato grown in U ganda. Ann.
Appl. Biol.132:399-411.
Aritua , V., T. A li cai, E. A dipala, E.E . Carey,
and R.W. Gibson. 1998. Aspects of
resistance to sweet potato virus disease
in sweet potato. A nn . App l. Biol.
132:387 -398.
Blas, R. , C. Arbizu, and G. Rodri guez .
1998. Numero d e c romo somas de la
arracacha (Arracacia xa nth orrhiza
Bancroft). Anal es Cientificos (XXX ll):4454
Bo wen, W.T. and W.E. Baethgen. 1998.
Simulation as a tool for improv in g

CIP Program Repoll 199/.98

439

nitrogen managem ent. In: Tsuji, G.Y., G.

Hoogenboom, and P.K. Thornton (eds.).
Understanding options for agricultural
production. Kluwer Academic Publishers , Dordrecht, The Netherlands. p. 193-

208.
Bowen , W.T. , P.K. Thornton, and G.
Hoo ge nboom. 1998. The simulation of
croppin g sequenc es using DSSAT. In:
Tsuj i, G.Y. , G. Hoogenboom , and
P.K.Thornton (eds.). Understanding
options for agricultural production.
Kluw er Academic Pub li shers, Dordrecht,
The N etherlands.
Campi Ian, D. and G. Pra in. 19 97. Integrated pest management, the v iew from
below: The role of users' perspectives
and participation in agricu ltur al R&D. In:
Pradhanang, P.M. and J.C. Elphingston
(eds. ). Integrated management of bacteri al w ilt of potato: Lessons and experiences for the hills of Nepal. LARC,
Lum le, Nepal. p . 87-100.
Castillo , A. , H. Quisp e, D. Mora les, and R.
Quiroz. 1997 . Los sistemas de
producci6n Agricol a en lo s Andes de
Bolivia. In : Mujica E. and J.L. Ru eda
(eds.) . La sostenibil idad de los sistemas
de producci6n campesina en los andes.
CONDESAN, Lim a, Peru. p. 163-214.
Cipriani, G., A . Golmirzaie, and D.P.
Z hang. 1998. The readers talk: Developing weev il resistance in sweetpotato w ith
genetic transformation. Sweetpotato
Research Front (SPOR F) Magaz in e
7(0ct.):6.
Collins, W.W., E.E . Carey, l.G. Mok, P.
Thompson , and D .P. Zhang. 1998.
Uti li zation of sweetpota to genet ic
resources to develop insect resistance .
In : Clement, S.L. and S.S. Quisenb erry
(eds.). Global pl ant ge net ic resources for
in sect-resistant crops. CRC Press, Boca
Raton, Florida, USA. p. 193-205.
Condori, B., A. Devaux, P. Mamani, J.
Va l lejos, and J. Blajos. 1997. Efecto
residual de la fertilizaci6n del cu lti vo de
papa sob re el cultivo de haba (Vicia faba
L.) en el sistema de rotaci6n. Revista
Latinoamericana de la Papa 9/ 10(1 ):171 -

187.

440 Selected Publirntions

Crissman , C.C. and J.M. Ant le. 1998.

Physical and Economic Mod el Integration for the M easu rement of En v ironmen tal Impacts of Farming. In: Lal , R. (ed.).
So il quality and agricultura l
sustainab il ity, A nn Arbor Press, Chelsea ,
Ml, USA. p. 319-334.
Cris sman , C.C., J. M . Antle, and S.M.
Capalbo (eds .) . 1998. Quantify in g
tradeoffs in th e environment, health and
sustai nab le agricu lture : Pesticid e use in
the Andes. Kluwer Academic Press,
Boston , MA, USA . 281 p.
Cri ssman , C.C. and P. Espinosa. 1997.
Papas , pesticida s y polfticas:
ln vesticac i6n y la agricutlura soste nibl e.
In: Martinez, L. (ed.). El desarrollo
so stenib le en el medio rural. Bib li oteca
de Ciencias So c iales (FL ACSO , Quito ,
Ecuador) 2:191 -219.
Crissman, C.C., P. Esp inosa, C. Ducrot, D.C.
Cole, and F. Carpio . 1998. Th e case
stud y site: Ph ysical, health and potato
farming system s in Carchi Prov ince. In:
Cri ssman, C.C. , J.M. Antle, and S.M.
Capa lbo. (eds .). Qua ntifying tradeoffs in
the env ironm ent, health and sustainable
agricultu re: Pe sticide use in the An des.
Klu w er Acad emic Press , Boston , MA,
USA. p 85-120.
Dayal, T.R. and C.R. Yada v. 1998. Evaluation of sweetpotato hybrids for yield and
quality traits. Recent Hort. 4:101-104.
del Rfo , A. , J. Bamberg, and Z. Huaman.
1997. Assess ing changes in the genet ic
diversity of potato gene banks. 1 . Effects
of seed increase. Theor. App l. Genet.

95:191 -198 .
de l Rfo, A. , J. Bamberg, Z. H uaman , R.
Hoekstra, A. Salas, and S. Veg a. 1997.
Assess in g changes in the genetic diversity of potato gene banks. 2. In sit u vs. ex
situ. Th eor. App l. Genet. 95:199-20 4.
Devau x, A. , J. Val lejos, R. Hijmans, and J.
Ramos. 1997 . Resp ue sta agron6mica de
dos variedade s de papa (ssp . tuberosum
y andigena) a d iferentes niveles de
fert ili zac i6n mineral. Rev.
Latinoamerican a de la papa 9/ 10 (1 ):123-

139.

Downham, M.C.A., N .E.J.M . Smit, P.O.

Laboke, D.R. Hall , D.I. Farman, A.
Braun, and B. Odongo. 1 998. Specific ity
of response to sex pheromon es among
the sweetpotato weevils, Cy/as
puncticollis Bohe. and C. brunneus F.
(Co leoptera: Apionidae). J. Chem. Ecol.

25(3):591 -609 .
Esca lada, M ., D. Campilan, and E. Cagasan.
1998. Diffusion mechanisms for improved sweetpotato var ieties in upland
commun iti es. In : Palomar, M. and E.
Gundaya (eds.). In tegrated rootcrops
research and development in the Philippines, IDR C-IRCP/ViSCA, Leyte, Philippines. p. 34-54.
Ewell, P.T. 1997. Internationa l cooperation
for the improvement of potato and
sweetpotato in sub-Saharan Africa . In:
Proceedings, 4'" Triennial Congress of
the African Potato Association. Pretoria,
South Africa. p. 157-1 64.
Forbes, G.A., X.C. Escobar, C.C. Aya la, J.
Revelo, M .E. Ordonez, B.A. Fry, K.
Doucett, and W.E. Fry. 1997. Popu lati on
genetic structure of Phytophthora
infestans in Ecuador. Phytopathol.

87:375-380.
Forbes, G.A. , S.B. Goodwin, A. Drenth, P.
Oyarzun, M.E. Ordon ez, and W.E. Fry,
1998 . A global marker database fo r
Phytophthora infestans. Plant Dis.

82:811-818.
George, M.L.C., R.J. Nel so n, R.S . Zeigle r,
and H . Leung. 19 98. Rapid population
ana lysis of Magnaporthe grisea by using
rep-PCR and endogenous repetitive DNA
sequences. Phytopathology 88(3):223-

229.
Ghislain, M. , R. Nelson , and T. Walker.
1997 . Resistance to potato late blight: A
globa l research priority. Biotechnol. and
Dev. Monitor 31 :14-1 6.
Ghislain, M. , M . Querci, M. Bonierbale, A.
Golm iarziae, and R. Nelson. 1997.
Biotechno logy and the potato: Appl ication for the developing wo rld . CIP, Lim a,
Peru. 18 p.
Ghislain, M ., M. Querci, M. Bonierbale, A.
Golmirzaie, and P. Gregory. 1998. The
app li cation of biotechnology to potato.

In: Ives, C. and B. Bedford (eds.). Agricultural Biotechnology in Intern ational

Deve lopm ent (Biotechno logy in Agriculture Series, 21) CAB Publishing, UK. p.

