Genet Resour Crop Evol (2008) 55:115131 DOI 10.

1007/s10722-007-9219-4

R E S E A R C H A RT I C L E

Diversity, distribution and management of yam landraces (Dioscorea spp.) in Southern Ethiopia
Muluneh Tamiru Heiko C. Becker Brigitte L. Maass

Received: 26 September 2006 / Accepted: 24 January 2007 / Published online: 11 April 2007 Springer Science+Business Media B.V. 2007

Abstract Yam (Dioscorea spp.) is widely grown in many parts of Ethiopia and plays a vital role in local subsistence. Nevertheless, its diversity has not been studied in detail. A survey covering 339 farm households and eight districts was conducted in the major yam growing regions of Southern Ethiopia to investigate the diversity and distribution of yam landraces using structured and semistructured questionnaires. A total of 37 named landraces were recorded, with a range from one to six (mean 2.9) on individual farms. Farmers decisions regarding the number and type of landraces maintained was inuenced by tolerance of the landraces to drought, their maturity time and market demand. Most landraces had limited abundance and distribution, and only a few dominant landraces were widely grown. There was also variation amongst districts with respect to diversity, distribution and abundance of the
M. Tamiru B. L. Maass (&) Department of Crop Sciences: Agronomy in the Tropics, Georg-August-University Goettingen, Grisebachstr. 6, Goettingen 37077, Germany e-mail: bmaass@gwdg.de H. C. Becker Department of Crop Sciences: Plant Breeding, GeorgAugust-University Goettingen, Von-Siebold-Str. 8, 37075 Goettingen, Germany M. Tamiru Hawassa University, P.O. Box 05, Awassa, Ethiopia

landraces found. In the majority of the localities surveyed, farmers reported a decreasing trend in the number of landraces maintained on individual farms and in the overall yam production. Besides, in those limited areas where yam production is expanding, farmers are increasingly relying on a few selected landraces that mature early. Findings of this study suggest that local farmers in Wolayita and Gamo-Gofa zones maintain considerable yam diversity that remains to be further explored for sustainable utilization and conservation of the available genetic resources. Keywords Dioscorea Ethiopia Genetic resource Landrace diversity Yam

Introduction Yam (Dioscorea spp.) belongs to the genus Dioscorea, representing more than 600 species worldwide (Coursey 1967). The Dioscoreales are believed to be amongst the earliest angiosperms that originated in Southeast Asia, but followed a divergent evolution in three continents separated by the formation of the Atlantic Ocean and desiccation of the Middle East (Hahn 1995). As a result, the major food species occur in three isolated centers: West Africa, Southeast Asia and tropical America (Alexander and Coursey 1969). These centers are also considered areas for

123

116

Genet Resour Crop Evol (2008) 55:115131

independent yam domestication, and represent considerable diversity (Asiedu et al. 1997). Yam is a crop of major economic and cultural importance in sub-Saharan Africa that accounts for about 95% of the world production (FAO 2004), the so called yam belt of West and Central Africa being the principal area of production (Coursey 1967; Hahn et al. 1987). Following the establishment of research institutions such as the International Institute of Tropical Agriculture (IITA), yam has attracted considerable research attention in recent decades. Consequently, substantial progress was made in understanding the origin, domestication, phylogeny, diversity and production of the major food species. Orkwor et al. (1998) give a review of the recent advances in yam research. However, much of the studies so far concentrated in the yam belt, whereas little is known about the status of yams in the other parts of Africa. This has led to the perception that yam is an important food crop only in parts of West Africa, a view that triggered concerns decades ago but still is largely valid (Ayensu and Coursey 1972; Quin 1998). In Ethiopia, which is the center of origin and diversity of a large number of crop species (Engels et al. 1991; Harlan 1969; Vavilov and Chester 1951), studies into genetic diversity have mainly focused on cereals. Other crops, including the widely consumed root and tubers, have been relatively neglected by research and conservation efforts. Yams in Ethiopia are hardly known to the scientic community. The country is only referred to as an isolated center of yam cultivation (Norman et al. 1995), where a number of Dioscorea species are grown in complex cropping systems together with cereals, and other root and tuber crops (Westphal 1975). There has been no systematic study on diversity, production and use of the crop. Although brief and passing remarks are available in the more general references (Engels et al. 1991; Westphal 1975), most of these materials contain only lists of one or a few of the yam species found in the country. About 23 indigenous yam types belonging to at least four Dioscorea species were reported in Sheko, Southeast Ethiopia (Hildebrand et al. 2002). The importance of yam for local subsis-

tence and its indigenous knowledge, as well as priorities for conservation and improvement of the crop were highlighted. Miege and Demissew (1997) described eleven Dioscorea species, both wild and cultivated, found in the country. These reports indicate that yam is widely distributed in Ethiopia, and is amongst the main root and tuber crops grown by subsistence farmers in the Southern, Southwestern and Western parts of the country. Nevertheless, the extent and distribution of the available inter and intraspecic diversity is poorly investigated. In situations where documented data are hardly available, the local farmer is the rst source of information to initiate diversity studies. Farmers perception of local varieties is of utmost importance because it is not only the unit of diversity they recognize but also the unit they actually manage and conserve (Hoogendijk and Williams 2002). Yam is a traditional crop that has long been cultivated in Wolayita and Gamo-Gofa zones as co-staple with enset (Ensete ventricosum (Welw.) Cheesman), cereals, and other root and tuber crops (Westphal 1975). As the crop is adapted to dry season planting (mainly at the onset of the dry season in October) early harvests in May ll a seasonal gap in food supply. The fact that it is preferred to the other root and tuber crops means yam is also an important cash crop, generating additional income for farm households (personal observation). This study forms part of a project initiated with the main objective of characterizing the yam diversity in Ethiopia (Tamiru 2006), and aims to investigate farm-level diversity and distribution of yam landraces in Wolayita and Gamo-Gofa zones, the major yam production areas in Southern Ethiopia, and to describe how the landraces are selected, managed and utilized by local farmers.

Material and methods The study area The study area is located approximately between latitudes 646 and 726 N, and longitudes 3701 and 3808 E in the Southern Nations,

123

Genet Resour Crop Evol (2008) 55:115131

117

Nationalities, and Peoples Regional State (SNNPRS) of Ethiopia (Fig. 1), including Wolayita zone and Kucha district from the neighboring Gamo-Gofa zone (Table 1). Wolayita zone is composed of seven districts and 273 peasant associations (PAs), the lowest administrative unit in Ethiopia. The zone is one of the most densely populated areas in the country, with an estimated size of about 4,500 km2 inhabited by around 1.5 million people. This corresponds to an average density of 355 people per km2, which ranges from 141 to 629 people per km2 in Humbo and DamotGale districts, respectively (CSA 2000). The district of Kucha, with an estimated area of 1,384 km2, was considered in the study to investigate the distribution of yam landraces beyond Wolayita. The Wolayita language belongs to the Omotic family, and is closely related to Gamo, a language of the same family spoken by neighbor-

ing farmers in Kucha district. The two languages have lexical similarity of 7993% (Girard 2002). This provides a good setting for studying crop diversity in traditional agriculture based on named landraces with a minimum inuence of language polymorphism.

