Professional Documents
Culture Documents
Tropical Forestry
Volumes Already Published in this Series
Tropical Forest
Genetics
ISSN: 1614-9785
ISBN-10 3-540-37396-9 Springer-Verlag Berlin Heidelberg New York
ISBN-13 978-3-540-37396-4 Springer-Verlag Berlin Heidelberg New York
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is con-
cerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, repro-
duction on microfilm or in any other way, and storage in data banks. Duplication of this publication or parts
thereof is permitted only under the provisions of the German Copyright Law of September 9, 1965, in its
current version, and permission for use must always be obtained from Springer. Violations are liable to pros-
ecution under the German Copyright Law.
springer.com
The use of general descriptive names, registered names, trademarks, etc. in this publication does not imply,
even in the absence of a specific statement, that such names are exempt from the relevant protective laws and
regulations and therefore free for general use.
available for tropical forest trees; tropical forest genetics is still in a prege-
nomic era. Thus, the focus of this book in on population genetics processes in
natural and managed tropical forests rather than on genomics and related,
newly emerging fields of research in forest genetics.
Acknowledgements
The authors are grateful to Oliver Gailing for critically reading Sect. 12.5.4 to
Barbara Vornam and Martin Ziehe for assistance in literature search and in
word processing, and to Wolfgang Tambour for preparing Figures 7.1, 7.2, 7.5,
and 12.4.
6 Mating Systems 69
6.1 Introduction 69
6.2 Random Mating and Panmixis 70
6.2.1 Heterogeneity of Pollen Allele Frequencies 70
6.2.2 Panmixis and HardyWeinberg Structures 72
6.3 Selfing and Outcrossing Rates 73
6.3.1 Estimates of Selfing Rates Based on Rare Alleles 74
Example 6.1: Estimates of Selfing Rates in Teak (Tectona grandis)
Populations 74
6.3.2 Estimates of Selfing Rates Based on a Mixed Mating Model 76
6.3.3 Estimates of Selfing Rates Based on Nonmaternal Alleles 76
6.4 Inbreeding and Inbreeding Depression 78
6.4.1 Genetic Consequences of Inbreeding 78
6.4.2 Inbreeding Depression 80
Example 6.2: Inbreeding Depression in Eucalypts 80
6.5 Incompatibility and Self-Sterility 82
6.5.1 Incompatibility 82
6.5.2 Self-Sterility 84
6.6 Environmental Effects on Mating 84
6.7 Recommended Literature 85
References 267
Index 307
Introduction Genetics of Tropical Forests 1
. . . the significance of forest genetics lies not only in squeezing extra gain out of mar-
ginal investments, but more importantly in understanding the genetics of forests, in
appreciating what forest ecosystems are and how they operate, and in better stewards of
forest resources (Namkoong 1989).
Population geneticists spend most of their time doing one of two things: describing the
genetic structure of populations or theorizing on the evolutionary forces acting on
populations. On a good day, these two activities mesh and true insights emerge
(Gillespie 1998).
2.1
Introduction
2.2
The Population
events between plants (or animals) from different populations in this context.
However, an assessment of reproductive isolation usually requires a fairly
detailed understanding of the reproduction system of a species. Furthermore,
the definition of a population is operational only if an occasional (but rare)
exchange of genetic information among populations is taken into considera-
tion. Thus, interfertile, spatially clustered plants are often regarded as belonging
to the same population, while spatially separated plants are regarded as parts of
different populations.
The delineation of populations is particularly difficult for immobile organ-
isms with poorly described spatial distribution patterns and largely unknown
means of gene dispersal through seed and pollen. In most cases, it is far from
obvious which trees in a species-rich tropical forest belong to the same popu-
lation and how big tree populations are. This has far-reaching consequences for
levels of genetic variation of tropical forest trees, as will be discussed in later
chapters.
Analyses of spatial patterns of genotypes without a priori information on the
delineation of populations or partially isolated subpopulations have recently
been proposed as an alternative to the traditional approach of dividing the
overall variation into components within and among demes or populations
(Diniz-Filho and Telles 2002; Manel et al. 2003). Spatially explicit analyses of
genetic variation patterns might be particularly informative for plant species
occurring in low density in tropical forests; however, currently available infor-
mation is mainly confined to comparatively small areas (Ng et al. 2004).
2.3
Variation at Gene Loci
2.3.1
The Molecular Basis of Genetic Variation
2.3.1.1
Molecular Markers
The most important methods to assess patterns of genetic variation within and
among species rely on the amplification of short DNA fragments by means of
polymerase chain reaction (PCR). PCR allows short fragments of the DNA of
a target organism to be selectively amplified (multiplied) for further analyses.
The sequence of two short oligonucleotides (primers), each usually between
ten and 25 base pairs in length, is decisive for the amplified DNA region. The
following marker types, all based on the PCR method, have gained particular
importance in the study of genetic variation patterns of tropical forest trees.
195 200 205 210 215 220 225 230 235 240 245 250 255 260
1000
500
1000
500
600
400
200
2000
1000
1500
1000
500
900
600
300
900
400
200
2000
1500
1000
500
Fig. 2.2. Length variation for eight Prosopis spp. trees at a microsatellite gene locus
(Mottura et al. 2005) separated in an automated sequencer (ABI 3100). Each line refers
to one tree. Different sizes of amplified fragments are visualized as peaks at different
positions in the chromatogram. Trees with two amplified fragments are heterozygous;
tree 5 is homozygous with only one amplified fragment. (From Mottura, unpublished)
Fig. 2.3. Low variation at a random amplified polymorphic DNA marker in teak
(Tectona grandis). (From Finkeldey, unpublished)
300
200
100
200
100
300
200
100
221.05
300
200
100
222.04
400
300
200
100
200
150
100
50
200
150
100
50
were developed for Acacia mangium to assess genetic diversity and differentiation
among populations (Butcher et al. 1998) and with the objective to incorporate
marker-based approaches in breeding programs (Butcher et al. 2000; Butcher
2004).
More frequently, the RFLP techniques is combined with a PCR step and is
used to assess variation of cpDNA. For example, Tsumura et al. (1996) and
Indrioko et al. (2006) (Fig. 2.5) analyzed phylogenetic relationships among
Southeast Asian dipterocarps using the PCR-RFLP technique (Example 8.1),
and Cavers et al. (2003) observed strong genetic differentiation among five
cpDNA haplotypes of Cedrela odorata in Central America (Example 3.6).
12 C HAPTER 2 Population Genetics an Overview
2.3.1.2
Biochemical Markers Isozymes
Tree: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
A1
A2
A2A2 A2A2 A1A2 A2A2 A1A2 A1A2 A1A2 A2A2
A1A2 A2A2 A1A2 A1A2 A1A1 A1A2 A1A2 A2A2
Fig. 2.6. Zymogram showing variation of 16 Dalbergia sissoo trees from Nepal at the
enzyme system alcohol dehydrogenase. The variation is controlled by a gene locus
ADH-A. The inferred genotypes of the respective trees are indicated. (From Pandey,
unpublished)
inventories and the comparatively low costs have made them useful tools for
population genetics studies, although they were often substituted by various
types of molecular DNA markers during the last decade.
2.3.2
The Gene As a Unit of Heredity
differences between the observed and expected values are not significant
(Hattemer 1991). Suitable statistical tests are, for example, described by Sokal
and Rohlf (1998, p. 686 ff.; c2 test or goodness-of-fit test, G test).
The formal identification of a genetic marker locus by an observation of seg-
regation is possible only for environmentally stable, genetic traits. It is also
advisable for the identification of gene loci based on the observation of bio-
chemical or molecular markers. For example, Moran and Bell (1983) observed
segregation at the isozyme system GOT in progenies of Eucalyptus regnans after
controlled pollination and found no evidence to reject the hypothesis of the
genetic control of the variation by a single gene locus.
Forest trees, in particular those of the tropics, are extraordinarily difficult
to cross owing to their large size, short viabilities of pollen, unknown repro-
ductive biology, and other obstacles. Thus, the identification of a gene locus
by inheritance studies is preferentially conducted on the basis of alternative
approaches such as the observation of segregation in the haploid megagame-
tophyte (the endosperm) of gymnosperms (Bergmann 1974) or the obser-
vation of segregation in the progenies of putatively heterozygous seed trees
after open pollination (Gillet and Hattemer 1989). The first method has been
used to identify enzyme gene loci in Pinus merkusii from Thailand
(Changtragoon and Finkeldey 1995b); the latter approach was used to clarify
the inheritance of isozymes in Pterocarpus indicus from the Philippines
(Finkeldey et al. 1998).
2.3.3
The Mode of Inheritance
2.3.4
Definition and Classification of Gene Markers
A gene marker (synonyms gene locus and marker locus) in a wide sense is
an environmentally stable trait; thus, the variation is determined by genetic
factors only.
A gene marker in a narrow sense is an environmentally stable, biparentally
inherited, codominant trait.
Table 2.1. Classification of genetic markers in a broad sense commonly used in studies
on genetic variation of tropical forest species with regard to their mode of inheritance
and their variability
Variation Uniparental Anonymousa Codominantb
2.4
Genetic Structures Within Populations
The final result from a laboratory study to assess the genetic constitution of a
plant at a particular biochemical or molecular gene marker system is a specific
pattern, which can be used:
2.4.1
Allelic and Genotypic Structures
where Nij is the (absolute) frequency of genotype AiAj in the population (or
sample) and N is the population size (or sample size).
The frequency vector of all (relative) frequencies of genotypes is defined as
the genotypic structure of the population at the respective marker gene locus.
It obviously holds that // Pij = 1.
i j
Each genotype of a diploid species consists of two alleles; thus, the total
number of alleles in a population (or a sample) is twice the number of plants
investigated, and the frequency of a particular allele Ai can be calculated as
p i = Ni / 2N, (2.2)
where Ni is (absolute) frequency of allele Ai in the population (or sample) and
N is the population size (or sample size). The respective allele Ai is counted
twice in case of homozygous plants AiAi.
The frequency vector of all (relative) frequencies of alleles is defined as the
allelic structure of the population at the respective marker gene locus. Again,
it holds that / p i = 1.
i
If the genotypic structure of a population is known, it is easily possible to
compute the allelic structure at the respective gene locus. The frequency of
allele Ai (pi) can be computed by calculating the sum of the frequency of the
corresponding homozygote (Ai Ai ) and half of the frequencies of all heterozy-
gotes where the respective allele occurs (Ai Aj , Ai Ak , . . .). However, the calcula-
tion of genotypic structures based on allele frequencies requires additional
information or assumptions (e.g., Sect. 6.1.2).
2.4.2
Variation at Uniparentally Inherited Markers
Table 2.2. Variation of chloroplast DNA haplotypes in five natural populations and
five plantations of Dalbergia sissoo in Nepal. (Adapted from Pandey et al. 2004)
Region Population N Chloroplast DNA haplotypes
1 2 3 4 5 6 7 8
Natural populations
A Hetauda N 10 9 1
B Shivapur 10 8 2
C Hattisar 6 5 1
D Godawari 10 4 6
E Pipariya 10 9 1
Plantations
A Hetauda P 10 4 6
B Surai 10 1 7 2
C Thakurdwara 10 4 3 3
D Attaria 10 6 4
E Shuklaphanta 10 5 5
N sample size
2.5 Evolution and Evolutionary Factors 19
2.5
Evolution and Evolutionary Factors
Mutation is the prerequisite for any genetic variation. Mutations are ran-
dom changes of genetic information of an organism which might affect a
single nucleotide (a spontaneous change of a single nucleotide resulting in
a SNP; Sect. 2.3.1), the chromosomal structure (e.g., fissions, translocations,
inversions), or the number of genes (e.g., duplications, polyploidy). Most
mutations are detrimental or neutral for an organism. Only few advanta-
geous mutations increase the fitness, i.e., the capacity to produce offspring.
20 C HAPTER 2 Population Genetics an Overview
Mutations are rare events; mutation rates are estimated to be in the range
between 106 and 108 for most genes. However, the frequency of mutations
is not uniform. For example, considerably higher estimates of mutation
rates were reported for microsatellites (Goldstein and Pollock 1997).
Gene flow and migration increase the genetic multiplicity of a population,
if genes which were previously restricted to certain populations migrate to
other populations. Transport of pollen (gene flow in a narrow sense) and
seeds (migration) are the main processes involved in the dispersal of genes
for most plant species (Chap. 5). By definition, gene flow and migration
among populations are rare events.
The mating system is decisive for the fusion of male and female gametes.
Gene flow through pollen is an obvious prerequisite for mating between
two seed plants. Thus, gene flow and the mating system are closely con-
nected and are the two most important aspects of the reproduction system
of a plant species. The mating system decides on the combination of alle-
les to genotypes. The genotypic structure of the progeny generation is
influenced by the mating system, which primarily does not cause changes
of allelic structures. Inbreeding due to selfing or mating among relatives is
a particularly important aspect of the mating system (Chap. 6).
Random fluctuations of genetic structures are described as genetic drift.
Genetic drift is a consequence of limited population size and becomes
stronger in small populations. The strength of genetic drift can be described
as a function of the population size, although its final outcome is unpre-
dictable. Alleles, in particular rare alleles, are likely to completely disappear
from populations as a result of genetic drift, and previously polymorphic
(variable) gene loci may become monomorphic (fixed). A sudden reduction
of the population size for one or several generations is often described as a
genetic bottleneck. Bottlenecks result in reduced genetic variation even
after later population expansion owing to the effects of genetic drift while
populations are small. Genetic differentiation among partially or completely
isolated populations is expected to increase as a result of genetic drift.
Selection is a consequence of the differing abilities of organisms to produce
offspring (fitness). Selection does not increase multiplicity, but often con-
tributes to the maintenance of genetic diversity by stabilizing polymor-
phisms, for example, owing to a selective advantage of heterozygous
genotypes. In general, selection results in an improved adaptedness (Chap.
7). The basis of selection is the different survival of genotypes (viability
selection) and the different contributions of genotypes to the formation of
a progeny generation (fertility selection).
2.6
Phenotypic Variation
The vast majority of directly observable traits of trees are an outcome of the
interaction between the genetic constitution of a plant at a large number of gene
loci (genotype, G) and the environmental conditions (E). Thus, it holds for the
phenotype (P) that P= GE, where symbolizes the interaction between envi-
ronmental and genotypic effects, which often differs from simple additivity of
both components.
Phenotypic traits deserve particular interest if they are important for the
adaptation of plants to their environment (adaptive traits) or for the value of a
tree (economic traits). Adaptive traits such as the tolerance against biotic
(pests) or abiotic (e.g., draught) stress and economic traits such as volume
growth are controlled by both genetic and environmental factors for most
plant species. The analysis of phenotypic traits in a breeding context will be
discussed in Chap. 12 in more detail. It is also described for forest trees in other
textbooks (Wright 1976; Zobel and Talbert 1984; Williams et al. 2002).
The majority of the gene markers described in Sect. 2.3.1 are unlikely to have
strong, recognizable effects on adaptive traits or traits of economic importance.
They are regarded as neutral or nearly neutral markers with regard to an
adaptation to the environment; however, this does not preclude their applica-
tion to monitor adaptive processes (Ziehe et al. 1999; Sect. 7.2).
In line with the definition of genetic structures, the frequency distribution
of particular phenotypes is defined as the phenotypic structure of a popula-
tion. Direct and indirect effects of the environment on the phenotypic struc-
ture are illustrated in Fig. 2.7. Apart from the direct impact of the environment
22 C HAPTER 2 Population Genetics an Overview
Phenotypic
Mode of gene structure
action
Migration Genotypic
structure
Modification
Viability selection
Mutation
Population size
Fertility selection
Gene flow Allelic
structure
Fig. 2.7. The environmental impact on phenotypic structures. (Adapted from Hattemer
and Mller-Starck 1990)
2.7
Recommended Literature
. . . the main difficulty one faces in a subject like population genetics (. . .) is not the
mathematics itself, or the biology itself: it is how to fit them together (p. ix in Maynard
Smith 1998).
3.1
Introduction
3.2
Genetic Inventories
3.3
Measurement of Genetic Variation
3.3.1
Genetic Variation Within Populations
3.3.1.1
Genetic Multiplicity
3.3.1.2
Genetic Diversity
The genetic diversity of a population does not only depend on the number of
genetic types (alleles or genotypes), but also on their (relative) frequencies (pi , pij).
1#v= 1 # n. (3.3)
/p i
i
2
The effective number of alleles v reaches its maximum value for a given n if
the frequencies of all alleles are identical (1/n). In this case, v = n. The allelic
diversity v decreases with increasing frequency of one allele and consequently
decreasing frequencies of the remaining alleles.
The effective number of alleles of the gene pool is the harmonic mean of the v
at single loci.
The effective number of alleles from Example 2.1 is
v= 1 = 1 = 1 . 1.82.
/p
i
2
i
p 12 + p 22 0.34375 2 + 0.65625 2
Expected Heterozygosity
The expected heterozygosity He is the most frequently used measure of genetic
diversity within populations. The expectation is based on the assumption
that alleles are randomly combined to genotypes, i.e., that HardyWeinberg
structures are realized (Sect. 6.2.2). In this case, it is possible to calculate the
frequency Pii of a particular homozygote with genotype Ai Ai as pi2 (pi is the fre-
quency of allele Ai). The frequency of the expected homozygosity is calculated
by summing over all alleles (pi2), the remainder being the expected heterozy-
gosity. Accordingly, the expected heterozygosity is calculated as
H e = 1 - / p i2 . (3.4)
i
Gene pool values of the expected heterozygosity are the arithmetic mean of
single locus values.
The expected heterozygosity from Example 2.1 is
H e = 1 - (p 12 + p 22 ) = 1 - (0.34375 2 + 0.65625 2 ) . 0.451.
Observed Heterozygosity
A simple measure of genetic diversity is the proportion of all heterozygotes
from all plants (homozygotes and heterozygotes) investigated in a population:
n n n
Ho = / / Pij = 1 - / Pii . (3.5)
i = 1j = 1 i = 1
i ! j
3.3.2
Genetic Differentiation Among Populations
3.3.2.1
Genetic Distances
Out of a large choice of commonly used genetic distances only the minimum
genetic distance according to Nei (1973) is mentioned here. For two popula-
tions k and l it is defined as
d min (k , l ) = 12 / 7 p i (k) - p i (l )A ,
2
(3.6)
i
where pi(k) and pi(l ) are the frequencies of the ith allele in populations k and l,
respectively. This distance reaches its minimum value if the allelic structures of
both populations are identical [pi(k)=pi(l) for all alleles i], and has a maximum
value of 1 if the populations are fixed on different alleles. More commonly used
distance measures such as Neis standard genetic distance are described by Nei
(1987).
If more than two populations are investigated, it is possible to compute a
matrix with all pairwise genetic distances, and to use this information as a basis
for the illustration of genetic differentiation patterns by means of cluster analy-
ses (e.g., Fig. 3.4). Populations are grouped on the basis of their genetic distances
by different algorithms. A simple and commonly used method for grouping in
population genetics is the unweighted pair-group method with arithmetic
means (UPGMA) method (Sneath and Sokal 1973). Similar methods are used
to illustrate the relatedness of higher taxa in molecular phylogenies (Sect. 8.3).
3.3.2.2
Genetic Differentiation
alleles (pi) calculated on the basis of their occurrence in the total population.
The average gene diversity within populations (HS) is calculated as mean gene
diversity (He) within the subpopulations. It can be shown that the difference
HT HS equals the average minimum genetic distance of Nei (see before)
between all population pairs (DST = HT HS) (Finkeldey 1994). GST (or FST) is
the relative contribution to the total gene diversity which is due to differences
among populations: GST = FST = DST /HT = (HT HS)/HT.
3.4
Genetic Variation in Tropical Forest Species General Trends
Levels and patterns of genetic variation at single loci depend on numerous fac-
tors. Apart from species-specific or population-specific variability, the type of
the marker(s) investigated is of crucial importance. An assessment of genetic
variation levels is meaningful only if results are related to those of comparable
studies. Most available data on patterns of genetic variation of tropical forest
plants were based on isozyme inventories, although the number of studies
using molecular markers is constantly increasing.
In view of the multitude of tropical forest plants it is hardly surprising that
levels and patterns of genetic variation considerably vary among species
(Table 3.1). In tropical forests, plant species with low, intermediate, and high
levels of genetic diversity within populations are encountered.
Mean levels of genetic variation of woody plants, mainly forest trees, from
different climatic regions were compared by Hamrick et al. (1992) (Table 3.2).
At isozyme gene loci, the measures of genetic multiplicity (PPL; number of
alleles per locus) are on average slightly lower for woody plants from the trop-
ical region as compared with temperate and boreal species. However, this does
not hold for the average expected heterozygosity (He) as the most commonly
used measure of genetic variation. Thus, there is no evidence for fundamental
differences with regard to levels of genetic variation of tropical and temperate
trees, even though the majority of tropical trees occur in low density, while
many temperate and boreal species form large, pure stands.
The impact of density on levels of genetic variation was studied in more
detail by Hamrick and Murawski (1991). Genetic variation of 16 uncommon
tropical forest tree species (density less than 0.5 trees per hectare) occurring on
Barro Colorado Island, an artificial island in the Panama Channel, was assessed
at isozyme gene loci, and compared with variation levels of 16 common
species (density more than four trees per hectare). Variation levels within
uncommon species varied from low (PPL = 14.3%, He = 0.026 for Tricanthera
gigantea) to high (e.g., PPL = 76%, He = 0.257 for Myrospermum fructescens).
The average variation of the uncommon species is lower than the variation of
more common species (Table 3.3); thus, there is evidence for losses of genetic
3.4 Genetic Variation in Tropical Forest Species General Trends 29
Table 3.1. Examples of tropical tree species showing low and high levels of genetic
variation within populations, respectively
Species Country/region Marker type He Reference
Low variation
Acacia mangium Australia, Papua Isozyme 0.017 Moran et al. (1989a)
New Guinea
Acer skutchii Mexico SSR 0.112 Lara-Gomez et al. (2005)
Moringa oleifera India, East Africa AFLP 0.067 Muluvi et al. (1999)
Pentaclethra Costa Rica Isozyme 0.074 Hall et al. (1994a)
macroloba
Pinus merkusii Thailand Isozyme 0.058 Changtragoon and
Finkeldey (1995a)
High variation
Acacia albida West Africa Isozyme 0.442 Joly et al. (1992)
(=Faidherbia albida)
Calycophyllum Peru AFLP 0.249 Russel et al. (1999)
spruceanum 0.349
Gliricidia sepium Central America Isozyme 0.238 Chamberlain et al. (1996)
Pinus caribaea Caribbean Islands Isozyme 0.26 Zheng and Ennos (1999)
Pterocarpus Thailand Isozyme 0.246 Liengsiri et al. (1995)
macrocarpus
Shorea leprosula Malaysia Isozyme 0.406 Lee et al. (2000)
Swietenia macrophylla Brazil SSR 0.781 Lemes et al. (2003)
He average expected heterozygosity within populations, SSR simple sequence repeat, AFLP amplified
fragment length polymorphism
Table 3.2. Average genetic variation at isozyme gene loci in long-lived woody species
from different climatic regions. (Adapted from Hamrick et al. 1992)
Climatic zone No. of species Average no. PPL (%) A/L He
of populations
Table 3.3. Mean genetic variation at isozyme gene loci for 16 uncommon (rare;
density less than 0.5 trees per hectare) and 16 common (density more than four trees
per hectare) tree species on Barro Colorado Island in comparison with average
variation of 486 plants and 115 long-lived woody species. (Adapted from Hamrick and
Murawski 1991)
Group No. of species Average no. PPL (%) He
of gene loci
Forest trees are among the genetically most variable of all organisms. This
result was first obtained for a few species of the temperate and boreal zone
and later confirmed for tropical tree species. The genetic variation within
populations at isozyme gene loci is similar for temperate and tropical tree
species (Table 3.2).
Several important tropical tree species are characterized by low levels of
genetic variation within populations, while other species are highly diverse
(Table 3.1). It is not possible to reliably predict genetic variation levels for
a particular species on the basis of its life-history characteristics.
Endemic species are, on average, less variable within populations than
species with a wide distribution range (Loveless 1992).
Uncommon species, i.e., those occurring in very low population densities
in tropical forests, are, on average, less variable than species occurring in
3.5 Case Studies on Genetic Variation Patterns 31
higher density (Table 3.3). However, the expected heterozygosity is high for
many tree species even if the population density is low.
The degree of genetic differentiation among populations (GST) is slightly
higher for tropical forest tree species than for tree species of temperate
forests. Tropical tree species with abiotic means of seed dispersal (baro-
chorous and anemochorous species) show, on average, a much higher dif-
ferentiation among populations (GST = 0.138) than zoochorous species
(GST = 0.050) (Sect. 5.3). Seed dispersal by animals is usually very efficient
and results in small genetic differentiation among populations of zoo-
chorous tree species (Loveless 1992).
3.5
Case Studies on Genetic Variation Patterns
Since predictions of genetic variation patterns are very unreliable for a partic-
ular species, experimental investigations on the distribution of genetic
variation within and among populations are mandatory for a detailed under-
standing of variation patterns of a target species. Six case studies were selected
as examples to illustrate genetic variation patterns in tropical forest species out
of a wide variety of published reports. The case studies were chosen since they
represent different:
The focus is on genetic variation patterns and variation levels in forest tree pop-
ulations without or with only minor human impact. Genetic effects of human
disturbance of tropical forest ecosystems will be discussed in later chapters.
islands (Fig. 3.1). A genetic inventory was conducted at 22 isozyme gene loci
coding for 17 enzyme systems in 18 natural populations of the species from
Papua New Guinea and Australia (Wickneswari and Norwati 1993). Clear
differences with regard to the level of genetic variation measured as expected
heterozygosity were observed (Fig. 3.1). Populations from Papua New Guinea
revealed the highest level of genetic variation (average He = 0.133 or 13.3%),
which is comparable to the average for long-lived woody plant species
(Table 3.3). Intermediate levels of genetic diversity were observed in popula-
tions from Queensland, and low levels in populations from the Northern
Territory of Australia. Similar results were also reported for other measures of
genetic variation (PPL; alleles per locus; data not shown).
Thus, Papua New Guinea is a center of genetic diversity of the species.
Wickneswari and Norwati (1993) explained the observed patterns of genetic
variation by the evolutionary history of the species. They suspect that the
species survived less favorable climatic conditions in small, scattered refuges in
Papua New Guinea, and expanded only recently via land bridges connecting
Papua New Guinea and Australia until the end of the last glaciation. According
to this hypothesis, only a few trees colonized Queensland and, later, the
Northern Territory. Thus, the lower levels of genetic variation in Australia as
compared with Papua New Guinea are explained by bottleneck effects (genetic
drift) during recent recolonization (From Wickneswari and Norwati 1993).
Fig. 3.1. Natural distribution and genetic variation of Acacia auriculiformis. He expected
heterozygosity. (Adapted from Wickneswari and Norwati 1993)
3.5 Case Studies on Genetic Variation Patterns 33
0.25
r 0.725
(P=0.003)
Genetic distance (Nei)
0.20
0.15
0.10
0.05
0.00
0 50 100 150 200 250
Geographic distance (km)
Fig. 3.2. Comparison of genetic distances (Nei 1972) and geographic distances between
ten populations of Eugenia dysenterica. (From Telles et al. 2003)
34 C HAPTER 3 Genetic Variation of Tropical Forest Plants
of 18.4 alleles per locus observed in all populations and a mean expected het-
erozygosity of 0.781 (Table 3.4). The inbreeding coefficient was positive and
significantly different from 0 in four populations, indicating a slight excess of
homozygotes relative to HardyWeinberg proportions (Sect. 6.4.1). The total
genetic differentiation among populations was estimated as 0.097. This is only
slightly lower than the average reported for tropical forest trees (GST = 0.109;
Sect. 3.4) and may be regarded as a moderate level of genetic differentiation
among populations.
High levels of genetic variation were observed since microsatellites (SSRs)
were chosen as marker types. SSRs are hypervariable markers showing higher
levels of genetic variation in comparison with other codominant markers such
as isozymes (Sect. 2.3.1). However, genetic variation in the Amazon basin was
higher than in a directly comparable study on mahogany in Central America
(Novick et al. 2003). On the other hand, the differentiation among populations
was slightly higher among Central American populations. The significant
genetic differentiation among populations in both regions stresses the impor-
tance of conserving as many populations as possible rather than of concentrat-
ing on one or a few populations in a region to be conserved in situ (Chap. 14)
(From Lemes et al. 2003).
Table 3.4. Genetic variation at eight microsatellite gene loci in seven populations of
Swietenia macrophylla from the Brazilian Amazon. (Adapted from Lemes et al. 2003)
Population A/L He Ho F Deviationa
SB
(S. lepr.
S. parv.)
BB
(S. parv.)
PS
TS (S. lepr.
(S. lepr. S. parv.)
S. parv.)
NS
(S. lepr.
S. parv.)
AS TB
(S. lepr. (S. lepr.)
S. parv.)
HA
(S. lepr.)
Java
Fig. 3.3. Location of sampled populations of Shorea leprosula and S. parvifolia in
Indonesia. Abbreviations as in Table 3.5. (From Cao et al. 2006)
Table 3.5. Genetic variation at AFLP loci in two dipterocarp species (Shorea leprosula
and Shorea parvifolia) in Indonesia. (Adapted from Cao et al. 2006)
Species Population Abbreviation Location PPL (%) He
S. lepr._AS
S. lepr._HA
S. lepr._NS
S. lepr._PS
S. lepr._TS
S. lepr._TB
S. lepr._SB
S. parv._SB
S. parv._BB
S. parv._TS
S. parv._AS
S. parv._PS
S. parv._NS
Fig. 3.4. Unweighted pair-group method with arithmetic means tree based on Neis
standard genetic distance illustrating genetic differentiation among populations of two
dipterocarps from Indonesia. Abbreviations as in Table 3.5. (From Cao et al. 2006)
Panama
Gualaca 5
Las Lajas 5
San Francisco 1 4
Costa Rica
Horizontes 20
Puriscal 20
Canas 20
Palo Verde 20
Hojancha 20
Jimenez 20
Talamanca 20
Upala 19 1
Pacifico Sur 20
Perez Zeledon 20
Nicaragua
Ometepe 20
Masatepe 20
Wabule 20
La Trinidad 20
Honduras
Meambar 20
Taulabe 20
Comayagua 20
La Paz 20
Guatemala
Los Esclavos 4 1
El Idolo 5
Tikal 5
San Jose 5
Mexico
Calakmul 20
Bacalar 20
Zona maya 20
Escarcega 20
Total 99 1 259 106 9
3.6
Recommended Literature
Flowering plants have evolved one of the most complex and sexiest life cycles on
earth (Armstrong 1996).
4.1
Introduction
4.2
Sexual Reproduction
compared with the chromosome number in somatic cells of the parental plant.
The consequences of this process for genetic structures of a progeny generation
derived from a parental stand (t) and the possible causes of deviations of the
genetic structures in the progeny generation (t +1) from the population of adults
are illustrated in Fig. 4.1. Trees of generation t produce female and male gametes.
Differences in the ability to produce gametes result in fertility differences among
trees (fertility selection). Male gametes (pollen) are distributed during the flow-
ering period and external pollen from other populations is carried in the popu-
lation (pollen influx). Mating preferences or incompatibilities possibly result in
differences between allelic structures of the pollen cloud reaching a stigma and
the fertilizing or effective pollen, for example, owing to differences in pollen tube
growth. Viability selection starts early after fertilization and results in differences
of genetic structures between fertilized ovules and fully developed seeds. Viability
selection continues during seed germination, seedling growth, and throughout
stand development. For practical reasons, it is not easily possible to observe the
genetic structures of female and male gametes (pollen and ovules) or zygotes.
Thus, changes of genetic structures are often monitored by comparing the adult
population (generation t) with the mature seeds (generation t +1; comparison 1 in
Fig. 4.1), by comparing the seeds with the population of seedlings, saplings, or
adults (comparison 2 in Fig. 4.1; Example 7.1), or by comparing two generations
of the same or, in case of long-living organisms like forest trees more frequently,
of different age (comparison 3 in Fig. 4.1).
4.2.1
Sexual Types and Sexual Systems
Only gametes of opposite sex can fuse and jointly produce the zygote which
will further develop by mitotic cell divisions into an embryo, a seedling, and
eventually into an adult plant capable of producing gametes itself. The pro-
duction of these two different types of gametes (female and male) is a require-
ment for successful sexual reproduction in a population. The sexual type of an
organism depends on the types of gametes it produces: It is either cosexual
(female and male gametes are produced) or unisexual female or male. Plants
not capable of producing any functioning gametes are sterile.
The determination of the sexual type of an organism is usually under strict
genetic control although an environmental influence on the sexual type has
been observed for some animals. The sexual type of most animals and some
plants is determined by particular chromosomes (gonosomes, i.e., sexual chro-
mosomes). Females often have two identical gonosomes: they are homoga-
metic. Males of such species have two different gonosomes: they are
heterogametic. An example of sex determination by gonosomes are the X and
4.2 Sexual Reproduction 43
production of
and gametes
gametes
fertility selection distribution of
flowering over time
trees of pollen transport
pollen influx
generation t
pollination
mating preferences
incompatibility
comparison 1
fertilization
viability selection
(germination, mortality)
comparison 2 stand
n development
c tio
sele
l ity
bi
via
trees of
generation t + 1
Fig. 4.1. Causes for changes of genetic structures during and after reproduction in
consecutive generations. (From Hattemer 1999)
chromosomes have been identified for a few dioecious plants, for example, in
the herbaceous genera Silene and Rumex (Dellaporta and Calderon-Urrea
1993). A particular region on the LG1 chromosome has been identified as
carrying sex-determining genes in papaya (Carica papaya) (Charlesworth
2004). Sex expression is regulated by hormones in many dioecious plants
(Dellaporta and Calderon-Urrea 1993), but the mechanism of sex determi-
nation is still unknown for most woody plant species and for virtually all
tropical forest species.