73-88.
Gibson , R.W. , R.O.M. Mwa nga, S. Kasul e,
I. Mpembe, and E.E . Carey. 1997.
Apparent abse nce of vi ruses in most
symptom less field-grown sweetpotato in
Uganda. Ann. Appl. Bio l. 130:481-490.
Gibson, R.W., I. Mpembe, T. Alicai, E.E.
Carey, R.0 .M. Mwanga, S.K. Seal, and
H .F. Vetten, 1998. Symptoms, aetio logy
and se rol ogica l analysis of sweetpotato
virus disease in Uganda. Plant Path.

47:95-102.
Golmirzaie, A.M., K. Bretsc hneider, and R.
Ort iz. 1998. Inbreed in g and tru e seed in
tetrasomic potato. II. Se lfin g and sibmat ing in heterogen eous hybrid populations of Solanum tuberosum. Theor.
Appl. Gen et. 97:1129-1132.
Golmirzaie, A.M. and A. Panta. 1997.
Tissue culture methods and approaches
for conservation of root and tuber crops.
In: Razdan , M.K. and E.C. Cocking (eds.).
Conservat ion of plant genetic reso urc es
in vitro. Volume 1: General aspects.
Sci ence Publishers, In c. USA. p. 123-

152.
Golmirzaie, A., A. Panta, and J. Toledo.
1998. Biotec hnological advances in the
conservation of root and tuber crops. In :
Benson, E.E. (ed.). Plant conservation
biotechnology. Taylor and Franc is Ltd.,
London, UK. p. 165 -178.
Golmirzaie, A.M., and J. Toledo. 1998.
Non-cryogenic, lon g-term germplasm
storage. In : Hall, R. (ed.). Plant ce ll
culture protocols. Humana Press,
Netherlands. p. 95-101.
Golmirzaie, A.M., R. Ortiz, G.N. Atlin, and
M. lwanaga. 1998. Inbreeding and true
seed in tetrasomic potato . I. Selfing and
open pollination in Andea n landraces
(Solanum tuberosum Gp. Andigena).
Th eo r. Appl. Genet. 97:1125-1128.
Griffon, D. and G. Scott. 1998. Avantpropos. In Scott, G. and D. Griffon (eds.).
Prix, Produits, et Acteurs. Methodes pour
ana lyser la commercia li sation agricole

CIP Progrom Repotl 1997-98

441

dans les pa y s en deve loppement. U ne

co-pub li catio n du CIP et le Centre de
cooperation internationale en rec herche
agronomique pou r le deve loppement
(CIRAD ). 1998. Ed iti on Karthala (France) .

p. 7- 10.
Ha genimana, VE ., E.E. Carey, S.T. Gi chuki,
M.A . Oyunga, and J.K. lmun gi. 1998.
Carotenoid contents in fresh , dri ed, an d
pro cessed sweetpotato produ cts. Ecol.
Food N utr. 37(5):455-474.
Ha genimana, V., E.E. Carey, S.T. Gi chuki ,
M.A. Oyunga, and J.K. lmun gi. 1998.
Rep lac in g pills w ith sweet potatoes to
combat Vitamin A deficiency. Improvi ng
po stharvest systems. 1ITA/ GTZ i nternati onal newsletter (October).
Ha gen im ana, V., E.G . Karuri , and M.A.
Oy un ga. 1998. Oil co ntent in fri ed
sweetpotato processed produ cts. J. Food
Proces. Preserv. 22:123-137.
Hagenimana, V. and C. Owori. 19 97.
Sweetpotato in ch apati processin g:
Feas ibility and acceptabi lity in rur al
areas . Ed. J. of Fo od Tec hn ol. in Africa
(Nairob i, Ken ya) 2( 1 ): 4-8 .
Hermann, M. and J. H ell er (e ds.). 1997.
And ea n roots and tub ers: Ah ip a,
arrracacha, mac a and yacon . In stitute of
Pl ant Genet ics and Crop Plant Researc h,
Gaterslebe n/ lntern at ional Plant Genet ic
Reso urces In stitu te, Rome , Italy. 256 p.
Hu aman, Z . and D.P. Zhang. 19 97.
Sweetpotato. In: Fu cc i Ilo, D., et al. (e ds .).
Bi od ive rsity in tru st - Co nservation and
use of plant genetic resources in CGIAR
Centers. Ca mbrid ge Univers ity Pre ss,
U K. p. 29 -38 .
Huam an, Z. 1998. Coll ection , maintenance
and eva luation of potato genetic re sources. Plant Var. Seeds 11 :29-38.
lragavarap u, T.K. , J.L. Pos ner, and G.D.
Bu be nz er. 19 98. Th e effect of var ious
crops on bromid e leac hing to sha ll ow
groundwater und er natural rain fa ll
cond itions. J. Soil Wate r Con se rv.

53:2:146- 151.
Farr ington, J. and G. Thiele. 1 99 8. Inn ova ti ve approaches to tech nol ogy genera tion
and di ssemination for low-in come
farmers in Ag ri culture and th e En viron-

442 Selected Publications

ment. In: Lutz, E. (ed.). Perspectives on

sustainab le rur al deve lopm ent. The
World Bank, Was hington , D.C. , USA.
Jacobse n, S.-E. 1997. Ad aptation of quinoa
(Chenopodium quinoa) to northern
Europect n agriculture: Stud ies on deve lopmenta l pattern. Euphytica 96:41-48 .
Jacobse n, S.-E. and A. P. Bach. 1998. The
influ ence of temperature on seed
germination rate in quinoa (Chenopodium qu ino a W illd .) . Seed Sci. Tec h.

26(2):515 -52 3.
Jacobse n, S.-E., B. Skadhauge, and S.E .
Jacobse n. 199 7. Effect of di etary in cl usion of qu in oa on broiler growth performance. A nim . Fee d Sci. and Tech . 65:5 -

14.
Jacobsen, S.-E. 1998. Developm ental
stabil ity of quin oa und er Eu ropean
conditio ns. Indu s. Crops Prod ucts 7: 169-

174.
Jayasi nghe, U. and L.F. Salazar. 1998.
Prese nt status of co ntrolling Potato
Leafro ll Virus. In: Hadidi , A., R.K.
Khetar pal , and H . Koganezawa (eds.).
Pl ant v iru s di sease co ntro l. APS Press, St.
Paul, MN, US A. p. 584-592.
Jo nes, J.W. , G. Tsuji , G. Hoogenboom, L.A.
Hunt, P.K. Th ornton , P.W. Wilk ens, D .T.
Imamura, W.T. Bowen , and U. Si ngh.
1998. D ecision support system for
agrotec hnol ogy transfe r: DSSAT v3. In :
Tsuji, G.Y., G. Hooge nboom , and
P. K. Th ornton (eds.). Understandin g
options for agricultural produ cti on.
Kluwer Academic Publ ishers, Dordrecht,
Netherland s.
Kaimo w itz, D ., P. Pacheco, and G. Thi ele .
1999. The effects of structural adjustment
on d efo restati on and forest degradation
in lowla nd Bo li v ia. Wo rld D ev.

27 (3):5 05-520 .
Korva, J.T. and G. A . Forbes. 1997. A simple
and low -cost method for leaf area
masurement in detac hed lea ves. Exp.
Agri c. 33:65-72.
K'osambo, L.M., E.E . Carey, A.K. Misra, J.
Wi lkes, and V. Hagen ima na . Influ ence of
age, farm in g site, and boiling on prov itam in A content in sw eetpotato

(/porn oea batatas (L .) Lam .) stora ge roots.