Sampling and data collection A household-level survey covering eight districts was conducted from October 2003 to September 2004. A stratied sampling procedure was followed to dene the sampling unit. The area was rst stratied in terms of geographic distance and elevation to cover the approximate ecological range of yam so that valid generalizations can be drawn from the results. Four to six peasant associations (PAs) were selected from each

400

400

800

1200 Kilometers

SNNPRS

Bolosso-Sore D a mo t Gale Kind o Koyisha Soddo Soddo-Zuria D a mo t Woyde

Wolayita Zone
Offa Hu m b o

Kucha

(Gamo-Gofa Zone)

Gamo-Gofa Zone

Fig. 1 Location of the study area in Southern Ethiopia, indicating administrative districts and the administrative capital of the Wolayita zone, Soddo (SNNPRS = Southern Nations, Nationalities and Peoples Regional State)

123

118

Genet Resour Crop Evol (2008) 55:115131

Table 1 Description of the districts included in the study of Wolayita and Gamo-Gofa zones, Southern Ethiopia District Areaa (km2) Elevation range (m asl) Meana population density per km2 491 629 236 141 224 234 528 91 No. of PAsb surveyed No. of households interviewed

Wolayita Bolosso-Sore Damot-Gale Damot-Woyde Humbo Kindo-Koyisha Offa Soddo-Zuria Gamo-Gofa Kucha Total
a b

633 429 783 846 776 588 481 1,384

1,8301,980 1,7652,200 1,7772,220 1,6001,832 1,6601,730 1,6001,950 1,8501,950 1,6902,100

4 4 6 4 4 4 4 4 34

40 42 56 42 39 40 40 40 339

Data source CSA (2000) Peasant Associations

district in consultation with district agricultural ofcers and key informants knowledgeable in the area. Then, 10 households were randomly selected from each PA, bringing the total number of PAs and households covered by the study to 34 and 339, respectively (Table 1). The elevation ranges where yam farmers were interviewed varied among districts (Table 2). Lower elevations fall within the warm semiarid climatic zone of Ethiopia, traditionally known as Kola, that is characterized by longer dry seasons and a mean annual rainfall of 200800 mm (MoA 2000). The
Table 2 Number of yam landrace growing farms surveyed at three different altitude ranges, mean farm size per individual farmer, and ratio of land allocated to yam District Number of farms according to elevation Low Wolayita Bolosso-Sore Damot-Gale Damot-Woyde Humbo Kindo-Koyisha Offa Soddo-Zuria Gamo-Gofa Kucha Total
a a

intermediate and higher elevations correspond to the cool sub-humid (Woinadega) climatic zone with the mean annual rainfall varying from 800 to 1,200 mm (MoA 2000). Data were collected through individual interviews with the member(s) in each household responsible for the management of yam elds, using structured and semi-structured questionnaires. The semi-structured questionnaire was included to enable full consideration of openended questions such as how farmers evaluate and identify the different landraces. Most of the
cultivation in Wolayita and Gamo-Gofa zones, Southern Ethiopia (figures in parenthesis refer to minimum and maximum values) Mean farm size (ha) Ratio of land allocated to yam

Intermediate 40 18 40 32 0 9 40 23 202

High 0 24 16 0 0 0 0 10 50 0.70 0.54 0.68 0.69 0.54 0.66 0.68 (0.382.00) (0.251.00) (0.132.00) (0.381.25) (0.131.00) (0.130.55) (0.102.00) 0.16 0.12 0.15 0.08 0.18 0.11 0.15 (0.010.50) (0.030.33) (0.030.50) (0.020.25) (0.040.50) (0.040.50) (0.050.50)

0 0 0 10 39 31 0 7 87

0.99 (0.252.00)

0.30 (0.060.50)

Low (1,5501,750 m asl), intermediate (1,7502,000 m asl) and high (2,0002,225 m asl)

123

Genet Resour Crop Evol (2008) 55:115131

119

respondents were men even though women farmers were also interviewed in places where they were head of the family or responsible for yam production. Since yam is a crop of much economic and social signicance and involves a laborious production system, it is generally considered a mans crop. The number of landraces grown by individual farmers was recorded on farm where farmers were asked to distinguish and name the different landraces. This was conducted during the time of the year when the plants were still growing in the eld to assist identication of the different morphotypes. Data were also recorded on elevation, total farm size, size of land occupied by yams, cultivation practices and uses of the landraces. Besides, farmers were asked to verbally report names of landraces they knew and/or heard about other than the ones they were currently growing on their farms. Data analysis For the purpose of this research, a landrace refers to a morphologically distinct population of yam that farmers recognize, name and manage. Accordingly, a list of all the landraces recorded throughout the study area was summarized after grouping known synonyms with the help of elderly farmers. All data were calculated on a district basis assuming that they could reect a certain geographic pattern. As richness of the district, the overall number of distinct landraces recorded, without accounting for the number of farms where they were found, was considered. As measures of diversity that take into account the proportional abundance of landraces (richness and evenness), Simpson and Shannon diversity indices were calculated for all the districts. Simpsons diversity index (D) basically measures the probability that two individuals randomly selected from a sample belong to the same category (Simpson 1949) and, hence, as D increases, diversity decreases. The index was, therefore, transformed as 1 D so that greater diversity corresponds to higher values:

Simpsons diversity index 1 D 1

ni =N2

where ni represents number of farms where landrace i was found, and N sum of the number of farms where individual landraces were found. Shannon diversity index (H) combines both number and evenness of categories considered, and can be increased either by greater evenness or more unique species (landraces in our case). The index is dened as Shannon diversity index H 0
S X i1

pi ln pi

where s is number of landraces, and pi frequency of landrace i (ni/N). Evenness (E) was also calculated separately as a measure of the ratio of the observed diversity to the maximum diversity. It is dened by the function E = H/ln s, where H is Shannon index and s refers to the number of landraces recorded in each district. High evenness resulting from all landraces having equal abundance is normally equated with high diversity (Magurran 1988). To assess differentiation or beta (b) diversity (Magurran 1988), Srensons similarity index was employed. This index estimates how different or similar habitats are regarding diversity of the categories under consideration, using similarity measures of pairs of sites. The index was computed based on the presence or absence of landraces (qualitative data) to estimate landrace similarity between all possible pairs of districts as follows: S(rensons similarity index 2c a b

where a represents number of landraces in district A, b number of landraces in district B , and c number of landraces common to both districts. Frequency distributions, descriptive statistics, correlations and other relevant data analyses were carried out applying SPSS statistical software (SPSS 12.0.1, SPSS Inc. 2003).