A possible determination of the sex by age (or size) of trees has been
described for the neotropical tree Iryanthera macrophylla. The plants of this
species produce only one type of flower (female or male) in a particular repro-
duction period; however, Ackerly et al. (1990) found evidence that small, young
trees of this species flower male, but later produce female flowers. Thus, the sex
expression may not be determined by genetic factors, but by the age or size of
plants, and the species may not be dioecious, but diphasic (Ackerly et al. 1990).
4.2.2
Sexual Systems
Cosexuality: all plants produce both types of gamete (female and male)
Dioecy: plants produce only one type of gamete they are female (female)
or male
Trioecy: some plants produce both types of gamete (female and male),
others are female or male
Gynodioecy: some plants produce both types of gamete (female and male),
others are female (no males)
Androdioecy: some plants produce both types of gamete (female and
male), others are male (no females)
Cosexuality is the archetypal and most common sexual system for plants. Two
forms of cosexuality are distinguished: Hermaphroditic plants have complete
flowers producing both types of gametes in a single flower. Monoecious plants
produce both types of gamete in different flowers. Thus, single flowers are
either male or female in monoecious species, but these flowers are produced by
the same plant (Table 4.1).
Different types of pollination can be distinguished for hermaphroditic species
depending on the origin of the fertilizing pollen (Fig. 4.2). Autogamy is the
4.2 Sexual Reproduction 45
Allogamy
Xenogamy
Autogamy Geitonogamy
Single flower
Selfing Outcrossing
Fig. 4.2. The origin of fertilizing pollen for a flower for a hermaphroditic tree species.
(Adapted from Hattemer et al. 1993)
fertilization of an ovule by pollen from the same flower. It is just like geito-
nogamy, the fertilization with pollen from a different flower of the same tree,
a form of allogamy resulting in selfing (self-pollination; Sect. 6.3). Xenogamy is
pollination from a different plant and results in outcrossing. Autogamy is not
possible for monoecious species, and no form of allogamy, neither autogamy
nor geitonogamy, is possible for dioecious species, which are, in consequence,
obligatory outcrossed.
46 C HAPTER 4 Sexual and Asexual Reproduction in Tropical Forests
Table 4.2. Frequency (%) of common sexual systems in tropical and temperate forests.
(Adapted from Bawa and Opler 1975)
Location Cosexual: hermaphroditic Cosexual: monoecious Dioecious
Tropics
Costa Rica 68 10 22
South Nigeria 47 13 40
Sarawak 60 14 26
Florida 61 12 27
Temperate zone
Ohio 6 83 11
Ohio 27 60 13
Arkansas 0 83 17
Kentucky 15 70 15
New Hampshire 13 81 6
4.2 Sexual Reproduction 47
assumed to be the main reason for the evolution of dioecy for many originally
self-compatible species of tropical forests (Bawa and Opler 1975).
The frequency of sexual types in a forest is related to main pollinators (Sect.
5.2.2). Monoecious species are mainly wind-pollinated. The complete flowers
of hermaphroditic species are predominantly pollinated by animals.
4.2.3
Sexual Structures and Sexual Function
Table 4.3. Sexual structures of tree species in a tropical forest in Costa Rica. Test for
significant deviation from an even sex ratio (male-to-female ratio of 1; com. X2) and
test for significant heterogeneity of the sex ratio among populations (int. X2). (Adapted
from Opler and Bawa 1978)
Family Species No. of No. of No. of Male-to- Com. Int.
Popu- males females female X2 X2
lation ratio
Table 4.3. Sexual structures of tree species in a tropical forest in Costa Rica. Test for
significant deviation from an even sex ratio (male-to-female ratio of 1; com. X2) and
test for significant heterogeneity of the sex ratio among populations (int. X2). (Adapted
from Opler and Bawa 1978)Contd
Family Species No. of No. of No. of Male-to- Com. Int.
Popu- males females female X2 X2
lation ratio
** **
Polygonaceae Coccoloba caracasana 6 55 136 0.41
Coccoloba floribunda 1 23 27 0.85 NS
Ruprechtia costata 1 21 30 0.70 NS
**
Triplaris americana 8 157 220 0.71 NS
* **
Sapindaceae Allophyllus 4 74 50 1.48
occidentalis
Simaroubaceae Simarouba glauca 4 111 102 1.08 NS NS
**
Rubiaceae Albertia edulis 4 83 49 1.69 NS
Genipa caruto 6 76 77 0.99 NS NS
** **
Randia spinosa 9 154 57 2.70
**
Randia subcordata 7 157 77 2.04 NS
**
Zanthoxylaceae Zanthoxylum 4 115 60 1.92 NS
setulosum
NS not significant
*
P < 0.05; **P < 0.01
various taxa. A significant deviation from an expected even ratio of male and
female plants was observed for ten out of the 23 species (com. X2 in Table 4.3).
Significantly more males than females (male-to-female ratio greater than 1
and significant com. X2) were observed for eight species. A significant excess of
females (male-to-female ratio less than 1 and significant com. X2) was observed
for only two species. Significant heterogeneity of the sex ratio among popula-
tions (int. X2) was observed for four species with P < 0.01.
Deviations from a balanced sex ratio were also observed in populations of
functionally dioecious figs (Ficus spp.) and rattan (Calamus spp.). In most
cases an excess of male plants has been reported (Table 4.3). Possible reasons
for an unbalanced sex ratio are:
likely for many dioecious plants. Female plants need to supply energy and
nutrients for the production of female flowers and seeds which are often
embedded in large fruits. These resources are not available for vegetative
growth. This possibly results in reduced viability as compared with that of
male plants producing only pollen in male flowers.
Earlier maturity of one sex, usually males, as compared with the opposite
sex, or higher frequency of flowering of one sex. For example, annual flow-
ering for males but a biannual cycle for females results in an excess of males
in each flowering period.
Preferential asexual reproduction of one sex. Strict genetic control of the sex
results in identical sex expression of all plants from vegetative reproduction
of a particular (female or male) tree.
4.3
Asexual Reproduction
Many plant species and some animal species have the potential for asexual
reproduction. In this case, offspring is produced from tissue of a single organ-
ism by mitotic cell divisions only; meiosis and recombination do not occur.
50 C HAPTER 4 Sexual and Asexual Reproduction in Tropical Forests
Thus, the genotypes of asexually produced offspring are identical copies of the
genotype of the original plant. Obviously, asexually produced progenies of a
single plant are also genetically identical among each other. A collective of
genetically identical organisms is called a clone.
4.3.1
Vegetative Reproduction
4.3.2
Apomixis
embryos. Only one or two of these develop into seedlings. In most cases, the
sexual embryo dies during seed development since adventive embryos grow
faster. Thus, seed production requires regular fertilization, but embryos of
most mature seeds have an apomictic origin (Baker 1960). This form of seed
development (pseudogamy) requiring pollination but resulting in apomictic
embryos has also been described for fruit species of the genus Citrus.
At least facultative apomixis has been described for several dipterocarps,
including Stemonoporus oblongifolius (Murawski and Bawa 1994), Hopea odor-
ata (Wickneswari et al. 1995), and the autotetraploid Shorea ovalis ssp. sericea
(Ng et al. 2004). Apomixis is suggested to be an evolutionary strategy allowing
the maintenance of high genetic diversity in populations of Shorea ovalis ssp.
sericea (Ng et al. 2004). Apomixis in H. odorata is possibly related to the occur-
rence of triploid trees in this species (Jong 1976).
Kaur et al. (1978) suggested that apomixis may be common not only for
dipterocarps, but also for other tropical forest plants. The observation of abun-
dant seed crops of spatially isolated trees of dioecious or self-incompatible
species suggests apomixis as a likely explanation. However, marker-based stud-
ies rarely confirmed apomixis in tropical forests. In many cases, long-distance
transport of effective pollen seems to be responsible for the production of out-
crossed seeds even for trees isolated by distances of several hundred meters
from other flowering trees of the same species serving as potential pollen
donors (Sect. 5.2.3).
4.4
Recommended Literature
Early studies of Bawa and Opler (Bawa 1974, 1980; Bawa and Opler 1975;
Opler and Bawa 1978) on the sexual system of neotropical trees revealed the
high frequency of dioecy and self-incompatibility. Kaur et al. (1978) suspected
that apomixis may be widespread among tropical forest plants. This idea stip-
ulated in-depth studies which, however, only rarely supported their hypothesis.
Gene Flow and Migration 5
5.1
Introduction
5.2
Gene Flow Through Pollen
5.2.1
Pollination and Fertilization
Sexual reproduction is based on the fusion of a male and a female gamete. The
transportation of the male gametophyte to the female plant is required in seed
plants. This requires a series of events in angiosperms: First, pollen is trans-
ported to the stigma of the same (autogamy) or a different (geitonogamy or
xenogamy; Fig. 4.2) flower. The pollination is followed by the growth of a
pollen tube through the style to the ovule. Then, after entering the ovule
through the micropyle, fertilization of the embryo sac takes place (p. 32 ff. in
Proctor et al. 1996).
The genotypic structures of the progeny generation are influenced both by
the dispersal of pollen and by genetic differences of the pollen with regard to
germination, pollen tube growth, and fertilization success. Pollen dispersal
from anthers to stigmas or, in the case of gymnosperms, to the micropyle is dis-
cussed in this chapter. Aspects related to the processes of the germination of
pollen, pollen tube growth, and fertilization are covered in Sect. 6.5.1.
5.2.2
Pollen Vectors in Tropical Forests
5.2.2.1
Pollination by Wind (Anemogamy)
Wind is the most important pollen vector of trees in the temperate and boreal
zone. For example, the principal stand-forming species in central Europe belong-
ing to the families Fagaceae (Fagus sylvatica, Quercus spp.) and Pinaceae (Pinus
sylvestris, Picea abies) are wind-pollinated (anemogamous). Wind pollination is
rare in the humid central tropics (Regal 1982); only 12% of plants in the central
5.2 Gene Flow Through Pollen 55
tropics disperse their pollen by the wind (Bawa 1990). However, this proportion
increases in semidry and dry tropical forests. Tropical pines (Pinus spp.) and
many other gymnosperms such as Podocarpus species are wind-pollinated (Regal
1982). Pioneers and species commonly found in secondary forests such as some
Cecropia spp. are also anemogamous. Wind has been suggested as a pollen vector
for some eucalypts, but these statements are not supported by experimental
evidence (p. 190 in Eldridge et al. 1994).
The success of pollen parents in pollinating a particular seed tree depends
on numerous factors, including the spatial structure and the distance between
trees, environmental factors, including the preferential wind direction, velocity,
and competing vegetation acting as a pollen filter, species-specific criteria
such as pollen morphology and pollen size, and individual characteristics,
including the number of pollen produced during a flowering period. For exam-
ple, there is evidence for limited pollen dispersal and preferential matings
among near neighbors for the wind-pollinated Araucaria angustifolia (Sousa
and Hattemer 2003; Sect. 11.4.3).
A large number of studies conducted on wind-pollinated tree species in
temperate forests revealed both preferential pollen transport from directly
neighboring trees (e.g., for Pinus sylvestris; Mller 1977) and a large proportion
of successful long-distance pollination (e.g., for Quercus macrocarpa; Dow and
Ashley 1998).
Observations in clonal seed orchards (Sect. 12.8.1) of wind-pollinated
conifers in temperate and boreal climates proved a high frequency of contam-
ination by external pollen (Wheeler and Jech 1992); thus, spatial isolation of
clonal seed orchards from conspecific forest stands does frequently not result
in reproductive isolation owing to efficient long-distance dispersal of pollen by
the wind. A high frequency of matings after long-distance pollen dispersal has
been suggested by Lindgren et al. (1995) for Scots pine (Pinus sylvestris). The
authors assume that viable pine pollen can be transported over hundreds of
kilometers before becoming effective in the boreal climate of Scandinavia.
The high species diversity and, hence, the low population density is likely
to be the main reason for the rarity of wind-pollination in evergreen moist
tropical forests, since wind-pollination is usually not as effective as animal
pollination where inter-individual distances are large (Regal 1982).
5.2.2.2
Pollination by Animals (Zoogamy)
Vertebrates
Only small flying vertebrates (birds and bats) are important pollinators for
tropical plants. Nonflying mammals such as opossums and monkeys have been
reported to act as pollinators for a few plant species with robust flowers
(Janson et al. 1981), but most flowers are severely damaged by visiting
nonflying mammals.
Bats
Many Bombacaceae such as the baobab (Adansonia digitata) and Ceiba
pentranda (Lobo et al. 2005), but also some Parkia spp. (Mimosaceae) and
Bauhinia spp. (Caesalpinaceae; Heithaus et al. 1974) are predominantly pol-
linated by bats. Flowers visited by bats are usually very big and white or light
yellow. They open at dawn or during the night, smell strong, and produce large
quantities of nectar (p. 36 in Whitmore 1975). Inflorescences either stand
upright over the crown or hang from long twigs. Sometimes they are produced
by modified twigs originating directly from the stem (cauliflory, e.g., the can-
nonball tree Couroupita guianensis; Fig. 5.1). The production of flowers
suitable for bat pollination is energy-demanding for a tree. On the other hand,
most bats fly long distances. Accordingly, they are efficient vectors for the
transport of pollen over long distances. The efficiency of bats as pollinators
critically depends on the behavior of each species as well as preferred resting
areas (Heithaus et al. 1974; Stuart and Marshall 1976).
Fig. 5.1. Flowers of the cauliflorous cannonball tree Couroupita guianensis (Lecthidaceae).
(Photo: R. Finkeldey)
5.2 Gene Flow Through Pollen 57
Birds
Only comparatively few small bird species act as pollinators for tropical forest
plants. Hummingbirds are well-known pollinators of forest plants in tropical
America. Their function is fulfilled by various other species belonging to the
Zosteropidae, Dicaeidae, Nectariniidae, and Meliphagidae in other parts of
the tropics.
Many flowers visited by birds are bright red and big. They produce large quan-
tities of nectar, which is the main energy source for the pollinators, but the flow-
ers rarely have strong smells. Many inflorescences are brushlike with long
filaments. The conspicuous flowers of Erythrina spp. and Bromeliaceae are polli-
nated by birds. Heliconia spp. (Musaceae) are frequently planted ornamentals
originating from America. Their beautiful flowers are adapted to pollination by
birds. Several taxa, such as species of the genus Calliandra, are pollinated by birds,
bats, and possibly also insects. The production of bird-flowers is as energy-
consuming as the production of bat-flowers. The size and the behavior of birds
suggests that they are efficient vehicles for long-distance transport of pollen;
however, the actual distances between two trees which can be bridged by birds
critically depends on the sizes of their territories.
Pollination by Insects
Insects are the most important pollen vectors in the tropics. Many generalist
species visit flowers of a multitude of plants. Other pollinating insects coe-
volved with particular plant species. The obligatory symbiosis between figs
(Ficus spp.) and their pollinators (fig wasps; Agaonidae) is a particular well-
known example of a close coevolution between tropical plants and insects
(Janzen 1979a; Example 7.2).
Hymenoptera
The Hymenoptera (bees, wasps, bumble bees, etc.) are an extraordinary
species-rich group of pollinators. Although most bees forage on flowers of
numerous plant species, they often prefer a particular species at a given time.
The flight behavior of bees and hence their efficiency as pollinators depends on
numerous factors, including their size, their sex (male or female; Ackerman
et al. 1982), and their social or solitary life cycle.
There are few uniform features of flowers which are exclusively or pre-
dominantly pollinated by Hymenoptera. The flower morphology greatly
varies even within a single tree family such as the neotropical Lecythidaceae,
which is predominantly pollinated by bees (Prance 1985). Some plant species
attract numerous bees and wasps. For example, more than 70 different bee
species were counted visiting the flowers of a single Andira inermis tree (Bawa
1990).
58 C HAPTER 5 Gene Flow and Migration
Bees prefer to fly short distances between visited flowers. Accordingly, most
visits are within the crown of a single tree, and geitonogamy is expected to be
common (Sect. 4.2.2). Strategies to minimize geitonogamy and hence self-
fertilization include the evolution of incompatibility mechanisms (Sect. 6.5.1)
and a particular flowering phenology (see Example 6.1 for teak).
Other Insects
Other, frequently inconspicuous groups of insects serve important functions
for forest ecosystems since they are the main pollinators of keystone plant
species. Tiny thrips, cicadas, and bugs belong to this group. Dipterocarps from
the central tropics show the phenomenon of simultaneous mass flowering in
irregular intervals. Many of them, in particular Shorea and Hopea spp. with
small flowers, are mainly pollinated by tiny thrips (Thysanoptera) (Ashton
1988). Another example of the enormous diversity of pollinating insects is the
climber Uvaria elmeri (Annonaceae), which is pollinated by cockroaches
(Nagamitsu and Inoue 1997).
5.2.3
Pollen Dispersal Assessed by Marker-Based Studies
5.2.3.1
Dispersal of Rare Alleles
Example 5.1: Gene Dispersal in a Clonal Seed Orchard of Teak (Tectona grandis)
A tree in a clonal seed orchard of teak (Tectona grandis) comprising 16 clones
was found to be the unique carrier of a rare allele A2 at the isozyme gene locus
PGM (genotype A1A2). Since no other tree in the investigated compartment of
the seed orchard exhibited this allele, it is unlikely that the tree was a ramet of
5.2 Gene Flow Through Pollen 61
1997
40%
20%
10%
20 40 60 80 100
Distance from marker tree (m)
Fig. 5.2. Success of a teak (Tectona grandis) tree as a pollen parent for progenies of seed
trees growing at different distances in a clonal seed orchard in Thailand. (Adapted from
Finkeldey 2006)
5.2.3.2
Paternity Analyses
a particular progeny if the tree does not carry an allele which was paternally
transmitted to the progeny at any gene locus. Obviously, the reliability of pater-
nity exclusion procedures increases with an increasing number of observed
gene loci and with increasing variability at the observed loci. Hypervariable,
codominant microsatellites (simple sequence repeats; Sect. 2.3.1) are partic-
ular suitable tools for paternity analyses (Example 5.2). The calculation of
paternity likelihoods by special software such as FAMOZ (Gerber et al. 2003)
is often preferable to simple exclusion procedures.
Example 5.3: Pollen Dispersal in Three Neotropical Tree Species with Different
Spatial Distribution Patterns on Barro Colorado Island
The impact of population density and population structure on pollen disper-
sal was studied for three insect-pollinated species growing on Barro Colorado
Island, an artificial island in the Panama Canal (Stacy et al. 1996). Paternity
exclusion was conducted for each species on the basis of the observation of
variation at five to nine polymorphic isozyme gene loci. All adult or subadult
trees in an area of 50 or 84 ha were sampled, and seeds were collected under
reproducing trees. Minimum estimates of pollen dispersal were calculated
assuming that the nearest not-excluded potential pollen parent according to
the paternity analysis was the true pollen parent.
No selfing was observed for any of the populations investigated, and con-
siderable distances between pollen parents and seed trees were bridged by the
pollinators; however, the pollination patterns differed among the species
studied and depended on the spatial distribution of trees.
The distribution of the 25 adult trees of Calophyllum longifolium
(Clusiaceae)was rather uniform throughout the area investigated (84 ha). The
nearest neighbor was rarely identified as a pollen parent in the case of spatially
isolated seed trees, and the successful pollen parent was separated from the seed
tree by a distance of at least 210 m for 62% of the progenies investigated. If
trees were widely separated, pollen flow ranged up to several hundred meters.
However, cross-pollination seemed to be frequent within the only cluster of
two directly neighboring trees, which could not be excluded as a pollen parent
for 72 and 73% of the progenies investigated from the other tree in this cluster,
respectively.
Fourteen adults of Spondias mombin (Anacardiaceae) were counted in the
same area of 84 ha. Pollination was mainly within clusters, and the nearest
neighbor could not be excluded as a pollen parent for roughly 90% of the prog-
enies investigated for this species. On the other hand, a low level of gene flow
through pollen over distances of at least 300 m was also confirmed, and pollen
flow from outside the area studied was confirmed by the occurrence of rare
alleles in progenies which were not observed in any adult in the plot.
Only five of the 30 potentially reproducing trees of Turpinia occidentalis
(Staphyleaceae) in the 50-ha plot studied produced abundant seeds in the sea-
son studied (1993). These five trees occurred in two small clusters of two and
three trees with a diameter of 35 and 57 m, respectively. The clusters were sep-
arated by a distance of approximately 235 m. Pollen flow was mainly within
clusters, and pollen flow over a distance of more than 130 m was confirmed for
only one out of 115 progenies investigated.
These results point towards a considerable importance of the spatial population
structure on pollen dispersal by small insects in tropical forests. Predominance of
64 C HAPTER 5 Gene Flow and Migration
pollination among near neighbors was observed mainly within clusters and
small groups of trees. Pollen dispersal is more far-reaching and not restricted
to nearest neighbous in the case of a more random distribution of reproducing
adults in the forest (From Stacy et al. 1996).
5.2.3.3
Model-Based Estimates of Pollen Dispersal
5.2.4
Efficiency of Pollen Vectors for Gene Flow
5.3
Migration of Seeds
5.3.1
Seed Vectors
Seeds are dispersed by biotic and abiotic means. Water and wind are the prin-
cipal abiotic dispersal agents. Biotic seed dispersal is facilitated by animals.
Humans play an enormous role in todays distribution of plants and patterns
of genetic variation within species.
5.3.1.1
Abiotic Seed Dispersal
The diaspores of many seed plants contain structures promoting the dispersal
of seeds by the wind. The name of the tree family Dipterocarpaceae refers to
two wings attached to the fruits of many dipterocarps. However, some diptero-
carps have more than two wings, and the wings of several species are only
66 C HAPTER 5 Gene Flow and Migration
5.3.1.2
Biotic Seed Dispersal
Vertebrates, in particular birds and mammals, are by far the most important
biotic seed dispersers of tropical forest plants (Loiselle et al. 1996).
Endozoochorous seed dispersal requires a passage of seeds through the gas-
trointestinal tract of an animal. Otherwise, dispersal of seeds by animals is exo-
zoochorous. Seed dispersal of tree species with fleshy, juicy fruits is typically
endozoochorous. Exozoochorous dispersal is typical for nuts.
Most bird species and both flying (bats, flying foxes) and nonflying (rodents,
pigs, primates, and others) mammals disperse seeds. The seeds of the small
mountain tree Dunalia arborescens (Solanaceae) are dispersed by at least 16
different bird species in Jamaica (Cruz 1981).
5.3.2
Efficiency of Seed Dispersal
The type of seed vector and, in the case of biotic seed dispersers, their behav-
ior are crucial factors for the seed shadow of forest plants. A comparison
between seed dispersal agents in temperate and tropical American forests
revealed a dominance of endozoochorous seed dispersal only in the tropics.
Endozoochorous, exozoochorous, and wind dispersal are of similar impor-
tance in temperate forests of North America (Howe and Smallwood 1982). In
general, endozoochorous seed dispersal results in a wider dispersal of seeds as
compared with that from other mechanisms since the fruits containing seeds
are digested before the seeds are defecated. Endozoochorous seed dispersal may
be wide but clustered, for example, under the rest trees of seed-dispersing birds
5.4 Long-Distance Gene Flow and Migration in Tropical Forest Species 67
and bats, and may result in nonrandom dispersal patterns (Loiselle et al. 1996).
The comparatively low differentiation among populations of the rare
European wild service tree (Sorbus torminalis; Rosaceae) at maternally inher-
ited chloroplast DNA (cpDNA) has been attributed to the endozoochorous
dispersal of seeds by birds resulting in frequent migration among spatially sep-
arated stands (Oddou-Muratorio et al. 2001).
Limited dispersal of seeds results in complex relations between spatial dis-
tribution patterns and kinship relations. In general, neighboring trees are
expected to be more closely related to each other and hence to be genetically
more similar than spatially separated trees of the same population. The exis-
tence of family structures can be studied if spatially explicit genetic data are
available. Ng et al. (2004) found a significant spatial aggregation in small- and
medium-diameter classes for two dipterocarp species (Shorea leprosula and
Shorea ovalis ssp. sericea) in a 50-ha plot in the Pasoh Forest Reserve, Malaysia.
They explained the result by limited dispersal of both seed and pollen for
Shorea leprosula, and an apomictic mode of reproduction in the tetraploid
Shorea ovalis ssp. sericea.
Strong genetic differentiation among populations separated by less than 10
km from each other has been found for the Amazonian tree Corythophora alta
at cpDNA haplotypes (Hamilton 1999). Seed dispersal is very restricted in
Corythophora alta, resulting in areas as large as 10 ha founded by a single
maternal lineage.
5.4
Long-Distance Gene Flow and Migration in Tropical
Forest Species
5.5
Recommended Literature
Gene flow in seed plants has been thoroughly reviewed by Levin and Kerster
(1974), Ellstrand (1992) and, with a focus on forest trees, Adams (1992). The
book by Proctor et al. (1996) describes many aspects of the reproductive biol-
ogy of plants with regard to pollination.
The availability of uniparentally inherited markers and hypervariable gene
loci has greatly promoted experimental studies in this field as pointed out by
Cruzan (1998), by McCauley (1995) for cpDNA polymorphisms, and by Chase
et al. (1996a) for biparentally inherited microsatellites with emphasis on trop-
ical forest trees. The impact of palaeoclimatic conditions on migration and
patterns of genetic variation has been reviewed by Hewitt (2001).
Mating Systems 6
6.1
Introduction
6.2
Random Mating and Panmixis
6.2.1
Heterogeneity of Pollen Allele Frequencies
II
Fig. 6.1. Mating of two trees in a population. Only conspecific trees in the forest are
shown. Tree I is fertilized by only two other trees contributing effective pollen in some-
what unequal proportions; therefore, the effective number of its male mating partners
is close to unity and the majority of its offspring are full-sibs. Neither can the effective
pollen cloud of tree II be considered representative of that of the virtual pollen cloud
of the population. The two trees share only one common male mating partner. Their
effective pollen clouds are expected to be markedly differentiated
GST
0.30
0.25
0.20
0.15
0.10
0.05
0.00
Picea abies I
Picea abies II
Beilschmedia pendula
Quararibea asterolepis
Sorocea affinis I
Sorocea affinis II
Pterocarpus indicus
Trichilia tuberculata
Brosimum alicastrum
Cavanilessia platanifolia I
Cavanilessia platanifolia II
Ceiba pentandra
Platypodium elegans I
Platypodium elegans II
Tachigali versicolor
>50 Ind/ha 1-10 Ind/ha <1 Ind/ha
Fig. 6.2. Differentiation (FST = GST) among pollen allele frequencies for species occur-
ring in different densities in temperate and topical forests. (Adapted from Finkeldey
2002)
6.2.2
Panmixis and HardyWeinberg Structures
6.3
Selfing and Outcrossing Rates
6.3.1
Estimates of Selfing Rates Based on Rare Alleles
Table 6.1. Genotypic structures at the gene locus PGM-A in progenies of a teak tree
(genotype A 1A2) in a clonal seed orchard in Thailand in 1996 and 1997. (Adapted
from Finkeldey 2006)
N P11 P12 P22
a b
Fig. 6.3. Single flower (a) and inflorescence (b) of teak (Tectona grandis). (Photos:
R. Finkeldey)
receptive each day. The extended flowering of teak inflorescences reduces the
probability of selfing due to geitonogamy since pollinators find only a limited
number of receptive flowers available within single inflorescences and hence
are forced to move among inflorescences, and possibly trees.
Observations of the mating system in natural and breeding populations
confirmed that teak is a predominantly outcrossing species. The existence of a
system of self-incompatibility has been suggested for teak (Sect. 6.5.1), but
considerable levels of selfing were observed in all populations studied in
Thailand. High variation of the selfing rate of individual trees was observed.
Some trees are (almost) completely outcrossing, while surprisingly high esti-
mates of selfing rates were estimated for other trees. Significantly different self-
ing rates were estimated for the same tree in different seed years, suggesting an
impact of the environmental conditions during flowering (temperature,
activity of pollinators, etc.) on selfing rates in teak (Finkeldey 2006).
76 C HAPTER 6 Mating Systems
6.3.2
Estimates of Selfing Rates Based on a Mixed Mating Model
6.3.3
Estimates of Selfing Rates Based on Nonmaternal Alleles
6.4
Inbreeding and Inbreeding Depression
Plants (and animals) are genetically related, if they have at least one common
ancestor. If two related plants mate, they possibly transmit a copy of the same
allele, which both of them may have inherited from their common ancestor, to
their progeny. In result, mating among relatives leads to a certain probability
that two alleles in the progeny are identical by descent (p. 179 ff. in Hedrick
2000); thus, mating among relatives results in an increased probability that the
two alleles transmitted to a progeny are identical, i.e., that the progeny is
homozygous at the respective gene locus. The same applies in the case of a
progeny resulting from gametes produced by the same parent after selfing.
Inbreeding is defined as mating among relatives, including selfing, which is
regarded as the strongest form of inbreeding. It results in increased homozy-
gosity as compared with matings among nonrelated plants. The increase in
average homozygosity due to inbreeding depends on the degree of relatedness
of the parents.
6.4.1
Genetic Consequences of Inbreeding
(Sect. 6.2.2) and the observed heterozygosity Ho (see Sect. 3.3.1 for a computa-
tion of He and Ho). It holds for the inbreeding coefficient F
F = 1 Ho/He. (6.3)
In case of HardyWeinberg structures, Ho equals He, resulting in F = 0. If the
observed heterozygosity exceeds the expected heterozygosity, F becomes nega-
tive. Negative values of F cannot be explained by inbreeding but have other
causes, for example, selection favoring heterozygotes at the gene locus investi-
gated. If fewer heterozygotes were observed than expected for corresponding
HardyWeinberg structures, F becomes positive. If positive inbreeding coeffi-
cients F are observed at several gene loci, and if the values for F are homoge-
neous across loci, inbreeding is a plausible explanation for the observed
genotypic structures.
Examples of inbreeding coefficients F observed in populations of tropical
forest plants are shown in Table 6.3. Weak or no impact of inbreeding on geno-
typic structures is suggested by low inbreeding coefficients for Pterocarpus indi-
cus, D. condorensis, Dipterocarpus cf. condorensis, and Swietenia macrophylla.
The inbreeding coefficient was low, but significantly different from zero in four
out of seven populations of Swietenia macrophylla studied (Lemes et al. 2003;
Sect. 3.5.3; Table 3.4). The genotypic structures of Ancistrocladus korupensis
(Foster and Sork 1997) and the five Pinanga species studied by Shapcott (1999)
are strongly influenced by inbreeding. Very high selfing rates and preferential
mating among neighboring, closely related trees are likely to account for the
high inbreeding coefficients observed for both species.
6.4.2
Inbreeding Depression
Inbreeding depression during an early life stage was studied for E. dele-
gatensis (Moran and Brown 1980). Fruits of this species remain on seed trees
for several years, allowing the investigation of seed cohorts of different age. The
outcrossing rate tm was estimated in 1-, 2, and 3-year-old seed cohorts on the
basis of isozyme studies and a mixed mating model with the following result:
tm = 0.66 (1-year-old seeds), tm = 0.78 (2-year-old seeds), tm = 0.85 (3-year-old
seeds). Thus, outcrossing seems to increase and, hence, selfing to decrease with
increasing age. The most likely reason for the effect is early mortality of inbred
seeds. Preferential survival of outcrossed progenies is a plausible explanation
for the apparent increase of the outcrossing rate.
The growth of progenies from controlled selfing, open pollination, and con-
trolled outcrossing was compared for seed parents of E. regnans in a field trial
in Australia. Survival and height growth were only slightly affected by selfing
during the early development of the trial, but strong and highly significant dif-
ferences were observed 12 years after planting. Only 31.2% of the selfed trees
survived; their average diameter was 20.4 cm. This compares with 71.9% sur-
vival with an average diameter of 24.2 cm for the open-pollinated trees. The
completely outcrossed family was even superior to the trees from open polli-
nation (Griffin and Cotterill 1988).
Many studies proved strong inbreeding depression in E. grandis. Inbreeding
results in inferior growth of young plants and has a negative impact on quality
traits like stem form (Hodgson 1976a). Inbreeding depression in E. grandis is
not restricted to selfing, but has also been described for matings among full-
and half-sibs (van Wyk 1981). Its strength depends on the relatedness of the
parent trees.
The growth performance of E. grandis progenies from single seed trees after
open pollination is correlated to the outcrossing rate tm of the seed parents esti-
mated by a mixed mating model on the basis of isozyme investigations
(Burgess et al. 1996). The mean family heights of 20 open-pollinated families
with high estimates of outcrossing rates (tm) were quite uniform for 2.5-year-
old plants (Fig. 6.4). Some families with low estimates of outcrossing rates
reached comparable mean heights, but the mean height growth of most fami-
lies with low tm was considerably below the mean for the reference populations
with high tm. Thus, the mean growth rate of a family is negatively affected by a
low outcrossing rate, presumably owing to inferior growth of selfed progenies.