J. Food Comp . Analy. 11 :305-321.
Kha ta na, V.S, A.S. Chilver, S.G.
ll anganti leke, and M.D. Upadh ya. 1998.
Use of TPS technol ogy for potato
produ ct io n in Indi a: Prioritization of
re search and diffu sion age nda. Int. J. of
Trop . Ag. 16:1-11.
Khatan a, V.S, S.G. ll angant il eke, M.D.
Upad hya, and G .J. Scott. 199 8. Hi story
of pota to storage in Ind ia w ith specia l
reference to farm ers' storage practi ces .
Asi an Agr i-Hi story 2:195-207 .
Lagnaoui, A. 1997. Interfere nce of fungi c id es w ith entomopath oge ns: Effects on
entomophthoran path oge ns of green
pea ch aphid . In: Andow, D.A., D.W.
Ragsd ale, and R.F. Nyv al (eds.). Eco logical in teractions and biologica l contro l.
W estv iew Press, CO, USA. p. 301-320.
Lagnaou i, A. and E.B. Radcliffe. 199 8.
Potato fungic id es interfere with
entomop at hogenic fungi impa cting
popul at ion dy namics of Green Peac h
Aphid. A mer. J. of Potato Res . 75 : 19-2 5.
Leon -Ve lad re, C.U., B. A rce, and R. Quiroz.
1997. Mode laci6n de sistemas de
produ cc i6 n de leche; descr ipci 6n de sus
comp o nentes e in teracci6nes para el
diseno de modelos de simul aci6n. In :
Conceptos y metodologfas de
in vestigaci6 n en fin cas co n sistemas de
produ cc i6n anima l de doble prop 6s ito.
CIAT, Ca li , Co lomb ia. p. 95 -11 6.
Lery, X., J.- L. Zeddam, J. Giannott i, and S.
Abo l- Ela. 1998. Ev id ence for two small
viruses persiste ntl y in fec ting establi shed
cell lines of Ph th orimaea opercu lella
derivin g from embryos of the potato
tuber moth. Microbi o log ic a 21:10 3-107.
Ma ld onado, L. , S.G. Wri ght, and G. Scott.
1998. Co nstraints to prod uction and use
of potato in Asia. Am. J. of Potato Res.
75:1-9.
Mujica , E. and J.L. Ru eda. (eds.) 1997. La
sostenibi lid ad de los siste mas de
produ cc i6 n campesina en los Andes .
CONDE SA N , Lima , Peru. 227 p.
Mutombo, T., P. Ph emb a, N.B. Lutalad io,
and H .M . Kidanemariam . 1998. Evalu ation des genotypes de pomme de terre

dans differentes zones ag ro eco logi qu es

I' est du Congo. Afri ca n Crop Sc i. J.

6(4):359-368.
Ord onez, M. E., G.A. Forbes, and B.T.
Trognitz. 1997. Resistance to late b li ght
in potato. A putative ge ne th at suppresses R genes and is eli c ited by
spec ifi c isol ates. Euphytica 95 :1 67 -172.
Ord onez, M.E., G.A. Forbes, and B.T.
Trog nitz. 1998. Re lat ionship between
ineffective R- genes and expansion rate of
les io ns on potato ca us ed by
Ph ytophthora infesta ns. Plant Pat h.

4 7: 130- 136.
Orti z, R., 0. St0 1en, J.L. Chri st iansen, S.
Madsen, and S.-E . Jacobse n. 1998.
Geograph ic al and morp hol og ica l
patte rns of va ri ation defined a co re
co ll ect io n of Peruv ian ge rm plasm . In :
Macintosh, D .J. and T. Nielsen (eds.).
Biodiversity and Deve lopment. Ei gtveds
Pak hu s. Resea rc h Counc il for Deve lo pment Resea rc h & Centre for Tropi ca l
Ecosyste ms Researc h, Aa rhu s,
Copenh age n, Denmark. p. 75-77.
Otazu , V., M. Rodri guez, C. Arbizu , and
Ch . Li za rra ga . 1998. Arrosetamiento del
ullu co (Ullucus tuberosus Cald as)
causa do por Verticillium dahliae.
Fitopato logia 33(3):153-159
O yarz un, P.J. , A. Pozo, M .E. Ordo nez, K.
D oucett, and G.A. Fo rbes. 19 98. Ho st
spec ificity of Phytop hthora infesta ns on
tomato and potato in Ec uad or. Phyto path o logy 88:265 -271.
Oyarzun, P.J., A. Pozo, M.E . Ordonez, K.
D oucett, and G.A. Forb es 19 98. H os t
spec ificity of Phytophthora infestans on
tomato and potato in Ecuador. Epid emio logy 88(3):2712-2765
Pand ey, J., R.P. Sharma, M. Sharma, and
T.R. Daya l. 199 8. Potato and
sweetpotato. Pivota l fo r food security.
Indi an Farming 47(11 ):12 -14.
Patel, P.K., P.C. Pand e, B.N. Sark ar, and
M.S. Kadian. 1998. Multilocation
eva lu ati o n of seed l ings and seed lin g
tubers der ived from true potato seed. J.
Indi an Potato A ssoc. 25: 18 2.
Peters, Dai. 1998 . Imp rovi ng small-scale
li vestoc k produ ct ion in developin g

CIPProgram Report 1997-98

443

countries: The case of pig production in

Northern Vietnam . World Animal Re v iew
91 (2): 2-12 .
Peters, D. and C. Wheatley. 1997. Small
scale agroenterprises provide opportunities for income generation: Sw eetpotato
flour in East Java, Indonesia. Quarterl y J.
Int. Ag. 36 (4):331-352.
Phemba, P., T. Mutombo, N.B. Lutaladio ,
and E.E. Carey. 1998. Performance et
stabi I ite de rendement des genotypes de
patate douce dans divers
en v ironnements a l' est du Congo.
African Crop Sci. J. 6(2):109-11 8.
Priou , S., M.A. Triki, M. El Mahjoub, M.
Fahem, and M. Souibgui. 1997. Assessing potato cultivars in Tunisia for susceptibility to leak caused by Pythium
aphanidermatum . Potato Res. (40):399406.
Querci, M., R.A. Owens, I. Bartolini, V.
Lazarte, and L.F. Salazar. 1997. Evidence
of heterologous encapsidation of potato
spindle tuber viroid in particles of potato
leafroll virus. J. Gen . Virol. 78:12071211.
Querci, M. and L.F. Salazar. 1998. Molecular probes for the detection of viruses
and v iroids in potato. In: Khurana,
P.S.M., R. Chandra , and M.D. Upadh ya
(eds.). Comprehensi ve potato biotechnology. Malhotra Publishing House,
Bombay, India. p. 187-206 .
Quiroz, R., D . Peso, D . Rearte, and F. San
Martin. 1997. Dynamics of feed resources in mixed farming systems in
Latin America. In: Renard, C. (ed.). Crop
residues in mixed crop/livestock farming
systems. CAB Int., London, UK. p. 149180.
Quispe, C., A . Devaux, S. Gonzales, C.
Tourneux , and R. Hijmans. 1997.
Evaluaci6n comparativa del desarrollo y
crecimiento de papa , oca e isaf\o en
Cochabamba, Boli via . Rev.
Latinoamericana de la Papa 9/ 10(1 ):140155.
Ritchie , J.T., U. Singh , D.C. Godwin, and
W.T. Bowen. 1998 . Cereal growth,
development and yield . In: Tsuji, G.Y. , G.
Hoogenboom, and P.K.Thornton (eds. ).

444 Selected Publications

Understanding options for agricultural

production . Klu w er Academic Publishers, Dordrecht, The Netherlands.
Scott, G. 1997. Le marche de la pomme
de terre : Paradoxes et previsions pour les
annees. Agriculture+De veloppement
Rural (German y) 4(1):57-59.
Scott, G. 1998. Methodes d' evaluation du
marche potentiel des produits
transformes. In Scott, G. and D. Griffon
(eds .). Prix, Produits, et Acteurs.
Methodes pour analyser la
commercialisation agricole dans les pa ys
en developpement. Une co-publication
du CIP et le Centre de cooperation
internationale en recherche
agronomique pour le developpement
(CIRAD ). Edition Karthala (France). p.
228-255.
Scott, G. , R. Basa y, and L. Maldonado.
1997. Comercio exterior de papa en las
Americas: Crecimiento , potencial y
peligros. Comercio Exterior (Mex ico)
7(12 ):984-996.
Scott, G . and D. Griffon. (eds.). 1998. Prix,
Produits, et Acteurs. Methodes pour
analyser la commercialisation agricole
dans les pays en developpement. A copubl ication of CIP and the Centre de
cooperation internationale en recherche
agronomique pour le developpement
(CIRAD ). 1998. Edition Karthala. Montpelier, France. p. 11-29.
Scott, G. and L. Maldonado. 1998. Metodos
de evaluacion de factibilidad para el
procesamiento simple de productos
agricolas : Algunas experiencias en el
Peru. lndoagro (Peru) 5:71-78.
Smit, N.E.J.M. 1997. The effect of the
indigenous cultural practices of inground storage and piecemeal harvesting
of sweetpotato on yield and quality
losses caused by sweetpotato weevil in
Uganda. Ag. Ecosys . En v. 64(3):191-200.
Smit, N.E.J .M ., M .C.A. Dow nham, B.
Odongo, D.R. Hall, and P.O. Laboke.
1997 . Development of pheromone traps
for control and monitoring of
sweetpotato weevil ' s C. punctico//is and
C. Brunneus in Uganda. Entomologia
Experimentals et Applicata 85:95-104.