123

120

Genet Resour Crop Evol (2008) 55:115131

Results Landrace diversity Overall, local farmers described a total of 37 recognized yam landraces (Table 3). Of these, two landraces (bola-boye and bunde-buchi) belong to a species of aerial yam (D. bulbifera L.), and are apparently identied based on variations

in shape and size of the bulbils (aerial tubers). However, the same characters vary within a landrace or even among bulbils of the same plant. The remaining landraces belong to a yet unidentied species or group of species (Tamiru 2006) that are distinct from the Dioscorea species widely cultivated in West Africa (Tamiru et al. 2007). Most of these landraces (70%) are earlymaturing types, and are harvested twice

Table 3 Yam landraces described in Wolayita and Gamo-Gofa zones of Southern Ethiopia and respective number of farms in the various districts where they were encountered Landrace Number of farms per district BolossoSore Afrad Arkiyad Ayino or Ayinas Banchuwad Barcha or Barchyad Barchahuwad Bolaboyea Botaboyed Buha, Buhed Buluwad Buna, Bune, or Buniyad Bundebuchia Chamias Chawulas Chichiyad Fara, Furad Gajelas Gasad Genad Hatiye or Hatiyad Lohuwad Machad Maleho or Malehuwad Martabod Molchad Mortawa or Mortabuwas Natrad Olama or Alamad Ochied Ohad Sasas,w Suyitiyad Wadalas Welluwad Wolabua, Walabo, or Walabuwos Woyichas Zorewuwad
d

DamotGale 7 8 1 1 1 20 1 22 1 1 23 10 7 5
a

DamotWoyde 4 1 53 3 56 31 2 1
s,w

Humbo 4 2 1 2 40 1 1 23 33

KindoKoyisha 1 1 14 2 35 38 7 2 38 8

Offa 2 5 1 1 8 1 12 40 4 37 1

SoddoZuria 1 1 3 1 1 4 1 35 1 1 2 26 1 29

Kucha 6 1 1 23 2 3 40 3 1 15 40

9 1 2 3 28 25 3 11 12 11 13

double-harvested;

single-harvested;

aerial yam;

single-harvested and wild

123

Genet Resour Crop Evol (2008) 55:115131

121

(double-harvested). The remaining 30% mature late and are harvested only once. Wild yam, referred to by the name sasa, was encountered only in some localities especially of KindoKoyisha where forest patches still exist. The number of landraces recorded on individual farms ranged from one to six with a mean and standard deviation of 2.9 and 1.1, respectively. The variation among districts with respect to number of landraces per farm across all farms visited is summarized in Table 4. A relatively high number of farms with four or more landraces were encountered in Kindo-Koyisha, Offa and Kucha districts. All the farms surveyed in KindoKoyisha and about 78% of those surveyed in Offa districts were located at elevations below 1,750 m asl., whereas Kucha farms were mostly located at intermediate elevations (Table 2). The total number of landraces recorded in each district (richness) varied from eight at DamotWoyde to 14 at Soddo-Zuria and Damot-Gale districts with a mean and standard deviation of 11.0 and 2.1, respectively (Table 5). Both Simpson and Shannon diversity indices revealed that the neighboring districts of Bolosso-Sore and Damot-Gale were the most diverse, while Damot-Woyde was the least diverse, despite its adjacency to Damot-Gale. As expected, Shannon diversity index was signicantly correlated with number of total (r = 0.69) and unique (r = 0.70) landraces. A similar relationship was observed between Simpson index of diversity and number of total (r = 0.60) and unique (r = 0.62) landraces. Although Damot-Gale and Soddo-Zuria were similar in terms of richness, the latter was found

less diverse partly due to the relatively lower number of unique landraces. The difference between the two districts could also be due to the variation in the abundance of landraces, which was evident from their respective evenness values. The lowest number of landraces, none of which was unique, represented the least diverse district of Damot-Woyde. The similarity between all possible pairs of districts with respect to named landraces was assessed using Srensons similarity index (Table 6). Overall, the similarity between two districts varied from 0.16 to 0.67. Damot-Woyde and Kindo-Koyisha were the most similar districts, followed by Damot-Woyde and BolossoSore, and Humbo and Kucha. On the other hand, the most dissimilar district pairs were Damot-Gale and Offa, Bolosso-Sore and Offa, Damot-Gale and Kucha, and Damot-Woyde and Offa in ascending order of similarity. Both the most similar and dissimilar pairs of districts were among those located relatively farther apart, and this suggested that similarity of districts did not entirely correspond to their geographic distance. Distribution and abundance of landraces There was a considerable difference among the landraces with respect to their distribution across the districts covered (Fig. 2). Eighteen (49%) of the landraces had a narrow distribution and were specic to single districts. The remaining 21 (51%) were recorded in more than one district. But, only two (5%) were ubiquitous, being found in all the districts surveyed. These were the early-maturing

Table 4 Variation in the number of landraces planted per farm across the districts of Wolayita and Gamo-Gofa zones in Southern Ethiopia No. of landraces per farm Number of farms per district BolossoSore 2 11 14 13 0 0 40 DamotGale 13 8 10 7 3 1 42 DamotWoyde 0 11 32 2 1 0 56 Humbo 4 15 19 4 0 0 42 KindoKoyisha 0 2 20 4 12 1 39 Offa 3 22 3 4 5 6 40 SoddoZuria 7 11 14 4 4 0 40 Kucha 0 10 14 9 5 2 40 Total 29 100 126 47 33 4 339

1 2 3 4 5 6 Total

123

122

Genet Resour Crop Evol (2008) 55:115131

Table 5 Yam landrace diversity in the various districts of Woalyita and Gamo-Gofa zones in Southern Ethiopia expressed as richness, Simpson (1 D) and Shannon (H) diversity indices, and evenness District Richness %a of the total 29.7 37.8 21.6 24.3 27.0 29.7 37.8 29.7 Number of unique landraces 2 6 0 2 1 2 3 2 1D H Evenness