The strong inbreeding depression of eucalypts has considerable practical
importance. Eucalypt plantations are established worldwide on a large scale
(Eldridge et al. 1994), and the origin of reproductive material is often
unknown. Most eucalypts flower and fruit prolifically and regularly even if they
grow as isolated individuals or in small stands. Uncontrolled harvest of seeds
from isolated trees or small stands poses a risk of producing inbred reproduc-
tive material which is expected to suffer from strong inbreeding depression.
82 C HAPTER 6 Mating Systems
12
Families with low outcrossing rates Families with high outcrossing rates
11
Mean family height (m)
10
6
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6
Family outcrossing rate
Fig. 6.4. Family mean height of 2.5-year-old Eucalyptus grandis seedlings in relation to
an estimate of the family outcrossing rate tm for 40 families with contrasting estimates
of tm. (From Burgess et al. 1996)
6.5
Incompatibility and Self-Sterility
Since most tropical forest trees are hermaphroditic, selfing by means of auto-
gamy and geitonogamy is possible. However, selfing is the strongest form of
inbreeding and is associated with strong inbreeding depression. In consequence,
mechanisms preventing selfing offer considerable evolutionary advantage. The
most important mechanisms acting prior to fertilization are incompatibility sys-
tems. Cosexual species without incompatibility reactions often evolve systems of
self-sterility which result in early abortion of embryos after fertilization.
6.5.1
Incompatibility
The genotypes of two plants which are able to produce a common progeny are
compatible. If plants belong to different biological species, they are incompatible
6.5 Incompatibility and Self-Sterility 83
since they cannot produce common progenies. On the other hand, incompati-
bility also often occurs between gametes produced by a single cosexual plant. For
example, controlled pollination of Gmelina arborea (Verbenaceae) is not
successful if pollen from the seed parent is used (Bolstad and Bawa 1982). This
suggests that G. arborea is completely self-incompatible. Observations of the
flower development after pollination suggest that teak (T. grandis Tectona
grandis), another important timber species of the Verbenaceae, is also at least
partially self-incompatible (Tangmitcharoen and Owens 1997); however, con-
trolled self-pollination in teak results in low success rates (Bryndum and
Hedegart 1969), and marker-based analyses of the mating system after open
pollination revealed that teak is a mixed mating species (Example 6.1).
Self-incompatibility is controlled by a single locus or only a few incompati-
bility loci (S-loci). The incompatibility usually results from the inability of
pollen to germinate on the pistil of particular plants, or from inhibition of the
growth of the pollen tube. The incompatibility system does not only exclude
self-pollen from successful fertilization since the genetic information at the
S-locus is crucial for the reaction.
Two main types of incompatibility systems are distinguished (Boshier
2000). In the case of gametophytic self-incompatibility, the genetic constitu-
tion of the male gametophyte at the S-locus is decisive for the incompatibility
reaction. Pollen is excluded, if it carries any of the two alleles of the seed par-
ent. Thus, self-pollination is impossible just like pollination between two
related or unrelated trees with identical genotypes at the S-locus. If two trees
share only one allele at the S-locus, they are semicompatible since half of their
pollen is compatible but the other half is excluded from pollination success.
Obviously, all plants are heterozygous at the S-locus, and populations need at
least three alleles at this locus in order to be able to produce progenies. Usually,
the allelic diversity at the S-locus is very high.
The genotypic information of the pollen parent at the S-locus, and not only
the allelic constitution of haploid pollen, is decisive for the incompatibility
reaction in the case of sporophytic systems. Two plants with identical geno-
types at the S-locus are incompatible; thus, selfing is obviously impossible.
Sporophytic systems are often heteromorphic with regard to flower morphol-
ogy. For example, two types of flowers are observed in case of distyly. Flowers
have either a long style and short stamina, or they are short-styled with long
stamina. Pollen from short stamina is excluded in long-styled flowered and vice
versa. Since plants produce only one of the two flower morphs, they are self-
incompatible. Many other mating combinations are also excluded. Distyly is
controlled by a sporophytic incompatibility system with only two alleles and
complete dominance of one of the alleles.
An incompatibility system is regarded as cryptic if selfing and other forms of
incompatible matings are disadvantageous, but not completely excluded. For
84 C HAPTER 6 Mating Systems
example, cryptic incompatibility has been reported for Brazil nut (Bertholetia
excelsa) (Moritz and Ldders 1993). Cryptic incompatibility results in a strong
dominance of outcrossing, if foreign, fully compatible pollen from other trees
is available, but does not preclude the production of seeds by sexual processes in
the case of the absence of foreign pollen.
Incompatibility systems are very common among hermaphroditic plants in
tropical forests, but the genetic control has rarely been investigated. Gametophytic
incompatibility has been reported for Inga spp. (Koptur 1984) and Acacia retinodes
(Kenrick and Knox 1985). Cordia pringlei and Cordia dentata are distylous (Opler
et al. 1975). Other species of the genus Cordia (Boraginaceae) are dioecious
(C. inermis, C. collococca, and C. panamensis), and Cordia alliodora is self-compat-
ible, but predominantly outcrossing (Boshier et al. 1995; Table 6.2). The system of
self-incompatibility is particularly complicated in cacao (Theobroma cacao).
Several loci are involved; the genetic control is both sporophytic and gametophytic.
Most trees in the South American region of origin of cacao are self-incompatible,
but there are also self-compatible varieties (Warren et al. 1995).
Bawa (1974) estimated a frequency of 54% of self-incompatible hermaph-
rodites among 130 species investigated in a tropical lowland rain forest in Costa
Rica. Since 22% of the species in this community are dioecious, 76% are likely
to be obligatory outcrossing. The remaining trees are either self-compatible
hermaphrodites (14%) or monoecious (10%).
6.5.2
Self-Sterility
Postzygotic effects of selfing have been studied in detail for many conifers
(p. 67 f. in Stern and Roche 1974; Savolainen et al. 1992). The flower morphol-
ogy of gymnosperms does not permit the selective discrimination of particular
pollen types. The early abortion of zygotes from selfing is not controlled by a
single locus or a few gene loci, but is the effect of many interacting genes. Thus,
self-sterility may be regarded as a very early form of inbreeding depression act-
ing on young zygotes and embryos and resulting in strong viability selection
against inbred genotypes. The dominance of angiosperms and the frequent
observation of incompatibility systems in tropical forests suggest that systems of
self-sterility are of lesser importance.
6.6
Environmental Effects on Mating
6.7
Recommended Literature
The contribution of Boshier (2000) to the book edited by A. Young et al. (2000)
gives a concise introduction to many aspects of plant mating systems, includ-
ing inbreeding and incompatibility. The genetic consequences of inbreeding in
plants were discussed by Ritland (1996a). Inbreeding depression has been
reviewed with emphasis on conifers by Williams and Savolainen (1996) and
with a focus on tropical forest trees by Griffin (1990). Incompatibility is
discussed in detail in the book of de Nettancourt (2001).
Adaptation and Coevolution 7
The results of natural selection, the evolutionary force most responsible for adaptation
to the environment, are evident everywhere, yet it is remarkably difficult to observe the
time course of changes brought about by selection (p. 49 in Gillespie 1998).
7.1
Introduction
7.2
Physiological and Evolutionary Adaptation
Different norms of reactions are not only observed among clones, but are also
the rule among genetically heterogeneous demes such as families or prove-
nances. The use of genotypes in plantation forestry requires their adaptedness
to the respective site conditions. Thus, traditional tree improvement (Chap. 12)
is based on differences of the physiological adaptedness among demes com-
pared in field trials although the domestic fitness rather than the natural fit-
ness of genotypes and demes matters in applied tree breeding (p. 94 ff. in
Eriksson and Ekberg 2001).
Drastic environmental change results in the loss of adaptedness of particu-
lar genotypes. Preferential survival or higher reproductive capacities of
adapted genotypes result in changes of genetic structures owing to viability
selection or fertility selection. Evolutionary adaptations are caused by selective
changes of genetic structures; hence, the population, and not the single plant,
is the basic entity to investigate evolutionary adaptations, which can take place
only in genetically diverse populations.
There is no doubt that evolutionary changes of genetic structures due to via-
bility and fertility selection contributed to the adaptation of trees to their
respective environments in boreal, temperate, and tropical forests. Most tropi-
cal tree species are highly variable, and most of the genetic variation at single
loci resides within populations (Chap. 3). Thus, evolutionary adaptations were
not only involved in the creation of the high species diversity in tropical forests,
but they continue to play an important role for the reactions of forest ecosys-
tems to environmental change. The urgent need to conserve genetic resources
of tropical forest trees (Chap. 14) is motivated by the importance to maintain
evolutionary adaptive potentials in tropical forests.
Selection primarily acts on single gene loci; thus, it is much more difficult to
investigate in natural ecosystems than evolutionary factors affecting all gene loci
in principally the same way. For example, drift results in losses of genetic varia-
tion at all loci, and inbreeding reduces heterozygosity at all biparentally inher-
ited, nuclear genes. The dynamics of genetic structures at putatively neutral
gene loci can be used to monitor selective processes due to associations between
selected loci and marker loci (Finkeldey and Ziehe 2004). Comparisons of
genetic structures at isozyme gene loci between tolerant and sensitive
adult beech (Fagus sylvatica) trees in environmentally stressed forests in
Germany suggest a selective response of trees to drastic environmental change
(Mller-Starck 1985; Ziehe et al. 1999).
The potential for drastic changes of genetic structures due to viability selec-
tion without encountering critically low population sizes is highest during early
life stages of trees since most trees produce a large number of progenies.
Changes of genetic structures were observed during the development of beech
seedlings in greenhouses and under field conditions (Ziehe et al. 1999). Viability
selection was also observed during early life stages of the tropical tree
7.2 Physiological and Evolutionary Adaptation 89
Example 7.1: Viability Selection During Early Life Stages of Platypodium elegans
Genetic structures at four hypervariable microsatellite loci were observed at
three different early life stages (aborted fruits, mature seeds, and seedlings from
the same cohort) of progenies from 12 P. elegans (Fabaceae) trees and a popu-
lation of adults. The trees grow on Barro Colorado Island, an artificial island in
the Panama channel. A slight increase in heterozygosity was observed from
mature seeds to seedlings at two loci. The average observed heterozygosity
increased from Ho = 0.750 in aborted fruits to Ho = 0.761 in mature seeds,
90 C HAPTER 7 Adaptation and Coevolution
7.3
Species Interactions and Coevolution
Plants do not only need to be adapted to their chemical and physical environ-
ment, but also to interacting organisms. The biological environment is partic-
ularly complex for plant species in species-rich tropical forests. It consists of
other higher plants, animals, fungi, and microorganisms which interact either
positively (symbioses) or negatively (pests and diseases) with plants.
Interactions with animal pollinators are particularly important from a genetic
point of view (Chap. 5). The coevolution between tropical figs (Ficus spp.) and
fig wasps (Agaonidae) is an interesting and well-studied example of a coevolu-
tion between plants and their animal pollinators (Example 7.2).
The impact of pests and diseases on the distribution of forest trees and their
regeneration has been suggested as one of the main explanations for the high
species diversity in tropical forests (Baker 1970; Janzen 1970; Hill and Hill
2001). Breeding for disease resistance is an important goal in tree improvement
(Namkoong 1991). It is based on genetic variation with regard to the suscepti-
bility to particular diseases, which is commonly observed in forest trees. The
preservation of disease-resistant genotypes is an important objective in con-
servation of forest genetic resources (Byrne 2000).
Symbiotic associations between trees and mycorrhiza constitute an important
aspect of biotic interactions in tropical forests and possibly explain the domi-
nance of particular species in some low-diversity tropical forests (Connel and
Lowman 1989). Little is known about mutual preferences between genotypes of
7.3 Species Interactions and Coevolution 91
Fig. 7.1. Inflorescence (A) and single male flowers (B1), short-styled gall flowers (B2),
and long-styled female flowers (B3) of a monoecious fig (Ficus spp.). (Adapted from
Hill 1967a)
The Reproduction Cycle of Figs and the Life Cycle of Fig Wasps
The reproduction cycle of figs and the life cycle of the pollinators are closely
connected (Janzen 1979a; Wiebes 1979; Bronstein and McKey 1989; Fig. 7.3).
Female wasps are attracted by pheromones and enter the syconia of a recep-
tive, female flowering syconium, through the ostiole. They fertilize the
long-styled female flowers with pollen brought from another fig of the same
species and lay eggs through the style of short-styled gall flowers. Afterwards
they usually die within the syconium; only a few wasps manage to leave a syco-
nium they entered (Giebernau et al. 1996). The emerging larvae feed on the
developing seeds of their gall flowers, thereby destroying it. The heterostyly
within syconia allows the simultaneous development of seeds and larvae of
wasps in different flowers of the same inflorescence.
A new generation of wasps develops within a few weeks. Male insects appear
first, fertilize the females, and die shortly thereafter inside the syconium. The male
fig flowers shed their pollen only after the female fig wasps have emerged. Thus,
figs are distinctly protogynous, i.e., female flowering is earlier than male flowering.
The fertilized female fig wasps take up pollen actively or are passively loaded with
pollen. They leave the syconium and start searching for another syconium in the
receptive phase. The seeds continue to develop after the female wasps have left
the syconium. The shape and the taste of a syconium change when the seeds are
ripe. The seeds are endozoochorously dispersed by mammals and birds.
Fig Wasp
pollination, wasp
fertilization lays eggs
imago imago
fertilization
Fig. 7.3. The reproduction cycle of monoecious figs, and the life cycle of fig wasps
94 C HAPTER 7 Adaptation and Coevolution
Flower Phenology
The flowering cycle of a single fig plant is usually strictly synchronized; all syco-
nia of a plant are in the same stage of floral development. Only a few fig species
growing in a seasonal climate show a certain degree of individual asynchrony
(Cook and Power 1996). Thus, emerging wasps do not find receptive syconia on
the same plant but have to search for other trees of the same species. The wasps
survive only for a few days outside of syconia. In consequence, asynchronous
flowering of the fig plants of a population is a necessary requirement for the
preservation of the pollinators (Fig. 7.4). Asynchronous flowering results in the
production of fruits throughout the year. The importance of figs as a more or
less continuously available energy source for vertebrates is explained by the need
to produce flowers as breeding grounds for their pollinators, and hence also
fruits, throughout the year.
Gynodioecious Figs
The previous description of the life cycle of figs refers to monoecious species.
The eatable fig F. carica and many Asian fig species are gynodioecious (Berg
1989). In these species, the cycle for the preservation of the pollinating wasps
is confined to male and female flowering plants. All female flowers in these
plants are short-styled gall flowers; hence, the pollinators are able to lay eggs in
all female flowers, and no mature seeds develop in this type of syconia. In con-
sequence, gynodioecious figs are functionally dioecious.
...
...
Fig. 7.4. A hypothetical example on the phenology of figs and its impact on pollinat-
ing wasps. (Adapted from Bronstein et al. 1990)
7.3 Species Interactions and Coevolution 95
If a wasp enters the syconium of a solely female flowering plant, the flowers
are fertilized by the pollen carried by the insect. All female flowers in these
syconia are long-styled; thus, the wasp cannot deposit its eggs and seeds
develop from all fertilized flowers. Since the wasp is irreversibly trapped in the
syconium, it will not be reproductively successful if it enters a solely female
flowering inflorescence without gall flowers (Fig. 7.5).
Fig. 7.5. Reproduction of gynodioecious figs. (Adapted from Hill 1967b). Further
explanation in the text
96 C HAPTER 7 Adaptation and Coevolution
conserving even widely isolated fig trees which possibly belong to viable
populations owing to the high efficiency of fig wasps as pollinators.
7.4
Recommended Literature
The wealth of species in humid tropical and subtropical forests has always fascinated
botanists and not only taxonomists but also ecologists, geneticists, and students of
evolution. Much research remains to be done in these forests,. . . (p. 210 in Stern and
Roche 1974).
8.1
Introduction
One of the most obvious and fundamental differences between tropical and
temperate forests refers to their floral diversity and in particular the diversity
of tree species (Whitmore 1975). The number of tree species in a single
hectare in a tropical lowland forest is often considerably higher than the num-
ber of trees native to central Europe. Public and scientific discussions are often
focused on the conservation of this diversity rather than its evolution (Lugo
1988; Raven 1988; Sayer and Whitmore 1990). Conservation of biological
diversity continues to deserve the highest priority in view of the undiminished
speed of forest destruction in many parts of the tropics (FAO 2002;
http://www.fao.org/forestry/site/fra/en). However, improved knowledge
of the evolution of the high species diversity in tropical forests and its
maintenance in undisturbed forest ecosystems will help us to develop sound
strategies for its conservation.
8.2
The Evolution of Species Diversity in the Tropics
8.2.1
Species Diversity in Tropical Forests
The enormous floral diversity was noticed during early expeditions to all tropi-
cal regions and quickly became a field of intensive research (e.g., von Humboldt
1808). The rich biological diversity in the tropics stipulated evolutionary
100 C HAPTER 8 Phylogenies and Evolution Above the Species Level
8.2.2
Evolution Above the Species Level
The floral diversity of tropical forests is fairly well described; however, its evo-
lution and the mechanisms ensuring its maintenance are only poorly under-
stood. The most important theories trying to explain the species diversity of
tropical forests were reviewed by Baker (1970) and, more recently, by Hill and
Hill (2001) and Montagnini and Jordan (2005, p. 27 ff.). Most of these theories
are chiefly concerned with the ecological requirements for the coexistence of
many species in a given area (ecological niches, environmental heterogeneity,
disturbance) and species interactions, including competition and predation.
Two hypotheses with a focus on the factors ruling the emergence of the high
species diversity in tropical forests are concerned with the duration of periods
of an undisturbed ecosystem development.
A long period of cumulative evolution undisturbed by annual periods of
unfavorable environmental conditions such as low temperatures and long dry
seasons is regarded as an important prerequisite for the diversification of species.
A more or less constant climate for extended periods of millions of years is likely
to be even more important (museum model). The evolution of angiosperms in
tropical forests lasted at least since the beginning of the Tertiary for approxi-
mately 65 million years.It is likely that in some at least of the regions where Rain
forest is now the climatic climax, similar vegetation has existed uninterruptedly
since a very remote geological period. (p. 14 in Richards 1981). Thus, even a
slight excess of newly evolving species in comparison with presumably low
species extinction rates resulted in the accumulation of many species during this
long era. According to Richards (1981, p. 16), the floristic richness of tropical
forests is largely due to its great antiquity.
Multiple evidence suggests a strong impact of the climatic variability during
the Pleistocene not only on the flora in temperate regions, but also on tropical
rain forest vegetation (Haffer and Prance 2001). Allopatric speciation in iso-
lated refugia during dry periods possibly increased species diversity since popu-
lations isolated from each other in separate refugia eventually developed
reproductive barriers (Hill and Hill 2001); however, the size and the location of
pleistocene refugia are under dispute. The fossil pollen record suggests stability
and continuity of the forest cover over the Amazonian lowlands (Colinvaux and
De Oliveira 2001). Sequence variation patterns at the mitochondrial
cytochrome b gene in different mammals from North America and Amazonia
suggest rapid population growth for the North American species after the end
of the last glaciation, but no or only minor demographic expansion for small
mammals in Amazonia (Lessa et al. 2003). This result also suggests only a lim-
ited impact of climatic variability on the population sizes of small mammals in
102 C HAPTER 8 Phylogenies and Evolution Above the Species Level
the Neotropics and hence contradicts the refugia hypothesis. However, a molec-
ular phylogeny of the species-rich genus Inga suggests a recent speciation,
possibly as a consequence of the climatic variability during the pleistocene
(Richardson et al. 2001; Example 8.2).
Ecological research on the dynamics of tropical forests, including their regen-
eration, greatly profited from the establishment and repeated analysis of inten-
sively studied plots on Barro Colorado Island (Panama), Pasoh Forest Reserve
(Malaysia; Lee 1995), and in other tropical regions (Hubbell and Foster 1992).
Numerous genetics studies were conducted in these plots; several of them are
described in this book (Table 3.3, Fig. 6.2; Examples 5.2, 5.3). However, the focus
of these investigations was on the maintenance of species diversity, the ecological
and genetic consequences of a high species diversity, and the role of species inter-
actions in this context. The limited time for observations does not permit the evo-
lution of species diversity for long-living organisms like forest trees to be studied.
Taxonomic studies are the most important basis for the description of the
biodiversity in tropical forests. The relationships among existing species as
revealed by the differentiation of morphological traits of taxonomic significance
offer opportunities not only to classify taxa, but also to understand their origin.
During the past few decades, traditional taxonomic surveys were complemented
by molecular approaches to reveal phylogenies. Early caryological studies used
chromosome polymorphisms and in particular polymorphisms of chromo-
some numbers to infer relationships and the evolution of plants in the tropics
(Mehra and Bawa 1969). More recently, molecular genetics markers based on
variation of DNA became commonly applied tools not only to reveal genetic
variation within species (Sect. 3.4), but also to clarify phylogenetic relations
among more or less closely related taxa occurring in tropical forests (Sect. 8.3).
8.3
Molecular Phylogenies
which repeatedly evolved from woody ancestors, dioecy (Sect. 4.2.3), or leafless
plants in adaptation to a dry environment in Neotropical Cactaceae and
Euphorbiaceae of the Old World. Thus, similar trait expression may not only
be due to a common origin (homology), but characters may also be similar
although they are not derived from a common ancestor (homoplasy).
The observation of variation at molecular traits offers obvious advantages in
comparison with morphological characters. This applies in particular to variation
of DNA, which is not influenced by environmental conditions. Thus, some of the
methods briefly described in Sect. 2.3.1 are not only commonly used to investigate
genetic variation within species, but are also used to clarify differentiation among
species and to elucidate evolutionary relationships by phylogenetic analyses.
The observation of variation at nonrecombining chloroplast DNA (cpDNA)
is of particular importance for plants. Low mutation rates of cpDNA are respon-
sible for, in general, low variation within species, but ample and easily inter-
pretable differentiation among species at various levels. Sequences of cpDNA
were used to investigate phylogenetic relationships among plants at virtually all
taxonomic levels. Variation of cpDNA is informative to study splits during the
earliest evolution of land plants (Goremykin and Hellwig 2005), and also to
reveal phylogenetic relations within single genera (Gernandt et al. 2005).
Phylogenies are not only based on cpDNA, but also on nuclear DNA (nDNA)
and anonymous markers. Commonly used nDNA sequences for phylogenetic
analyses are ribosomal DNA internal transcribed spacer (ITS) regions. For exam-
ple, molecular phylogenies of the pioneer tree genus Macaranga (Euphorbiaceae)
were based on ITS sequence data (Blattner et al. 2001) and amplified fragment
length polymorphisms (Bnfer et al. 2004). The evolution and phylogeny of
flowering plants has recently been reviewed largely based on molecular data (The
Angiosperm Phylogeny Group 2003).
Two approaches are commonly used to construct phylogenetic trees based
on DNA variation (Hall 2001). The most important algorithmic methods are
the unweighted pair-group method with arithmetic means (UPGMA) and the
neighbor joining (NJ) method. Both are distance methods based on the com-
putation of a matrix of all pairwise differences between samples (Sect. 3.3.2).
Parsimony is the most commonly used tree-searching method. Many trees are
constructed and compared. The most likely (maximum parsimonious) tree is
the one requiring the lowest number of evolutionary steps, including
homoplasies, to explain the data. The reliability of phylogenetic trees is often
statistically tested by bootstrapping (p. 58 ff. in Hall 2001).
consisting of about 470 species in 13 genera. Nine genera with more than 250
species are native to Indonesia. The Dipterocarpoideae are commonly divided
into two tribes (Ashton 1982). Tribe Dipterocarpae, with a basic chromosome
number of x = 11, consists of the genera Anisoptera, Cotylelobium, Dipterocarpus,
Upuna, and Vatica. Tribe Shoreae (x = 7) consists of Dryobalanops, Hopea,
Parashorea, and the species-rich genus Shorea. Genus Dryobalanops has a basic
chromosome number of x = 7, but shows intermediate characters between the
tribes concerning flower morphology (sepal aestivation) and wood anatomical
features (vessels, resin canals).
Variation of cpDNA was studied by means of polymerase chain reaction
restriction fragment length polymorphisms and chloroplast simple sequence
repeats (Sect. 2.3.1). A total of 129 samples from 58 species were investigated.
The African dipterocarp Monotes kerstingii (subfamily Monotoideae) was used
as an outgroup to root the phylogenetic tree (Indrioko et al. 2006).
No variation was observed within species. The differentiation among species
at cpDNA is in good agreement with the classification based on morphological
characters (Fig. 8.1). One major clade consists of Upuna, Cotylelobium,
Anisoptera, Vatica, Dipterocarpus (tribe Dipterocarpae), and Dryobalanops (tribe
Shoreae) in a basal position. A second clade consists of Hopea, Parashorea, and
Shorea (tribe Shoreae). The genera are, in general, well resolved and strongly
supported. The intermediate position of genus Dryobalanops is confirmed by
the analyses. Parashorea groups with Shorea in a single big, poorly resolved
clade. Both genera have close botanical affinities and the phylogenetic
relevance of their distinction is questionable.
Similar results were obtained by previous studies based on cpDNA variation
of dipterocarps (Tsumura et al. 1996; Kajita et al. 1998; Kamiya et al. 1998;
Gamage et al. 2003). Phylogenetic analyses based on cpDNA failed to consis-
tently group the species-rich genus Shorea into well-supported clades. However,
the variation at the nuclear PgiC gene locus is consistent with the distinction of
different timber groups (balau, red meranti, white meranti, yellow meranti) and
sections according to Ashton (1982) within genus Shorea (Kamiya et al. 2005).
(From Indrioko et al. 2006)
Cotylelobium lanceolatum
Upuna borneensis
Anisoptera costata
Anisoptera marginata
68 99 Anisoptera reticulata
Vatica bantamensis
Vatica bella
99 Vatica granulata x=11;
83 Vatica pauciflora valvate
94 Vatica rassak Diptero-
Vatica venulosa carpeae
Dipterocarpus grandiflorus
Dipterocarpus oblongifolius
71 Dipterocarpus retusus
100 Dipterocarpus rigidus
Dipterocarpus tempehes
Dryobalanops aromatica sub
99 Dryobalanops lanceolata valvate
Hopea bancana
Hopea celebica
85 Hopea odorata
Hopea sangal
Hopea dryobalanoides
Hopea nigra
Hopea griffithii
66 Hopea mengarawan
100 Parashorea lucida
Parashorea globosa
Shorea blumutensis
Shorea guiso
Shorea montigena
Shorea scaberrima
Shorea seminis
Shorea acuminata
Shorea andulensis
Shorea javanica x=7;
67 Shorea johorensis imbricate
Shorea leprosula Shoreae
95 Shorea macroptera
69 Shorea mecistopteryx
63 Shorea ovalis
Shorea palembanica
Shorea parvifolia
Shorea platyclados
Shorea xanthophylla
Shorea balangeran
59 Shorea selanica
Shorea splendida
Shorea macrophylla
55 Shorea pinanga
Shorea stenoptera
Shorea acuminatissima
Shorea dasyphylla
77 Shorea faguetiana
Shorea multiflora
Shorea fallax
98 Shorea materialis
60 Shorea virescens
Monotes kerstingii valvate x=7
a h
b
c
d
e
f
g
b h
with a recent speciation within the genus Inga. Rough estimates of the diver-
gence time suggest that speciation occurred within the last ten million years,
and approximately 30% of the species diverged from their closest relative dur-
ing the Pleistocene (within the last two million years).
Shorea (Dipterocarpaceae) contains only slightly fewer species than Inga.
The molecular phylogeny within the genus Shorea is also poorly resolved
(Example 8.1), likewise suggesting comparatively recent speciation within this
genus, at least in Southeast Asia. However, the diversity of tropical forests does
not only result from a few species-rich genera like Inga, Shorea, Ficus, and
others, but is also reflected in a large number of genera, families, and higher
taxonomic units present in most tropical forests (Bermingham and Dick 2001).
Thus, the species diversity of tropical forests is presumably a consequence of
both their great antiquity (museum model) and recent speciation for certain
taxa like Inga. (From Richardson et al. 2001)
8.4
Recommended Literature
The causes of the species richness of many tropical forests were reviewed
by Baker (1970), Hill and Hill (2001), and Montagnini and Jordan (2005,
p. 27 ff.).
8.4 Recommended Literature 107
Hart (1990) pointed out that not all tropical forests are characterized by a
high species diversity.
Phylogenetic methods are described among other topics related to phylogeog-
raphy in the book by Avise (2000). Their application to plants was reviewed by
Schaal et al. (1998). An update of the classification of the orders and families of
angiosperms largely based on molecular phylogenetic methods has recently been
given by The Angiosperm Phylogeny Group (2003). The rationale behind
molecular phylogenies and the commonly used software PAUP for phylogenetic
analyses are described in the book by Hall (2001).
Applications of Genetics to Tropical Forestry Part B
The majority of tropical tree species possess high genetic diversity (Sect. 3.4).
The elements of their genetic systems that are responsible for the evolution of
this genetic variation and its temporal and spatial dynamics under natural con-
ditions were the topics of Part A.
Part B is oriented towards application. It concentrates on both genetic impli-
cations of human activities and on planned utilization of genetic variation.
Tropical forests are being cleared, burned, fragmented, logged, and overhunted
at rates that lack historical precedent (Laurance and Bierregaard 1997a). The
consequences of the large-scale conversion of forests to other types of land use
and the fragmentation of contiguous forests require analysis and the develop-
ment of guidelines for appropriate treatment of remnant forests (Chap. 9).
Forestry involves logging and the silvicultural management of diverse types of
forests. Silvicultural management has far-reaching consequences for the
genetic system of tree species and their genetic structures (Chap. 10).
Tree breeding including provenance selection (Chap. 11) aims at changing
genetic structures of populations in order to increase the plantation yield. Tree
improvement and planting activities in tropical forestry commenced a long
time before population genetic research on the genetic system of tropical for-
est trees was taken up (Chaps. 12, 13).
Last but not least, in view of the genetic impact of human interference, pro-
jects on the conservation of tree genetic resources are an imperative application
of genetics. Genetic information is not a renewable resource. These projects
require concomitant research in order to be successful (Chap. 14).
In most tropical countries forest genetics plays only a minor role in forestry
education. In most curricula emphasis is put on tree improvement, if subjects
related to genetics are discussed at all. In courses concentrating on breeding,
students become acquainted with working methods of quantitative genetic
research and relevant statistical methodology. Basics of population genetics, an
analysis of the genetic systems of tropical forest trees, the role of genetics in sil-
viculture, and the conservation of forest genetic resources have conventionally
been regarded as topics of minor importance as compared with tree improve-
ment. Not too much at variance from the conditions in the temperate zones,
110 PART B Applications of Genetics to Tropical Forestry
9.1
Introduction
9.2
The Genetic Status of Fragmented Tree Populations
With some time-lag, the extinction of at least some tree populations or species
in fragmented tropical forest is likely in view of their reproduction system.
Most tropical trees are characterized by predominant outcrossing (Sect. 6.3),
strong inbreeding depression (Sect. 6.4.2), and reproductive strategies relying
on the presence of other tree and animal species (Ashton 1981). As already
indicated, if a forest has been fragmented by human action, its tree populations
are suddenly put into an environment to which they are not necessarily fully
adapted; therefore, it is the formerly widespread and common plant species
that suffer most from (complete) fragmentation of their populations (den Njis
and Oostermeijer 1997). The remnant forests possess less biodiversity. They
represent disturbed ecosystems comprising merely part of the tree species
belonging to the respective forest community. The moderate numbers of
individuals of the former keystone species form populations of reduced
reproduction-effective size. Even if tree species were intimately mixed, a certain
minimum area is required for all species to occur. Consequently, many species
are locally lost if the respective fragment is small. The same is true with
genotypes belonging to one species. The relict populations are in danger of
lacking formerly existing genetic variants due to genetic drift and those losses
are no longer compensated by the influx of external effective pollen. Isolation
is likely to result in increased genetic differentiation among relict populations
owing to nonadaptive changes of genetic structures. It must be emphasized
that reduced genetic variation in small populations is presumably a symptom
of endangerment rather than its cause (Namkoong et al. 1996; Holsinger 2000).
With progressing disturbance of forest ecosystems coadapted other species
of plants and of animals are eventually no longer present. Pollen vectors may
not survive or may no longer visit remote forest fragments (but see Example
8.2). However, some dispersers of pollen and seed respond positively to forest
clearings and to the edge effects implied by fragmentation (Laurance et al.
2002). Pollen vectors support a critical portion of the life cycle of tropical trees
and must have a permanent food basis. It is interrupted, if the flower rewards
are not offered all-year round. If also the seed vectors have become rare or
absent, the occupation of patches of the former habitat becomes increasingly
difficult for tree species. First of all, rare widely dispersed tree species will suf-
fer from altered pollinatorplant assemblages (James et al. 1998); therefore,
one has to anticipate complex, cryptic responses to fragmentation (Aldrich and
Hamrick 1998). The trees in those relict forests are less adapted to the new
environment, since they have evolved in obligate associations with groups of
pollen and seed dispersers that either are now missing or have greatly changed.