Tara Lam a, M.P. Chaturvedi , R.L. M and al,

S.P. Dh akal, and M.S. Kad ian. 1998. Onfarm tru e potato seed (TPS) tec hn o logy in
Nepal. J. Ind ian Potato Assoc. 24(3/

4):13 1-1 33 .
Thi ele, G. and R. Barber. 1998. Linkin g
with ag ricu ltural inpu t sup pl iers for
tec hn o logy transfer: Th e adoption of
vertica l tillage in Bolivi a. J. Soi l Wate r
Co nse rv. 53( 1 ):5 1-56.
Thi ele, G. , G. Gardner, R. Torrez, and J.
Gabri el . 1997. Farm er in vo lve ment in
se lectin g new va ri eties: Potatoes in
Bo liva. Exp. Ag. :275-290 .
Thi ele, G., 0. Na v ia, and E.N. Fern and ez North co te. 1998. Anali sis eco n6mico de
la estrateg ia d e control quimi co del tiz6 n
tardio (Phytop hth ora infes tans) para
culti va res de papa susceptibl es en
Coc habamba, Bol iv ia. Fitopato log ia

33(3): 176-181 .
Thi ele, G. and F. Terrazas. 1998. Las
carcavas. ProCampo 8 1 :18-20.
Thi ele, G. and F. Terraz as. 1998. Th e
wayq'os (g ulli es) are eat in g eve rythin g!
Indi ge nou s knowledge, PRA and so il
co nse rv ation . PLA Notes 32: 19-23.
Thi ele, G., J. W adsworth , and R. Velez .
1998. M akin g the link: Lesso ns fro m
agri cultural res ea rch and exte nsion in
low land Bo li v ia . Europea n J. Ag. Ed uc.
Exte n. 4(4):213 -223
Thi ele, G. 1998. In forma l potato seed
system s in th e Andes: Why are th ey
imp ortant an d what to do about them?
Wor ld D ev. 27( 1 ):83-99 .
Thi ele, G., J. Bustamante, J. Man si Il a, an d
G. Scott. 1 998 . lQue pasa co n el prec io
de la papa? Un ana li sis del periodo
1980-95. Proca mpo (Bo li via) . 1 3 p.
Triki , M.A. and S. Priou . 1997. Usi ng
chemi ca l and b io logical treatments to
redu ce th e potenti al of potato lea k
ca used by Pythium aphan idermalwn in
Tuni sia. Potato Res. 40: 39 1-398.
To ledo, J. , P. Dehal, F. Jarrin , J. Hu , M.
Herma nn , I.A. A l-Sh ehbaz , and C.F.
Qu iros. 1998. Genetic variab ility of
Lepidium meyenii in refe rence to other
And ea n Lep idium speci es (Brass icaceae)

assessed by mo lec u lar markers. An n.

Bot. 82:523-530.
Trognitz, B. 1998. Inh eritanc e o f res ista nce
in potato to les io n ex pan sion and
sporu lati on by Phytophthora infesta ns.
Pl ant Path. 47 :7 12-72 2.
Trognit z, B., M. H erm ann , and S. Carr ion.
1998. Germp las m co nse rvat ion of oca
(Oxa lis tub erosa Mol .) throu gh botanica l
seed. Seed fo rm at ion under a system of
po lymorphi c in co mpatibi I ity. Euph yti ca

101:13 3-141.
Turya mu reeba, G., R.O.M . Mwanga, and E.
Carey. 1997. Sweetpotato clo nal
evaluat ion for res istance to sweet potato
di seases (S PD) and hi gh y ield . In :
Proceed in gs of th e 4'h Tri enni al Co ngress
of th e Afri ca n Potato Assoc. , Pretoria,
South Afri ca. ARC, Roodepl att, So uth
Afri ca . p. 108 -11 2.
Va n de Fl ierl, E. 1997. Integ rated pest
manage ment: Sp rin gboa rd to sustain ab le
agr icul ture. In : Dhali wa l, G.S. , and E.A.
Heinri chs (ed s.). Crit ical issu es in ins ec t
pest manage ment. Natio nal Ag ri c ultu ra l
Technology In fo rm at ion Centre,
Ludhi ana, Indi a.
Van de Fli ert, E. 1998 . Integ rated pest
management: Sprin gboard to sustain abl e
agri culture. In : Dhaliwal , G.S. and E.A.
H ei nri chs, (eds .). Critica l iss ues in insect
pes t management. Com monwea lth
Publ ishers, New Delhi , Ind ia. p. 250 -

266.
Va n de Fli ert, E. and A.R. Braun . 199 8.
From probl em to impact: A mod el for
in tegrati ve, partic ip atory resea rch and
deve lopment for sustain abl e ag ri culture.
In : Su stain ab le li ve lih ood for rural
house ho ld s: Co ntribution from rootcrop
agric ulture. UPWARD, Los Banos,
Lagun a, Philippines. P. 81-98.
Va n de Fl ierl, E. A.R . Braun, R. As mun ati ,
Wiyanto, and Y. W id odo. 1997. On e step
bac k, two steps forward: Sweetpota to
integ rated crop manageme nt deve lopment in In do nes ia. In: UPWARD.
In st itu tionali zin g innovat ion s in root crop
resea rch and deve lopment. UPWARD,
Los Banos, Phi lippines. P. 127-140.

CIP Pt0gt0m Report 1997-98

445

Van de Fliert, E., A.R. Braun, S.R. Ghimire,

and J. Bron s. 1998 . Three cases and a
model: Th e app li catio n of an in tegrative,
participatory R&D framewo rk to UP WA RD projects in Ind o nes ia, Nepa l, and
the Philippin es . In: Sustainabl e li velih ood
for rural hous eholds: Contribution fr om
rootcrop agricu lture, UPWARD, Los
Banos, Laguna, Philippines. p. 99-110 .
Va n de Fliert, E., W iyanto, and A .R. Braun.
1998. The farme r fie ld sc hool as an
are na for farmer participatory research.
In: Reijn tj es, C., M. M inderh oud, and P.
Laba n (eds. ). LEIS A in persp ect ive: 15
years ILEI A . ILEI A, Leusden , Neth erla nds.
p. 40-41.
W heat le y, C. , G. Scott, R. Best, and S.
Wiersema . 1997. Meto dos para agrega r
valo r a raic es y tuberculos alim enticios:
Manual par a el desarrollo de productos.
Centro Internacional de Agricultura
Tropical, Cali , Colomb ia. 187 p.
Winters, P. , P. Espinosa , and C.C. Cr issma n.
1998. Manejo de Recursos en lo s Andes
Ecuatorianos: Revision de la liteatura y
eval uaci6n del proyecto Man ej o del Uso
Sostenible de Ti err as A ndina s
(PROMUSTA) de CARE. Editorial Aba
Ya la, Quito , Ecuador. 67 p.

446 Selected Publications

Yencho, G.C., S.P. Kowalski, R.S.

Koba yas hi , S.L. Sinden , M.W.
Bo nierbale, and K.L. Dea hl , 1998. QTL
mapping of foliar glycoa lkalo id agl yco nes in Solanurn tub erosurn x 5.
berth aulti i potato progenies: Quantitative
var iat ion and pl ant secon dary meta boli sm . Theor. Appl. Gen et. 97(3):563-574.
Zedd am, J. - L. , A . Poll et, S. Ma ngoendiharJO,
T.H. Ramadhan , and M. Lopez -Fer ber.
1998. Occ urren ce and v irulence of a
gran ul osis viru s in Phthorirnaea
opercule/Ja (Lep. , Ce lechiidae) pop ul ations in Indone sia. J. In verte brate Pathology 74 (1 ):48-54.
Zhan g, D.P. , M. Ghislain , Z . Huaman , A.
Go lm irzaie, and R.J . Hijm ans . 1998.
RAPD variation in sw ee tpo ta to [lpomoea
batatas (L. ) Lam ] cu lti va rs from South
Ame ri ca and Pap ua N ew Guinea.
Genetic Resources Crop Ev. 45 (3):27 1277.
Zhan g, D.P. , W.W. Co llin s, and M .
A nd rade . 1998 . Genotype and fertilization effects on trypsin in hib itor activity in
sweetpota to. H o rtS c i. 33:2 25-228.