Bolosso-Sore Damot-Gale Damot-Woyde Humbo Kindo-Koyisha Offa Soddo-Zuria Kucha

a

11 14 8 9 10 11 14 11

0.85 0.85 0.70 0.72 0.79 0.74 0.76 0.78

2.08 2.14 1.36 1.46 1.76 1.66 1.71 1.75

0.87 0.81 0.65 0.67 0.76 0.69 0.65 0.73

Calculated on the basis of the 37 landraces described throughout the study area

hatiye (hatiya) and the late-maturing wadala. The other widespread landraces included oha, gena, ayino (ayina), and gasa. A similar trend was observed with regard to the abundance (proportion of farms where the landraces were found) of individual landraces. Hatiye and wadala were the most abundant landraces as they were recorded on 86% and 67% of the farms surveyed, respectively (Fig. 3). Most of the landraces (70%) were encountered on less than 3% of the farms surveyed. Furthermore, 12 (32%) landraces were recorded on single farms. Landrace abundance also varied across the districts (Table 3). Few landraces were well represented in some districts, but virtually missing from the others. For example, gajela was encountered on more than 45% of the farms visited in DamotGale. But outside this district, it was only found in the adjacent Bolosso-Sore with a very low abundance. Landraces walabua (walabo) and woyicha, buna (bune), and fara (fura) and zoreuwa showed similar patterns in Bolosso-Sore,

Kucha and Offa, and Kindo-Koyisha districts, respectively. In general, there was a signicant correlation between the distribution and abundance of the landraces (r = 0.85, P < 0.01). Distribution of the landraces throughout the study area and in two selected districts was summarized by the abundance and frequency matrix given in Fig. 4. Most landraces described in this study were local (found in limited districts) and rare (encountered on a limited number of farms in each district) (Fig. 4a). The trend in the least diverse district of Damot-Woyde was similar to the overall study area (Fig. 4b). The landraces described in this district were either local and rare (63%) or widespread and common (37%). In the most diverse district of Damot-Gale, the majority of the landraces was fairly distributed with a relatively lower but comparable abundance (Fig. 4c). This was also reected in the relatively higher evenness of landrace abundance in Damot-Gale (Table 5). In addition to the landraces grown on their farms, farmers verbally reported 46 landrace

Table 6 Srenson similarity estimates of yam landrace diversity between the different districts in Wolayita and Gamo-Gofa zones of Southern Ethiopia on the basis of presence and absence of landraces BolossoSore Bolosso-Sore Damot-Gale Damot-Woyde Humbo Kindo-Koyisha Offa Soddo-Zuria Kucha 1.00 0.56 0.63 0.50 0.57 0.27 0.48 0.45 DamotGale 1.00 0.45 0.35 0.42 0.16 0.36 0.32 DamotWoyde Humbo KindoKoyisha Offa SoddoZuria Kucha

1.00 0.59 0.67 0.32 0.45 0.42

1.00 0.53 0.40 0.52 0.60

1.00 0.48 0.50 0.48

1.00 0.48 0.55

1.00 0.48

1.00

123

Genet Resour Crop Evol (2008) 55:115131

20 16

123

48.6%

12
24.3%

8
18.8%

4
2.7% 2.7%

5.4%

5.4%

0 1 2 3 4 5 6 7 8

landraces encountered on farms of other households visited. The widely distributed landraces, such as hatiye, wadala, oha, gena and gassa were also among those verbally reported as missing. The remaining 19 vernacular names (41%) were new in the sense that they were never encountered on farmers elds during the survey. These additional landraces were mostly reported by a single or two and, at most, by six (about 2%) of the households interviewed. Management of yam diversity and its determinants In Wolayita and Gamo-Gofa zones, yam is cultivated on an annual cycle of planting that commences at the onset of the dry season. The majority of the farmers interviewed (90%) carried out planting in October; whereas very few

Number of landrces

Number of districts where the landraces were found

Fig. 2 Distribution range of yam landraces across the districts surveyed in Wolayita and Gamo-Gofa zones of Southern Ethiopia

names that were no longer found in their community and thought to be lost. Twenty-ve (59%) of these names corresponded to those

% of farms visited
0
Hatiye, Hatiya Wadala Oha Gena Buna, Bune, Buniya Ayino, Ayina Gajela Fara, Fura Wolabua, Walabo, Walabuwo Woyicha Suyitiya Gasa Zorewuwa Arkiya Maleho, Malehuwa Barcha, Barchya Natra Bota-boye Afra Bundebuchi Mortawa, Mortabuwa Banchuwa Buha, Buhe Olama, Alama Sasa Barcheuwa Bola-boye Buluwa Chamia Chawula Chichiya Lohuwa Macha Martabo Molcha Ochie Welluwa

20

40
47.5 24.2

60
67.0

80
86.4

100

9.4 8.6 6.5 5.6 5.3 3.8 3.2 2.9 2.7 2.1 2.1 1.8 1.5 1.2 0.9 0.9 0.9 0.6 0.6 0.6 0.6 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3 0.3

Fig. 3 Relative abundance of yam landraces recorded throughout Wolayita and Gamo-Gofa zones of Southern Ethiopia

Landraces recorded

123

124

Genet Resour Crop Evol (2008) 55:115131

Common

1. 0

0.8

0.6

0. 0

0 .2

0. 4 0.4

0 .6

0 .8

1 .0

ABUNDANCE

0.2

0.0

1.0

1.0

0.8

0.8

0.6 0.0 0 .2 0.4 0.4 0 .6 0.8 1 .0

0.6 0 .0 0.2 0 .4 0.4 0. 6 0.8 1 .0

0.2

0.2

0.0

0 .0

Rare L oca l Widespread

FREQUENCY

Fig. 4 Frequency and abundance matrix of yam landraces found throughout the study area in Southern Ethiopia (a), in the least diverse district of Damot-Woyde (b), and most diverse district of Damot-Gale (c)

farmers delayed planting till December (4%) or January (1%). Factors such as soil moisture content, anticipated severity of the dry season, and tuber sprouting were considered in timing eld planting. Land is prepared while the soil is still moist enough from the preceding rainy season to meet the requirements of yam for loose and deep soils, as well as permitting planting before the onset of the dry season so that early growth can make use of the residual soil moisture. There is no formal seed supply system for yam in the study area nor do farmers specialize in the production of yam planting-materials. Farmers mostly rely on their own planting-materials saved from the previous cropping season. Some farmers partly meet their demand for seed tubers through purchases from local markets or exchanges with neighbors (Fig. 5). About 60% of the farmers

interviewed used pieces of tubers, while others depended on whole tubers (3%) or both types of tubers (37%) for planting. The type of tubers used corresponded to type of landrace grown. For single-harvested landraces that normally produce a single tuber per plant, the head region (proximal end) is retained for propagation (while the remaining part is consumed), and is planted either as a single piece or further divided into smaller pieces. With double-harvested landraces, a single plant produces multiple tubers following the rst harvest, and these small whole tubers serve as ideal planting materials. Yam is mainly established as a sole crop in the eld (Table 7). The late and early-maturing landraces occupy separate rows on the same plot, and those with similar maturity time are planted in mixtures with no regular patterns. It is widely