114 C HAPTER 9 Fragmentation of Forests
least the effective size of the parental pool relative to the census number may
be estimated. The diminished size is due to imbalances in the production of
effective gametes (review by Hattemer in 2006). Stern and Gregorius (1972)
have shown that in monoecious populations a further reduction is due, since
the individuals may produce gametes of either sex, which represents a further
source of kinship. Depending on the degree of the imbalance of gamete contri-
butions and the sexual function of the individual trees, the effective number is
more or less strongly reduced. The reproduction-effective population size has
two aspects, as Kjr (1996) discussed with reference to tree populations. For
instance, a population fragment of 50 trees with an inbreeding-effective size of
half the census number is expected to produce the same number of inbred pro-
geny as a panmictically reproducing population of size 25. Furthermore, a collec-
tive of 50 trees with a variance-effective number of half the census number is
expected to produce progeny with the same amount of nonadaptive gene fre-
quency change as a panmictically reproducing population of size 25. Although
the assumption of panmictic reproduction raises the impression of being highly
unrealistic, the panmictic population still represents a useful reference (Sect.
6.2.2). Depending on the number of progeny produced, the variation in the effec-
tive number of female and male gametes, and the proportion of self-fertilization,
the two types of effective population size may differ but they are mostly less than
the census number. Kjr (1996) indicated also methods of estimation. The the-
ory of the two parameters introduced was explained and reviewed under a range
of circumstances by Caballero (1994). These parameters possess relevance also for
breeding and are addressed in Chap. 13. A situation where presumably both
parameters are drastically reduced is described in Example 9.1.
forest and 2.8 in fragmented forest owing to the lower seed migration rate in the
latter. These estimates are extremely low. The reproductive imbalance becomes
even clearer if also the respective area sizes are accounted for: the effective
population size was much smaller on 38 ha of remnant forest than on only 1 ha
of continuous forest. Parentage analysis by means of microsatellite markers
revealed that in remnant forest by far most of the progeny were produced by a
few large trees left on nearby pastures. The large reproductive success of the
pasture trees was positively correlated with their increased crown size owing to
the release from competition following fragmentation. This was interpreted by
a shift in the foraging pattern of the humming-bird pollinators from traplining
towards more pronounced territoriality, i.e., from moving among tree crowns
towards moving within the same large tree crown for the sake of saving energy.
The size difference between saplings and seedlings reflected their difference in
age, which helped in classifying progeny individuals as seedlings or saplings. The
adult trees did not show any homozygote excess (F 0; see Sect. 6.4.1 for an
explanation of the inbreeding coefficient F ). Saplings in the remnant forest had
but a slight excess of homozygotes. The moderate inbreeding coefficient of the
saplings which was presumably due either to mating events prior to fragmenta-
tion or to viability selection since then was greatly exceeded by that of the
seedlings (F = 0.21). Bats dispersed fruits of the pasture trees over a considerable
distance into the remnant forest, thus adding notably to the reproductive imbal-
ance but little, if anything at all, to the variance-effective population size and the
genetic variation in the remnant forest. It must be added that the seedlings were
established in a suboptimal habitat, and that they belonged to a small number
of families, i.e., they were closely related or even inbred. This is consistent with
statements made by Ledig (1992) on the decrease of heterozygosity to be
expected in future generations assuming restricted population size. (From
Aldrich and Hamrick 1998; Aldrich et al. 1998)
Hall et al. (1996) have pointed out that the effective population size at a
given time may be diminished by variation in flowering phenology and may
thus become critical in small fragments. This expectation was supported by
results of Fuchs et al. (2003) for bat-pollinated Pachira quinata (Bombacaceae).
These authors found no reduction in reproductive output, because the lower
percentage of flowers developing into fruit was overcompensated by the larger
number of flowers produced by trees in fragments. However, they also esti-
mated lower outcrossing rates and a lower number of sires, leading to a higher
degree of relatedness of the seed produced in fragmented forest, particularly by
asynchronously flowering trees.
The degree of isolation cannot necessarily be judged on a geographic scale
alone (Nason and Hamrick 1997). Rather, dispersal delimits the patches. It
should have become clear by now that population fragments are reproductively
118 C HAPTER 9 Fragmentation of Forests
isolated only if they are separated by distances that greatly impede the transfer
of effective pollen or that cannot even be bridged by pollen or seed vectors at
all (Chap. 5). Their behavior and abundance ultimately maintains links among
forest fragments. To some degree, low tree density expands pollinator foraging
ranges and pollen dispersal distances. It has been shown that tree pollen can
travel far even in fragmented terrain (Chase et al. 1996b; Dawson et al. 1997;
Nason and Hamrick 1997; Aldrich and Hamrick 1998; Aldrich et al. 1998; James
et al. 1998; White and Boshier 2000; Dick 2001; Example 7.2). Nonetheless, one
has to assume absolute limits beyond which no effective pollen can be trans-
ferred. However, as was shown both in wind-pollinated (Starke 1996; Wang
2001) and in insect-pollinated tree species of the northern temperate zone,
fragmentation may also facilitate pollen dispersal. Ballal et al. (1994) reported
up to 17% influx of pollen into small population fragments of Acer saccharum
close to contiguous forest. This species has a dual system of pollen transfer,
because in the absence of pollinators wind-transferred pollen is sufficient for
normal seed set (Gabriel and Garrett 1984). Surveying seedlings in samples
from a natural park and in patches (comprising 90646 individuals), Young
et al. (1993) found no evidence of reduced genetic variation, reduced gene flow,
or increased selfing in the latter, although fragmentation had taken place more
than a century before. These authors explain superior diversity of population
fragments by increased gene flow. Also Young et al. (1996) came to the conclu-
sion that increased gene flow may counteract if not even compensate or
overcompensate the effect of reduced population size.
Nason et al. (1997) discussed the results of numerous studies of pollen dis-
persal in tropical trees and concluded that the absolute reach of transfer
distances in species adapted to low population density in species-rich commu-
nities is several hundred meters to some kilometers (Sect. 5.2). As Wright (1931)
derived from his island model, few migrants or effective pollen per generation
are sufficient to counteract loss of genetic diversity and drift-related genetic dif-
ferentiation. The probability of mating between trees at a certain distance in
adjoining fragments is higher than expected on the basis of that distance in a
contiguous population. The reason is the absence of pollen-producing trees
growing in-between. Bierregaard et al. (1992) reported that in Amazonian
forest breaks as little as 80 m wide may diminish interfragment movement
of some insects and birds. There must clearly be a certain limit to the gap-
crossing ability of any pollinators; however, this gap may be rather wide depend-
ing on several circumstances, as shown for figs in Example 7.2 and Swietenia
humilis in Example 9.2. This study also shows that fragmentation-induced
changes in the foraging pattern of dispersal agents and the concomitant genetic
processes in a tree population can differ among fragmented habitats. Apsit et al.
(2001) described another situation of substantial pollen immigration into a
population fragment of the insect-pollinated Enterolobium cyclocarpum (Jacq.)
Griseb. (Fabaceae). A small number of trees received pollen from many outside
9.2 The Genetic Status of Fragmented Tree Populations 119
pollen donors growing hundreds of meters away. Similar findings were reported
by Cascante et al. (2002) studying isolated trees of Samanea saman
(Mimosaceae). Isolated trees received pollen from far-distant trees; however, the
styles of these trees received lower quantities of pollen and a greater proportion
of the pollen tubes were arrested at the bottom of the style, presumably owing
to genetic incompatibility. In spite of high reproductive capacity of the isolated
trees, they received pollen from fewer pollen donors, which led to an increase in
the degree of the relatedness among their progeny.
Fig. 9.1. Locations of Swietenia humilis trees sampled. Each tree is denoted by a dot; the
17 trees selected for progeny analysis are circled (five in Las Tablas, three in Jiote, six in
Butus/Jicarito, and the isolated tree 501). No trees were sampled in the spotted enclosed
area adjoining the Las Tablas site marked by U. (From White and Boshier 2000)
120 C HAPTER 9 Fragmentation of Forests
The total number of trees was 98 in the largest fragment, Las Tablas, and was
smaller in Butus/Jicarito (44), Jiote (22), and Tablas Plains (7). Owing to
immediate losses of rare alleles of highly polymorphic simple sequence repeats,
their numbers of low-frequency alleles encountered in the trees amounted to
54, 29, 19, and 8, respectively.
The simple sequence repeats provided for complete paternity exclusion except
for one, so paternal assignment by exclusion was 100%. Observed maximum
pollen transport distances were at least 0.6 km for Las Tablas, 0.9 km for Tablas
Plains, 1.5 km for Jiote, 3.6 km for Butus/Jicarito, and 4.5 km for tree 501. The
true maximum distances may have been underestimated. As impressive as reports
on maximum reach of animal-mediated pollen transfer may ever , the totality
of the distribution of the quantities of transferred effective pollen is crucial.
Pollen flow had a contrasting pattern. Besides the predominance of near-neighbor
matings within 300-m distance, a large proportion of the pollen donors grew
outside each fragment in the study area. In Las Tablas, the largest fragment, 64%
of the matings were between near neighbors. Within-fragment matings domi-
nated also in Jiote (62%) and Butus/Jicarito (53%). In the smallest fragment,
Tablas Plains, only 31% of the effective pollen was produced within that
fragment. The lone tree 501 received less than 30% of its effective pollen from
distances between 1.2 and 2.4 km but over 70% of its pollen from unidentified
male partners at least 4.5 km away It turned out that restricted mating between
near neighbors was due to their asynchronous flowering. The flowers of the
Meliaceae are as a rule functionally monoecious, i.e., either the male or the
female organs are rudimentary (Styles 1972). The male flowering period of a
given tree may be much shorter, so one of two synchronously flowering trees
may fertilize the other but not vice versa. This reduces the number of near
potential mates over and above asynchronous flowering but opens the way to
some far-distance pollen becoming effective. Both the identity and the number
of pollen donors varied greatly among seed trees.
The number of individuals producing seed increased with increasing distur-
bance in this species. In general, a large proportion of the effective pollen came
from the largest fragment rather than the nearest one. Long-distance pollen
flow increased with a decrease in population size of the fragments, leading to
restoration of alleles lost in the original fragmentation, if not even to an
increase in the level of genetic variation within the local subpopulations. This
was facilitated by the ability of pollinators to move between the spatially
isolated stands.
The proportion of pollen flow was governed by the size of the pollen source
(large crowns or many trees, respectively) rather than its distance to the seed
tree. The latter was also true for a remnant tree in pasture. The levels of
heterozygosity among the seeds produced in the various fragments were only
little influenced by the sizes of the respective fragments. This was interpreted by
9.2 The Genetic Status of Fragmented Tree Populations 121
tree, from small islands. Seed viability was up to more than 30% less than in
contiguous forest.
The use of genetic markers is indispensable in measuring tree population
size as determined by the activity range of insects, birds, and bats (Nason et al.
1997; Sect. 5.2.3). In the case of only recent fragmentation, the search for
private alleles as indicators of reproductive isolation makes little sense. The
fragments are expected to share most genetic variants except the rarest ones,
unless they are very small and short-distance gene flow prevailed before human
interference. In general, genetic divergence may be due to pre-existing spatial
genetic structure (Aldrich et al. 1998).
Forest fragmentation does not only affect tree populations. For instance, the
decline in the number of animal species with decreasing size of forest frag-
ments has been studied in four Malagasy forest birds. However, owing to the
relatively large size of the fragments surveyed and the high migratory potential
of the birds studied, human interference was not reflected in genetic differen-
tiation of populations (Andrianarimisa et al. 2000).
Autopolyploidy is expected to partially offset the risk of genetic erosion and
the decay of heterozygosity. The incidence of polyploidy is high in some trop-
ical tree families such as the Bombacaceae (Bawa 1973). It is, however, not
known how many of these species are autopolyploids.
Several authors presented empirical data on the inverse relationship
between the size of isolated conspecific natural tree populations and the
amount of their genetic variation. Nason et al. (1997) focused our attention on
the differential short-term and long-term genetic consequences of bottlenecks.
Under the assumption of equal initial variation in the past, smaller genetic
multiplicity due to the loss of rare genes is to be expected in smaller popula-
tions. After several generations, overall genetic diversity of smaller populations
is affected more strongly (Nei et al. 1975). It has to be awaited whether empir-
ical studies will provide the evidence for these genetic changes in completely
isolated fragments. These authors also analyzed the quality of pre-existing
spatial genetic structure and of the spatial scale of fragmentation. Data presented
by Moran and Hopper (1983), McClenaghan and Beauchamp (1986), and
Billington (1991) largely confirmed the negative relationship between popula-
tion size and genetic variation. However, the establishment of a strict correlation
takes many generations of continuously small size of the reproduction-efficient
population until the results of multiple bottlenecks become manifest in diver-
sity, because in every generation with finite size a renewed size-dependent risk
exists that genetic variants will become lost. This risk also depends on the ini-
tial frequency structure of those variants. One may derive from this that the
sizes of man-made fragments of a spatially structured population are moder-
ately appropriate predictors of their genetic variation, as was found by Young
et al. (1996), Sork et al. (1998), and Aldrich et al. (1998). The findings of
9.3 Genetic Preconditions for Restoration and Persistence 123
Young et al. (1999) at allozyme gene loci in a perennial herb are consistent with
those in trees. Further to tropical trees, an inverse relationship was also found
by Hall et al. (1996) on the basis of their study of Pithecellobium elegans and by
White and Boshier (2000) studying highly variable microsatellites of Swietenia
humilis in recent forest fragments (Example 9.2). Little impact of fragmenta-
tion on the genetic system of another tree of the Meliaceae family, Swietenia
macrophylla, was reported by Cspedes et al. (2003). Also Prober and Brown
(1994) found moderately strict relationships between diversity and population
size; however, after discussing incomplete isolation and other reasons, they
considered the relationships surprisingly strong. Be the relationship called
strong or only moderate, the tendency involved in the results is beyond any
doubt. Among natural populations with more or less unknown past history,
the estimated degree of determination of genetic variation by population size
cannot possibly be expected to be as high as in artificial populations of model
insects in the laboratory.
When summarizing the results of the genetic impact of forest fragmenta-
tion, it has to be stated that there exists, of course, no empirical information on
the genetic change during a full tree generation. Nonetheless, according to the
estimates of various population genetics parameters, the reproduction system
of some tree populations is subject to major change with regard to mating sys-
tems, gene flow, and migration. Spatial genetic structures are affected by frag-
mentation both on a landscape scale and on a regional scale.
9.3
Genetic Preconditions for Restoration and Persistence
The attempt to precisely restore a forested area to its presumed original vege-
tation would certainly be all but promising. This applies even more to the
genetic structure of tree populations. However, taking account of some basic
preconditions for an adequate population size of tree species is helpful in
making decisions.
Since fragments of tropical tree populations are in a critical situation, man-
agers have to consider the size of a population that is worthwhile for preserva-
tion and possible later enlargement. The persistence of tree populations in
tropical forests over generations has certain preconditions, such as a minimum
population size, since small populations are prone to extinction sooner or later.
Besides the reduction of evolutionary adaptability due to the loss of genetic
variation (Sect. 7.2) caused by drift, the accumulation of detrimental inbreed-
ing effects is another genetic reason for increased vulnerability of small isolated
populations. Information on the size and structure of neighboring fragments
and their locations is important for making decisions on the restorability of
124 C HAPTER 9 Fragmentation of Forests
concept refers to two aspects of populations. One is the totality of critical factors
of the genetic structure of populations above a level below which inbreeding
depression and the loss of adaptive capacity would become a problem for their
continued survival. The other one is the size of an isolated population of a given
species in a given habitat that keeps the risk of extinction low. The MVP is to
ensure at some acceptable level of probability that the isolated population
persists in a viable condition for a given interval of time. Some authors have
specified these parameters as 95% probability and 1,000 years, respectively. This
period is rather extended at first sight; however, it is equivalent to a small
number of generations of forest trees.
Apart from genetic considerations, other factors such as demographic sto-
chasticy, i.e., random fluctuations of the sizes and perturbations of the age
structures of populations without associated changes in their genetic structure,
may also pave the way towards extinction (Lande 1988). The long-term growth
rate of a population is determined by the geometric mean of the growth rates
over time. The geometric mean is less than the arithmetic mean. Consequently,
the long-term growth of a population may be negative, although it presently
shows positive growth. Several nongenetic factors share in the determination of
the expected persistence time of small populations. Shrinking populations
require constant support and management. Also deterministic effects such as
the incidence of herbivores, pests and disease, competition, and the coloniza-
tion by invasive nonindigenous species represent threats to persistence. The
long-term conservation of forest fragments is more promising, the better the
habitat quality of pollinators and frugivores as seed dispersers (Nason et al.
1997). Reduced density of fruit-eating animals in small forest fragments leads
to lower recruitment of large-seeded animal-dispersed trees, although the trees
bear fruit (Cordeiro and Howe 2001). If the animal disperser of a tree species
is absent owing to subsistence hunting in fragmented forest, the seeds are only
gravity-dispersed. This leads to a closer relationship of the seedlings under
parental canopies. Therefore, defaunation also has an impact on the dynamics
of spatial genetic structures in trees which will become crucial at reproduction
(Pacheco and Simonetti 2000).
Fragmentation of tropical forests creates conditions that are similar to a
metapopulation structure. In unfragmented forests, the habitats of extinct local
subpopulations can be recolonized and mating contact between temporarily
isolated local subpopulations can sometimes be regained (Cspedes et al.
2003). In permanently or temporarily isolated subpopulations an increase of
inbreeding is expected, and genetic drift causes local losses of genetic variation
and evolutionary adaptability. Cspedes et al. (2003) concluded that metapop-
ulation models could be used to describe the genetic structure of plants in
highly fragmented landscapes. However, as Gilpin (1991) pointed out, not only
the sizes of the subpopulations but also the size of the total metapopulation
matter greatly. The question of MVP is then extended to the question of the
126 C HAPTER 9 Fragmentation of Forests
size introduced in Sect. 9.2 rather than the census number of reproducing trees.
In contiguous populations of anemophilous species this parameter can be
estimated from the variation in flower counts (Kjr 1996). More reliable in
tropical forest fragments is counting effective gametes on the basis of the
reconstruction of descent of seeds or even seedlings by means of highly vari-
able genetic markers such as microsatellites (Aldrich et al. 1998). This requires
much effort, since relevant genetic surveys have to be repeated in order to
account for variation in time (Murawski and Hamrick 1992). Eventually it is
sufficient to sample progeny of a restricted number of adult trees and to esti-
mate the size of the reproduction-effective population relative to the census
size. This ratio may serve as an appropriate reduction factor.
In view of the multitude of unforeseeable events, crisis-oriented immediate
decisions on long-term viability and preservation are required. Uncertainty
must be tolerated and managers be prepared to make best guesses on available
data (Menges 1991). Although the MVP concept is easier to apply to animal
populations, it is still useful also in connection with plants for several reasons
(Gilpin and Soul 1986). It reminds us that the focus of conservation must be
on populations rather than individual trees, and that only minimum require-
ments for persistence during a minimum period of time can be planned. Still
today, not all managers consider genetics an important factor in planning
preservation of fragmented forests. This is partly due to ignorance in a field of
biology conventionally felt to be difficult.
9.4
Recommended Literature
In his classic paper, Ledig (1992) discussed the various types of human impact
on forests and addressed their genetic implications. During the past 25 years
many books on conservation have analyzed the impact of fragmentation of
plant and animal populations. Six introductory chapters of the book edited by
Young and Clarke (2000) provide an overview of genetic and demographic
issues related to fragmentation. Two among the subsequent chapters covering
case studies deal with tropical forest trees. In the book edited by Brigham and
Schwartz (2003), many authors analyzed the preconditions for persistence in
various plant populations and designed management procedures. This book
demonstrates that the biological characteristics of the respective species must be
accounted for, if planning is to be reliable. In the book edited by Laurance and
Bierregaard (1997b), many aspects of the peculiarities of tropical forest frag-
ments and their appropriate management are viewed and discussed. A contri-
bution of Young and Boyle (2000) to the book edited by A. Young et al. (2000)
presents a brief but rigorous discussion of the genetic aspects of fragmentation.
Genetic Aspects of the Management 10
of Natural Tropical Forests
10.1
Introduction
Research that is somehow related to the genetics of tropical forest trees has long
been largely confined to breeding of a few tree taxa that are suitable for plan-
tation forestry. Owing to the restriction of forest genetics to breeding work, the
genetic aspects of the management of natural forests in the tropics have
received only little attention and relevant studies are scarce. Numerous reports
on progress achieved through breeding of these taxa have been published but
only a few publications deal with genetic implications of the management of
natural forests.
Research on silvicultural systems for the sustainable management of tropi-
cal forests and their conversion to ecologically valuable species-rich production
forests has for several decades been supported by, for instance, German devel-
opment agencies (Lamprecht 1986). However, genetic aspects have been
neglected in this context. Although genetic diversity is one of the keystone
components of biological diversity, forest genetics research is still considered to
be of secondary importance. Only fairly recently has the preservation of bio-
logical diversity been accepted as an integral part of sustainable management
systems for tropical forests (p. 119 ff. in Deutscher Forstverein 1986; p. 50 f. in
Enquete Kommission Vorsorge zum Schutz der Erdatmosphre des deut-
schen Bundestages 1990; Ministerial Conferences on the Protection of Forests
in Europe, Strasbourg 1990: Resolution S2; Helsinki 1993: Resolution H2).
Since then, there has been growing public awareness of the fact that the
destruction of tropical forests leads to an irreversible loss of genetic informa-
tion of many species. Few studies on the genetic impact of the management of
natural forests were made in view of the need for the conservation of genetic
resources of tropical tree species (Chap. 14).
More recently, the number of genetic research projects on an international
level has increased. Some of the results will be addressed in the present chapter.
The subject deserves our interest, since much the surviving part of a tropical
forest is likely to be timber-production forest (Lamprecht 1986; Whitmore 1998).
130 C HAPTER 10 Genetic Aspects of the Management of Natural Tropical Forests
10.2
Selective Logging
10.2.1
Effective Population Density After Logging
to describe the totality of conditions for the dispersal of sufficient pollen (Sect.
5.2). This is more relevant in species-rich tropical moist forests than in other
climatic zones, because the pollen of tropical trees is dispersed by pollinators
adapted to food trees occurring at a certain density. Increased distances to con-
specific neighbors are then less likely to be bridged by pollinators. Logging is
also likely to have an effect on the species spectrum of pollen vectors at least of
the keystone species, since these are the primary target of logging operations.
Logging is expected to have an influence on the mating system of the
remaining trees in general, since the balance between cross-pollination and
self-pollination is changed in favor of the latter. If the genetic barriers
against self-pollination are not fully efficient, self-fertilization as the most
severe form of inbreeding is implied. Since the largest trees produce most flow-
ers, their extraction has the most severe impact on the quantity of seed pro-
duced and its genetic structure.
Reduced mating contact between trees within populations due to selective
cutting results in decreased fertility of dioecious and self-incompatible species
(Ghazoul et al. 1998) and promotes inbreeding in species with a mixed mating
system. This is not just an expectation, as Murawski et al. (1994b) have shown
in Shorea megistophylla, a canopy-emergent species of the Sri Lankan diptero-
carp forest. These authors compared the estimated outcrossing rate in undis-
turbed forest with a density of ten trees of 20 cm or greater breast height
diameter per hectare and on plots that were logged 20 years before with a den-
sity of only two trees per hectare. This reduction is equivalent to an increase in
average tree distance from about 30 to about 70 m. It was unclear whether trees
that were left at the time of logging have since then entered the reproducing part
of the population. A genetic survey of seed at several polymorphic allozyme loci
provided for a comparison of an important parameter of the mating system on
the two plots. The multilocus estimate of the outcrossing rate t (Sect. 6.3.2)
amounted to 0.86 in the primary and 0.71 in the less dense forest; hence, the
outcrossing rate was significantly reduced owing to the disturbance.
Lee (2000) studied the impact of reducing tree density on the mating system
of another canopy-emergent dipterocarp species, Dryobalanops aromatica, in
two different types of forest in Peninsular Malaysia. In response to lower den-
sity and greater distance to the nearest conspecific neighbor (Table 10.1), the
outcrossing rate in the logged forest was remarkably reduced. Since even less
cross-fertilization in this species was estimated elsewhere (Chap. 13), it may be
concluded that self-incompatibility is presumably only partial, so the difference
of roughly 15% more self-fertilization after logging reflects the effect of less-
efficient pollen dispersal. If there once existed a family structure in the primary
forest, the consanguineous neighbors have been extracted, so after logging less
biparental inbreeding arises in comparison with that in the primary forest. Part
of the inbreeding due to the increased proportion of self-fertilization would
then be compensated by this condition. The striking difference in rp, the prob-
ability that seeds produced by the same seed tree are full-sibs, is difficult to
explain. The reduced density would be expected to reduce also the number of
male mates of the seed trees (i.e., the reciprocal of rp) instead of raising it from
less than 3 to 10 (Table 10.1). Eventually the increase in atmospheric turbu-
lence was favorable to pollinator movement and thus pollen dispersal. The
author also pointed out that the density of the flowering trees rather than that
of all trees had a major influence.
Kitamura et al. (1994) did not find a significant reduction in the outcross-
ing rate of this species in Borneo. These authors reported estimates of t = 0.86
in primary forest as compared with t = 0.79 in secondary forest, where the trees
larger than 60 cm in diameter had been logged 20 years before. However, the
physical population density may not simply be taken as the effective density, as
was indicated at the beginning of this subsection. Consequently, the estimates
of tm vary among populations in response to their numerous other determi-
nants. The authors interpreted the results by indicating high flowering density,
although in the secondary forest this was around 7 ha1, i.e., half of that in the
primary forest.
The results of Lee (2000) show that the increased intertree distances implied
by logging disturb the sensitive balance between mating system parameters at
least in particular situations. Besides pollination possibly being less complete,
the proportion of biparental inbreeding is supposedly lower but the proportion
of self-fertilization is raised. In a heavily logged population of Pinus caribaea,
Zheng and Ennos (1997) did not only find a marked reduction in the out-
crossing rate (tm = 0.894) but also substantial incidence of consanguineous
biparental mating. This was obviously due to highly irregular spacing of the
remaining trees.
The impact of logging was also studied by Dayanandan et al. (1999) in three
Costa Rican populations of Carapa guianensis, a species of the Meliaceae. In a
biological reserve used as a reference this species had a density of ten repro-
ductively mature trees per hectare. In two managed forests approximately four
and three large trees, respectively, had been removed about 10 years earlier. The
density of medium-sized trees (diameter at breast height greater than 35 cm)
after logging was still 11 and 15 per hectare, respectively. One of these two
134 C HAPTER 10 Genetic Aspects of the Management of Natural Tropical Forests
managed forests was surrounded by degraded forests and pastures; the other
was part of a large contiguous block of forest. Adult genotyped trees had an age
of more than 100 years; saplings were estimated to be between 5 and 10 years
old. In all three populations the genetic distance between the adult and sapling
cohorts was small; however, the average distances among the three sapling
cohorts (D = 0.15) was twice that among the adult cohorts (D=0.08), indicat-
ing reduced influx of external effective pollen prior to the establishment of the
sapling cohorts. Since no other reason for an allele frequency change of
microsatellites is thinkable, the hypothesis of reduced gene flow must be
accepted. It indicates incipient genetic divergence. The degree of logging must
be considered mild, anyway. There was no evidence of inbreeding in any of the
three populations.
Once the density of a population has been strongly reduced, the distances
among the trees approach or even exceed the distances of seed dispersal. This
induces spatial genetic structures in the form of family structures, i.e., neighbor-
hoods of related trees. The existence of family structures is most likely to exist in
barochorous species (Sect. 5.3.1) with their inefficient mechanisms of seed
dispersal. Hence, by logging close neighbors are removed. However, new and
more pronounced family structures are expected to arise, since seed shadows do
not overlap. The seeds in the same seed shadow are also sired by fewer pollen
parents and are expected to be closely related. This implies future inbreeding
owing to preferential mating of related neighboring trees. Indications of a fam-
ily structure were found in the dioecious species Altingia excelsa on the basis of
surveys of morphological traits and isozymes (Sakai 1985).
The negative consequences of inbreeding on phenotypic trait expressions,
i.e., inbreeding depression (Sect. 6.4.2), are difficult to observe in natural forests;
however, there is no doubt that inbreeding affects naturally regenerated popu-
lations just as much as it affects planted species. Inbreeding depression was
demonstrated by numerous studies also in tropical trees. The mating contact
among unrelated trees of predominantly outcrossing species must not be
unduly interrupted by management practices. This obviously requires not only
keeping minimum sizes but also keeping minimum densities of tree popula-
tions and preserving the food basis of pollen vectors. Owing to our restricted
knowledge of the mating system and of patterns of gene flow in most tropical
tree species, it is difficult to estimate threshold sizes and threshold densities of
populations. The long-term genetic effects of lowered density during subse-
quent generations is largely unknown. More research in this field is urgently
needed. Wickneswari and Boyle (2000) stress the desirability of research on the
response of tree species with different life-history strategies to logging disturbance
in general. Bruenig (1996, p. 166 ff.), Montagnini and Jordan (2005, p. 133 ff.), and
Wickneswari and Boyle (2000) describe various logging regimes mainly
developed in Southeast Asia that presumably differ in their genetic impact.
10.2 Selective Logging 135
10.2.2
The Question of Dysgenic Selection
10.3
Natural Regeneration
The assumed ideal of naturally regenerated forest being maximally stable, maximally
diverse, and minimally susceptible or maximally resistant to unanticipated threats, is a
commonly held paradigm. The three properties are assumed to be simultaneously
achieved by the normal processes of forest regeneration without human intervention
(Namkoong 1999).
The sporadic exploitation of a forest is usually not followed by artificial regener-
ation. Even though a logged or exploited forest may be naturally regenerated, it
is still not a natural forest anymore. By definition, natural forests have never been
influenced by humans. Since this refers in particular to their reproduction, natu-
ral forests have to be discriminated from those that were naturally regenerated
after silvicultural treatment, particularly after selective logging. Since the inten-
sity of human influences varies, we have to consider degrees of naturalness.
Other human influences exerted as man-made environmental change of the
global water, energy, and nitrogen budget should not be considered marginal.
Namkoong (1999) and Wickneswari and Boyle (2000) state that virtually all
forests in the world are directly or indirectly affected by human activities.
However, it is difficult to decide, whether a given forest has escaped human
influence completely. The three properties mentioned in the prefatory quota-
tion of the present subsection are likely to be exacerbated with increases in
human intervention. Therefore, both climax forests and secondary forests are
still considered natural, if they were predominantly or completely regenerated
without planting or sowing. They are then likely to possess a family structure
(Sect. 5.3.2). Loyn and MacAlpine (2001) used the term natural to indicate
lack of discernible impact of modern technological society. Of particular
importance is that many if not all mature trees have contributed their share to
the subsequent generation. Since not all trees reproduce at a given time, the
extension of the regeneration period over several years up to several decennia
helps to achieve this condition. However, maximizing diversity does not neces-
sarily lead to maximizing stability (Altukhov 2006).
10.4 Genetic Aspects of the Manipulation of Dynamic Processes in Natural Forests 137
10.4
Genetic Aspects of the Manipulation of Dynamic Processes
in Natural Forests
Von Gadow and Kleinn (2005) distinguish just two archetypes of forest man-
agement besides sporadic exploitation. Under rotation forest management the
totality of the standing volume is removed at the end of the rotation period
(Fig. 10.1). The stand is then artificially regenerated (Sect. 11.7), whereby
genetic structures are changed completely. This system is typical of intensive
silviculture as found primarily in the southern hemisphere.
Under a continuous-cover system the standing volume is never harvested
completely (Fig. 10.1). The age of the stand is undefined and forest develop-
ment does not follow a cyclic harvest-and-regeneration pattern. Instead, it
oscillates around a certain level. Natural regeneration is possibly going on peri-
odically depending on overwood retention.
138 C HAPTER 10 Genetic Aspects of the Management of Natural Tropical Forests
age time
0 R R t1 t2
Fig. 10.1. Development of biomass (v) over age or time for two archetypical systems
of forest management. Left: rotation management (RFM) systems; R denotes the end
of a rotation period. Right: continuous-cover forestry (CCF); the biomass is largely a
function of the time elapsed since the most recent thinning or harvest. (From von
Gadow and Kleinn 2005)
Both archetypes may be sustainable (Sect. 10.5), although more risks are
involved in rotation forestry. A wide variety of silvicultural regimes have been
developed for the two systems. A forest development path or treatment sched-
ule (von Gadow and Kleinn 2005) comprising a sequence of operations related
to thinning, harvesting, and regeneration on the basis of growth models or
experience can be designed for any forest type or individual forest (Bruenig
1996). Continuous-cover forestry in species-rich tropical forests challenges
refined methods of controlling density, the species composition, the allocation
of trees belonging to the various species, and the regeneration at least of the
keystone species (Sect. 7.6 in Whitmore 1998).
Thinning as such in conifer stands without applying any phenotypic criteria
of tree removal such as described by Buchert et al. (1997) and El-Kassaby
(1999, 2000) is expected to lead to a reduced number of genetic variants and
reduced diversity. This reflects only the decrease in the number of individuals
left on a given area. As Hosius (1993) has shown in Picea abies, thinning from
above and thinning from below may have different implications for the
genotypic structure of the remaining stand.
Density and mating system parameters such as outcrossing rates are to be
considered sensitive indicators of a changed mating system owing to the
removal of trees for regeneration. Referring to western North American
conifers, El-Kassaby (1999) reported the surprising result that the incidence of
mating among relatives was higher in stands treated under the shelterwood sys-
tem (around 200 trees per hectare and 30% of the basal area are removed) than
in old growth (used as a control). One should have expected that the removal
of possibly related neighbors would force mating among less related trees.