Staff in 1997-1998
Impact on a C ha ng ing World

Director General-HUBERT ZANDSTRA,

PHO
Deputy Director General for Finance/
Adm in istration-J osE VALLE- R1 EST RA,
PHO
Deputy Director General for
Research - WANDA COLLINS, PHD 5
Deputy Director General for
Research - PETER GREGORY , PHD 6
Director for International
Cooperation - ROGER CoRTllAOUI, PHO
Director General's Office
Christine Graves, MA, Senior Advisor 1
Edward Sulzberger, MS, Senior
Advisor 6
Mariella Altet, External Re lations
Manag e r
Ruth Arce, Administrative Assistant
Marcela Checa, Administrative
Assistant
Maria Elena Lanatta, Bilingual
Secretary
Lilia Salinas, Administrative Assistant
Gladys N ey ra, Administrativ e Assistant
Haydee Zelaya, Administrativ e
Assistant, Internationally Recruited
Staff Officer
Office of the Deputy Director General
for Research
Edward French, PhD, Associate
Director for Research 6
Jose Luis Rueda, PhD, Coordinator,
Andean Natural Resources 6

* Project leader
1 Jo ined du rin g 1998
2 Left during 19 98
3 Funded by special project
4 Joint Appointment
5 Joined duri ng 1997
6 Left during 1997

Crop Improvement and Genetic

Resources Department
Merideth Bonierbale, PhD, Senior
Potato Breeder, Head s
Carlo s Arbizu, PhD, Andean Crnps
Sp ec iali st
Dap e ng Zhang, PhD, Plant Breeder *
Nelly Espinola, MS, Nutritionist,
Res ea rch Associate
Ma re G hi s I a in, PhD, Mo I e cu I aiBio Io g i st
Ali Golmirzaie, PhD, Geneticist
Michael H ermann , PhD , Andean Crops
Speci a li st5*
Miguel Holl e, PhD, Andean Crops
Coordinators*
Z6simo Hu aman, PhD, Germplasm
Curator*
Sven Jacobse n, PhD, Plant Breeder 5
Juan Land eo, PhD, Plant Bre eder
Carlos Ochoa, MS, Taxonomist,
Scientist Emeritus
Alberto Salas, Agronomist, Research
Associate
Maria Scurrah, PhD, Adjunct Scientist 1
Bodo Trognitz, PhD, Geneti c ist
Mah es h Upadhya, PhD, Plant Breeder,
Principal Scientist*
Cesar Aguilar, Agronomist, Resea rc h
Assistant, Field / Greenhou se
Super visor (San Ramon )
Walter Amoros, MS, Agronomist,
Res ea rch Associate
Jorge Benavides, Biologist, Research
Assistant
Raul Blas, Research Assistant 1
Rolando Cabello, Agronomist,
Re sea rch Assistant
Patrici a Cipriani, Biologist, Research
Assistant
Lorena Danessi, Bilingual Secretary
Silvia d e la Flor, Bilingual Secretary

CIPProgcamReporl 1997-98

447

Luis Diaz, Agronomist, Resea rc h

Assi sta nt
Jorg e Espinoza, MS , Agro nomi st,
Researc h Ass i stan t
Ma nu e l Gastelo, MS , Ag ronomi st,
Res ea rc h Assistant
Ren e A . Gomez, A g r o n o mist , Resea rch
Assistant
Ma ri a Lu isa Gu evara Fujita , Biol og i st 1
Carm en H e rrera , Biologist, Research
Ass ist a n t 1
Maria del Rosario H er r era, Biolo g i st,
Re sea rc h As sist a n t
Tere sa l coc h ea, PhD , Pathologist 6
Maritza Luqu e, Bilin gu a l Sec ret ar y
Mariana Martin, Bilin g ual Secretary
Elisa M ih ov ilovich, MS, Biol og ist ,
Rese a rc h As sistant
Luis H . Nopo, Biolo gis t, Resea rc h
Assistant
Matild e Orrillo, Biologist, Resea rch
Assistant
Ana Lu z Panta, BS, Biologist, Researc h
Assistant
Letici a Po rt al, BS , Biologist, Resea rc h
Assis tan t
Dani e l Rey noso , MS, Ag ro nomist,
Res ea rc h As sistant
Flor de Maria Rodri guez, BS , Researc h
Assistant
Rosa Salazar, Bilin g u a l Secretary
Judith Tole d o, BS, Bio l og i st, Research
A ss i sta nt
Fann y Vargas, BS, Ag ronomi st,
Resea rc h Ass istant

Crop Protection Department

Luis Salazar, PhD , Viro l og ist, Prin c ipal
Scientist , Head*
Tere sa Ames, PhD , Scientist Em eritu s
Manuel Ca nt o, PhD , Ne mat o lo g i st
Faus to Cisneros , PhD , Ento mo lo gist*
Ed wa rd Fr e n c h , PhD , Scientist
Em e ritu s 1
Segund o Fuentes, MS , Plant
Pat h o l ogist, Resea rc h Associate
Guil leme tt e Garr y, PhD ,
Phytopathologist, Ass ociat e Expe rt 1
Stefan Keller, MS, Agro nomist,
Ass oc i at e Scientist 1
Aziz Lag naoui , PhD , Entomologi st*

448 Stott

Charlotte Lizarr aga , MS, Plant

Pathologist, Research Associate
Rebecca Ne ls o n , PhD , Mo le c ul a r
Pathologist*
Ma ri a Pa l ac io s, Bi olog i st, Resea rc h
Associate
Sy lvi e Priou, PhD , Bacteriologist 3
Maddal e na Qu erc i , PhD , Molecular
V ir o l og i st *
Ma rc Sporleder, MS, Agronomist ,
Associate Sci e ntist 1
H ebe rt Torres, MS, Plant Path o l og ist,
Research Associate
Lod J. Tur ke nst een, PhD , A djun ct
Scientist (Ne th e rl a nds )
Jea n-L ouis Z edd am, PhD,
En tomov irolo g i st (IRD , formerly
ORSTOM) 4
Jesus Alcazar, MS, Ag ronomist ,
Research Associate
Pedro A l ey, MS , Plant Pathologist,
Research Assistant
Jeni Barboza, Researc h Assistant 1
Id a Bartolini, MS, Biochemist,
Research Ass i sta nt
Ju an Cabre ra , Agro n o mist, Re sea rc h
Ass i sta nt2
Ve ro ni ca Caned o, Biologist, Research
Assistant
Carlos Chuquillanqui, BS, Agronomist ,
Research Assistant
Christian Delg ado, MS, Bioch em ist ,
Research Assistant
Judith Echegara y, Research Assistant 2
Violeta Fl ores, Biologist, Researc h
Ass i st a nt
Soledad Gamboa, Biologist, Research
Assis t a nt
Vero niqu e, Ger a rd , Biologist, Research
Ass ist an t 2 3 5
Rosa Ghilardi, Bilingual Secretary
Erw in Guevara, Ag ronomist, Research
Ass i stan t
Lili a m Gutarra, Agro nomist, Resea rch
Ass i sta nt
Ana Hur ta do , BS , Biolo gis t , Res ea rc h
Assistant
Angela Ma to s, Agronomist, Resea rch
Assistant 2
Norma Mujica, Agro nomist, Research
Ass i sta nt

Giovanna Muller, Biologist, Resea rch

Ass i st an t 3
Ri ca rd o Or rego, Agronomist, Resea rch
Ass i stant
Wi l mer Perez, Agronom i st, Research
Assista nt
Karina Petrovich, Bilingual Secretary
Magno I ia Santa Cruz, Biologist,
Research Assistant
Mirt h a Soldevilla, Secretary
Ana Ma r ia Tab oada, BS, Biolo gist,
Resea rch Assistant
Jorge Tenorio, Biologist, Research
Assista nt
Ernesto Ve l it , B i olog i st 6
A l c ir a Vera, Biolo g i st, Research
Assistant
Juli a Zam udio , Bilingual Secr eta ry
Octav i o Zega rr a, Bi o l og i st, Resea rch
Assistant