123

Genet Resour Crop Evol (2008) 55:115131

125

Own Harvest (47%)

Own Harvest + Market (32%)

M arket (2.7%)

Market + Exchange (0.3%) Exchange (0.3%)

Own Harvest + Exchange (1.5%)

Own Harvest + Market + Exchange (16%)

Fig. 5 Major sources of planting-materials (seed tubers) for eld planting of yams as reported by farmers in Wolayita and Gamo-Gofa zones of Southern Ethiopia (gures in parenthesis are percentage values based on the total 339 farmers interviewed)

perceived that intercropping reduces yield and complicates cultural practices (Table 7). There is also a common belief in the area that yam does not appreciate frequent visits, which apparently reduce yield. Accordingly, frequency of entrance to yam elds is kept to a minimum, and monocropping is one way to achieve this. Even those farmers who practiced intercropping shared these opinions but adopted the system due to shortage of land. In intercropping, the crops planted with yams included maize (Zea mays L.), sweet potato (Ipomoea batatas (L.) Lam.), cabbage (Brassica spp.), beans (Phaseolus spp.) and, to a lesser extent, coffee (Coffea arabica L.). Yam is usually planted on relatively fertile plots, or gets the most
Table 7 Methods of yam establishment in the eld and reasons for their preference given by farmers in the major yam growing areas of Wolayita and Gamo-Gofa zones, Southern Ethiopia Method of eld establishment and reasons for preference Responses by farmers No. Monocropping 326 Lowers competition 223 Convenient for cultural practice 76 Intercropping 7 Shortage of land 7 Both 6 Total
a

%a 96 95 22 2 100 2 100

329

Sums over 100% are due to multiple responses

attention during applications of manure that is incorporated into the soil during land preparation. There is no use of commercial fertilizers in yam production. Rotation of yam plots on intervals of 14 years, depending on land availability, is practiced to achieve sustainable yields. Yam is chiey cultivated along rows of stakes, except for wild yams that are planted near trees for support. Young Eucalyptus, and maize and sorghum (Sorghum bicolor (L.) Moench) stalks are among the materials widely used for supporting yam plants. Staking commences after the tubers have sprouted and produced vines of considerable size. Every plant is supplied with a vertical stake and trained along it. Individual staking is the only method of staking encountered in the survey area. The majority of the farmers interviewed (74%) obtained staking materials from surrounding forests and trees planted on their farms, while about 11% bought the materials on local markets. The others depended on both sources to secure the materials required for staking. Two practices exist with respect to yam harvesting. The late-maturing landraces are harvested only once at full senescence, whereas the early-maturing types are harvested twice (doubleharvested). For yam planted in October, double harvesting involves a rst harvest in May or June, when the tubers are severed at their point of attachment to the corm with maximum care to avoid damage to the root system. Visible onset of senescence is used as a guide for timing harvesting of the late-maturing landraces and the second harvest of double-harvested landraces. However, there is no easy way of determining the optimum time for the rst harvesting of double-harvested landraces. Farmers in the study area are guided by different, largely phenological signals to subjectively judge the rst harvesting (Table 8). The aim here is to avoid harvesting too early (lower yield) or too late that can compromise the second harvest due to insufcient time for re-tuberization. According to farmers account of trends over the last 2030 years, yam production and the number of landraces maintained on individual farms are on the decrease in most localities (Table 9). Even those areas where yam production has been an increasing business (for example, the case in Damot-Woyde) are characterized

123

126 Table 8 Criteria employed by farmers in Wolayita and Gamo-Gofa zones of Southern Ethiopia for timing the rst harvesting of double-harvested yam landraces Criteria Senescence of inorescence Senescence of inorescence + ower scent Senescence of inorescence + wilting of vine tips Wilting of vine tips Senescence of inorescence + digging and checking of tubers Senescence of inorescence + ower scent + wilting of vine tips Time from planting + wilting of vine tips Senescence of inorescence + ower scent + soil cracking Senescence of inorescence + soil cracking + wilting of vine tips Senescence of inorescence + time from planting Time from planting Othersa Total
a

Genet Resour Crop Evol (2008) 55:115131

6

Mean number of landraces per farm

5 4 3 2 1 0 1500

r = - 0.40

Proportion of farmers (%) 19.8 13.0 9.7 5.9 5.0 4.7 3.8 3.5 3.2 3.2 2.7 25.5 100.0

1600

1700

1800

1900

2000

2100

2200

2300

Elevation (m asl.)

Fig. 6 Mean number of yam landraces per farm related to elevation in Wolayita and Gamo-Gofa zones of Southern Ethiopia

Include the use of the above criteria in various forms of combinations

either by a low level of landrace diversity (Table 5) or a decreasing trend in the number of landraces maintained on individual farms (Table 9). The distribution pattern of yam landraces revealed that the type and number of landraces grown by individual farmers were inuenced by
Table 9 Trends in the number of landraces maintained on individual farms and in the overall yam production in Sidama and Gamo-Gofa zones of Southern Ethiopia as District Number of landraces Increasing Bolosso-Sore Damot-Gale Damot-Woyde Humbo Kindo-Koyisha Offa Soddo-Zuria Kucha Total 27 (68) 10 (24) 2 (4) 5 (12) 8 (21) 2 (5) 7 (18) 26 (65) 87 (26) Decreasing 4 21 17 28 31 34 13 14 162 (10) (50) (30) (67) (79) (85) (33) (35) (48)

factors such as elevation. Overall, the number of landraces grown per farm was negatively correlated (r = 0.40; P < 0.05%) with elevation (Fig. 6). However, elevation might be confounded with drought tolerance. At relatively lower elevations with extended dry season, landraces that are perceived to be drought-tolerant (for example, wadala) are widely cultivated. Besides, farmers plant double harvested landraces to ensure early harvests although some of these landraces are less adapted to drier conditions. There was no signicant correlation between farm size and number of landraces per farm. On the other hand, the proportion of land allocated for yam production was negatively correlated with total farm size (r = 0.13, P < 0.05), indicating that even those farmers with smaller landholdings
perceived by farmers (number in parenthesis refer to percentage values based on total number of farmers interviewed in each district) Total production