Under this system also the correlation of paternity (rp) exceeded that of old
growth. However, under a patch-cut system with patch sizes ranging from 1.5
to 2 ha the incidence of correlated mating was typically reduced. Later analyses
of the same replicated field experiment (El-Kassaby et al. 2003) yielded the
result of small, significant treatment effects in only a few cases, although the
10.4 Genetic Aspects of the Manipulation of Dynamic Processes in Natural Forests 139
most drastic reduction was to fewer than 25 adult trees per hectare. Number,
density, and spatial arrangement of the seed-bearing trees are important deter-
minants of genetic parameters of the offspring.
Glaubitz et al. (1999) found no difference of genetic diversity and multiplic-
ity between native stands of Eucalyptus sieberi and their natural regeneration
achieved under the shelterwood system with 10% of the adult trees left. In a
study of regeneration in the closely related E. considensiana Glaubitz et al.
(2003) concluded that clear-felling and aerial resowing with local seed was more
efficient in genetic diversity conservation than natural regeneration under the
seed tree system. The number of seed trees must have been much smaller than
the number contributing genetic information to the seed collected.
According to Finkeldey (2002) continuous-cover forestry has the prime
impact on the population density of target species and on kinship of conspecific
neighboring trees. Experimental results from forests in both the temperate zone
and the humid tropics suggest that the latter are more sensitive because of their
a priori lower population densities; therefore, the consequences for inbreeding
and lowered heterozygosity are more pronounced in tropical forests.
Information on threshold densities is urgently needed.
Silvicultural management practice may considerably change environmental
conditions to an extent that populations lose their adaptedness. Adaptational
processes are then required to ensure the survival of the population. Evolutionary
adaptations are a manifestation of changed genetic structures (Sect. 7.2). Genetic
structures of populations are likely to be affected by silvicultural practices
although it is often difficult to prove adaptive or nonadaptive changes of genetic
structures in response to the human impact on forests.
The genetic dynamics during the reproduction of species-rich tropical forests
is rather complex. Frequently a succession of different dominant plant species is
observed (p. 377 ff. in Richards 1981). The reoccupation of a habitat by species
of late successional stages is only possible from surrounding populations if early
stages of forest regeneration are dominated by species of primary forests. The
life cycle of tropical forests is characterized by the disappearance not only of
trees but also of local populations. This holds in particular if forests are simul-
taneously regenerated naturally in a large area, for example, after a catastrophic
event. Extinction and recolonization are natural events both for pioneer species
and for species of later successional stages, including species of the climax
forest, although some species with less efficient systems of seed dispersal and
without a good seed or seedling bank are less tolerant to local extinctions.
Partially isolated populations experiencing a dynamics of local extinctions
followed by recolonization are defined as metapopulations (Hanski and
Gilpin 1991).
The population genetics consequences of a metapopulation structure cannot
be analyzed if only single populations are observed (McCauley 1991). A habitat
140 C HAPTER 10 Genetic Aspects of the Management of Natural Tropical Forests
10.5
Genetic Aspects of Sustainability in Natural Tropical Forests
oriented the world over. Unfortunately, the rural poor in the respective parts of
the world suffer most from the forest decline. Today, the ethical imperative
of intergenerational fairness and justice deserves most weight among the
motivations for sustainable forest management. This anthropocentric vision
has been accepted by the Brundtland Commission (World Commission on
the Environment and Development 1987). The present generation of men and
women should not impair the potential of future generations to cover their
demands. The idea of intergenerational justice might well be extended to
intercontinental justice, since the per capita consumption of natural resources
still varies greatly among continents but the worlds climate relying strongly
on globally sustainable forestry is shared by all living beings. Humans are not
only an integral part of forest ecosystems but play an ever more prominent
role in them.
Primack and Corlett (2005) analyzed the decline of tropical rain forests and
identified its complex causes. They described the international economic con-
nections and stressed the responsibility of countries outside the tropics for the
decline. Among possible solutions the authors discussed also certification of
forests and forest products (see also p. 223 in Whitmore 1998; p. 144 f. in
Montagnini and Jordan 2005; Bennett 2001; Putz and Romero 2001). In future,
genetic fingerprinting methods might play an important role in the prevention
of illegal felling in tropical forests, forest certification, and observation of the
chain-of-custody. The development of molecular genetics tools to test state-
ments concerning the origin of tropical wood from certified forest enterprises
is ongoing (Finkeldey et al. 2006).
Recent definitions of sustainable forest management account for the dynam-
ics of the systems and try to consider all factors necessary for the long-term
preservation of a balanced status. An assessment of the stability of a forest ecosys-
tem and the sustainability of wood and timber production requires not only
detailed analysis of nutrient cycling (Ulrich 1981) but also of the information
budget of tree populations as triggered by their genetic system. Sustainable man-
agement of forests implies not only the maintenance of a balanced nutritional
status and other chemical aspects. The functioning of a forest ecosystem rests also
on information, which is predominantly stored as genetic information of the
animals and plants that are part of the ecosystem.
Sustainability does not imply the preservation of a fixed status, for example,
of certain genetic structures. The objective is the continued existence of the
ecosystem itself rather than the preservation of certain components such as
individuals, genetic variants, or species. A necessary condition for the long-
term persistence of an ecosystem is the preservation of the evolutionary adapt-
ability of its keystone species (Sect. 7.2). Thus, apart from economic necessities,
the sustainable management of a forest requires the conservation of the genetic
resources at least of some keystone species (Ziehe et al. 1989; Chap. 14).
10.5 Genetic Aspects of Sustainability in Natural Tropical Forests 143
land uses are examples of human influences on forests that are also expected to
change genetic structures (Namkoong et al. 2002). These authors proposed the
following four general indicators in the context of genetic variation as the basis
of adaptability: genetic drift, selection (in the widest sense), migration, and the
mating system. Some relationships are presented in Fig. 10.2.
The most fundamental ecological indicator is the amount of available habi-
tat in the landscape. Prevalent forms of forest destruction are conversion for
food production by clearing and fire. They are not considered selective in the
genetic sense. Loyn and MacAlpine (2001) have proposed the formulation of
ecological indicators of fragmentation. Their discussion is very helpful for
describing the physical conditions for gene flow with due detail. The mating
system, although part of the reproduction system, cannot be observed directly
but must be reconstructed from findings on the latter by routine methodology.
It includes also selection processes. Therefore, the closeness of the reproduc-
tion system to panmixia is presented as an indicator of the mating system. The
difference between the census number of populations and their parental pool
size was addressed in more detail in Chap. 9. Sexuality refers to sexual function
of monoecious genotypes and is thus used as a genetic verifier. On the other
something worth conserving at all. Not much tropical forest will survive that is
not logged sooner or later. Therefore, the integration of appropriate logging
regimes into forest management will be of more help to forests than mere con-
cern about the inevitable. However, the avoidance of major disturbances is
imperative. If the new environments differ much from the predisturbance
environments, then adaptation to novel environments requires a large adaptive
capacity. This can be carried only in a large pool of potential parents available
for reproduction. In this case, natural selection may still not be able to increase
ecological security or economic productivity rapidly enough (Namkoong et al.
2000).
It should have become clear by now that it is not justified to consider any
change in genetic structures at any marker gene locus induced by a certain
management practice as a violation of the principle of genetic sustainability
(Finkeldey and Ziehe 2004). Genetic sustainability in the sense of Mller-
Starck (1996) is impaired only if the state of adaptedness or the adaptive capac-
ity (Ziehe et al. 1999) is endangered.
10.6
Recommended Literature
Stern and Roche (1974, p. 233 ff.) have presented a fairly general description of
how humans affect forest ecosystems. Finkeldey and Ziehe (2004) reviewed a
large number of studies made on the genetic impact of forestry in both temper-
ate and tropical forest. The focus of their paper is on forestry operations such as
artificial regeneration, soil amelioration, thinning and harvesting, and the regu-
lation of forest density and species composition. They come to the conclusion
that forestry induces most severe and detrimental genetic change in tropical
forests. Finkeldey (2002) has formulated a basic hypothesis on the genetic impli-
cations of continuous-cover forestry. Its implied assumptions are discussed and
some pieces of experimental evidence from both moderate and tropical forests
are presented. This hypothesis applies to any type of continuous-cover forest.
A joint test of all of these assumptions in one or several representative examples
is encouraged.
Provenance Research 11
11.1
Introduction
11.2
Definitions
is again spongy under a genetic aspect and may illustrate the need for clear
definitions that might help to advance the discussion on the choice of appro-
priate forest reproductive material.
A race is an aggregate of several similar populations that are differentiated
from other races of the same species genetically or by particular expressions of
genetically controlled phenotypic traits. As an intraspecific division, races are
still connected by common descent. Just as populations, also races are difficult
to delineate. This is particularly true if they occupy the same habitat and are
thus sympatric. Owing to the effect of gene flow, also allopatric races occupy-
ing different territories present problems of clear-cut delineation. The pheno-
typic traits used for the description of races should be under strict genetic
control, i.e., their expression should show no or only minor environmental
influence. In plant taxonomy, races are often described morphologically as sub-
species or varieties. In the northern temperate and boreal zones, tree races
often originated in glacial refuges and have persisted during their postglacial
reimmigration. Also strains of animals or plants developed under the control
of humans are called races (Sect. 12.2).
An allochthonous population that after its transfer to its new growth site has
developed particular genetically controlled trait expressions owing to adaptation
is commonly called a land race (p. 89 in Zobel and Talbert 1984). The genetic
change is attributed to natural selection, artificial selection of trees with the best
performance as in breeding, or silvicultural selection. Eldridge et al. (1994,
p. 251) consider it essential that the respective population is exotic. Owing to the
low number of generations since then, it is hardly justified to expect substantial
adaptation of the genetic structure of a land race to its new environment. Zobel
and Talbert (1984), Eldridge et al. (1994), and Ladrach (1998) consider a popu-
lation with the quantity of genetic change as achieved during even only one gen-
eration a land race. Since a population originating from seed collected in its
progenitor population elsewhere can then have achieved improved adaptedness
only through viability selection, the customary definitions imply strong effects
of this type of selection that are linked to high mortality. The aforementioned
authors stress that the quality of the provenance used in the first planting is
crucial for the quality of the developing land race. This means that virtually all
allochthonous populations, particularly all plantations of exotic tree species,
deserve the attribute land races. For instance, first-generation stands of North
American conifers such as Douglas fir (Pseudotsuga menziesii) and eastern
white pine (Pinus strobus) planted early in the twentieth century in central
Europe have been called land races if they show good growth and their off-
spring show better health. In field tests with Pinus sylvestris, an exotic species in
the Netherlands, Squillace et al. (1975) found that the progeny of trees from
foreign sources planted before 1900 grew faster and were more resistant against
the needle cast disease than the progeny of trees from Dutch sources planted
150 C HAPTER 11 Provenance Research
later. A geographic pattern also existed, since offspring of trees in the northwest
of the country grew faster and were more resistant than those from the south.
The authors considered it highly likely that the combined effect of natural
adaptation and artificial selection was responsible for this gradual geographic
gradient. Eldridge et al. (1994, p. 31 f.) discussed the status of exotic eucalypt
plantations and stated that there is not much evidence in writing about the
improvement in eucalypts and other trees one generation after introduction
from natural stands. They attributed the superior growth of offspring of
Eucalyptus urophylla planted in the Congo in comparison with the growth of
the offspring of the progenitor population in Timor to the eventuality that the
neighborhood inbreeding effects of the natural stand had been broken down in
the plantation. Ladrach (1998) discussed several reports on land races in trop-
ical tree species. The critical factor in their formation seems to have been arti-
ficial selection rather than natural adaptation. Consequently, some land races
that have arisen on one continent showed superiority over natural provenances
elsewhere. In some instances, the poor seed production at the new growing site
had to be compensated by intervention of humans.
In future the importance of land races will certainly increase. The close
connection between their emergence and domestication will be discussed in
Sect. 12.2. Processes inducing immediate adaptedness may without doubt yield
arguments for the development and the existence of land races in trees.
11.3
Historical Development
The performance of forest tree plantations critically depends on the origin of the
reproductive material used for their establishment. Uncritical use of reproductive
material regardless of its origin has led to adverse experiences made with conifers
planted in several European countries. This insight gained considerable attention
from scientists and foresters concerned with the establishment of conifer planta-
tions. The foundation of the Verein deutscher forstlicher Versuchsanstalten
(Association of German Forestry Research Institutes) comprising the German,
Austrian, and Swiss forestry research institutes in 1892 and later, in 1929, the for-
mation of the International Union of Forestry Research Organizations (IUFRO)
have been intimately linked to research on the growth of different provenances of
economically relevant conifer species. From its very beginning this research field
has strongly relied on international co-operation and the exchange of research
material among countries. The history of the research on geographic variation in
forest trees was described by Morgenstern (1996, Chap. 1).
The first internationally co-ordinated provenance trials for tropical forest
tree species were established during the 1960s. Their main focus again was on
11.4 Provenance Tests 151
11.4
Provenance Tests
A survey of genetic marker gene loci covering different tree populations can
readily be made using plant parts or seeds collected in those populations them-
selves. Comparisons between populations in phenotypic trait expressions are
not as easy to accomplish. The differentiation in phenotypic traits, particularly
the differential adaptedness, of provenances can be tested either in long-term
field experiments or in short-term tests under controlled conditions in phy-
totrons, greenhouses, or nurseries. The results o the latter type of experiments
are valid only if some overriding environmental factors can be readily simu-
lated in the controlled environment. For instance, Bekessy et al. (2003) tested
Chilean and Argentinean provenances of Araucaria araucana in the greenhouse
and found much higher drought resistance in those from the drier regions east
of the Andes.
Field trials require much more time, cost, and effort. They are nevertheless
indispensable experiments for genetically efficient selection of populations
152 C HAPTER 11 Provenance Research
suitable for the establishment of plantations and for the utilization of genetic
variation within populations by breeding. The main practice-oriented objec-
tive of the single field trial is to compare average trait expressions of the exper-
imental entries in economically relevant traits in a uniform environment. As
will be further described in Sect. 11.4.3, exposure of the experimental entries
to a variety of environmental conditions increases the information gained con-
siderably. Provenance tests represent also a collection of material that can be
utilized for more intensive breeding purposes.
Provenance trials are only one type of genetically motivated field trials.
Their establishment is an integral part of most breeding programs (Chap. 12).
Their methodology is not much different from that in forest tree breeding, and
many basic considerations of planning apply to any type of genetically moti-
vated field trials. Methods for their design, establishment, and analysis have
been covered more thoroughly in books such as those by Wright (1976) and
Williams et al. (2002). The latter publication contains worked examples and
reports on several experiments in tropical trees. Davidson (1995) prepared a
useful manual for this type of experimentation with tropical trees.
It is open to debate and to empirical research with model plants whether
part of the observed differentiation of provenances is the result of imprinting,
since the seed used for the experiments was produced under possibly very dif-
ferent environmental conditions. In experiments requiring great effort,
Skrppa and Johnsen (1999) found that offspring of Scandinavian Norway
spruce, Picea abies, possessed considerably less tolerance to frost if the seeds
were produced in a heated greenhouse (Johnsen et al. 2005a, b) . Although this
species is adapted to a cool climate, it obviously loses its adaptedness owing to
raised temperatures during seed maturation. Conversely, it acquires cold
resistance only during embryo development under lower temperatures in the
field. Hence, the behavior under different temperature regimes is controlled
by the expression of the genes responsible rather than the accumulation of
adaptive genes during many generations. The observed change is achieved
during a few months at a crucial stage of the life cycle. This result is undoubt-
edly relevant also for trees in warmer climates, because drought stress
(Vornam et al. 2003) had a similar effect as frost stress. Since corresponding
experiments with tropical trees have not yet been made, genetic variation has
still to be considered the main basis of their evolutionary adaptability. The
conservation of adapted, genetically variable and thus adaptable populations
is of particular importance, since rapid and pronounced environmental
change is predicted. Examples of expected environmental change in forest
ecosystems are air pollution and emissions in general, global warming, the
conversion of pristine forests to production forests, and other forestry prac-
tices. The conservation of genetic variation within and among populations is
11.4 Provenance Tests 153
11.4.1
Types of Field Experiments in Provenance Research
In view of the cost, provenance trials aimed at the study of ecological genetics
problems might also serve practical goals. It has to be kept in mind that prove-
nance tests are planted at regular, wide spacing on cleared land without the
protection of a canopy in order to create an environment as uniform as possi-
ble. These highly artificial conditions are exactly the same as in plantations;
therefore, the results are expected to yield information primarily on the behav-
ior of trees in plantations rather than in the forest communities to which they
belong. Although provenance research ranks high in priority, many foresters
154 C HAPTER 11 Provenance Research
are only little aware that the implementation of its results ultimately leads to
stands to be established with allochthonous populations.
Several types of practice-oriented field trials have evolved:
11.4.2
Traits Studied in Field Trials
In view of the cost involved in field work, periodically several traits determin-
ing plantation yield might be observed simultaneously. Several traits are usu-
ally recorded that characterize growth speed, such as tree dimensions reached
during given periods of time. At a young age the most important trait is plant
height. In later stages of the experiment, tree diameter (at breast height) is
recorded for the sake of easier measurement. Viability and tree growth indicate
harvestable yield per unit area, if not even stem volume or the total biomass
produced above ground can be measured directly.
Occasionally traits characterizing the health condition of the trees such as
leaf or needle color or the resistance against an insect or pathogen are relevant.
The presence or absence of symptoms of disease makes sense only if there is
notable infection pressure. The same applies to resistance against insect attack.
Other traits codetermine the use of provenances for commercial planta-
tions, i.e., the quality of the harvested product. Examples are morphological
traits such as stem straightness and branching habit (number, size, or angle of
branches). Depending on the planned utilization of plantations, wood charac-
teristics such as fiber length and cellulose content are measured. Sprouting
ability is an important trait if future plantations are to be coppiced (Quaile and
Mullin 1984). Palmer (1994) has presented a refined catalogue of desirable
properties of tropical forest trees.
Only some of these traits are measured on a continuous scale and are quan-
titative in the close sense of the word. Most traits characterizing growth (height
and diameter growth etc.) are quantitative; however, the viability is a trait with
only two possible trait expressions in the individual tree (survived or dead).
Many traits characterizing the health status are also qualitative with two
expressions only (infected or not) or are recorded in discrete classes. Some
traits such as growth form and stem straightness are actually not measured but
scored and are therefore subject to personal bias.
Some of the traits recorded in field trials are closely associated with the state
of adaptedness of the trees to the environmental conditions at the trial site.
This obviously holds true for viability and for traits characterizing tree health,
particularly under environmental stress or exposure to infection. Growth traits
are not necessarily positively correlated to fitness. Other economically impor-
tant traits, for example, stem straightness and length of branch-free bole, are
hardly associated with fitness. Trees showing superior growth may even possess
lower viability. Fitness components connected to fertility and reproduction are
only rarely observed in field trials. Thus, the provenances that are identified as
superior in field trials are not necessarily the best-adapted to the environmen-
tal conditions at the experimental site.
156 C HAPTER 11 Provenance Research
11.4.3
Design, Conduct, and Analysis of Provenance Trials
11.4.3.1
Collection of Material for Provenance Trials
The modalities of seed collection are a largely neglected element of the method-
ology of provenance trials. Principles of seed collection must be observed in
order to prevent the induction of undesirable genetic change.
Many details of the reproduction system that are of relevance for seed collec-
tion have been studied by means of genetic marker investigations. Above all, the
genetic structures of the seeds produced by a tree population vary over time. In
order to achieve a seed sample that possesses genetic structure representative of
its basic material, seeds produced in several years have to be mixed. This in turn
would require seed storage until material from all seed sources could be collected
repeatedly, an impossible undertaking. The seed of many tropical tree species, for
example, the dipterocarps, cannot be stored for long since they are recalcitrant
(Schmidt 2000). Consequently, seed should be collected in a year of abundant
seed production or just after most of the trees have flowered abundantly, as
Hodgson (1977) recommended with regard to seed orchards (Sect. 12.8.1).
The seed lots collected from individual seed trees are genetically differenti-
ated. This is due to the variation among the individual genotypes of the seed
trees themselves and to the differentiation of their individual effective pollen
clouds (Sect. 6.2.1). The less the seed lots differ, the easier it is to arrive at a rep-
resentative sample of the genetic structure of the reproductive output of the
seed source.
Figure 11.1 shows factors responsible for the differentiation of the individ-
ual effective pollen clouds. The higher the density of the population, the more
efficient the system of pollen dispersal, and the more homogeneous the physi-
cal structure of the population are, the larger is the expected effective number
of trees involved in cross-fertilization and the less are effective pollen clouds
expected to be differentiated (Sect. 6.2.1). Excessive self-fertilization that might
have an impact on phenotypic trait expressions may easily be tested in small
seed samples before a major collection. Genetic incompatibility (Sect. 6.5.1)
and the degree of self-sterility (Sect. 6.5.2) are beyond human control.
However, sufficient size of the seed source helps to capture many S-alleles and
provides for high fertility of stands to be derived from a seed source. The effect
of the genetic differentiation of neighborhoods conditioned by spatial genetic
structure, correlated matings, and hence a low number of effective male mates
of seed trees contributes directly to genetic differentiation among the seed lots.
The more pronounced the mass flowering of the trees, the less may flower phe-
nology contribute to this differentiation.
11.4 Provenance Tests 157
Physical stand
Differentiated structure
Influx of external
effective pollen individual effective
pollen clouds
Spatial genetic
structures
Flower phenology
Correlated
Incompatibility and matings
self-sterility Excessive
self-fertilization
Fig. 11.1. Causes of genetic differentiation of the pollen that becomes effective in indi-
vidual trees of a population (individual effective pollen clouds)
rather than genetic effects, it hardly makes sense to collect amounts that are
roughly proportional to their seed production in the given year.
A minimum distance between the seed trees (at least double tree height or
100 m) helps to avoid harvesting seeds from related trees. This recommenda-
tion is particularly important if the gene flow system suggests the existence of
a family structure in a native population. Seed collection from close neighbors
also increases the risk of sampling progeny from reciprocal matings.
Owing to scarce reproduction in parts of the distribution range of the
species, the seed required for a provenance trial cannot be collected in 1 year.
Storage until completion is not possible for many tropical tree seeds. An expe-
dient solution of this conflict is to split up the experiment and compare sub-
sets of the provenances on the basis of the same age reached in different years.
The storage of seeds as practiced for many tree species of the temperate zones
is impossible for many tropical trees with recalcitrant seeds. One has to be
aware that the losses of germinability that are unavoidable even in conifers of
the temperate and boreal zones lead to genetic change of the stored seed. The
mortality during seed storage is selective as shown by Melchior (1986a) for
aspen (Populus tremula) seeds.
Very common in the past have been mail-order collections. In these cases
foresters in the region of origin of the seed sources were contacted and were
asked for shipment of local seed lots. The colleagues sent a sample of seed,
mentioning at best the number of seed trees, the distances between those trees,
and the proportions of seed collected from them. Many provenance experi-
ments with conifers of the northern hemisphere started out from seed lots
representing unspecified mixtures of seed from normally at least ten trees. It
is easy to comprehend that trait expressions observed in such provenances
are possibly biased considerably. The moderate demand for seed tempts us into
seed collection from a small number of trees only.
After all, seed must be harvested from a minimum number of trees. Several
authors recommended at least ten to 25 seed trees as sufficient to represent the
genetic structures of a population (p. 30 in Eldridge et al. 1994; p. 10 in
Williams et al. 2002; p. 84 in Davidson 1995). Population genetics considera-
tions suggest that these numbers are insufficient to obtain a representative
sample of the genetic structure of a population, since not only the genotypes of
the seed trees but also the genetic structures of their effective pollen clouds are
differentiated. Certain moderately frequent alleles and allele combinations are
expected to be severely underrepresented or overrepresented if seeds are har-
vested from only a small number of seed parents. Consequently, these numbers
represent absolutely lower bounds. However, as Williams et al. (2002) note,
even 25 seed trees might be hard to find in small local populations. It is then
open to debate whether populations of this size are eligible for entry into
provenance experiments.
11.4 Provenance Tests 159
11.4.3.2
Production of Planting Stock
11.4.3.3
Experimental Design
Regrettably enough, even many years after the principles of the planning of
experiments as conceived by R.A. Fisher had been introduced in research,
costly provenance trials with forest trees have been established without paying
160 C HAPTER 11 Provenance Research
attention to appropriate design. Although older provenance trials were laid out
without adequate design, they are still an important source of information on
the growth of trees of different origin. This is particularly true for the first
provenance trials set out by P.A. de Vilmorin in Pinus sylvestris as early as
between 1820 and 1840, and early trials established with tropical trees such as
teak (T. grandis) (Coster and Eidmann 1934; Coster and Hardjowasono 1935).
However, small differences between provenances can hardly be entrusted and
generalization of the results requires caution.
Only some general suggestions for the design and the analysis of genetically
motivated field trials in forestry will be made. This topic is covered with due
detail in other publications (e.g., Williams et al. 2002).
Principal elements of experimental design now considered indispensable are
replication of the entries, the random layout of the experimental units (plots),
and the control of the experimental error. The plots may comprise just one tree
but mostly comprise many more trees. Single-tree plots are suitable for short-
time experiments only. Larger plots provide for later thinning when the indi-
vidual trees require more growth space. They are the method of choice for
medium-term tests. Commonly used are square plots with 4, 9, 16, or 25 trees.
Larger square plots reduce the effect of competition between adjacent plots
with possibly widely different growth speed of the trees growing there. In four-
tree plots every tree has more neighbors in other plots than in its own plot. In
a square plot with 25 trees, nine of them have no neighbor belonging to an
adjacent plot. In row plots the effects of competition between adjacent plots are
maximized.
Four replicates of an experiment with a plot size of 25 trees leads to 100 trees
per provenance. Cotterill (1990) found that for estimating the provenance
means with sufficiently low error a total number of 200 would be required.
Depending on the heritability of the trait (Chap. 12), five to ten families should
be represented among those. Considerations of genetic representativity call for
larger numbers of families.
The randomized complete block design as proposed by R.A. Fisher 80 years
ago has become the most frequently used design for genetically motivated field
experiments in forestry, if not even for just any experiments. In this design,
a block contains just one plot of every entry. The spatial arrangement of the
plots of the various entries within the block is random; therefore, the design is
called random and complete. The design comprises several complete blocks (or
replicates). Statistical considerations call for a certain number of replicates.
If p provenances are to be tested in r replicates, the experiment contains pr
plots. If on every plot t trees are planted at a spacing of k meters, say, the
experiment requires an area of prtk2 square meters plus the space for one or
two border rows. If, for instance, only 20 provenances are to be compared in
a short-term experiment containing four replicates with nine trees per plot at
11.4 Provenance Tests 161
11.4.3.4
Choice of Experimental Sites
11.4.3.5
Number and Distribution of Locations of an Experimental Series
The ecological conditions prevailing at a given site induce certain growth dif-
ferences between the provenances. Apart from the genetically conditioned
growth potential of the provenances, these differences may be larger or smaller,
since the provenances respond to the environmental conditions in different
ways. Therefore, it is strongly recommended to plant an experiment on more
than one site as an experimental series. Statistical considerations suggest that
the design of the various trials should be identical.
According to experience with many series of experiments, not only the
degree of differentiation changes with the planting site but also the ranking of
the provenances. For instance, a provenance displaying fastest growth in one
location may be among the slow-growing provenances in another (Example
11.1). This phenomenon is described as interaction between provenances and
environments. The existence of interaction between genetically defined exper-
imental entries and the environment or G E interaction is commonly
experienced. Interaction in the close sense, i.e., changes in the ranks, calls for
the selection of different provenances at different locations. The importance of
interactions relative to the variation between provenance means can be meas-
ured by the relative sizes of variance components estimated by analysis of vari-
ance. Since only changes in the ranks are crucial, the consistency of the
rankings may be measured by the concordance statistic (Hattemer 1969). In
recent years the evaluation of interactions has been considerably improved
with the help of refined statistical models. Raymond and Lindgren (1990) have
discussed procedures of deploying provenances over planting sites. They pro-
posed a new approach based on the reaction of provenances to given changes
in ecological severity between the origin and the planting site.
The locations of the experimental series should have climatic and edaphic
conditions covering the typical spectrum of those in the region of future plant-
ing. In order to avoid serious setbacks, only those provenances with good per-
formance on several sites should be recommended for planting. The superior
performance of such provenances is considered environmentally stable.
Recommendations of provenances for planting involve a high risk of failure if
they are based on only one or a few field trials. This obviously holds even more
true, the more the environmental conditions at future planting sites differ from
the environmental conditions at the experimental sites.
The establishment of an experimental series requires more planting stock
and more area; however, the number of replicates in the single experiments can
be reduced. Experimental series require also much more effort in coordination
and conduct.
11.5 Provenance Differentiation and Geographic Variation Pattern 163
11.4.3.6
Observation Period of Field Experiments
The effort required for recording the data from a single tree can be low (e.g.,
measurement of breast-height diameter) or considerable (e.g., certain physio-
logical variables such as water-use efficiency). In most cases data are recorded
repeatedly in order to get both early results and information on their temporal
dynamics. The estimated means of the various traits are used for ranking the
provenances. The ranks may change considerably during the duration of a trial.
For example, in a trial in northern Thailand a land race of Pinus merkusii from
Papua New Guinea proved to grow much faster than provenances from
Thailand (Example 14.1). However, with increasing age of the trial this
land race lost its superiority in height growth over provenances from north-
easternThailand. No significant relation between the height of trees at the age
of 3 and 13 years was observed in this experiment (Changtragoon 1984). Thus,
with increasing age of the trial, conclusions on provenances with superior
growth had to be repeatedly revised.
The length of the observational period depends on the objective of the
experiment, the species, and the traits to be assessed. Species-elimination trials
in tropical countries are often restricted to a duration of 34 years. This time is
considered sufficient to assess the suitability of a species for plantation forestry.
Provenance trials should have a duration of at least half of the planned rotation
period. The fast juvenile growth of most tropical plantation species and their
short rotation periods allow the length of the observation period to be reduced
considerably in comparison with that for field trials with temperate or
even boreal species. For instance, in the context of selecting provenances of
E. grandis for high pulpwood production, Ladrach (1998) mentions a rotation
period of 6 years and a test period of 3 years.
11.5
Provenance Differentiation and Geographic Variation Pattern
Two basic types of geographic variation have been described (Langlet 1963;
Stern 1963; p. 89 ff. in Stern and Roche 1974). Since the differences in trait
expressions mainly reflect the results of adaptation to their habitats (origins),
the pattern of selective environmental factors has to be considered. Some of
these factors, such as day length and average annual temperatures, vary con-
tinuously. Edaphic or biotic factors such as the occurrence of competing
species may vary more or less abruptly. The totality of these effects finds its
common expression in the variation pattern of the populations.
164 C HAPTER 11 Provenance Research
11.6
Choice of Provenances in Tropical Forestry
through the Australian Tree Seed Centre (Arnold and Midgley 1995). The
availability of reproductive material for both field-testing and large-scale plant-
ing has significantly contributed to the wide use of Australian tree species as
exotics throughout the tropics. Evans and Turnbull (2004 p. 113 ff.) discussed
some examples of successful species introductions.
Programs for afforestation and plantation forestry have often been pro-
moted by international organizations and/or bilateral development agencies
that also supported the establishment of field trials (Barnes and Burley 1990;
Palmberg and Esquinas-Alcazar 1990; Arnold and Midgley 1995; Ladrach
1998). The respective provenance trials and other long-term experiments were
often initiated with considerable effort and input (Examples 11.1, 11.2).
However, the long-term maintenance of the experiments and their adequate
analysis sometimes became impossible, when the respective projects expired.
However, meaningful outcomes take their time. Provenance research in tropi-
cal forest tree species is more complex than in trees of the temperate zone even
though rotation and observation periods are shorter.
Conifer plantations have not achieved the important role once forecasted by
the FAO and other international agencies during the 1960s. Thus, the consid-
erable amount of results available on provenance differentiation of pines is of
limited use. However, for certain species of pines and some eucalypts as well as
a few valuable hardwood species such as teak, results of long-term observation
of provenance trials are available.
Provenance research is closely associated with plantation forestry and con-
centrates on important plantation species and a few others with potential
importance. Practice-oriented provenance research in a tree species should
therefore be made only if it is of prospective use as a plantation species. New
provenance trials have recently been established for agroforestry species and
multipurpose tree species. There exists a strong argument for exploring prove-
nance variation in regular tests before major plantation projects are taken up.
If provenances of a species are differentiated in spite of inferred widespread
gene flow, immigrants must be part of the genetic load. Provenances that are
not adapted to the planting sites do not only imply the risk of failures. If they
are able to reproduce, the respective populations increase the genetic load of
the surrounding autochthonous populations through their emitted effective
pollen. Therefore, the large-scale transfer of arbitrary, possibly maladapted,
reproductive material adversely interferes with ecological genetics principles.
Field trials with several temperate conifers yielded the information that the
local autochthonous source is not always the fastest-growing or otherwise
best provenance (Namkoong 1969). For example, local sources of Pinus taeda
and Pinus palustris were frequently inferior to provenances transferred over
large distances (Wells and Wakeley 1966, 1970). A frequent experiencee in
reciprocal transplant experiments is that the local seed source is always among
11.6 Choice of Provenances in Tropical Forestry 169
the top-ranking but is not necessarily the best-adapted nor the most produc-
tive. There may exist obvious adaptation lags. However, limits of seed transfer
may be derived from mortality data of transferred provenances (Mtys and
Yeatman 1992).