Production Systems and Natural

Resources Management Department
Rob e rto Qu i roz, PhD , Land Use
Systems Specialist, H ead 3 *
Walter Bowen, PhD, Nutrient Cyc lin g
Specialist (IFDC) 4
Cat h er in e Brabet, PhD, Food Scient i st 3
Eli as Mujica, MS, Anthropologist ,
Adjunct Scientist , CONDESAN 3
Robert J. Hijm a n s, MS, Geographic
In format ion Scient i st3
Car l os Leon-Ve l arde, PhD, Animal
Production Systems Speci a li st (ILRl )4
Aart Osma n , MS, Associate Exp e rt 3 5
Noe l Pallais, PhD, Physiologist, H ea d
of Seed Unit
Joshua Posner, PhD, Agronomist,
Coord inator, CONDESAN 3 5
Mario Tapia, PhD, Agroecologist,
CON DESAN 3
Marian va n H a l , MS , Assoc i ate
Ex pe rt 3 5 2
Marfa de l os A n ge l es Laura , Bilingua l
Secre t a r y, CONDESAN
Gu ill e rmo Baigorria, MS ,
C lim ato lo gis t , Research Ass i stant'
C ir o Barrera, MS, Plant Pathologist,
Resea rch Assistant
Aurora Cornejo, Bilingual Secr eta ry

Rosari o Fa l co n , BS , Biologist,
Research Ass i stant
Carl a Fe rr adas , Bilingu a l Sec r etary 2
Enriqu e Grande , Tec hnici a n
Aldo Gutarra, BS, Resea rch Assistant '
Lui sa Hu acc ho, MS, Reseat"ch
Assistant 1
Ana Marfa Ponce, Eng., ln foA ndin a,
CONDESAN 3
Jorge Reinoso , MS, Agric. Ecnomist ,
Puno , Per u 3
Jorge Roca, Bi o l ogist, Research
Assistant 2
Ivonn e Va ldi zan, Bilingual Secretary
Percy Zorogastua, MS, Research
A ssista nt '

Social Sciences Department

Thomas Walker, PhD, Economist,
Prin c ipal Scientist, Head*
Th omas Bernet, MS, Economist, Sw i ss
Associate Expert 3 5
Rub e n Dario Estrada , MS, Nat ur a l
Reso ur ces Economist ( C/AT) 3
Hu go Fano, MS, Economist, Resear c h
Associate
Oscar Ortiz, PhD, Special Project
Coordinator
Sonia Sa l as, MS, Food Technologist,
Research Assoc i ate
Gre gory Scott, PhD , Eco nomist *
Paul Winters, PhD , Econom i st 3 5
Ro sario Ba say, BA, Economist 6
Beatriz Cabrejos, Bilingual Secretary 2
Cristina Fons eca, MS, Agronomist ,
Re searc h Assista n t
Luis Ma ld onado, BA, Economist,
Res ea rch Assistant
Eliana Mogni, Bilingual Secretary'
Joann e Sea r s, MA, ICRTCR Assistant'
Victor Suarez, Statistician, Assistant
Zandra Vasquez, Bilingual Secretary
Latin America and the Caribbean
(LAC)
Regional Office
Lima, Peru
Fern ando Ezeta, PhD, Agronomist,
Regiona I Repr ese ntative
Alberto Gonza le s, MS, Research
Assoc i ate, Phytopathologist

CIPP1ogram Repo11 I9979B

4 49

Isabel Mel , Bilingua l Sec retary

Sve n V illagarcia , PhD 6

Sub-Saharan Africa (SSA)

Regional Office
Nairobi, Kenya

Cochabamba, Bolivia

Peter Ewe ll , PhD , Economist, Reg i o n a l

Repr ese ntati ve
Edward Carey, PhD, Regional
Sweetpot ato Bree der
Ramzy El-B ed ewy, PhD , Plant Bre ede r
V ital H age nimana , PhD, Food
Sci e nti st 3
H ai l e M. Kid an e-Ma riam , PhD ,
Re g ion al Pot ato Seed Specia list,
Plan t Bree d er 2
Simon G i c hu k i , MS, Breeder 6
B erha n e Kif I ewa h id , PhD , AS A R ECA/
C IP, Coord in ator, Technolog y
Tr ansfer Proj ec t
V ir g ini a Kirumba , Administrati ve
Secr eta r y
Thom as Mc har o, MS, Swe etpotato
Br ee d ing, Re sea rch Assis t ant
Humb e rt o Mendoza, Ph D , Assoc i ate
Reg i o nal Repr ese ntati ve 6
Ros e ma ry M uttun gi, Secretar y
G eo rge Ng und o, BS, Tec hni cian , PQS
Br ee din g, Resea rch Ass istant
A li ce Nj o rog e, Sec r etary
Simon Obaga, Accou nts
Peter Oj ia mbo , MS, Patholog i st,
Rese arc h Ass i stant

Andr e D evaux, PhD , Pr o j ect

Coo rdinator, Agronomist , Reg ion a l
And ean Program / SOC (formerly
PROINPA )3
Enri que Fernande z-No rt hcote, PhD ,
V ir o l ogis t 3 6
Jav i er Fra n co, PhD , Ne m atolog i st 3 6
G ra h am Th i e l e, PhD , A nt hropo logist ,
Regio n al A ndean Program / SOC
(fo rm er l y PROINPA) 3

Quito, Ecuador
Charles Crissman, PhD, Econom i st,
Li a i so n Scientist
Gregory A. Forbes, PhD , Pl ant
Patho lo g i st
Pedro Oyarzun, PhD , M yco l og i st,
Reg i ona l An dean Program / SOC
(forme rl y FORTIPA PA )3
Steve She r w ood , MS , Tr a in ing
Specia li st
Magaly Asp i az u , Admini str at iv e
A ss istant
Susana Ba rri g a, Account a nt
Lili an Basantes, Trainin g, Researc h
Assista nt
Mari a Gab r ie l a Chac o n, MS,
Patholo gy, Research Ass i sta nt
Ly nn Er e l y us , PhD , Path o l ogy,
Resea rch Ass i stant
Lui s Esc ud ero, Agron o mi st, Research
As sistant
Patricio Espinoza, MS, Ec o n o mist ,
Resea rch Ass istan t
Corrin e Fankhauser, MS, Assoc i at e
Ex pert "
Fran c isco Ja rrin , Patho l ogy, Resea rch
A ssis t a nt
Jul io Mol in ero, Pathology, Research
Assista nt
Fab i a n Mu n oz, Statisti cs, Resea rc h
Ass i stan t
Mari a na Perez, Nurs e, Re sea rch
Assista nt
Ri ca rdo Rodriguez, Eng. Agr.,
Exp e riment Station Superint e nd e nt

450 Stoff

Bamenda, Cameroon
Josep h Koi , MS, Ag ronomist , Liaison
Sci e n tist

Kampala, Uganda
Nicole Smit, PhD , Regional
Entomolo g ist, Li a ison Scienti st
Be rga Le maga, PhD, Pathologi st,
Reg i onal Resea rch Fellow, Af ri ca n
Hi gh l an ds lniti ative / CIP
N.B. Lu ta l adio , PhD , ASAR EC A / C IP,
Coord ina t or, PR A PA CE
Joseph Ot i eno , PhD , Geograph er,
Rock efe ll er Fo undation Fe ll ow 1
Vi nc ent Og ir o, BS, Research Ass i st ant
(Soroti) 1

South and West Asia (SWA)

Regional Office

New Delhi, India

Sarath ll angantileke, PhD , Post h arves t
Sp ec ia li st , Region a l Representati ve
Su sh m a Arya , MA/ MS , Program
Coordin ator, Accountant
T.R. D aya l , PhD , Plant Breede r
M.S . Kadi a n , PhD , Agronom i st
V.S. Khat ana , PhD , So c ioeconomi st
K.C. Th aku r, PhD, TPS Breeder
L. Mony, BS, Secretary

Islamabad, Pakistan 3
Oscar A. Hidalgo , PhD , Patholo g i st,
Pr oject Lea der, Proj ec t CIP / SDC
Atif Man zoo r, BS, Ac co untant
Zareen Siddiqi , BA, Sec retary 1

Nin a Li sna Ningsih, MS, Plant Breeding,

Resea rch Assistant
Ru sma di , BS, Agr o nom y, Re sea rc h
Ass i sta nt
Tjint oko h ad i , BS , Socioeconom i cs,
Resea rch Ass i stant
War sit o, BS, Pest M a n age m ent,
Resea rch Assistant '
Ca ec ili a Afra Wid yas tuti , BS ,
Ag ri cu ltur a l Ex t en si on