No change 9 11 37 9 (23) (26) (66) (21) 0 (0) 4 (10) 20 (50) 0 (0) 90 (26)

Increasing 13 (33) 10 (24) 34 (61) 2 (5) 2 (5) 4 (10) 3 (8) 14 (35) 82 (24)

Decreasing 27 32 22 40 37 36 37 26 257 (67) (76) (39) (95) (95) (90) (92) (65) (76)

No change 0 0 0 0 0 0 0 0 0 (0) (0) (0) (0) (0) (0) (0) (0) (0)

123

Genet Resour Crop Evol (2008) 55:115131

127

allocate a signicant share of their land for yam cultivation in order to get a reasonable production and meet family needs. Signicant proportions of farmers interviewed in Kindo-Koyisha (87%), Kucha (70%), Offa (63%) and Humbo (33%) districts stated the presence and use of wild yam in their area. The gure was relatively lower in Soddo-Zuria (8%) and Bolosso-Sore (3%), while there was no such report in Damot-Woyde. Wild yam was found predominantly in localities situated at lower elevations (mainly below 1,700 m asl.) that are sparsely populated, and where patches of forest could still be found. Wild yam tubers collected from surrounding forests are either consumed directly or planted on farms under big trees, where they are left to grow for up to 3 years. These tubers are normally consumed during periods of relative food shortage.

Discussion Status of yam diversity Farmers in the study area maintain diverse yam landraces with respect to attributes such as environmental adaptation and length of growing period. This nding conrms the salient feature of traditional farming systems in the tropics, where diverse crop species or varieties of the same species are maintained on a single farm (Boster 1983; Brush 1995; Clawson 1985) in response to economic, social, cultural and natural factors (Cox and Wood 1999). Similar observations were made in various traditional farming systems for clonally propagated crops such as enset (Tesfaye and Ludders 2003; Tsegaye and Struik 2002), cassava (Manihot esculenta Crantz) (Boster 1985; Salick et al. 1997), and potato (Solanum tuberosum L.) (Brush et al. 1981). Tsegaye and Struik (2002) recorded a total of 55 named enset landraces in Wolayita, where individual farmers on average maintained eight landraces. They also reported that landrace diversity was affected by factors such as household resources, cultural background, population pressure and agro-ecology. Two of the landraces described, bola-boye and bunde-buchi, belong to the species D. bulbifera.

However, the species identity of the remaining landraces is yet to be established (S. Demissew, personal communication). Preliminary observations based on morphological features seem to indicate that some of the landraces belong to the D. cayenensis/D. rotundata species complex (Tamiru 2006) as presently understood by most researchers working on yams (e.g. Dansi et al. 1999). However, this was not supported by molecular data (Tamiru et al. 2007). Most named yam landraces are morphologically distinct. Farmers consider a combination of characters viz. morphological, growth and organoleptic, as well as ecological adaptation to classify yam landraces (M. Tamiru et al. submitted to Genetic Resources: Characterization and Utilization). Tuber esh color is the most important selection criterion to distinguish the so-called female (macha) landraces. Nevertheless, macha does not seem to be a distinct landrace. The same name is used as a reference to a group of the socalled female yams (macha boye) that include the early-maturing landraces such as hatiye and oha (M. Tamiru et al. submitted to Genetic Resources: Characterization and Utilization). But, there are no peculiar characteristics that distinguish the landrace macha from members of the group macha. The extent of landrace diversity detected in this study is comparable to an earlier report from Sheko, Southwest Ethiopia, where 23 separate indigenous yam types belonging to at least four species of Dioscorea were described (Hildebrand et al. 2002). But, it is considerably lower than the level of diversity reported from some West African countries. For Example, about 300 different named yam landraces were described across 10 different ethnic groups throughout Benin (Dansi et al. 1997, quoted by Dansi et al. 1999), while Baco et al. (2004) recorded 88 varieties in the Sienende district of Benin. However, such records are not entirely comparable. First, some reports cover an entire region or country, whereas others, including our study, deal with relatively smaller areas. Furthermore, when conducting such studies across ethnically diverse regions, like in the above reports, linguistic polymorphism may lead to an overestimation of diversity based on named landraces.

123

128

Genet Resour Crop Evol (2008) 55:115131

Yam production in Wolayita and Gamo-Gofa is mainly based on a limited number of widespread landraces such as hatiye, wadala and oha (Table 3 and Fig. 3). The majority of landraces recorded have a rather limited distribution and abundance. This hierarchical nature of spatial distribution, where a limited number of landraces or cultivars are dominant, has been documented for several crop species (Boster 1985; Louette et al. 1997; Tesfaye and Ludders 2003). The widespread distribution of some landraces also challenges the view that traditional farming systems are isolated and closed, with limited exchange of germplasm. Our nding and those of others mentioned above depict these systems rather as open and dynamic, where local networks exist for moving planting materials across wider areas and heterogeneous environments. Yam farmers in the study area acquire part of their planting materials either through purchases from local markets or exchanges with neighbors (Fig. 5). Such networks can cover relatively larger areas, as getting to the next market often entails long-distance travels. One advantage of double-harvesting is that the rst harvesting induces the formation of multiple tubers. Tubers from the second harvest are mostly lignied and brous, and possess several visible buds even at harvest (Onwueme and Charles 1994), making them ideal planting materials. Some farmers prefer to delay or forgo rst harvesting, opting for a single harvest of apparently higher yields to maximize income. These farmers purchase seed tubers from local markets for the following production season (Fig. 5). This has created a potential market for seed tubers, where middlemen who are now increasingly involved in the business move planting materials even over longer distances. This may offer some explanation as to why some districts (for example, Damot-Woyde and Bolosso-Sore), although located further apart, are relatively similar with regard to yam landrace diversity. Nevertheless, addressing this issue needs analysis of events in the past that might have inuenced movement of yam germplasm in the study area, an important data that is missing at the moment.