Petford
Lake Albacutya
plantations depends not only on growth rates but also on the quality of the
stems produced.
Recognizing the importance of teak as a plantation species, the DANIDA
Forest Seed Centre organized a series of international provenance trials (Kjr
and Verapong 1995). Forty-eight provenance trials were established in
19731974. A total of 75 provenances and land races from India, Laos,
Thailand, Indonesia, Africa, and Latin America were tested. The provenances
were assigned to eight main provenance regions. In the single trials only sub-
sets of provenances were represented. Unfortunately, material from Myanmar
was not available for the trials.
Only 21 out of the 48 once-planted trials were included in a first interna-
tionally co-ordinated evaluation at an age of 79 years. A second evaluation
was made at the age of 17 years (Kjr and Verapong 1995). By that time the
number of trials evaluated had decreased to only 8.
The main planting region for teak is tropical Asia. However, only one trial
in Asia, located at Pah Nok Kau in northeastern Thailand, was evaluated at the
age of 17 years. This trial comprised 25 provenances from seven provenance
regions. No significant differences among provenance regions were observed
in survival and growth (measured as basal area) at the age of 9 and 17 years.
Small but significant differences were observed in stem form and stem quality
at age 17. The stem form of provenances from Thailand is slightly superior to
that from most other provenances. Provenances from Indonesia showed
inferior stem form.
The results of African and Latin American provenance trials indicated
preferable provenance regions. However, pertinent recommendations must be
specified for the observed traits. Since fast-growing provenances may have
poor stem form, it is impossible to generally recommend certain provenances
on the basis of these international trials. No particular provenances or prove-
nance regions seem to be superior to others in the main planting regions.
The results of the series of internationally co-ordinated provenance trials on
teak are of limited value for immediate application for the following reasons:
anymore or could not be evaluated for some reason. The remaining trials
that could be evaluated are distributed unevenly. Three trials are located on
Puerto Rico, and only one represents the main growing region of teak in
Southeast Asia. No trials were evaluated at the age of 17 years in southern
Asia (India).
In many traits no significant differences among provenances or prove-
nance regions were found. Statistically significant patterns of variation
seem to be ecotypic. Neighboring provenances often showed contrasting
trait expressions.
The recommendation of superior provenances or provenance regions is
complicated by the contrasting variation patterns found in economically
relevant traits. Provenances showing good survival and growth are often
inferior in quality traits such as stem quality (From Kjr et al. 1995).
11.7
Recommended Literature
Stern and Roche (1974) analyzed the genetic variation in tree populations and
their genetic systems mainly in the context of their adaptation to the environ-
ment. Also the genetic implications of the influences of humans on tropical,
temperate, and boreal forest ecosystems have found the interest of those
authors. The book by Wright (1976) on forest genetics covers all aspects of this
field as seen 30 years ago, with emphasis on breeding. Methods and results of
provenance research are presented in three of the 19 chapters. The book by
Eldridge et al. (1994) deals with provenance research, breeding, propagation,
and resource conservation in eucalypt species. Morgenstern (1996) has pro-
duced the first book to concentrate on geographic variation of tree populations
in phenotypic traits. The results for the amount of variation and variation pat-
terns are implemented in silviculture and breeding in the major climatic
regions of the world, although the emphasis is on temperate and boreal species.
Williams et al. (2002) concentrated on the design, conduct, and analysis,
including data processing, of experiments in provenance research and breed-
ing. Much other literature on this subject has been published in languages
other than English.
Domestication and Breeding of Tropical 12
Forest Trees
12.1
Introduction
system of the target species. Breeding of forest tree species is still an important
application of forest genetics research.
The decision to embark on a breeding project can best be substantiated if
adequately sized, well-adapted, and productive populations with known his-
tory are available as base populations. A breeding goal must then be defined.
Starting selection right away in a population that happens to be locally avail-
able is not necessarily promising and hardly justifies the substantial future
investment. Other preconditions are trained personnel, sufficient land, and
continued financing.
Species of agricultural crop plants have long been subjected to breeding.
Inadvertent selection practiced by humans has been an integral part of their
evolution. Today, more deliberate and refined methods are in use. The main
principles of breeding were successfully used long before the molecular basis of
inheritance was discovered and even before genetics became a science. It is
likely that humans have influenced the genetic structures of forest tree popula-
tions for several thousand years. However, systematic breeding of forest trees is
a much more recent development. European and North American conifers and
poplars were the first forest trees subjected to intensive breeding in the twenti-
eth century. Methods involving gene transfer and other biotechnology cannot
be addressed in this introduction.
12.2
Domestication
Publications on the domestication of forest trees mainly deal with the genetic
consequences rather than the process itself (El-Kassaby 2000). Harvesting seeds
from the wild, natural state, germinating them, producing planting stock, and
planting stands is just the first step. Further and more crucial steps follow in
breeding.
In the introduction to his comprehensive account of domestication Darwin
(1875) wrote, From a remote period in all parts of the world, man has sub-
jected many animals and plants to domestication or culture. In the past,
foresters have taken seeds from the wild and used them without selection,
thereby practicing inadvertent sampling of tree populations. Libby (1973) con-
sidered as essential elements of domestication the selection of desired trees in
variable populations and their mainly vegetative propagation. Forest trees have
been subjected to a series of selection stages as the essential phases of the
domestication process. These involve an increase in frequency of desirable
physical attributes with economic relevance such as those listed in Sect. 11.4.2.
Certain procedures of seed collection, seed extraction, and storage and, last but
not least, the methods of planting stock production are linked to breeding.
12.2 Domestication 175
These procedures are necessary elements of bringing trees under the control of
humans. Allard (1960, p. 28) defines domestication as the bringing of a wild
species under the management of humans. Sooner or later the natural repro-
duction system is brought under control and converted into something highly
artificial.
Darwin (1875) did not mention changes in the reproductive system in our
modern sense. He emphasized that the principle of selection is very important:
No doubt man selects varying individuals, sows their seeds, and again selects
their varying offspring. But the initial variation on which man works, and with-
out which he can do nothing, is caused by slight changes in the conditions of
life, which often must have occurred under nature. Man, therefore, may be said
to have been trying an experiment on a gigantic scale; and it is an experiment
which nature during the long lapse of time has incessantly tried. Hence it fol-
lows that the principles of domestication are important for us. and continues:
Domestic races of animals and cultivated races of plants often exhibit an
abnormal character, as compared with natural species; for they have been mod-
ified not for their own benefit, but for that of man.
One may doubt that domestication of forest trees will ever progress to a
stage where they totally rely on humans and would become extinct if left alone,
although some authors consider the inability to survive in natural ecosystems
as part of the definition of full domestication (Clement and Villachica 1994).
These authors proposed also a nomenclature of domestication. Only semido-
mesticates, although they are significantly modified from the wild state, can
survive if abandoned. The lowest degree of domestication is that of species that
are only managed. Darwin (1875, Chap. 23) emphasized, Length of time is all
important; for as each character, in order to become strongly pronounced, has
to be augmented by the selection of successive variations of the same kind, this
can be affected only during a long series of generations. This judgment is sup-
ported when considering the long way from species-elimination trials to prove-
nance trials of the promising species, breeding experiments, and the
development of methods for propagation and mass cultivation.
The natural system of reproduction of populations is replaced step by step
by artificial matings following certain designs (Sect. 12.5) and the propagation
of the individuals selected. De Vecchi Pellati (1970) made the important point
that in conjunction with artificial selection just this change is an essential ele-
ment involved in domestication. She also stated that the new populations are
normally propagated in isolation, while native populations receive easily new
contributions from different original populations.
Hence, three crucial elements in domestication and breeding to increase fit-
ness for purpose or domestic fitness (p. 94 f. in Eriksson and Ekberg 2001) are
sampling, selection, and the change of the reproduction system. The domesti-
cation process has a strong social involvement. For this reason, Evans (1980)
176 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
12.3
Genetic Controlledness of Phenotypic Traits
The success of artificial selection for a phenotypic trait critically depends on its
genetic control. This is an absolute biological precondition for starting a breed-
ing project and one may only then predict some response of the population to
artificial selection. If trait expression proved to be different among provenances
of a species in a provenance experiment, then we may safely conclude that this
trait is under genetic control. If the test material has received equal treatment,
we cannot explain the observed phenotypic differentiation other than by dif-
ferentiation at the controlling gene loci. Planned experiments are the standard
procedure for inferring genetic controlledness. The genetic controlledness of
adaptive traits is also crucial in the context of genetic conservation (Chap. 14).
The statement of genetic controlledness alone is reliable only if the experi-
ment was based not only on an environmental design (Sect. 11.4) but also on
a genetic design (Sect. 12.5). Since an experiment takes effort and time, we have
a problem in testing the genetic controlledness of traits that are expressed only
late in the ontogeny of long-lived forest trees. Hattemer and Ziehe (1997) have
presented a review of pertinent procedures.
12.3.1
Individuals Related by Descent
Similarity between parents and their progeny has been extensively exploited in
the theory and practice of breeding. The enormous breeding progress made by
early humans was hardly based on comparison tests with a sophisticated design
but rather on the experience that a plant or animal with desirable properties
gave rise to progeny with the same or even better value for use. The phenotypic
similarity between parents and progeny must be based on the transmittance of
genes and implies genetic control; however, only little can be said about the
mode of inheritance.
Regression of offspring on a parent can be tested in seed trees and their prog-
eny from natural regeneration in the forest. Any observed similarity in trait
expression is then primarily due to the contributions of the female gametes to
the progeny. The seed parents and their progeny are compared in the same local
environment where reproduction took place. However, even if the peculiarities
of seed chorology admit the attribution of young trees to only one seed parent,
it is essential to check the descent of the progeny by genetic markers.
12.3 Genetic Controlledness of Phenotypic Traits 177
12.3.2
Pair Comparisons in the Field
tolerant
or not
infested comparison
Fig. 12.1. The method of pair-
wise sampling. In this example, of genetic
structures
stress symptoms are tested for sensitive
genetic controlledness. (From or
Ziehe and Gregorius 1996) infested
178 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
seed pool. Hence, the contrasting phenotypes should be located within seed
transport distances from a single potential seed parent.
After genotyping many pairs of trees at a set of marker gene loci, a statisti-
cal test of homogeneity between the genetic structures of phenotypes X and Y
is undertaken. If the null hypothesis is rejected, phenotypes X and Y are asso-
ciated with different genotypes and the trait must be under genetic control;
however, gene loci with significant differences are not necessarily involved in
controlling this trait they may merely be stochastically associated with con-
trolling gene loci. Such stochastic associations can originate under a system of
mixed self-fertilization and random mating, which results in a correlation of
heterozygosity over the gene loci (Ziehe 2006). The difference in genetic struc-
ture between the phenotypic groups is quantified by genetic distances. Gene
loci that are stochastically associated with controlling gene loci are more likely
to show substantial distances if they carry a larger amount of variation.
If carriers of contrasting phenotypes tend to show similar frequency devia-
tions in different populations, we may call the genetic differences consistent.
Then either stochastic associations with the controlling gene loci are of a sim-
ilar type, or the marker gene locus surveyed is involved in the genetic control
of the trait. In predominantly outcrossing species, an increasing number of
populations with different genetic backgrounds can substantially strengthen
this evidence.
The extension of pair comparisons to traits with more than two expressions
is straightforward. Several slight modifications of the method were discussed
by Hattemer et al. (1993, Sect. 16.4). The genetic control of traits in trees at any
age can efficiently be tested in the forest environment even after the stand has
been thinned repeatedly. Since the method may also be applied to nonrepro-
ducing trees, it has essential practical advantage.
The method has been applied to European beech (Fagus sylvatica L.) but did
not support the hypothesis that certain stem forms of this species are under
genetic control (Ziehe and Hattemer 2002). In various instances, as reviewed
by Geburek (2000), the method helped to present evidence that air pollution
induced selection in European forest trees. In stress-related traits also degrees
of heterozygosity in the tolerant and sensitive groups were compared.
This method is straightforward and simple to apply but does not yield any
information on the rigidity of the genetic control. Ritland (1996b) has devised
a different approach to inferring genetic control of phenotypic traits. The
method is based on phenotypic similarity of pairs of (not necessarily neigh-
boring) individuals measured in the field. The phenotypic similarities of tree
pairs are regressed on the degrees of their pairwise relationship as estimated
from the fraction of alleles shared by the two individuals at a set of marker gene
loci. This yields a quantitative measure of genetic control (narrow-sense heri-
tability; Sect. 12.4).
12.4 Linear Model of Genetic Effects on a Phenotypic Trait 179
12.3.3
Response to Natural Viability Selection
12.4
Linear Model of Genetic Effects on a Phenotypic Trait
12.4.1
Phenotypic Trait Expressions
Pheno-
type
a
d
a
A1 A1 A 1 A2 A2 A2 Genotypes
Fig. 12.2. Partitioning of the genotypic value. m is considered the average phenotypic
value of the trait produced by the homozygous genotypes in the absence of environ-
mental effects. Phenotypic values can therefore be called genotypic values. a is the effect
of the genotype possessing the advantageous allele A1 in homozygous condition.
Consequently, a measures the effect of the genotype possessing only A2. d is the domi-
nance deviation. Note that a and d are defined relative to m. (Adapted from Becker 1993)
Dominance, partial (0 < d < a) or absent (d = 0). If the breeder selects indi-
viduals with the greatest trait expression and excludes all others from
reproduction, the population will contain only allele A1 in the subsequent
generation. Owing to complete selection against A2 the advantageous allele
is fixed after one generation. The breeding goal has been reached.
Complete dominance A1 > A2 or d = a. In this situation the selected part of
the population contains genotypes A1A1 and A1A2. In order to find out
which individuals with the greatest trait expression have the desirable
genotype A1A1 and which are only A1A2, the breeder has to cross them with
individuals possessing the smaller trait expression and to test their progeny.
This involves a certain effort and also a serious setback, since all progenies
produced contain the undesirable allele A2.
Overdominance or d > a. Reproduction of the selected part of the popula-
tion will consist of heterozygotes only and will contain A1 and A2 with
equal frequency. The subsequent generations will contain also homozy-
gotes with nonoptimal phenotypes. The balanced selection equilibrium
prevents further breeding progress. A possible solution is to vegetatively
propagate individuals with the greatest trait expression.
It is also easy to see in Fig. 12.2 that any environmental effects would blur the
phenotypic differences between the three genotypes, so also individuals with
genotype A2A2 may be selected, although they possess an undesirable pheno-
type. This can only reduce breeding progress.
182 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
12.4.2
Genetic Variance Components and Heritability
12.5
Estimation of Breeding Parameters; Progeny Testing
Progeny tests have a dual function. They open the way to estimating genetic
components of the variation in economically relevant phenotypic traits.
Furthermore, it is possible to evaluate selected trees and to select them on the
basis of the performance of their offspring, or to select families and/or indi-
viduals within those families (Sect. 12.5.2). However, many related individuals
are produced by progeny testing. The degree of their relationship (usually full-
sibs or half-sibs) must be known for interpretations in terms of genetic vari-
ance components.
A variety of mating designs have been worked out and are being practiced.
They differ in efforts to be taken, and in the type and amount of information on
breeding parameters (van Buijtenen 1976; Otegbeye 1998). Otegbeye (1998)
included descriptions of the (environmental) design of the respective field tests.
12.5.1
Progenies of Open-Pollinated Trees
In this basic experimental design the progeny means can be used to estimate
the breeding values of the seed trees, i.e., the means of their offspring under the
conditions of random mating. Analysis of variance of the trait observations
allows then estimation of the additive genetic variance and heritabilities.
Finally, the breeding progress expected from selection of the seed trees or that
expected from the selection of trees within the progenies can be estimated.
However, these ideal conditions are rarely realized.
Before sufficiently variable genetic markers were available, the mating sys-
tem of trees could not possibly be analyzed in detail; therefore, it was consid-
ered random for the sake of convenience. However, the pollen clouds differ,
because only moderate numbers of male mating partners become effective for
particular seed trees (Sect. 6.2.1; Fig. 6.1), which induces a drift effect. If also a
family structure exists in the population, the differentiation of effective pollen
clouds is enforced by differentiation of the neighborhoods (Sect. 11.5).
Under random mating all offspring of a family are ideally sired by different
pollen parents; hence, they share only their female parent and may be consid-
ered half-sibs that are expected to hold one quarter of their genes in common.
However, in many species the families might contain also offspring from self-
fertilization. Owing to restrictions in the effective number of male mating part-
ners, certain portions of the family members are in fact full-sibs, since they share
both of their parents. Hence, the progeny of a tree after open pollination is a
mixture with different but unknown degrees of relationship. This condition is
highly likely to be the rule for forest trees. Thus, the genetic relationship among
the offspring of single trees is closer, i.e., they share more than one quarter of
their genes (Squillace 1974). Consequently, the variation among the families is
inflated. If it is used for estimating VA, heritabilities are overestimated. The esti-
mates of VA and h2 are also not reliable. Selecting the trees on the basis of the
performance of their open-pollinated progeny would be only a rough approxi-
mation, since the offspring of a seed tree have a more or less unique group of
male parents. Dvorak and Wright (1994) took the complicated kinship relations
of open-pollinated progenies into account by assuming the average degree of
relatedness to be one third. Eldridge et al. (1994, p. 203) reported problems in
interpreting the results of testing open-pollinated progeny.
12.5.2
Progeny Tests After Controlled Pollination
vented, the degree of relationship is then more homogeneous and closer to what
is expected from half-sib families. If pollen of the respective seed trees is not
contained in the otherwise identical pollen mix, self-fertilization is excluded. It
has turned out that the male mating partners contributing equal amounts to the
pollen mix still do not have equal proportions of offspring. The progenies aris-
ing from this experiment are, nonetheless, expected to yield unbiased estimates
of breeding values of the seed trees and the additive genetic variation.
Many other mating schemes have been developed (Otegbeye 1998). The
production of full-sib and half-sib families in the same experiment offers man-
ifold options for analysis and use of genetic variation.
In a factorial mating design (Fig. 12.3, top) a moderate number n of seed trees
are pollinated by a larger number m of other trees used as males. The objective is
to estimate the general combining abilities of those m trees, i.e., the average per-
formance of their offspring when mated with the three or four testers. Also
specific combining ability can be estimated. It reflects the ability of the trees to
transmit genotypic effects when crossed with specific other trees. The group of n
full-sib progenies of each tree are half-sibs among themselves, since they have just
their male parent in common. Consequently, the variation among all m half-sib
families yields an estimate of 0.25VA. The means of these groups reflecting their
general combining abilities are estimates of their breeding values in this experi-
mental population. Furthermore, the variation among the families having both
parents in common yields an estimate of 0.5VA + 0.25VD, so also the dominance
variance can be estimated separately and can be exploited by selection.
1 2 3 4 ... m
1 ...
2 ...
.
.
.
n ...
1 2 3 4 ... n
1 ...
2 ...
3 ...
Fig. 12.3. Complete mating schemes for 4 ...
progeny testing. Top: Factorial design com- .
prising m trees used as males and n trees
.
used as testers. Bottom: Half-diallel com-
.
prising n trees without selfs and reciprocal
crosses n 1
186 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
In the diallel cross all possible progenies are produced among a group of n
cosexual trees. In Fig. 12.3 (bottom) a variant of this design is shown (a half-
diallel) that includes neither selfed progenies nor the progenies resulting from
reciprocal crosses. If, for instance, trees 3 and 4 have to be crossed, either tree 3
is pollinated by tree 4 or tree 4 is pollinated by tree 3, or the cross is produced
in either direction and the seeds are pooled. Just as in the factorial design both
VA and VD as well as general combining abilities can be estimated.
Pollination in most zoophilous tree species requires much effort. Various
methods used in eucalypts have been described by Hodgson (1976b), van Wyk
(1977), and Eldridge et al. (1994, p. 205 f.). Species with different floral struc-
ture require very different methods (Sedgley et al. 1992; Oudraogo 1997).
Since work in tree crowns is an additional obstacle, pollinations are best
made in a clone bank or seed orchard. In view of the large number of proge-
nies to be produced [n m in the factorial design and n (n 1)/2 in the half-
diallel], there exist numerous incomplete designs requiring less effort (p. 234 ff.
in Namkoong 1981). Progeny tests on the basis of complete designs are hardly
applicable in just any tropical forest tree species. If not all crosses are success-
ful at the same time, missing families require advanced computational meth-
ods for evaluation. The same is true if mortality in the field experiments leads
to missing plot values or even family means. The single-pair mating design is
recommended for avoiding relationships between families (Otegbeye 1998).
The expected breeding progress is much faster if a large number of full-sib
families are tested as compared with the test of single tree progenies (Wright
1976, p. 176). Among tropical tree species, this approach has been confined to
the intensively bred genera Eucalyptus (Bouvet and Vigneron 1996) and Pinus
(Nikles 1995).
12.6
Methods of Selection
The decision for a method of selection is based on experimental evidence of its
breeding efficiency and on whether it fits into the given breeding strategy. For
instance, on the basis of the results of a progeny test, the genetic parameters are
estimated and used for exploring the breeding progress to be expected from
various selection regimes such as the selection among families or of single trees
within families. If the variation within families is much larger than that
between them, it is advised to retain more families and select with higher inten-
sity within families in a combined procedure. The variant maximizing expected
gain (Shelbourne 1992) is finally chosen as the most promising one.
For estimating the breeding progress expected from various selection regimes,
the heritabilities mentioned in Sect. 12.3 have to be adjusted to the particular
12.6 Methods of Selection 187
experimental design and the particular selection procedures. McKinley and van
Buijtenen (1998) have concisely reviewed the methods of selection in tree popu-
lations.
The closer the age at selection is to the rotation age, the more reliable is
selection in view of the desired target expression of the trait. However, the per-
formance at earlier ages, possibly at half of the rotation age, is often sufficiently
closely correlated to the performance at the rotation age. Periodic measure-
ments yield information on ongoing changes in the ranks. For breeding pro-
grams in several tree species of the northern temperate zone continued over
several generations, the selection age has been optimized.
12.6.1
Selection of Plus Trees
The selection of plus trees is a method of mass selection, i.e., the selection of
individuals in populations in the field for their subsequent propagation. The
selection intensity may be high.
Together with provenance testing, selection of trees in the wild (Ledig 1973)
represents the first step in many breeding projects. It differs much from the
selection in plantations. In natural forests the phenotype hardly indicates the
genotype of old trees closely (Sect. 10.2.2). Heritabilities are supposed to be low
owing to environmental heterogeneity (competition, etc.). Plantations consist
of more or less equally spaced individuals possessing the same age and prima-
rily belonging to the target tree species; therefore, the neighbors can be used as
comparison trees. In natural tropical forests, at best a regional average may be
used as a reference.
However, plus-tree selection is defined as a method based on the observa-
tion of the phenotypes of the selected trees themselves rather than those of
their progenies. The types of traits used for plus-tree selection are basically
those already mentioned in Sect. 12.4.2. Trees are selected on the basis of their
superiority in certain traits, for example, because of their good stem form
and/or volume growth. Practical guidelines for the identification of plus trees
have been summarized by Schmidt (1993a). The selection is often based on
indices which combine several traits characterizing the growth and quality of
trees (Zabala 1994; McKinley and van Buijtenen 1998).
Studies in European temperate conifers proved the efficiency of positive
plus-tree selection for increased height growth and even more so volume
growth and stem form (Cornelius 1994; Simpson 1998). Other investigations
on North American conifers detected only a small effect or failed to disclose any
effect of plus-tree selection (Wright 1976, p. 165). Eldridge et al. (1994) attrib-
ute the spectacular improvement of eucalypts to the selection of outstanding
188 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
12.6.2
Selection in Progeny Tests
Fig. 12.3. Following evaluation, the field test may be converted into a seedling
seed orchard (Sect. 12.8.2).
12.6.3
Selection of Clones
Clonal propagules derived from an individual are not progeny in the close
sense. They can nevertheless be tested in field experiments. Selection of clones
in field tests still leads to single individuals but their performance can be eval-
uated much more efficiently than that of plus trees, even though only a small
number of ramets per clone are required. During clonal propagation the mul-
tilocus genotypes are conserved. Breeding progress is thus faster than in con-
ventional selection based on sexual progeny, because in the selection of clones
also all nonadditive genetic variation, i.e., dominance and epistatic effects
(D and I; Sect. 12.4), is used. The advantages of clonal selection become more
prominent with increasing significance of nonadditive genetic variation
(Burdon 1989). However, only one or a few selection steps are possible in a
given finite aggregate, since the number of individuals would get smaller at
every step. The selection of clones is part of the most successful breeding pro-
grams with eucalypts that also involve species and provenance trials, hybridiza-
tion among species, and progeny testing.
12.6.4
Marker-Assisted Selection
Fig. 12.4. Linkage map and localization of quantitative trait loci in a Eucalyptus uro-
phylla clone. (Adapted from Grattapaglia et al. 1995)
12.6 Methods of Selection 191
The results for linkage among gene loci clearly apply to the respective
species but the results for the association of certain alleles at the marker loci
and QTLs apply to the particular full-sib family only. Since forest trees are
selected within populations rather than within one family only, it must be
possible to transfer the results to other families or even populations. Since
in outcrossing tree species linkage disequilibria even within genes rapidly
decrease with physical distance, even the transfer between families would
be possible only for closely linked markers.
The effect of single gene loci on quantitative trait expressions is often small
and difficult to detect by QTL mapping techniques. For instance, the suit-
ability of the markers must be tested by regressing the expressions of the phe-
notypic trait on the presence of no, one, or two marker alleles. Other test
procedures involve the phenotypic difference between carriers and non-
carriers of certain allelic variants. Also the phenotypic differentiation
between offspring carrying a codominant allelic variant in a homozygous or
heterozygous condition can be tested for significance. These methods were
used by Gailing et al. (2005) in several families of central European oaks for
a study of QTLs for flushing date, an adaptive trait in temperate climates. In
two species of the family Fagaceae, Quercus robur L. and Castanea sativa
Mill., QTL positions for this trait but not for the likewise adaptive isotope
discrimination were conserved (Casasoli et al. 2006). Putative candidate
genes for bud burst were at least located on the same chromosomes.
As follows from the test procedure, the identification of QTLs is compli-
cated if the observed quantitative variation is mainly due to nonadditive
gene effects.
192 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
Linkage maps have been developed for many agricultural species and breeding
techniques based on marker-assisted selection are in use. Saturated linkage
maps have also been developed for a few temperate conifers. A genetic map on
the basis of RAPDs has been developed for some eucalypt families within the
framework of the intensive breeding program of Aracruz Florestal (Grattapaglia
and Sederoff 1994). A linkage map was developed from the full-sib family of a
hybrid cross Eucalyptus urophylla E. grandis (Grattapaglia et al. 1995). Eleven
linkage groups corresponding to the gametic chromosome number of eucalypts
were identified. QTLs were mapped for three different quantitative traits: three
QTLs located in different linkage groups have an influence on the rooting per-
centage of cuttings, four QTLs influence the fresh weight of micropropagated
shoots, and two QTLs have an impact on the sprouting ability of stumps. Since
full-sib families are difficult to produce in some forest trees, also open-pollinated
progenies have been successfully used for detecting QTLs influencing the
volume yield of E. grandis families (Grattapaglia et al. 1996).
More recently, a linkage map was constructed for E. globulus, and candidate
genes (Sect. 7.2) for wood, fiber, and floral traits were mapped (Thamarus et al.
2002). The identification of QTLs and efforts to achieve marker-assisted selec-
tion are also ongoing for tropical acacias (Butcher and Moran 2000; Butcher
2004).
The individuals selected in this way are integrated into a population for
further breeding operations or for propagation.
12.7
Heterosis Breeding in Forest Trees
Full-sib families of E. grandis and the hybrid E. grandis E. urophylla were pro-
duced. Controlled pollination for artificial hybridization required the development
of techniques for the long-term storage of pollen, because different eucalypt species
flower at different times in Aracruz. Partial diallels were used as mating schemes.
Breeding success considerably increased by the vegetative propagation of
elite trees. Seventy-five clones originating from crosses between E. grandis,
E. urophylla, and E. pellita were selected in 1989. These clones proved to be
completely resistant against cancerous diseases and, when coppiced, showed
high volume growth and desirable wood properties. The trees performing best
in progeny tests are continuously selected according to a set of criteria and are
mass-multiplied in clonal archives in order to steadily increase the number of
clones used in the wood-producing plantations. The multiple-population con-
cept (Sect. 12.9) has been integrated with 20 subpopulations of E. grandis and
E. urophylla consisting of 20 trees each. The best-performing trees in ten full-
sib families produced in each subpopulation are used for making hybrid
crosses and selecting among those hybrids for clonal testing.
Principles of the breeding strategy of Aracruz Florestal are presented in
Fig. 12.5. Breeding is supported by research on in vitro techniques for the
Introduction of
species and V Vegetative propagation
provenances
Seed collection
Best trees in best provenances
Progeny test
Unimproved plantations
(with hybrids)
V
Selected trees
Best trees in best families
Individual tree
V selection
Controlled hybrids Breeding orchard Seed orchard
1st generation V
Clonal test V
V
V
Clonal multiplication Routine
Clone bank area plantation
Fig. 12.5. Breeding strategy for eucalypts in Aracruz. (Adapted from Campinhos and
Ikemori 1989)
196 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
12.8
Propagation of Breeding Products
A breeding project includes both the selection of superior trees and their sub-
sequent multiplication in order to capitalize on the breeding progress achieved;
hence, a certain progress achieved in breeding makes sense only if it is possible
to mass-propagate the selected material with as little genetic change as possi-
ble. Several options for sexual and vegetative propagation exist in tropical tree
species that were hitherto subjected to domestication. Seed production areas
are addressed in Sect. 13.5.
12.8.1
Clonal Seed Orchards
Propagating plus trees by harvesting their seed is hardly adequate for pro-
ducing the quantity and quality of reproductive material required for the
establishment of plantations. For the breeding progress involved in plus-tree
selection to become manifest, the individuals have to be brought into mat-
ing contact in a clonal seed orchard. The fertilization of ovules by external
pollen (pollen contamination) is avoided by maintaining sufficient distance
to conspecific populations in the region and by planting large orchards.
Remote pollen is then highly unlikely to prevail even if the density of the
orchard is low. In young seed orchards one has to envisage incomplete pol-
lination by the orchard clones, since the trees do not yet produce abundant
pollen and are planted with wide spacing. Consequently they are eventually
flooded by foreign pollen produced by unselected trees in surrounding
stands.
The reproduction in seed orchards is rarely random (Sect. 6.2) just as is the
case with forest stands. However, an equilibrium resulting in HardyWeinberg
structures (Sect. 6.2.2) and allele frequencies representing the frequencies
12.8 Propagation of Breeding Products 197
within the deme of the plus trees selected may be approximated by some pro-
visions made in the design of the orchard.
The clones are as a rule represented by the same number of copies (ramets).
In order to overcome implicit setbacks due to high variation in reproductive
output of the clones, additional ramets for the less prolific clones have been
recommended; however, this condition possibly induces additional kinship
among the seed produced depending on the variation in male mating success.
Vegetative propagation by cuttings is difficult if not impossible because of the
usually advanced physiological age of the plus trees selected (Hong 1975). The
trees selected are usually first incorporated in a clonal garden, a multiplication
garden, or a clonal archive. The scions for grafting are taken therefrom. Many
techniques for the production of grafts are in use (van Wyk 1977; p. 399 ff. in
Hartmann and Kester 1983). It must be checked repeatedly that the grafts in fact
develop from the scion rather than the unselected root stock (Example 11.2).
The number of clones represented in the orchard is important. In order to
maximize breeding progress, the selection intensity is high and the number of
retained clones moderate. However, if the orchard contains only a few clones,
the risk of phenological mismatch among them increases. The risk of uninten-
tional genetic change or even loss of alleles is obvious (Hattemer et al. 1982).
The size of an orchard is best determined by the number of clones rather
than the number of ramets per clone. Many old seed orchards contain only 20
or even fewer clones. Meanwhile, larger numbers of clones (up to 100) and,
consequently, lower numbers of ramets per clone are common (cf. Example
11.2). Problems related to losses of genetic variants due to small numbers of
orchard clones can be solved by reducing the intensity of plus-tree selection
that per se is very high. Ideal would be the reduction of the number of ramets
per clone to a few or just one grafted scion.
Each clone is represented with a limited number of ramets (grafts). Increasing
the number of ramets of a clone is genetically equivalent to the enlargement of
the crown of an individual tree; thus, the chance for self-pollination increases.
With reference to the conditions in a seed orchard, we have to consider pollina-
tor movement not only within but also between the crowns of the ramets; hence,
both individual and clonal self-fertilization are possible unless prevented by a
system of incompatibility. Both types are identical with regard to the genetic
consequences (Sect. 6.4).
The necessary spatial isolation from other populations of the same species
obviously depends on the system of pollen dispersal, and in particular on the
efficiency of the pollen vectors (Sect. 5.2). Clonal seed orchards are designed in
order to maximize the number and proportion of offspring from outcrossing
with pollen from other selected clones. Keeping the average distance between
ramets of the same clone large helps to lower the risk of clonal self-fertilization.
Many orchards are designed in completely randomized blocks. Some subsequent
198 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
corrections are made in order to prevent close proximity of ramets of the same
clone. Certain systematic designs are even more efficient in decreasing the pro-
portion of clonal self-fertilization (Nester 1994).
The density of seed orchards is much lower than in normal plantations in
order to get larger tree crowns. The trees are often sheared in order to develop
a short trunk and a wide crown to facilitate easy seed collection. Early and
gregarious flowering is promoted.