Lembang, Indonesia
Enriqu e Chu joy, PhD, Potato
Specia li st*
lnn e H erni ya ti , Offi ce Man age r/
Accounta nt
lstanti Surv iani, BS, Agric ultur al
En g in ee r, Resear c h Assistant

Kathm andu , Nepa/ 3

Deepak Ojha , PhD , Reg. Coord. / CR
Binod Saha, PhD , Re g. Coo rd ./ MWR
Basant Th apa, PhD, Reg. Coord ./ FWR
Bahadur Bhandari, BS, Accountant
Puspa Ra i , BA, Se c r eta ry

East, South East Asia, and the Pacific

(ESEAP)
Regional Office

Bogor, Indonesia
Gordon Pr a in , PhD , So c ial
Anthropologist , Reg ional
Repr ese nt ative
Ann Braun, PhD, Ecolo gist 6
Keith Fu g li e, PhD, Agri c ultural
Econ o mi st 1
II-Gin Mok , PhD , Plant Breed e r
Christopher Oates , PhD , V isitin g
Scienti st, Food Te c hnology
Dai Pet ers, PhD, Rural D evelopm ent
Sp ecia li st 3
Elske va n de Flierl, PhD , IPM
Spe c iali st 1
Christoph e r Wheatl ey, PhD,
Posth arvest Specialist 6
lrfan sya h, BS, Food Tec hnology,
Researc h Assistant 2
Dessy Ku sba ndi , Ex ec utiv e Secret ar y
Sukendr a Mahalaya, BS , Agronomy,
Inform atio n Manag e m ent
Coordinator
Kusye Naw aw i , Ac co unt ant

Los Banos, Philippines

Dind o Camp i Ian, PhD , Soci o l ogist,
Co o rdinator, UPWARD
Ch erry Leah Ba ga l ano n, MS, Human
Eco l og ist, Pro g ram Associat e
Lorn a D arlene Be luli a, BS , Acco untant
Raul Boncodin, BS , Ethnobotani st,
Pro gra m Associate
Fred erika Vogel, BS, Agricultur a l
Eco nomist, Dut c h Associat e Ex p e rt3

Baguio, Philippines
Up al i Ja yas in ghe, PhD, V irol ogis t ,
Li ai so n Scienti st *

Beijing, China
Yi W an g, PhD , Pl ant Physiologi st,
Li ai so n Sci e ntist 5
Son g Bofu, PhD, Li aiso n Sci e nti str'
Yupin g Bi, PhD, Mol ec ular Biolo g i st,
Resea rch Assist ant
Ka i yu n Xi e, PhD , Potato Sp ec i ali st ,
Resea rc h Ass i stant
Ji e Ji ang, BS, Secr etary
Lin Yu an, BS, Admini strative Assistant

Central and Eastern Europe, Transcaucasia and Central Asia (ECA)

Berlin, Germany
Pet er Schmied ich e, PhD, Pl ant
Br ee der, Coordin ator

CIPProgrornReportl99798

451

Egypt

( until November 1998)

Kafr El-Zayat, Egypt

Ramzy El-Bedewy, PhD, Plant Breede r,
Li aison Scientist

Training Unit
Patricio Malagamba, PhD, Head
Nelson Espinoza, Biologist, Training
Spec i alist
Martha Hu a nes , Training Coordinator
Mercedes Suito , Bilingual Secretary
America Va lde z, MS , Training Material
Specialist
Office of the Executive Officer
Jose Luis Rueda, PhD, Executive
Officer 6
Cesar Vittorelli, Eng.A g r. , Act in g
Exec utiv e Officers
Gloria Solis , A dmini strative Assistant
Veron i ca de Armero , Guest House
Supervisor
Human Resources
Lu cas Reano , CPC, Human Resources
Manager
Janneth Carbal I ido, Compensation &
Benefits Assistant
Monica Ferreyros, Auxiliary Services
Supervisor
David Halfin, MD
Sor Lapouble, Auxiliary Ser v ices
Ass i sta nt
Estan i slao Perez, Compensation &
Benefits Assistant
Martha Pierola, Social Worker,
Supervisor
Lu cero Schmidt, Nurse
Marfa Amelia Tavar a, Bilingual
Secretary
Yoner Varas, Comp ensation & Benefits
Assistant
Logistics & General Services
A ldo Ta n g, Comdr.(r. ), Logistic &
Genera l Services Manage r
Arturo A l varez, Purchasin g Supervisor 2
Pilar Bernui , Bilingual Secretary
Silvia Cordova, Bilingual Secretary
Hugo Davis, Ve hicl e Mainten a nc e
Off i ce r

452 Stoff

Ximena Ganoza , Purchasing

Super v isor
Ati li o Guerrero , Ve hicle Pr og r ammer
Jorge Locate ll i, Capt.( r.), Security
Super v isor
Jorge Lu que, MBA, Wareho us e
Supervisor
M i che lin e Moncloa, Fron t Desk
Super v isor
Anton i o Mo r il l o , Ma intenan ce
Super viso r
Jo se Pizarro , Purchasing Super viso r
Carmela Sa l azar, Bilingual Secretary
lt alo Solari, Electronic Technician
Sandra Va lve, Bi lin gual Secretary 2
Djordje Vel icko v ich, Pi lot
Pe rcy Zuzunaga, Pi lot
Ana Marfa Secada, Tra ve l Office
Supervisor

Office of the Chief Financial Officer

Carlos Nino-Ne i ra, CPA , CFO
Ama li a Lanatta , Adm in ist rati ve
Assistant
Accounting Unit
Sandra Albarrac in , Accou nt ant
Elian a Bardalez, CPA, Senior
Accountant
Ed gardo de l os Rios, CPA, Senior
Accountant
Rodmel Guzman , Accoun t a nt
Assistant 1
Blanca Joo, CPA , Acco untant
Silvia Loayza, Bilingual Secretary 1
Sandra Odrfa, Bilingual Secretary 2
Fer n ando O l aechea, Accountant 2
Ernesto Olive r a, Acco untant
Rosario Pastor, CPA, Senior
Accountant 5
Mi la gr os Patino , BA, Accountant
Eduardo Peralta , Accountant
Carmen Ramos , Bi lin gual Secr etary 1
Mig u e l Saavedra, CPA , General
Accoun t ants
Cesar Tapia, Accou ntant Ass is ta n t
Budget Unit
Alberto Mo nt eblanco, CPA, Senior
Accountant
Treasury Unit
Denise G i acoma, CPA , Treasur er
Sonnia So l ar i , Ch i ef Cashier

Communications Unit
Steven Kea rl , MS, Senior Writer/Editor,
H ead 5
Candelaria Atala ya, Photogr a ph e r 1
Mariella Corvetto, Communi cat ion
Servi ces Coordinator
Ruth D e lgado, Exhibits/Displ ay,
Assistant
Nini Fernandez-Concha, Graphic
D es i g ner, Assistant
Amparo Galindo, Bilingual Secretary
Milton Hidalgo , Graphic Designer,
Assistant
Ceci I i a La fosse, Chief Desi g n e r
Godofr edo Lagos, Print Chief
Victor Madrid, Graphic Desi g n e r,
Assistant
Susana Menacho, Bilingual Secretary 2
Ans e lmo Morales, Graphic Designer,
Assistant
Ana Luisa Munoz, Photogr a ph y
.A ssistant
Felix Munoz, Publications , Assistant
Zoraida Portillo, Writer/Spanish Editor
Alfr edo Puccini, Graphic D es ig ne r,
Assistant
Cesar Rossenouff, Photograph e r 2
Information Technology Unit
Anthony Collins, MS, Head
Monica Arias, BE, User Support
Lili a n a Bravo, BE, User Support
Andrea Caceres, User Support
Rob e rto Castro, Eng., System s
Dev e lopment2

Mois es Fe rn a nd ez, Systems Analyst,

Adm in. Systems ( DBA ) 1
Jose Navarrete, Systems Support
Pia Marfa 01 id e n, Databas e
Administrator
Giancarlo Rodriguez, User Suppon 1
Eric Rom ero, Eng., Systems
Administr at or
Edgardo Torres, Eng., Systems
Dev e lopm e nt Analyst
Alberto Velez, Eng., Network
Administrator