Yam production in the study area is constrained by several environmental and production factors (Tamiru et al. 2005). This has led to a decrease in production and yam diversity in the majority of the localities surveyed, except in some localities, mainly of Damot-Woyde, where production is on the increase and yam is establishing itself as an important cash crop. However, Damot-Woyde is the least diverse district in terms of total number of landraces found (Table 5). It seems that the increase in production is brought about at the expense of the overall landrace diversity, as farmers are increasingly growing a few selected landraces. Hatiye and oha are among the widely cultivated landraces in this district due to their early maturity and excellent culinary properties and, thus, are replacing the late-maturing landraces such as wadala. As noted by Frankel and Bennett (1970), besides the transition from landraces to advanced cultivars, selection for closely dened objectives can lead to a reduction in genetic variation. Although detailed information is lacking as to the extent of changes that might have occurred in yam genetic diversity and the implications of such changes, this study has provided a rst perception by farmers, and genetic vulnerability (Brown 1983) is a legitimate worry in Wolayita and Gamo-Gofa (Table 9). Farmers are expressing concerns that yam production is threatened by changing environmental conditions (erratic rains, increasing temperatures). This concern is particularly valid in view of the fact that most of the early-maturing landraces that are being used for expanding production are relatively more prone to drought than the late-maturing ones. Management and use of diversity Wolayita and Gamo-Gofa farmers are familiar with the diversity available in yams and attributes of each landrace, which they utilize accordingly to meet their needs. Unlike other crops, yam is adapted to dry-season planting, an attribute widely manipulated by local farmers to ensure household food security. For yams planted in October, the rst harvest of early-maturing landraces is expected around May or June. This is a period of relative food shortage in the area, as

123

Genet Resour Crop Evol (2008) 55:115131

129

most of the other crops are still in the eld. Thus, yam lls a seasonal gap in food supply. That is why the early-maturing landraces such as hatiye and oha are widely distributed throughout the study area (Fig. 3). Apart from early maturity, some landraces such as hatiye are popular due to their sweet taste and white tuber esh color, and are preferred for preparing chata, a popular dish made of boiled and mashed yam mixed with fermented milk and butter. The white tuber esh goes well with the milk during mixing. Thus, such landraces are widely distributed across different altitudinal ranges although farmers are aware of the fact that some perform poorly under drier and hotter conditions. Wadala is more common at lower elevations, and is highly valued for its sturdy growth, drought tolerance and bigger tubers. Its requirement for more stout staking materials, regular training and, hence, intensive management is usually tolerated because of its acceptable performance under sub-optimal conditions. The late-maturing nature of this landrace is exploited to extend harvesting into late seasons. There exists a striking similarity between the management and use of yams in the study area and other parts of Ethiopia such as Sheko (Hildebrand et al. 2002), as well as different West African countries (Asiedu et al. 1997; Hahn et al. 1987; Onwueme 1978). This provides an opportunity for sharing experiences mainly with West African countries, where the yam-based agriculture has been supported by research undertakings that have achieved technology delivery and adoption on farms (Quin 1998). On the other hand, Wolayita and Gamo-Gofa farmers employ unique practices with a certain degree of sophistication in managing yam. For example, double and single-harvestings are also common features of yam production in other African countries (Onwueme 1978). Among the main problems often mentioned in connection with double-harvesting is the lack of a reliable index of maturity to time the rst harvesting (Onwueme 1978), which farmers in the study area subjectively judge by using a range of criteria based on experience (Table 8). Such practices make the indigenous knowledge of local farmers an important aspect of the overall yam diversity.

Conclusions The high value that local farmers place on yam is expressed in its continued cultivation despite the lack of any form of support from researchers and policy makers. Due to its adaptability to dry season planting, yam ts well into the traditional cropping calendar, and this is widely utilized to ensure food availability during critical periods of the year. Given this practical importance of yam in the local livelihood, there is an urgent need for research programs to address the problems facing yam production and its diversity taking into consideration the multiple objectives of farmers and the importance of diversity in the physical, economical and cultural contexts of local agriculture. Findings of this study suggest that the majority of the landraces recorded in Wolayita and Gamo-Gofa face considerable risk of loss mainly due to their rare occurrence and local distribution. Besides, late-maturing landraces are becoming increasingly vulnerable to replacement by early-maturing ones. The diversity available in these landraces needs to be studied in detail in order to facilitate their conservation as well as utilization in crop improvement programs. Studies that address the problems currently faced by yam production can enhance the role that yam plays in household food security. Considering the importance of yam in local agriculture and tradition, such studies can also ensure continuous maintenance of yam diversity by farmers through increased utilization of available landraces. Describing the diversity in crop species based on named landraces, such as in this report, constitutes an essential step towards setting research and development priorities aimed at conservation and improvement of a traditional crop. In view of the current taxonomical confusion regarding the major African Dioscorea species and the lack of information on the status of yams outside the yam belt, broadening the knowledge base of yams in Ethiopia contributes substantially to our understanding of the diversity in African yams. To this end, further research must include other regions in Ethiopia that are not covered by the present

123

130

Genet Resour Crop Evol (2008) 55:115131 Cox TS, Wood D (1999). Nature and role of crop biodiversity. In: Wood D, Lenne JM (eds) Agrobiodiversity: characterization, utilization and management. CABI Publications, Oxon, UK, pp 3557 CSA (2000). Central Statistical Authority, Statistical abstract 2000. Addis Ababa, Ethiopia, p 403 Dansi A, Mignouna HD, Zoundjihekpon J, Sangare A, Asiedu R, Quin FM (1999). Morphological diversity, cultivar groups and possible descent in the cultivated yams (Dioscorea cayenensis/Dioscorea rotundata complex) in Benin Republic. Genet Resour Crop Evol 46:371388 Engels JMM, Hawkes JG, Worede M (eds) (1991). Plant genetic resources of Ethiopia. Cambridge University Press, Cambridge, UK, p 383 FAO (2004) Food and Agricultural Organization of the United Nations, statistical database, http://faostat.fao.org/faostat/collections?subset=agriculture, cited on 11 February 2006 Frankel OH, Bennett E (1970). Genetic resources. In: Frankel OH, Bennett E, Brock RD, Bunting AH, Harlan JR, Schreiner E (eds) Genetic resources in plants: their exploitation and conservation, International biological program handbook, No. 11. Blackwell Scientic Publications, Oxford, UK, pp 717 Girard T (2002) Lexico-phonostatistical analysis of Alemayehu Abebes Ometo word lists. SIL International, http://www.sil.org/silesr/2002/SILESR2002051.pdf, cited on 26 May 2006 Hahn SK (1995). Yams. In: Smartt J, Simmonds NW (eds) Evolution of crop plants, 2nd ed. Longman Group Ltd., London, UK, pp 112120 Hahn SK, Osiru DSO, Akoroda MO, Otoo JA (1987). Yam production and its future prospects. Outlook Agric 16:105110 Harlan JR (1969). Ethiopia: a center of diversity. Econ Bot 23:309314 Hildebrand E, Demissew S, Wilkin P (2002). Local and regional disappearance in species of Dioscorea L. (Yams) in southwest Ethiopia. In: Stepp JR, Wyndham FS, Zarger RR (eds) Ethnobiology and biocultural diversity. Proceedings of the 7th international congress of ethnobiology. University of Georgia Press, USA, pp 678695 Hoogendijk M, Williams DE (2002) Characterizing the genetic diversity of home garden crops: some examples from the Americas. In: Watson JW, Eyzaguirre PB (eds) Home gardens and in situ conservations of plant genetic resources in faming systems. Proceedings of the second international home gardens workshop, 1719 July 2001, Witzenhausen, Federal Republic of Germany, IPGRI, Rome, pp 3440 Louette D, Charrier A, Berthaud J (1997). In situ conservation of maize in Mexico: genetic diversity and maize seed management in a traditional community. Econ Bot 51:2038 Magurran AE (1988). Ecological diversity and its measurements. Croom Helm, London, UK, pp 125 Miege J, Demissew S (1997). Dioscoreaceae. In: Edwards S, Demissew S, Hedberg I (eds) Flora of Ethiopia &