Clonal seed orchards were established on a large scale for conifers of the
northern hemisphere (Pinus spp., Picea spp., and Pseudotsuga spp.) and for
pines, eucalypts, and teak in the tropics (Eldridge 1975; Feilberg and Segaard
1975; Example 13.2).
The mating systems of clonal seed orchards have been studied in both tem-
perate and boreal conifers. Estimates of the proportion of self-fertilization in
conifer orchards are usually low. They are thus comparable to the situation in
plantations and natural populations (Muona and Harju 1989). Deviations
from random mating were proven in orchards of Pinus sylvestris (Mller-
Starck 1982) and of teak (Finkeldey 2006). El-Kassaby and Askew (1998)
described some genetic characteristics of the reproductive process in clonal
seed orchads.
An excess of self-fertilization and other deviations from random mating as
well as pollen contamination have an impact on the genetic structures of the
progeny produced; thus, the genetic gain attainable by a production of seeds in
orchards is likely to be overestimated if it is based on the assumption of ran-
dom mating among the clones and their isolation (Mller-Starck 1982, 1991).
Estimates of the contamination rate with foreign pollen from outside the seed
orchard are often surprisingly high (Paule et al. 1993; Pakkanen et al. 2000;
Sect. 5.2.2).
In South Korea, Pinus rigitaeda, the hybrid between Pinus rigida and Pinus
taeda, has been mass-produced by handcrosses made on grafts. The equally
successful hybrid between Larix europaea and Larix kaempferi can be mass-
produced in specially designed seed orchards containing many grafts of a self-
sterile clone and many clones of the other species. Seed is then collected only
from the self-sterile clone.
Fig. 12.6. A clonal seed orchard of teak (Tectona grandis) in Thailand. (Photo:
R. Finkeldey)
in Lampang in northern Thailand. In total more than 300 clones were selected
on the basis of the phenotypic superiority of plus trees. A clonal multiplication
garden for the multiplication of selected trees was established at the Teak
Improvement Center. Single buds were grafted on stumps (Fig. 12.7). This
grafting method (budding) is quick and yields a high proportion of successful
grafts (Keiding and Boonkird 1960). Planting distances varied from 3 m 3 m to
12 m 12 m. The oldest orchard contains only 16 clones. Younger seed orchards
contain up to 100 clones; some of the clones are planned to be selectively
removed depending on the results of clonal and progeny tests.
200 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
a b
Fig. 12.7. Grafted teak (T. grandis) plants produced by the budding technique. (Photos:
R. Finkeldey)
Many seed orchards were established far from other teak forests in order to
ensure their reproductive isolation; however, several of the sites chosen were
unsuitable for growing teak. Thus, not all of the orchards established survived
and are productive. Seed harvest is below expectation in the 1830-year-old
orchards. In all orchards physiological incompatibilities between scion and
root stock are frequently observed on older grafts (Fig. 12.8). The low amounts
of fruits currently harvested in clonal seed orchards provide only a minor pro-
portion of the total reproductive material for teak plantations in Thailand.
Genetic marker studies based on isoenzyme gene loci were used to check the
clonal identity of the ramets in the clonal multiplication garden of the Teak
Improvement Center and in two clonal seed orchards. Clonal identity has not
been maintained for many trees of the multiplication garden and, consequently,
neither has it been maintained in the seed orchards. Possible explanations are a
mix-up of ramets during the establishment of the clonal garden and the seed
orchards as well as the occurrence of adventitious sprouts from the stump
which eventually develop into the main stem of the supposed graft. High levels
12.8 Propagation of Breeding Products 201
Fig. 12.8. Physiological incompatibility between root stock and scion in a grafted teak
(T. grandis) tree. (Photo: R. Finkeldey)
12.8.2
Seedling Seed Orchards
Also seedling seed orchards established with sexually produced trees serve
the production of seeds. In most cases seedling seed orchards arise from the
conversion of a progeny test aimed at the identification of superior families.
202 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
Fig. 12.9. Conversion of a progeny test with 16 families (left) into a seedling seed
orchard (right). One block containing four trees per family is shown. The eight families
with the numbers 411 are selected. In a two trees and in b only one tree per four-tree
plot is retained. In c and d the plots are not contagious and the trees are randomized
within the block. In c two trees per family are retained and in d only one tree per
family is retained. Note the uneven distribution of the remaining trees in a and b
12.8.3
Mass Multiplication of Clones
During clonal propagation the entire genotypes of the trees selected are cap-
tured for the establishment of plantations, since no recombination is involved.
This provides for one-to-one exploitation of an achieved level of breeding and
is a particular advantage for rare outstanding trees. Zobel and Talbert (1984,
Chap. 10) and Zobel (1992) strongly advocated operational vegetative propa-
gation in tropical forest trees. Park et al. (1998) reviewed the advantages and
explained the integration of clonal selection and vegetative propagation into
breeding programs.
Many forest tree species are easy to propagate vegetatively (Sect. 4.3.1).
Thus, single or a few superior trees may be selected in the forest or in a field
test and identically mass-multiplied. Asexual propagation is the basis of clonal
forestry (Chaperon 1984). Among practical advantages, clonal propagation is
useful, since the reproductive phase need not be awaited; however, the matura-
tion phase must be arrested by hedging. Clonal forestry decreases the hetero-
geneity of chemical and physical wood properties. A homogeneous product
can be beneficial in itself for certain end uses such as pulp and paper making.
Greater uniformity of plantations might be of particular value, where wood is
used in industry. Also, specific management procedures may be developed for
individual clones (Burdon 1989). Particularly clonal forestry critically depends
on specific production and management systems. The establishment of large-
scale clonal plantations requires intensive forest management techniques,
including forest protection and the maintenance of soil fertility, for example,
by fertilization (Example 12.1).
12.9 Multiple Population Breeding 205
The advantages listed previously once gave rise to clonal forestry. In contrast
to sexual propagation, clonal propagation reduces the genotypic multiplicity of
plantations drastically unless the heterozygosity of both parents is extremely
low. It is highly unlikely that two trees even of the same full-sib family share an
identical multilocus genotype. The recombination during sexual reproduction
results in a very large number of multilocus genotypes. In clonal forestry the
number of multilocus genotypes equals the number of clones represented in a
plantation. Although the spectacular yields of selected clones in plantations
have created much enthusiasm about clonal forestry in tropical countries, the
loss of genotypic variability within plantations is associated with certain risks
(Sect. 12.11).
Also in vitro propagation has received much interest (Khuspe et al. 1994).
These authors report on adapting techniques to a variety of tropical trees and
to the unpredictably flowering bamboos.
Summarizing experience with eucalypts in clonal forestry, Eldridge et al.
(1994, p. 2) stated that the mass vegetative propagation of plus trees selected in
plantations has been most successful.
Clonal forestry based on an intensive breeding program may be comparable
to the breeding of high-yielding varieties of agricultural crop plant species dur-
ing the 1960s and 1970s. The success of the green revolution has not simply
been due to isolated breeding programs but resulted from their integration into
intensified agricultural production systems in tropical countries.
In species hybrids vegetative propagation has a particular advantage, since
recombination during sexual reproduction would reduce an effect of heterosis.
Some species hybrids proved to be particularly suitable for vegetative mass
multiplication and the establishment of clonal plantations. Probably the most
successful breeding program for tropical trees was initiated in Aracruz, Brazil
(Example 12.1). A similarly comprehensive breeding and multiplication pro-
gram which provides clones for large-scale plantations of eucalypt hybrids was
initiated in the Republic of Congo with French support during the 1970s
(Delwaulle 1989). Clonal eucalypt plantations owned by private enterprises
exist also in India and other countries. The production and selection of clones
in these plantations is reported to be based on less intensive breeding programs
than those in Aracruz and the Republic of Congo (Lal 1993).
12.9
Multiple Population Breeding
The hitherto mentioned breeding methods are hierarchical since certain indi-
viduals or genotypes are selected from a base population, and these individuals
or their offspring are assembled to a breeding population of again smaller size,
etc. This approach has been adopted with certain modifications from traditional
206 C HAPTER 12 Domestication and Breeding of Tropical Forest Trees
breeding methods for agricultural crop plants. The expected breeding progress
is accompanied by a gradual loss of genetic variation in the breeding population
which does not only affect the gene loci selected but owing to possible drift
effects the totality of the genome.
The multiple population breeding concept of Namkoong is nonhierarchical
and aims at the preservation and utilization of genetically differentiated sub-
populations (Namkoong et al. 1980, 1988, p. 70 ff.; Namkoong 1984). In 1994,
Namkoong was awarded the Marcus Wallenberg Prize for his pathbreaking
contributions to quantitative population genetics, tree breeding, and manage-
ment of genetic resources, which form a solid scientific basis for the mainte-
nance of biological diversity in forests all over the world.
The objective is no longer breeding of a single population accompanied by
a reduction of genetic variation or the selection of only a few genotypes that
are expected to produce superior phenotypes in different environments.
Instead, the breeding population consists of a variable number of separate sub-
populations. Multiple population breeding is directed towards different objec-
tives, i.e., different traits are selected at different sites. Even if the selection
criteria are the same, sets of controlling genes are expected to be differentiated.
Both the adaptedness and the adaptability to different and changing environ-
mental conditions are promoted by the maintenance or enhancement of a
differentiated population structure. The concept sacrifices the maximization of
breeding progress in a few economically important traits. An important objec-
tive is the enhancement of the adaptedness of different subpopulations to
diverse environmental conditions. Adaptability is maintained by the resulting
differentiated population structure. The concept is highly flexible but requires
skill for managing the genetically effective population size, including pedigree
control. It is possible to adjust the breeding program to changed economic pri-
orities since different breeding objectives are defined in different subpopula-
tions. Both the number of replicate populations and the number of parents in
populations, their diversity and their differentiation can be used to control the
total genetic variation available for managers to use in deploying trees in
plantations (Namkoong 1999). Now as before, the total size of the subdivided
breeding population is crucial under the aspect of inbreeding avoidance.
However, the genetic variation can be managed more effectively.
Multiple population breeding is particularly suited for potentially important
plantation species of the tropics which still are in an early state of domestica-
tion. A main advantage of the concept is its flexibility both with regard to the
necessary expenditure and the possibility for modifications during its imple-
mentation (Namkoong et al. 1980). For instance, also heterosis breeding can
easily be integrated. Furthermore, breeding and conservation objectives are
regarded as equally important and may be profitably combined in a single
program (Namkoong 1984).
12.10 Breeding Strategy 207
12.10
Breeding Strategy
The main objective of forest tree breeding is the modification of average trait
expressions of economically relevant traits in plantations. However, breeding and
propagation of breeding products is usually done in small populations. Thus, it
is meaningful to distinguish the following populations as part of a breeding
strategy (Libby 1973; van Buijtenen 1975; Kang et al. 1998):
The populations serve different functions. For instance, a progeny test (breed-
ing population) may later be converted to a seedling seed orchard (propagation
population). Under a multiple population breeding strategy, the base, breed-
ing, and propagation populations are primarily separated, although there are
relationships between the populations (Kang et al. 1998). The breeding pro-
gram has the improvement of yield (in the widest sense) and the conservation
of genetic variation as two goals possessing equal priority. It is integrated into
the domestication process of the respective tree species.
The decrease of additive genetic variation (Young and Boyle 2000) in
breeding populations as closed systems is inevitable. One may think of a
somewhat simple relationship requiring little theory of genetic variances and
selection plateaus. In a given population a selection trait is assumed to be con-
trolled by several biallelic gene loci without dominance or epistatic effects. If
the favorable alleles possess low frequency in the base population, the genetic
diversity is going to increase under selection. Maximum diversity is reached
when the alleles possess equal frequency. The more the breeding goal is fur-
ther approached, the smaller becomes the diversity. The situation is, of course,
12.11 Genetic Consequences of Domestication and Breeding 209
12.11
Genetic Consequences of Domestication and Breeding
12.11.1
General Considerations
the long-term survival of the species. However, the actual danger of reduced
adaptedness through breeding has not been proven by experimental studies in
tropical forest trees. Some results have been reported for temperate and boreal
species. For example, selection for fast growth during early stand development of
Norway spruce (Picea abies) negatively affects the resistance against wind-break
at later age (p. 399 in Hattemer et al. 1993), an important element of stability. In
this case, breeding would lead to instability, i.e., the departure of realized from
predicted performance (Namkoong 1999). Among all organisms subjected to
breeding work, the environment of forest trees can be controlled least.
The less intensive the finite breeding populations are selected, the less is
population size reduced. Part of the breeding progress should be relinquished,
for instance, by retaining a few individuals also of culled families in the breed-
ing populations and the propagation populations such as seedling seed
orchards. The breeder must find a best compromise between selection objec-
tives and selection intensity.
Another form of renunciation is restricting the time during which seed is col-
lected from a given seed orchard. In view of the money spent for establishment,
there is a tendency to use the offspring for planting on vast areas. These consid-
erations may look uneconomic and paradoxical at first sight but they represent
sustainability imperatives. Immediate maximum economic benefit from tree
breeding implies setbacks during future generations of both trees and humans.
The technologies of breeding and selection offer opportunities for manag-
ing the genetics and guiding the evolution of forests that neither decrease nor
increase risks over that of uncontrolled, natural regeneration. Technology
involved in domestication and breeding is neither the problem nor the solu-
tion. It is a tool that humans have available. It can be used well or poorly
(Namkoong 1999).
12.11.2
Monitoring of Breeding Projects by Use of Genetic Markers
There is ample opportunity for the use of gene markers in breeding programs
(Yeh 1989; Forrest 1994; Butcher et al. 1999). A highly rewarding application is
the verification of the validity of crosses, as an example may show. A tree breeder
had established a clonal seed orchard. When it started to flower, crosses among
the clones were made for progeny testing. Years later it was detected by the appli-
cation of genetic markers that the progeny test was a complete mess, because
a muddle had happened already during production of the grafts. Part of the
financial loss may have been avoided by checking the clonal identity of the ram-
ets and by analyzing seed samples in the laboratory before sowing. Since possibly
12.12 Recommended Literature 211
12.12
Recommended Literature
The textbook by Wright (1976) covers forest genetics and tree breeding, includ-
ing provenance research. An introduction to quantitative genetics that has
found wide circulation is the book by Falconer and MacKay (1996). It deals
with population genetics and all aspects of quantitative genetics, including
simultaneous selection for multiple traits and the use of QTLs. Namkoong
et al. (1980, 1988) presented comprehensive accounts of breeding forest trees
on the basis of population and quantitative genetics. The book by Zobel and
Talbert (1984) covers all aspects of tree breeding, such as methods and results.
The book by Zobel and van Buijtenen (1989) is on wood and deals with the
genetic aspects, such as variation and breeding. The book edited by Fins et al.
(1992) covers all aspects of forest tree breeding and presents numerous exam-
ples from temperate species. Eldridge et al. (1994) cover breeding, with empha-
sis on eucalypts. In the book edited by Mandal and Gibson (1998), topics with
significant current interest in forest genetics and tree breeding are discussed. As
stated at the end of Chap. 11, the book by Williams et al. (2002) deals with
experimental design and analysis also in close-sense tree breeding.
Genetic Aspects of Plantation Forestry 13
in the Tropics
13.1
Introduction
In tropical countries ever more land is becoming available for tree plantations.
This is due to the destruction and degrading of natural forests or failures of
their management and growing problems of access to existing forest.
Plantations represent either reforestation, i.e., changing deforested land back to
forest land, or afforestation, i.e., creating forest where it historically did not
exist. The high productivity of tropical and subtropical plantations and their
numerous ecological, social, and economic benefits make plantations highly
attractive. Much attention has recently been devoted to research and planting
of nitrogen-fixing hardwoods such as Leucaena spp. (Brewbaker and Sorensson
1994) and Sesbania spp. (Owino et al. 1994). Planting is urgently required on
eroded sites or on sites threatened by erosion. It is also practiced with the
objective of restoring forest ecosystems. Planting is done by private companies,
in the framework of special developmental projects, and in national afforesta-
tion programs. Planting projects are also promoted by private companies and
nongovernmental organizations (Evans and Turnbull 2004). The total area of
forest tree plantations in tropical and subtropical countries has been estimated
to be close to 100 106 ha in 2000 (Table 13.1) and may have exceeded this area
since then. The annual loss of forest area is still estimated to exceed by far that
of newly established plantations (FAO 2002).
A comprehensive discussion of tropical tree plantations has to take into
account ecological, genetic, silvicultural, and socioeconomic aspects (Evans and
Turnbull 2004). The genetic aspects of man-made forests established by plant-
ing or sowing are centered around their establishment. Therefore, the most sig-
nificant genetic aspect is the choice of reproductive material used. Plantations
are much less competitive than natural forests. The competition among adult
trees for light, nutrients, and water is not accompanied by as much mortality as
during the early phases of the life cycle in natural populations. These critical
phases are replaced under human intervention by procedures designed towards
214 C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics
Table 13.1. Areas of planted forests including woodlots in tropical and subtropical
regions and their average annual increase from 1965 to 2000. Numbers refer to
thousands of hectares. (Adapted from Evans and Turnbull 2004)
Region 1965 1980 1990 2000 Annual
average
Table 13.2. Species used in tropical planted forest. Percentages refer to the main
groups only. Adapted from Evans and Turnbull (2004, p. 37)
Genus/group Percentage Species
13.2
Plantations of Exotic Tree Species
When an exotic species is to be planted, the forester has the choice among
a very wide spectrum of species that have the advantage of being free of their
native herbivores and parasites. It may take some time until local insects or
fungal pathogens become adapted to the new food source, or until the native
pests and diseases are transferred.
At present, exotic tree species are of prime importance for plantation
forestry throughout the tropics. Managers can build on international experi-
ence with these well-known species that have been successfully used for plan-
tations worldwide. Pure plantations using only a single or a few commercial
species are the rule. Particularly various fast-growing, mainly exotic, pioneer
tree species such as eucalypts, acacias, Gmelina, and pines are grown in single-
species plantations. The plantations existing in more than 100 countries are
used for pulp, woodchips, sawtimber, fuel wood, and other products.
Owing to extremely short rotation periods, ecological risks are expected to
be low. If plantations are grown on suitable sites, the wood production is
extremely high. However, the demand for nutrients and water may also be
high. More and more such plantations are established with breeding products
(Chap. 12), which provides for sometimes spectacular rentability and makes
plantation forestry attractive. These plantations cannot simply be considered
to be substitutes of natural forest (Chap. 8 in Bruenig 1996). Instead of replac-
ing existing primary or secondary forest, they should be planted on deforested
land and remove pressure from natural forest.
The introduction of exotic species has frequently happened in several steps.
The widest distribution originated from locations far away from natural pop-
ulations. For example, both the coffee (Coffea arabica) and rubber (Hevea
brasiliensis) trees were originally planted as exotics in fairly small populations
216 C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics
13.3
Plantations of Indigenous Species
Given the threat to the forest gene pool, the promise of a renewed private planting pro-
gramme, and our ignorance of propagation techniques and growth potential of many
potentially important species, it is more than ever necessary to co-ordinate national
germplasm collection and breeding efforts on native hardwood trees . . . (Lawson 1994).
The species spectrum of indigenous species of a country is, of course, not as
broad as the worldwide spectrum of exotic species. Indigenous species still pos-
sess the advantage of being adapted to the regional climatic conditions. Local
consumers are familiar with the properties of their wood and the various non-
timber products that they yield. The utilization of species for plantation forestry
is frequently impeded by the low level of knowledge of their basic biological,
ecological, and silvicultural features. Evans and Turnbull (2004, Chap. 8) dis-
cussed arguments for the choice of a given species and presented examples of
indigenous species that have recently been used for plantations. Many of these
species are rare in natural ecosystems. Their use in plantations containing a few
or only one species involves a drastic change in ecological conditions, particu-
larly the pressure of infestation by fungi and insects.
The importance of indigenous species is still low in plantation forestry of the
tropics (Table 13.2). However, the anticipated increase of the plantation area is
expected to be based on a larger number of species, including indigenous species
and lesser-known species. An increased number of species used for plantation
forestry promotes the establishment of mixed plantations (FAO 1992).
The growth of species in natural forests of the region gives some indication
of their value for use. Subsequent species-elimination trials allow us to assess
whether a species is suitable for being grown in plantations. The observation of
a few trees, for example, in a botanical garden or an arboretum alone, is not a
sufficient basis for the identification of potentially important plantation species.
A test of the suitability of a species or population for plantation establish-
ment is only meaningful if the population from which the seeds were har-
vested is conserved at least until the end of the trial. This represents a genetic
resource as a necessary component of any program for the domestication of a
species.
Procedures of seed procurement of exotics used during the early days of
tropical forest plantations may easily be avoided in indigenous species. This
provides for more genetic variation being retained in plantations. Nevertheless,
some basic principles of seed harvest and of experimental design must be fol-
lowed. Above all, seeds should never be harvested from isolated trees since
inbreeding is likely to occur and the genetic structures differ from those of nat-
ural populations.
13.3 Plantations of Indigenous Species 219
The most important exotic plantation species are characterized by their gre-
garious production of seeds. The copious small seeds of species such as pines,
eucalypts, acacias, and teak are easy to harvest in large quantities and to store
for several months or even years, i.e., the seeds are orthodox. This feature and
the ease of the production of seedlings in nurseries contributed to the impor-
tance of the species for plantation forestry.
Problems in seed procurement for afforestation projects are frequently
underestimated. It is difficult to produce and store suitable reproductive mate-
rial in sufficient quantities. This is due to irregular and sporadic seed produc-
tion and/or recalcitrant seeds, i.e., seeds which cannot be stored for long
periods. Other species have a seed dormancy that is difficult to break for large-
scale production of seedlings. The main problem with those species is frequently
not the choice of genetically suitable reproductive material but rather the pro-
duction of sufficiently large quantities of planting stock. Seed orchards (Sect.
12.8) help to solve some of these problems. Other help is seen in vegetative
propagation.
Leakey et al. (1982) emphasized the potential use of numerous species for
plantation forestry in the tropics. They described the options for an improved
seed procurement and relevant research in this context using the example of
the West African species Triplochiton scleroxylon, a species with an erratic
pattern of reproduction. By adequate treatment and the use of appropriate
facilities, the storage period of seeds could be extended. Besides the develop-
ment of methods for controlled pollination, an integrated procedure allows for
vegetative propagation and the early selection of tree clones with good stem
form. Ladipo et al. (1991) reported further on this research.
In West Africa, where the discrepancy between annual deforestation and
total tree plantations is particularly wide, Lawson (1994) identified an urgent
need for domestication of indigenous tree species such as Terminalia ivorensis,
Terminalia superba, Triplochiton scleroxylon, and Lovoa trichilioides. The clonal
propagation option is preferred in view of the problems involved in the pro-
curement of seed of good quality, and the increased yield. This may be an effi-
cient measure where forests disappear fast. Milimo et al. (1994) listed species of
semiarid West Africa that are worthy of breeding. In view of various difficulties
involved in propagation by seed, these authors stressed the importance of veg-
etative propagation and reported on experimental results. In their comprehen-
sive review of domestication of the genus Swietenia and other mahoganies,
Newton et al. (1994) described the development of molecular genetics tech-
niques, provenance research, selection, and the development of both conven-
tional and in vitro propagation. On this basis, appropriate silvicultural systems
accounting for the Hypsipyla problem could be developed.
The worldwide use of eucalypts may create the impression that these are the
only important tree species in Australia. Nevertheless, projects in provenance
220 C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics
13.4
Basic and Reproductive Material
on the market is restricted to approved basic material, only about 1% of the area
covered by the main tree species contributes offspring to plantings in some
member countries. The area where seed is collected changes with time, so there
is some change in the collective of approved basic materials. The percentage of
the area covered by the respective species is nonetheless very small. However, the
given degree of isolation hardly prevents influx of effective pollen. If no addi-
tional cost is involved, seed may be collected in selected stands, although expec-
tations of far-higher value for use are not necessarily justified.
Source-identified reproductive material does not meet the criterion of being
derived from selected base material. It is assured to have been produced in a
given region of provenance. Qualified reproductive material must be produced
by parent trees or in a seed orchard the components of which were selected on
an individual basis. It may also be propagules of a clone or clonal mixture.
Tested material is not just derived from phenotypically superior basic mate-
rial but has been tested for improved value for use in experiments that meet
certain criteria of design. The improved value is defined as the superiority over
officially prescribed standard entries included in the experiments. Until the
present, only a very small fraction of reproductive material has been marketed
in this category.
The amounts of reproductive material collected in selected, qualified ,or
tested base materials are not restricted. The advantage of the legislation is cer-
tainly the concomitant documentation of the seed collection site, even if this
may not represent relevant information on reproductive material (Jones and
Burley 1973).
13.5
Production and Collection of Seed
Natural forests as a seed source may present problems for seed collection.
Particularly the collection from tall trees in less prolific species is laborious in
spite of modern equipment. Fruits of many species are dehiscent. Fruit and
seed parasitism and fruit-eating animals mitigate harvested seed. Melchior
(1986b) described problems related to seed procurement for three important
and promising South American tree species in Peru and Venezuela (Cedrelinga
catenaeformis, Juglans neotropica, Bombacopsis quinata).
13.5.1
Seed Production Areas
Seed production areas arise either by establishment of a stand with the objec-
tive of seed production or by phenotypically superior stands being converted
13.5 Production and Collection of Seed 223
13.5.2
Provenance Resource Stands
Provenance resource stands are a particular type of seed production area. They
are frequently established simultaneously with the provenance trials, i.e.,
during an early stage of a breeding program. They represent a mixture of the
progeny of numerous seed trees of a single promising population and should
224 C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics
13.5.3
Seed Orchards
Plus trees are possibly older and more heterozygous at certain gene loci than
other trees in the same population (Bergmann and Ruetz 1991). They may have
more variable progeny. Comparisons of the inbreeding coefficient of the seed are
possibly less problematic. Since any comparisons are still subject to the vagaries
of sexual reproduction, repeated sampling of the seed output is required.
Also the less expensive seedling seed orchards are mostly linked to breeding
programs. Lee (2000) pointed to the aspect of pollination in seed orchards.
When studying the mating system of Dryobalanops aromatica, a dipterocarp
species with irregular flowering, he found an indication for a lack of pollina-
tors in a seed orchard being responsible for increased self-fertilization (see also
Sects. 10.2.1, 12.8.2). If a seed orchard does not consist of selected families to
be propagated under isolation, precautions may be taken for improving the
situation by locating orchards where they are surrounded by extensive forests
of dipterocarps or nondipterocarps providing pollinators. The appropriate
spatial configuration can help to minimize other consanguineous matings. The
incidence of the same species might also provide genes. Lee (2000) also warned
against placing seed orchards into other ecosystems in order to avoid undesir-
able molecular imprinting and maladaptation. Therefore, this author proposed
delineating planting zones on the basis of biological and ecological data and
establishing plantations only in their own planting zone. This procedure is
practiced in many countries of the northern temperate zone. It represents an
alternative to the principle of transfer rules as mentioned in Sect. 11.6.
The aspect of pollination was also discussed by Moncur et al. (1995). In seed
orchards of several eucalypt species these authors observed a considerable
improvement in both the amount and the quality of seed after setting up bee-
hives. The average outcrossing rate was raised from 0.76 to 0.91. In Europe,
beehives are also placed in remote stands of rare trees of the rose family in
order to increase the seed crop.
Also Moran et al. (1989b) and Chaix et al. (2003) (Sect. 12.8.2) reported
considerably lower rates of self-fertilization in seedling seed orchards than in
natural stands of eucalypts. The former authors emphasized the reduced
neighborhood inbreeding by less spatial clumping of relatives. In a seed
orchard containing both grafts and open-pollinated progeny of plus trees in
Douglas fir, Ritland and El-Kassaby (1985) found only little self-fertilization
but indications of consanguineous matings primarily due to the large and
varying size of the families. In a seedling seed orchard of Pinus merkusii Siregar
and Hattemer (2001) observed a relatively low outcrossing rate (0.87) with lit-
tle incidence of consanguineous biparental mating; however, during the
extended flower period the phenology differed between families and the indi-
vidual effective pollen clouds were genetically differentiated. In a far-distant
natural stand the outcrossing rate of 0.98 was similar to that of other pines and
226 C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics
13.6
Collection and Storage of Seed
Seed is either collected off the trees or picked up from the ground. Seed collec-
tion by climbing is expensive and dangerous in spite of modern equipment.
The seed should be collected in an appropriate way so as to minimize losses of
genetic variants. The genetic implications of seed collection from trees were
discussed in Chap. 11. Too few trees might be available for seed collection from
the crowns of felled trees. Seeds of other species are caught in nets laid out on
the forest floor. If a restricted number of nets are laid out under highly prolific
trees only, a spatial genetic structure becomes manifest in the seed crop and
induces genetic change. In stands with a family structure this seed is expected
to be markedly different from seed picked up by hand from the total area of the
stand as was shown by Ziehe et al. (1998) in a study of beech (Fagus sylvatica).
The seed of many tropical trees such as Triplochiton scleroxylon or Araucaria
spp. is also collected from the ground.
Also when collecting wildings, their genetic variation should be accounted
for. A patchy distribution of saplings indicates that they might represent off-
spring of a restricted number of trees.
The recalcitrant seeds of many tropical species are sensitive to desiccation
(mangrove species, neem, Araucaria angustifolia and other Araucariaceae, most
dipterocarps) and are thus rather short-lived. Seeds of many tree species of the
temperate zones are orthodox and can be stored for a considerable time. The
seeds of most Acacia spp., Eucalyptus spp., and teak are orthodox. Seed treat-
ment such as the careful reduction of the moisture content and the storage in
a strictly controlled environment can be used to extend the life span of the
seeds of most species (Schmidt 2000).
Air-conditioned cold-storage rooms are used for the long-term storage of
seeds in gene banks. Optimum conditions for long-term storage depend on the
seed physiology of the species. Seeds of some tropical species retain germinabil-
ity even at temperatures below the freezing point of water and thus can be
frozen without damage. The life span of seeds of other species is longest at low
temperatures just above the freezing point of water. The humidity in storage
rooms is also controlled and usually reduced compared with that outside. More
13.7 Planting Stock Production 227
details on optimum storage conditions for the seeds of tropical trees have been
summarized, for example, by Albrecht (1993), Tompsett (1994), and Schmidt
(2000, p. 223 ff.).
The germinability of many recalcitrant seeds cannot be extended to more
than a few months. The principal distinction between orthodox and recalcitrant
seeds is maintained even under optimum storage conditions. The long-term
storage of seeds is not an option for the conservation of genetic resources of
most species with recalcitrant seeds.
In any event, the loss of germinability represents a reduction of the popula-
tion size. As some studies in temperate tree species have shown, this reduction
is also connected with selection (Melchior 1986a; p. 340 f. in Hattemer et al.
1993; El-Kassaby 1999). The selection going on during germination and nurs-
ery operations rids the planting stock of adaptively weak individuals arising
from inbreeding. Seeds of certain families survive for different periods, so the
variation decays during storage. It is an open question whether this also implies
adaptation to the storage conditions. Evans and Turnbull (2004, p. 134),
Schmidt (2000), and Midgley (1996) reported new developments in seed col-
lection and seed technology of tropical trees. Research on seed germination of
fruit-bearing trees is under way (Maghembe et al. 1994).
13.7
Planting Stock Production
Artificial regeneration does not rely on the natural seed fall and the concomi-
tant seed dispersal. Seeds are collected off the trees, processed, and eventually
stored. The seed is either sown out directly or planting stock is raised in a nurs-
ery or greenhouse under intensive treatment so that mortality and thus the
chance of adaptive viability selection are minimized. Outplanting or seeding is
hardly ever done on the site of the parent stand. The implied habitat change
interrupts the long-term process of adaptation and enforces selection with
changed intensity and direction. The emerging artificial stand no longer has a
family structure.
13.7.1
Seedlings
system of trees (Ziehe et al. 1999). This type of selection is then postponed to
the period after outplanting.
Also grading of plant material has a genetic implication. Under highly favor-
able conditions of planting stock production, Eriksson and Lindgren (1975)
found less inbreeding depression in both Picea abies and Pinus sylvestris and
inferred that there may be more inbred offspring among tall plants. In experi-
ments done by Konnert and Schmidt (1996), the tallest plants in planting stock
of Picea abies turned out to possess the lowest degree of heterozygositya prop-
erty that is otherwise an advantage for field survival. In European white fir the
result was the reverse. In both species, the taller and shorter plants were genet-
ically differentiated.
13.7.2
Clonal Multiplication
difficult. The inevitable loss of genetic variation associated with the shift from
sexual to asexual propagation techniques is accepted in view of the pragmatic
advantages of vegetative multiplication techniques such as the uniformity of
the plantations. Seed production of other species such as most eucalypts is
prolific, seeds are orthodox, and sexual progenies, i.e., seedlings, are easily pro-
duced using standard nursery techniques. In these species asexual propagation
is more laborious and expensive. Vegetative propagation is applied only for the
mass multiplication of certain selected trees. The term clonal forestry is asso-
ciated with an anticipated increase of the yield by planting these selected clones.
The risks associated with the unavoidable reduction of genetic variation in
clonal plantations occur, of course, independently from the motivation for the
use of vegetative propagation techniques. The often-neglected genetic implica-
tions of clonal plantations are addressed in Sect. 13.9.