Library
Ceci I ia Fe rr ey ra, H ea d
Carm e n Arnillas, Bilingual Secr eta ry
Griseld a Lay, Librarian, Assistant
Glenda N eg rete, Librarian, Assi st ant
Field Research Support
Victor Otazu, PhD, Head
Lombardo Cetraro, Field / G ree nhou se
Sup e r v i so r (Sa n Rarn6n ) 2
Roberto Duarte, Eng. Agr., Fi e ld /
Gre e nhou se Supervisor (La Molina)
Hugo Goyas, Eng. Agr., Field/
Greenhou se Supervisor (Huancayo)
Carmen La ra, Secretary
Statistics
Alfredo Garcia , MS, Experiment a l
Stati stics 2
Felip e d e Mendiburu, Statisti cs
Engin ee r, Assistant

CIP PrngrornReporl 1997-98

453

Acronyms and
Abbreviations
lmpacl on o Changin g World

ADFA
AFLP
ARTC
ASAR
AUD PC
BU
CAPS
CEC
CIAT
C l MM YT
CON DE SAN
CPR I
CPRO
CpT I
CMD
DEM
DNA
ELISA
ESARC
FA O
FFS
FMV
FORTI PA PA
FU
GXE
GIS
GUS
Gp i
h.a. i.
HDP
HH
HV
IARC
ICAR
ICRAF
ICRISAT
IFAS
IFAD

avera ge of di seased fo li age area

amplifi ed fragment length polymorphi sm
Andean root and tuber c rops
A soc iac i6 n de Serv ic ios Artesanales y Rurales (Bolivi a)
area und er th e disease progress curve
bli ght unit
c leaved amplifi ed polymorph ic sequ ence
Cation exc hange capac ity
Centro Interna c ional de Agricu ltura Tropi ca l (Co lombia)
Internation al Maize and Wheat Improvement Center (Mex ico)
Cons o rt iu m for the Sustainab le D evelo pm ent of the A nd ea n Ecoregio n
Central Potato Resea rch In sti tute (Indi a)
Centre for Plan t Breeding and Rep rod ucti on Research (Netherl and s)
cowpea tryp sin inhibitor
cassava mo sa ic virus disease
di gital eleva tion m odel
deoxy ribonu c leic ac id
enzyme-linked immun oso rbent assay
Ea st and Southern Afr ica Reg ional Centre (Kamp ala, Uga nda )
Foo d and Agricu lture Organ ization of th e United Nations
farm er fi eld schoo l
feather y mottl e v iru s
Fortal eci mi ento de la ln ves ti gaci6n y Proclu cci6 n de Se mi Ila de Papa en
el Ec uador
fung ici de unit
ge notype by environm ent
geographic inform ation systems
b- glucuronidas e
glu cose-6-phosphate iso merase
hours afte r ino cul at io n
hi gh d em and and produ ct io n gro wth
li ke pol ari zed
cro ss polarized
intern ational agricu ltural research ce nter
Indi an Counci l of Agricultural Researc h
International Centre for Re search in Agroforestry
Intern ation al Crops Resea r-ch Institute for th e Semi-Ariel Tropics
immunofluoresce nce antibody stainin g
Intern ation al Fund fo r Agricultural D eve lo pm ent

CIP Program Report 1997-98

455

IFPRI
llTA
IMPACT
INERA
INFOANDINA
INIAP
INIA
INRA
INTA
IPGRI
IPM
IRRI
ISO
KARI
LB
LMF
MENA
MERCOSUR
NAFTA
NARO
NARS
NASH
NCM
NGO
N/m 2
OFE
OM
pep
PAGE
PARC
PCR
Pl
Pl CTI PAPA
PPO
PRAPACE
PR ECO DEPA
PROINPA
PSTVd
PVY
QTL
RAPD
rDNA
RFLP
RKN
RRA
SAAS

International Food Policy Research Institute (USA)

International Institute of Tropical Agriculture (Nigeria)
International model for policy analysis of agricultural commodities and
trade (IFPRI)
national potato and sweetpotato programs of the national agricultural
research institute of D.R. of the Congo
CONDESAN information netwo rk
lnstituto Nacional Aut6nomo de lnvestigaciones Agropecuarias (Ecuador)
lnstituto Nacional de lnvestigaci6n Agraria (Peru)
lnstitut National de la Recherche Agronomique (France)
Institute Nacional de Tecnologia Agropecuaria (A rgentina)
International Plant Genetic Resources Institute (Italy)
integrated pest management
International Rice Research Institute
International Organization for Standardization
Ken ya Agricultural Research Institute
late blight
leafminer fly
Middle East and North Africa (C IP region)
Southern Common Market
North American Free Trade Association
National Agricultural Research Organization (Uganda)
national agricultural research systems
nucleic acid spot hybridization
n itrocel lu lose membrane
nongovernmental organization
Newtons/m 2 (1 kg m per second squared )
overland flow element
organic matter
peptidase
polyacrylamide gel electrophoresis
Pakistan Agricultural Research Center
polymerase chain reaction
proteinase inhibitor
Programa Internacional Cooperativo del Tizon Tardio de la Papa (M exico)
project planning by objective
Programme Regional de I' Amelioration de la Culture de la Pomme de
Terre et de la Patate Douce en Afrique Centrale et de l'Est (CIP network )
Programa Regional Cooperati vo de Papa (CIP network in Central America
and the Caribbean)
Proyecto de lnvestigaci6n de la Papa, National Potato Research Program
(Bolivia)
potato spindle tuber viroid
potato virus Y
quantitative trait loci
randomly amplified polymorphic DNA
ribosomal DNA
restriction fragment length polymorphism
root-knot nematode
rapid rural appraisal
Shandong Academy of Agricultural Science (China)

456 Acronyms and Abbreviations

SAPPRAD
SARR NET
SBTI
SCRI
SDS
SEIN PA
SKTI
SLW
SPCFV
SPCSV
SPFMV
SPMSV
SPLV
SPMMV
SPSVV
SPVD
SWP
SSR
t

TAC
TPS
UNICEF
UPGMA
UPWARD
WEPP

Southeast Asian Program for Potato Resea rch and D eve lopment
Southern Africa Root Crop Research Network
soybean trypsin inhibitor
Scottish Crop Research Institute
sodium dodecyl sulfate
Semilla e lnvestigacion en Papa (Peru)
soybean kunitz trypsin inhibitor
silverleaf whitefly
sweetpotato chlorotic fleck virus
sweetpotato chlorotic stunt virus (see SPSVV)
sweetpotato feathery mottle virus
sweetpotato mild speckling virus
sweetpotato latent virus
sweetpotato mild mottle virus
sweetpotato sunken vein virus
sweetpotato virus dis ease
sweetpotato whitefly
simple sequence repeats
metric ton (1000 kg/2,200 lbs.)
Technical Advisory Committee of the CGIAR
tru e potato seed
United Nations International Children's Emergency Fund
unweighted pair group method with arithmetic mean algorithm
Users' Perspective with Agricultural Res ea rch and Development
(CIP network)
Water Erosion Pred iction Project

CIP Program RepO!l l997-98

457

Scientific Direction
W and a Colli ns

Managing Editor
Princess Ferguson

Editors
Caroline Arthur
Princess Ferguson
Bil l Sm ith

Production Coordinator
Cec i lia Lafosse
Alfredo Puccini, Layout design
Victor M adrid, Graphics
M ilton H idalgo, Text
Nini Fernandez-Concha, Cover design

Printing Coordinator
Godofredo Lagos
Photos in th is report were taken by
G. Aguirre, V. Canedo, G. Cipr ian i,
N. Esp inola, R. Haddad, M. Hermann,
R. O rtega, C. Rossenouff, and from
the CIP collection.

CGIAR

F UTURE
HAR'rfEST

The International Potato Center (CIP) is a

sc ientific, non profit institution dedicated
to the increased and more sustainable
use of potato, sweetpotato, and other
roots and tubers in the developing world,
and to the improved management of
agric ultural resources in the Andes and
other mountain areas.
CIP is part of the global agricultural
research network known as the
Consultative Group on International
Agri cu ltural Research (CGIAR).
CIP supports Future Harvest, an ini tiative
that builds understanding about the w ider
social benefits of improved agricu lture:
peace, prosperity, environmental renewal,
health, and the alleviation of human suffering.