study and give wider coverage to wild yams, as well as consider elite genotypes from West Africa to thoroughly investigate the available diversity.
Acknowledgments We are grateful to all farmers who participated in the study for their time, invaluable knowledge and hospitality. We also thank members of the regional, zonal and district agricultural ofces and, particularly the development agents who gave us all the help we needed during site selection and throughout the study. The assistance provided by the DU-Norad (Norwegian Agency for Development Cooperation) project of Debub University (Ethiopia) during the eldwork is highly acknowledged. The German Academic Exchange Service (DAAD) sponsored this study.

References
Alexander J, Coursey DG (1969). The origins of yam cultivation. In: Ucko PJ, Dimbleby GH (eds), The domestication and exploitation of plants and animals. Proceedings of a meeting of the research seminar in archaeology and related subjects held at the Institute of Archaeology, London University. Gerald Duckworth & Co. Ltd., UK, pp 405425 Asiedu R, Wanyera NM, Ng SYC, Ng NQ (1997). Yams. In: Fuccillo D, Sears L, Stapelton P (eds) Biodiversity in trust: conservation and use of plant genetic resources in CGIAR centers. Cambridge University Press, Cambridge, UK, pp 5766 Ayensu ES, Coursey DG (1972). Guinea yams: the botany, ethnobotany, use and possible future of yams in West Africa. Econ Bot 26:301318 Baco MN, Tostain S, Mongbo RL, Danou O, Agbangala C (2004). Gestion dinamique de la diversite varietal des ignames cultivees (Dioscorea cayenensisD. ro tundata) dans la commune de Sinende au nord Benin. Plant Genet Resour Newsl 139:1723 Boster JS (1983) A comparison of diversity of Jivaroan garden with that of tropical forest. Hum Ecol 11:4768 Boster JS (1985). Selection for perpetual distinctiveness: evidence from Aguarina cultivars of Manihot esculenta. Econ Bot 39:310325 Brown WL (1983). Genetic diversity and genetic vulnerability: an appraisal. Econ Bot 37:412 Brush SB (1995). In situ conservation of landraces in centers of diversity. Crop Sci 35:346354 Brush SB, Carney HJ, Huaman Z (1981). Dynamics of Andean potato agriculture. Econ Bot 35:7088 Clawson DL (1985). Harvest security and intraspecic diversity in traditional tropical agriculture. Econ Bot 39:5667 Coursey DG (1967). Yams: an account of the nature, origins, cultivation and utilization of the useful members of Dioscoreaceae. Longmans, Greens and Co. Ltd., London, UK, p 230

123

Genet Resour Crop Evol (2008) 55:115131 Eritrea, Vol 6, Hydrocharitaceae to Araceae. Addis Ababa, Ethiopia, and Uppsala, Sweden, pp 5562 MoA (Ministry of Agriculture) (2000) Agro-ecological zones of Ethiopia. Natural resources management and regulatory department. Ministry of Agriculture, Addis Ababa, Ethiopia, p 129 Norman MJT, Pearson CJ, Searle PGE (1995). The ecology of tropical food crops, 2nd ed. Cambridge University Press, Cambridge, UK, pp 305318 Onwueme IC (1978). The tropical root crops: yams, cassava, sweet potato and cocoyams. Wiley, Chichester, UK, p 234 Onwueme IC, Charles WB (1994). Tropical root and tuber crops: production, prospective and future prospects. FAO Plant Production and Protection Paper 126. FAO, Rome, p 228 Orkwor GC, Asiedu R, Ekanayaka IJ (eds) (1998). Food yams: advances in research. IITA/NRCRI, Ibadan, Nigeria, p 249 Quin FM (1998). An over view of yam research. In: Orkwor GC, Asiedu R, Ekanayaka IJ (eds) Food yams: advances in research. IITA/NRCRI, Ibadan, Nigeria, pp 215230 Salick J, Cellinese N, Knapp S (1997). Indigenous diversity of cassava: generation, maintenance, use, and loss among the Amuesha, Peruvian Upper Amazon. Econ Bot 51:619 Simpson EH (1949). Measurement of diversity. Nature 163:688 Tamiru M, Maass BL, Becker HC (2005) Traditional management and use of yams (Dioscorea spp.) in

131 Wolayita, Southern Ethiopia. Deutscher Tropentag 2005, The global food and product chain: dynamics, innovation conicts and strategies. International conference on research for development in agriculture and forestry, food and natural resource management, October 1113, 2005, University of Hohenheim, Stuttgart, Germany. Book of abstracts, p 433 Tamiru M (2006) Assessing diversity in yams (Dioscorea spp.) from Ethiopia based on morphology, AFLP markers and tuber quality, and farmers management of landraces. PhD thesis, Georg-August-University Goettingen, Germany. Cuvillier Verlag, Goettingen, Germany, p 155 Tamiru M, Becker HC, Maass BL (2007). Genetic Diversity in yam germplasm (Dioscorea spp.) from Ethiopia and their relatedness to the main cultivated Dioscorea species assessed by AFLP Markers. Crop Science (accepted) Tesfaye B, Ludders P (2003). Diversity and distribution patterns of enset landraces in Sidama, Southern Ethiopia. Genet Resour Crop Evol 50:359371 Tsegaye A, Struik PC (2002). Analysis of enset (Ensete ventricosum) indigenous production methods and farm-based biodiversity in major enset growing regions of Southern Ethiopia. Exp Agric 38:292315 Vavilov NI, Chester KS (trans) (1951) The origin, variation and breeding of cultivated plants. Chronica Botanica 13:1366 Westphal E (1975). Agricultural systems in Ethiopia. Center for Agriculture Publishing and Documentation, Wageningen, The Netherlands, p 278

123