However, vegetatively propagated plants are not necessarily superior to gen-
eratively produced progenies. For example, the growth of vegetatively propa-
gated E. grandis plants was compared with that of seedlings from the same trees
after open pollination in Brazil. For a variety of reasons, the growth of the
seedlings was superior to that of the vegetative offspring 30 months after plant-
ing (Kageyama and Kikuti 1989).
13.8
Establishment and Development of Plantations
From the genetic point of view the choice of the reproductive material is the
most important decision before plantation establishment; however, genetic
structures of artificially regenerated forest tree populations may change in time
during the reduction of the number of individuals owing to viability selection.
A temporal dynamics of genetic structures was observed in experimental popu-
lations of some temperate forest trees during early development of the seedlings
(Kim 1985; Ziehe et al. 1999). In order to provide ample room for adaptive
change, large numbers of individuals should be planted (Langner 1966; Ledig
and Kitzmiller 1992; Mller-Starck 1996). In rotation forestry with short rota-
tion periods, this safeguard is of less importance than in long-lived tree species.
Forest reproductive material can be transferred over large distances. The
capacity of artificial migration is obviously unlimited. The ecological gradients
overstepped by this transfer matter greatly. Sometimes reproductive material is
totally ill adapted. Uncritical use of just any reproductive material has long
been and still is hazardous to forestry; therefore, the careful choice of popula-
tions must be integral part of adequate forest management.
The number of plants per area or tree spacing is an important aspect of plant-
ing. In tropical plantations mostly fewer than 1,000 trees are planted depending
230 C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics
on the purpose of the plantation. The potential for selective change of genetic
structures and the purging of physiologically ill adapted individuals depends on
the genetic variation of the planted population, the number of trees planted, and
subsequent mortality. Evolutionary adaptations (Sect. 7.2) avoiding critical
population sizes are more likely in plantations established with a large number
of different genotypes planted per unit area. No evolutionary adaptation on the
basis of viability selection is possible if only a single clone is planted.
The temporal dynamics of the genetic structures of a population is not only
influenced by natural mortality. Silvicultural operations and particularly selec-
tive thinning have also the potential to change genetic structures. Different
thinning methods affect genetic structures in different ways as was shown for
Norway spruce (Picea abies) by Hosius (1993). Comparable experimental stud-
ies for plantations of tropical forest trees are not yet available.
The temporal dynamics of genetic structures in planted forests deserves par-
ticular attention if it is planned to harvest reproductive material in a plantation
either after its conversion to a seed production area or if it was planned to be nat-
urally regenerated. Changes of genetic structures in production populations that
will never give rise to a subsequent generation are of no evolutionary significance
although they may have an impact on economically relevant trait expressions.
13.9
Natural Regeneration of Plantations
the natural seed fall is the larger number of seeds and seedlings produced dur-
ing the regeneration period comprising several years. In ever-more countries a
mosaic of autochthonous and allochthonus stands exists, so the incidence of
hybridization between trees from stands of different origin must be expected to
increase. Particular features of natural regeneration were dealt with in Sect. 10.3.
13.10
Use of Breeding Products
All plantations are perceived to be in danger to some degree. They are threat-
ened by a massive buildup of pests and diseases (p. 387 in Evans and Turnbull
2004). Examples of devastating outbreaks have been reported from the tropics,
the temperate zone, and the boreal zone. Plantations are still of growing impor-
tance. They are the only means of deploying breeding products and obtaining
return from earlier investment. It was pointed out in Sect. 12.11 that the col-
lection of unrestricted quantities of seed in a given seed orchard or seed pro-
duction area implies that the progeny of the respective trees cover an area
exceeding that of the collection stand by a factor of several hundred. This is the
experience in countries with official approval of seed collection stands and
provenance recommendations. It is more costly to establish many orchards and
many seed production areas and to collect only moderate amounts of seeds
there; however, it is more in line with sustainability principles.
Even the selection of provenances is a form of breeding that usually precedes
more intensive selection steps. The large-scale planting of introduced species,
provenances, or other breeding products counteracts the preservation of locally
adapted populations that may contain elsewhere rare alleles or even unique alle-
les or allele combinations in high frequencies. Thus, important genetic resources
for future adaptation processes and breeding programs can get lost. The loss of
local genetic resources can be either direct by substitution or indirect through
gene flow from breeding products and other allochthonous populations into
surrounding natural populations of the same species.
The dangers associated with the loss of locally adapted populations through
their substitution by a few high-yielding varieties were first recognized for cul-
tivated species. They have been the main motivation for the establishment of
gene banks of many agricultural crop species. Tree plantations using breeding
products currently cover only a small part of tropical forests. The danger of the
loss of locally adapted forest tree populations is still much less severe than
damage caused by forest destruction.
Reduced genetic variation is a consequence of most breeding programs. This
increases the risk of severe attacks by insects and diseases that may rapidly
adapt to a homogeneous source of nutrients offered to them during several
232 C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics
proportions. This need not imply loss of uniformity of the wood and tim-
ber produced. The effect of mixture is the larger, the more the clones are
genetically differentiated from each other at adaptive gene loci or in resist-
ance. This quality of the mixture is more likely to be achieved if the clones
are of diverse descent. This spreads the risk in view of adverse environ-
mental factors and retards contagion during infestation by diseases and by
defoliating or other insects.
The question of the necessary number of components in mixtures has been
tackled by several authors. For instance, Hhn (1986) considered the risk
of losing the plantation through the loss of an unacceptable fraction q of
trees prior to final harvest. The risk is a function of this fraction, the num-
ber n of clones, the susceptibility x, and the intensity a of contagion. For
given q and a the risk decreases with increasing n and decreasing x. There
are also special cases of x where the risk increases with increasing n. In
many situations the risk function is asymptotic for larger n. Increasing this
n beyond 3040 equally represented clones is of little effect. Libby (1982)
used a binomial approach for constructing the risk function and recom-
mended between seven and 25 clones.
The genetic control of susceptibility was not considered in these models. In the
approach taken by Roberds et al. (1990) a single agent active in the plantation
and a single biallelic locus controlling resistance was assumed; hence, the fre-
quency of either a dominant or a recessive allele for susceptibility could be inte-
grated. These authors also assumed a particular mode of selecting given
numbers of clones from progenies of the base population for propagation and
found rather complex behavior of the risk function. It depends on the mode of
gene action, since under dominance the proportion of susceptible genotypes is
always larger. It also depends on the intensity of infestation whether increasing
numbers of clones reduce the risk. If the intensity of pest attack increases, the
number of clones necessary decreases. Simulation studies by Bishir and Roberds
(1999) revealed that situations requiring more than 40 clones are not prevalent.
Campinhos (1999) reported the use of 80 clones per production population in
a given cycle of clonal mass cultivation in Aracruz (Example 12.1). Several EU
and OECD countries have issued regulations on minimum numbers.
One of the strong simplifying assumptions for the validity of these models is
the static nature of susceptibility, i.e., the clones are either susceptible or not and
this property does not change during the single rotation period considered.
However, it is our experience that insects or pathogens, organisms with shorter
generation cycles and more flexible genetic systems, evolve mechanisms to
exploit the homogeneous substrate offered by genetically uniform plantations
for years. The clonal aggregates are in turn hardly able to develop evolutionary
adaptation in response to these attacks. This condition introduces uncertainty.
234 C HAPTER 13 Genetic Aspects of Plantation Forestry in the Tropics
Risk considerations are not intended to question or even disparage the mag-
nificent success of clone breeding in the tropics and subtropics. Some of the
13.11 Recommended Literature 235
worlds most productive tree plantations, primarily poplars and eucalypts, use
clones. Crucial advantages of clonal forestry are the independence of seed
shortages, the uniform stands, and the ease of their management. Problems are
seen in the deployment of clones. The concern is motivated by the growing
need for intensive measures of plant protection in agricultural crop plants, not
least in perennial crops such as avocado, banana, citrus, coffee, coconut,
mango, oil palm, papaya, rubber, tea, and vine. Sooner or later in all plant cul-
tures with reduced genetic variation a slow but steady increase in the incidence
of insects and diseases has to be envisaged and the pressure on tree plantations
will increase. Since forest trees are not a cash crop, intensive chemical control of
pests and diseases rules itself out for economic, not to speak for ecological rea-
sons; therefore, the matter is of importance both economically and environ-
mentally. It is therefore preferable to rely on prevention. Foster and Bertolucci
(1994) have discussed several independent approaches to the maintenance of
some level of genetic variation in large-scale clonal plantations.
For the time being, intensive monitoring is advised and is done (Campinhos
1999). The economic returns from highly productive plantations allow for this
active monitoring of pests and diseases (Evans and Turnbull 2004).
Repeated reference made to the clonal program of Aracruz Florestal should
not be perceived as hidden criticism. This program is under way in the most
widespread exotic tree genus and is also the most advanced and successful. In
1984, Leopoldo G. Brandao, Edgard Campinhos Jr., Ney M. Dos Santos, and
Yara K. Ikemori were awarded the Marcus Wallenberg Prize for their pioneer-
ing work leading to significant scientific and technological breakthroughs in
developing commercial forests based on cloned Eucalyptus. Their methods
have stimulated world-wide emulation. The high productivities achieved will
reduce pressures on natural forests.
13.11
Recommended Literature
The book by Evans and Turnbull (2004) covers all aspects of tropical tree plan-
tations. It represents comprehensive information on this subject and in Part III,
devoted to plantation silviculture, gives answers related to genetics that can
only be touched on in the present publication. Subjects related to genetics are
also covered in the book by Zobel et al. (1987).
Conservation of Genetic Resources 14
in Tropical Forest Trees
There is ample evidence that genetic impoverishment of many tree species has
occurred and is occurring as a result of the destruction of forest ecosystems of which
these species are a part. Furthermore, it is certain that in many countries forest genet-
ics and tree improvement programs will not be operative in time and on a sufficient
scale to insure the preservation in clone banks and seed orchards of an adequate sam-
ple of the spectrum of variation within the important tree species. Numerous foresters
are aware of this situation, and of the need for the conservation of forest gene resources
by insurance of the survival of the forest ecosystems in which the species are found, and
by the establishment of seed banks, clone banks, and seed orchards. Furthermore,
united action is occurring at an international level that has a direct bearing on this
problem (p. 274 in Stern and Roche 1974).
14.1
Introduction
About one quarter of the earths land surface has been converted to food
production. Only 10% of the original landscapes is estimated to have
remained unchanged. This condition and other forms of human impact on
plant and animal populations have led to the conservation imperative
accepted by ever-more people. The importance of the conservation of bio-
logical variation is undisputed. Genetic variation is a crucial component of
biological diversity, since it is a necessary precondition for the continuation
of biological evolution.
Furthermore, genetic variation within populations is a preconditon for het-
erozygosity, which according to empirical studies is correlated to fitness at the
population level. Losses of heterozygostity have a deleterious effect on fitness
(Reed and Frankham 2003). Although this correlation depends on the methods
of assessment used and is only moderately strict, it can be considered a second
argument for the commitment to genetic conservation as maintaining the basis
of both the long-term stability and the short-term productivity of forest
238 C HAPTER 14 Conservation of Genetic Resources in Tropical Forest Trees
Once certain priorities (Sect. 14.3) have been agreed upon, a conservation
project for a given tree species starts with the clear identification of the objec-
tives (Sect. 14.4). Genetic resources are subsequently selected mainly on the
basis of the available knowledge of spatial patterns of genetic variation (Sect.
14.5). The choice of the method of physical preservation of the genetic infor-
mation in the selected organisms (Sect. 14.6) is related to the final step, the
regeneration of the resource (Sect. 14.7).
14.2
Development of Tree-Conservation Genetics
Since 1960s, programs for the genetic conservation of a few species have been ini-
tiated in the framework of planned or ongoing breeding programs. Conservation
of forest genetic resources was considered an item of breeding strategies also in
more recent publications (Cossalter 1989; Eldridge 1990). However, the signifi-
cance of conserving genetic resources of tropical forest tree species not (yet)
incorporated in breeding programs was also realized (Kemp et al. 1976).
Scientific results and practical guidelines for the conservation of forest genetic
resources were first summarized in a report edited by Roche (1975).
During the 1980s, conservation biology emerged to provide the underlying
science needed to slow down the process of extinction by conserving biolog-
ical diversity at all levels, including the genetic diversity within species. This
crisis-oriented discipline (Soul 1991) builds on the fields of ecology, biogeog-
raphy, and genetics applied to small and declining populations. The crucial role
of tropical forests for the global conservation of biological diversity has been
repeatedly pointed out by conservation biologists (Myers 1986). Genetic
resources of forest trees are endangered for several reasons, including forest
destruction and fragmentation, global climate change, inappropriate manage-
ment practices, selective logging in natural forests, the conversion of species-
rich forests to plantations of low diversity, and inappropriate methods of
breeding. The destruction of tropical forests due to unsustainable management,
shifting cultivation, and the conversion to other land uses, for example, for graz-
ing, are regarded as the most serious threat to biological diversity on a global
scale. Estimates of the annual loss of tropical forest area are alarmingly high
(Chap. 10.8 in Whitmore 1998; FAO 2002). The extinction of tree species as a
result of forest destruction is poorly documented (Sayer and Whitmore 1990).
The causes of the destruction of tropical forests by humans are manifold,
complex, and closely related to the demographic and socioeconomic develop-
ment of human societies in tropical countries. Measures for the conservation
of biological diversity will be successful only if attention is paid to the socio-
economic situation of the local human population as the main target group for
240 C HAPTER 14 Conservation of Genetic Resources in Tropical Forest Trees
14.3
Defining Priorities
population breeding strategy combining breeding goals and the ex situ conser-
vation of genetic resources is particularly recommended for plantation tree
species of the tropics (Sect. 12.9).
The principal objective of a genetic management for the vast number of
tree species without immediate commercial importance is the preservation of
their genetic variation. This requires an improved understanding of the
genetic system of at least some keystone or indicator species. Most rare and
threatened species exist as small isolated populations. It is urgent that these
species be conserved. Viana et al. (1997) listed some nongenetic indicators of
vulnerable species that conservation efforts should be targeted at. Bierregaard
et al. (1997) extended this to a brief discussion of important aspects of frag-
mented tropical ecosystems and relevant research priorities. Species that were
once common approach this condition and become increasingly fragmented;
therefore, considerate handling of important features of the genetic system
and of sufficient population size (Sect. 9.3) is devised for any tree species
exposed to human interference (Young et al. 1999; A. Young et al. 2000). That
their populations are eventually harbored in small forest fragments might
help managers to adopt a holistic approach rather than limit their attention to
large forest tracts or simply the most pristine areas (p. 279 in Laurance and
Bierregaard 1997b).
14.4
Conservation Objectives
14.4.1
Objective 1: Preservation of the Potential for Particular Trait Expressions
14.4.2
Objective 2: Preservation of Maximum Variation
14.4.3
Objective 3: Preservation of Adaptability
14.5
Selection of Genetic Resources
The first step in any conservation project is to find out how the diversity we
wish to conserve is distributed in space (Bawa and Krugman 1990). We can
but preserve adequate samples of the total diversity and ecosystems, including
both centers of diversity and endemites. Since they run the risk of remaining
244 C HAPTER 14 Conservation of Genetic Resources in Tropical Forest Trees
14.5.1
Inventory of Genetic Marker Loci
Most efforts towards defining evolutionarily significant units have focused on genetic
distance measures based on markers to determine when populations are sufficiently dis-
tinct to warrant separate conservation. These methods, however, work poorly for many
tree species, as they are based on selectively neutral markers and thus reflect mutation,
genetic drift and migration, but not adaptive differences (Namkoong et al. 2000).
suitable tools for the detection of a geographic genetic pattern indicating the
selection response of tree populations; therefore, genetic variation of adaptive
significance may not be appropriately reflected at neutral marker loci. When
discussing the properties of alloenzyme loci, Finkeldey and Mtys (1999)
found that even though some of these gene loci under certain environmental
conditions possess relevance for fitness (Mller-Starck 1985; Ziehe et al. 1999;
Geburek 2000), they are still unsuitable for directly inferring the adaptive capac-
ity of populations and species. Reed and Frankham (2001) consider molecular
markers including alloenzyme loci unsuitable for inference on adaptive varia-
tion and differentiation (see also Navarro et al. 2005 for an example in Cedrela
odorata and Yang et al. 1996 for an example in a temperate pine species).
Two main objectives are evident with regard to the application of genetic
marker studies for the selection of genetic resources:
Genetic variation within populations and their differentiation has long been
observed most easily and reliably at alloenzyme loci. Their limited number is a
disadvantage for the selection of genetic resources. Many surveys have shown
remarkably low amounts of genetic differentiation among populations at these
loci (Sect. 3.4), although genetic differences in quantitative, adaptive traits is
suggested by the results of field trials (Chap. 11). These contrasting patterns of
genetic diversity are frequently explained by the presumed selective neutrality
of variation at all alloenzyme loci (Muona 1989).
Recently developed molecular tools allow the study of more types of genetic
markers (Sect. 2.3.1). Not all types of molecular markers are useful for the
selection of genetic resources (Szmidt and Wang 2000). Essential advantages
are access to large numbers of gene loci, the potential use of different markers
with different modes of inheritance (uniparental, biparental), and different
degrees of selective relevance. Approaches to relate variation at molecular
markers to adaptive variation patterns were mentioned in Sect. 7.2. In view of
the existing scarcity of available experimental data, an assessment of the gen-
eral potential of molecular genetics methods for the selection of forest genetic
resources would be premature.
A very large number of parameters have been developed that are suitable
for condensing the information gained from genetic inventories (Sect. 3.3).
14.5 Selection of Genetic Resources 247
14.5.2
Inventory of Adaptive Trait Expressions and Adaptive Markers
Phenotypic traits are expressions of genetic information at the level of the tran-
scribed genes. Phenotypic variation deliberately excludes variety at lower levels
of organization, such as within DNA base sequences where these differences
have no effect on phenotype.
The observation of phenotypic traits characterizing the health status, growth,
and reproduction of a population in a field experiment allows us to draw con-
clusions on its adaptedness to the given environment (Sect. 7.2). The adaptive
potential, i.e., the spectrum of environments that a population is able to
get adapted to, can be assessed in a series of field trials. In both situations a well-
performing population is usually considered adapted or adaptable, respectively,
although adaptation may be of a purely physiological nature. Strictly speaking,
inference on adaptedness in environments not included in the experimental
series cannot be made. Neither can adaptability to evolutionarily completely
novel environmental factors be foreseen. Nevertheless, populations performing
248 C HAPTER 14 Conservation of Genetic Resources in Tropical Forest Trees
well in experimental series are inferred to possess not only advantageous genetic
variants at relevant gene loci but possibly also enough variants that may locally
function as a genetic load but are advantageous under changed environmental
conditions.
Several options exist for assessing the genetic controlledness of phenotypic
traits (Sect. 12.3). Populations become prime targets for genetic conservation
measures if their superior average expressions of economically relevant traits
has a genetic basis. Traditional field trials in forestry are time-consuming.
Results from planned or recently established trials are not available for the
selection of genetic resources if the initiation of conservation measures is
urgent. Early tests in the greenhouse, climate chamber, laboratory, or nursery
are conducted to observe quantitative variation of seedlings. The conditions
can be modified as required. Environmental variation can be minimized or
even controlled. These experiments usually take only several months. It is usu-
ally impossible to identify the gene loci controlling a quantitative trait, their
mode of gene action, and their genetic structures; therefore, only very few
marker genes with relevance for survival or reproduction have so far been ana-
lyzed (Sherwin and Moritz 2000).
Some authors consider it possibly misleading to use data on neutral genetic
variation as the exclusive basis of decisions on selecting populations as gene
resources (Bekessy et al. 2003). The following case studies support this view:
In summary, it can be stated that genetic marker variation provides insight into
the genetic system, the Quarternary past of populations, and their centers of
variation. This indirect information of adaptability hardly possesses a causal
association with environmental variation but still makes a valuable contribu-
tion to the development of conservation strategies and might be even more
useful in combination with data on adaptive traits derived from experiments
done in judiciously chosen environments (Eriksson et al. 1993; Eriksson 1995;
Ennos 1996). Inventories of the variation at marker gene loci should then be
made if objective 2 deserves the highest consideration. Inventories of the vari-
ation in adaptive phenotypic traits should be made if objective 3 is to be pur-
sued (McKay and Latta 2002).
The integration of genetic and phenotypic information takes time and
effort; therefore, intensive studies have to be confined to species of prime eco-
logical or/and economic relevance. Results may be transferred to less impor-
tant species with similar life-history traits.
Immediate action for the protection of the genetic resources of a species
often becomes urgent in the absence of any data on spatial patterns of genetic
variation and without the possibility to conduct any of the experiments out-
lined previously. This situation applies to many lesser-known species in tropi-
cal forests. The conservation of at least one autochthonous population in each
of the main distribution areas of a species is recommended in this case.
Populations growing under both optimum and extreme environmental condi-
tions should be preserved. Both the geographic center of the species distribu-
tion and marginal populations should be represented in order to conserve
genetic information of populations that are possibly differentiated from each
other at adaptive loci.
Once a certain population has been found to be worthy of conservation, it
must have a sufficient size to be maintained over generations.
14.6
Conservation Methods
14.6.1
Dynamic Conservation in Situ
14.6.2
Dynamic Conservation ex Situ in Man-Made Forests
planning can be made, both the size and the design of genetic resources can be
chosen. It has to be kept in mind that the establishment of several small
resources is related to habitat fragmentation that might enhance the risk of los-
ing the resource (Wilcox and Murphy 1985); therefore, the issue is not splitting
up a given resource but rather establishing several of them in only partial isola-
tion. Long-term conservation focuses on persistence at spatial scales larger than
that of the single population (Thrall et al. 2000). As mentioned in Sect. 9.3,
a system of interconnected populations of moderate to large size has an optimal
chance of persistence. Persistence of a resource over generations must account
for other tree species providing a continuous food basis for pollen and seed dis-
persers, so that natural regeneration works and currently unoccupied habitat
fragments in the vicinity may be colonized. Thrall et al. (2000) reviewed both
the theory of metapopulations and its application to genetic conservation.
14.6.3
Conservation of Seeds in Gene Banks
Seeds can be transported easily and large numbers of genotypes can be stored
in a small facility. Their long-term storage in gene banks is the most important
static method for the conservation of crop plant genetic resources. The conser-
vation of seeds is static inasmuch as the genetic structures of the resource do
not change during storage.
The seed physiology of a tree species sets limits to the potential of long-term
storage of seeds in gene banks. Long-term storage of forest seeds as a genetic
resource is often combined with the establishment of forest seed centers in
tropical countries. The main task of seed centers is the procurement and pro-
vision of seeds for afforestation (Albrecht 1993). A concentration of tasks
related to forest tree seed handling and storage is meaningful in view of the
limited personal and financial resources available.
Long-term storage is impossible for recalcitrant seeds. The seed physiology
and optimum storage conditions of many tropical tree species are unknown.
In comparison with the lifetime of trees or their rotation periods, the maxi-
mum duration of seed storage is short even for species with orthodox seeds.
The seeds of annual agricultural crop plants are periodically sown out and the
seed of the new generation is put back into the gene bank. This cannot be
done for trees, so the germinability decays, which induces genetic change.
Since the main objective of a conservation programs for forest trees is the
preservation of evolutionary adaptability, static conservation methods are less
suitable than dynamic methods. Establishment and operation of a gene bank
are costly. Last but not least, gene banks are sensitive to short-term interrup-
tions of their operation, such as failure of the electricity network, which is
frequent in some areas.
256 C HAPTER 14 Conservation of Genetic Resources in Tropical Forest Trees
14.6.4
Vegetative Propagation for the Conservation of Forest Genetic Resources
14.6.5
Biotechnological Methods of Preservation
14.7
Regeneration of Forest Genetic Resources
the local population on the few remaining forests. Almost all pines in the gene
reserves are severely damaged by illegal resin tapping and/or the removal of
parts of the lower stem for the production of charcoal (Fig. 14.4). Ground fires
are frequent because the rural population wants to enhance the growth of
grasses for forest pasture.
Seed production of P. merkusii is poor in the gene reserves owing to overag-
ing, poor health, and low population density. Natural regeneration is almost
completely absent in the gene reserves. Young seedlings rarely survive owing to
yearly ground fires during the long dry season. In view of the absence of natu-
ral regeneration, seeds were harvested and sown in temporary nurseries. Three-
year-old seedlings of local origin were planted in the gene reserves. These
plants survive light ground fires (Fig. 14.3).
14.7 Regeneration of Forest Genetic Resources 261
Pinus kesiya
Pinus merkusii
Fig. 14.2. An approximately 3-year-old seedling of P. merkusii in the grass stage (front).
In the background is a seedling of P. kesiya, another pine native to Thailand. (Photo:
R. Finkeldey)
Fig. 14.3. A P. merkusii seedling after a ground fire. The long, fleshy needles are
burned, but the terminal bud survived and a new sprout has developed. (Photo:
R. Finkeldey)
Fig. 14.4. Heavily damaged P. merkusii tree in an in situ gene reserve in Thailand.
(Photo: R. Finkeldey)
outlined in Sect. 11.4.3. Seeds from the same region may be bulked, but the
reproductive material should not be exchanged among regions. The ex situ
conservation stands should be planted as pure forests in high density.
Outcrossing is expected to be promoted in the plantations owing to the high
population density and strong flowering of young trees after reaching maturity.
The double purpose of the plantations will be to conserve the genetic resources
of P. merkusii in Thailand and to serve as seed production areas for the
production of seeds eventually required for commercial plantings (From
Sa-Ardavut et al. 1989; Changtragoon and Finkeldey 1995a, 2000).
264 C HAPTER 14 Conservation of Genetic Resources in Tropical Forest Trees
14.8
Programs for the Protection of Forest Genetic Resources
National programs for the conservation of forest genetic resources are usually
planned and executed jointly by scientists and practical foresters trained in tree
breeding. Few researchers are experienced in the application of population
genetics and conservation-specific issues in tropical forests. Conservation of
forest genetic resources is often mistaken as subordinate elements of tree
improvement programs and the complex biological background is overlooked.
Apart from tree breeding, the conservation of genetic resources in tropical
forests is closely connected to the silvicultural management of forests (e.g., nat-
ural regeneration), nature conservation (conservation areas), legal regulations
(declaration of protected areas), political and socioeconomic issues (execution
of protection measures), and overall development planning. The biological and
genetic considerations outlined earlier must be incorporated into these subject
matters in order to initiate promising conservation programs. This applies in
particular to all planned activities for the in situ conservation of forest genetic
resources.
The local population is not only directly affected by many conservation
measures, but also by the main target group for the long-term beneficial effects
of a conservation program; thus, the rural populations must be integrated in
programs for the protection of forest genetic resources. Land is needed for
in situ and ex situ gene reserves and the utilization of forest products from con-
servation areas is restricted. The planning of conservation measures must take
into account legal and customary rights of forest owners and traditional users.
At an international level, several documents from the United Nations
Conference on Environment and Development (UNCED) during 1992 (the
Rio Summit) cover aspects of relevance for the conservation of tropical for-
est genetic resources. The Forest Principles, the Convention on Biological
Diversity, and Agenda 21 are of particular importance in addition to the 27
general principles of the Rio Declaration (Finkeldey 1996).
The importance of international scientific co-operation and financial aid for
the protection of natural forests in developing countries is stressed in the
Forest Principles and the Convention on Biological Diversity. They emphasize
the sovereign right of countries to utilize their genetic resources [see Article
8(g)]. The potential use of genetic resources by modern biotechnologies was a
controversial issue during the formulation of the Convention of Biological
Diversity. Technology transfer to developing countries is mentioned as being of
particular importance in this field.
Also Chap. 11 (Combating Deforestation), Chap. 15 (Conservation of
Biological Diversity), and Chap. 16 (Environmentally Sound Management of
14.9 Recommended Literature 265
14.9
Recommended Literature
Acher V, Faivre Rampant P, Pques LE, Prat D (2004) Chloroplast and mitochondrial
molecular tests identify European Japanese larch hybrids. Theor Appl Genet
108:16431649
Ackerly DD, Rankin-de-Merona JM, Rodriguez WA (1990) Tree densities and sex ratios
in breeding populations of dioecious Central Amazonian Myristicaceae. J Trop Ecol
6:239248
Ackerman JD, Mesler MR, Lu KL, Montalvo AM (1982) Food-foraging behaviour of male
Euglossini (Hymenoptera: Apidae): vagabonds or trapliners? Biotropica 14:241248
Adams WT (1992) Gene dispersal within forest tree populations. New For 6:217240
Adis J (1990) Thirty million arthropod species too many or too few? J Trop Ecol
6:115118
Albrecht J (1993) Forest seed handling. In: Pancel L (ed) Tropical forestry handbook,
vol 1. Springer, Berlin Heidelberg New York, pp 381462
Aldrich PR, Hamrick JL (1998) Reproductive dominance of pasture trees in a frag-
mented tropical forest mosaic. Science 281:103105
Aldrich PR, Hamrick JL, Chavarriaga P, Kochert G (1998) Microsatellite analysis of
demographic genetic structure in fragmented populations of the tropical tree
Symphonia globulifera. Mol Ecol 7:933944
Allard RW (1960) Principles of plant breeding. Wiley, New York
Altukhov YP (2006) Intraspecific genetic diversity. Monitoring, conservation, and
management. Springer, Berlin Heidelberg New York
Altukhov YP, Salmenkova EA (2002) DNA polymorphism in population genetics.
Russ J Genet 38:9891008
Alvarez-Buylla ER, Chaos , Piero D, Garay AA (1996) Demographic genetics of a
pioneer tropical tree species: patch dynamics, seed dispersal, and seed banks.
Evolution 50:11551166
Andersen M, Thornhill A, Koopowitz H (1997) Tropical forest disruption and stochas-
tic biodiversity losses. In: Laurance WF, Bierregaard RO (eds) Tropical forest rem-
nants: ecology, management, and conservation of fragmented communities.
University of Chicago Press, Chicago, pp 281291
Andrianarimisa A, Bachmann L, Ganzhorn JU, Goodman SM, Tomiuk J (2000) Effects
of forest fragmentation on genetic variation in endemic understory forest birds in
central Madagascar. J Ornithol 141:152159
Anonymous (2004a) Forest genetic resource conservation and management: vol 1:
overview, concepts and some systematic approaches. FAO-FGR-FLD-IPGRI, Rome
Anonymous (2004b) Forest genetic resource conservation and management: vol 2: in
natural forests and protected areas (in situ). FAO-FGR-FLD-IPGRI, Rome
268 References
Moritz C (1997) Key priorities for the study of fragmented tropical ecosystems. In:
Laurence WF, Bierregaard RO (eds) Tropical forest remnants: ecology, management,
and conservation of fragmented communities. University of Chicago Press, Chicago,
pp 515525
Billington HL (1991) Effect of population size on genetic variation in a dioecious
conifer. Conserv Biol 5:115119
Bishir J, Roberds JH (1999) On numbers of clones needed for managing risks in clonal
forestry. For Genet 6:149155
Blattner FR, Weising K, Bnfer G, Maschwitz U, Fiala B (2001) Molecular analysis of phy-
logenetic relationships among myrmecophytic Macaranga species (Euphorbiaceae).
Mol Phylogenet Evol 19:331344
Bolstad PV, Bawa KS (1982) Self incompatibility in Gmelina arborea L. (Verbenaceae).
Silvae Genet 31:1921
Bolstad SB, Kang H, Guries RP, Marty TL (1991) Performance of inter- and intraprove-
nance crosses of jack pine in central Wisconsin. Silvae Genet 48:261266
Booth TH, Turnbull JW (1994) Domestication of lesser-known tropical tree species:
the Australian experience. In: Leakey RRB, Newton AC (eds) Tropical trees: the
potential for domestication and the rebuilding of forest resources. Her Majestys
Stationery Office, London, pp 189194
Boshier D, Mesn JF (1989) Breeding population of Cordia alliodora in Costa Rica. In:
Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population struc-
ture and genetic improvement strategies in clonal and seedling forestry. Proceedings of
the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute,
Oxford, and Winrock International, Arlington, pp 406407
Boshier DH (2000) Mating systems. In: Young A, Boshier D, Boyle T (eds) Forest con-
servation genetics. Principles and practice. CSIRO, Collingwood, and CABI, Oxford,
pp 6379
Boshier DH, Chase MR, Bawa KS (1995) Population genetics of Cordia alliodora
(Boraginaceae), a neotropical tree. 2. Mating system. Am J Bot 82:476483
Bouvet JM, Vigneron P (1996) Variance structure in Eucalyptus hybrid populations.
Silvae Genet 45:171177
Boyle TJ (2000) Criteria and indicators for the conservation of genetic diversity. In:
Young A, Boshier D, Boyle T (eds) Forest conservation genetics. Principles and prac-
tice. CSIRO, Collingwood, and CABI, Oxford, pp 239251
Boyle TJB (1994) Introduction to the NATO advanced research workshop. In: Boyle
TJB, Boyle CEB (eds) Biodiversity, temperate forest ecosystems and global change.
Springer, Berlin Heidelberg New York, pp 17
Brewbaker JL, Sorensson CT (1994) Domestication of lesser-known species of the
genus Leucaena. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for
domestication and the rebuilding of forest resources, Her Majestys Stationery
Office, London, pp 196204
Brigham CA, Schwartz M (2003) Population viability in plants. Springer, Berlin
Heidelberg New York
Bronstein JL, McKey D (1989) The fig/pollinator mutualism: a model system for com-
parative biology. Experientia 45:601604
Bronstein JL, Gouyon P-H, Kjellberg F, Michaloud G (1990) The ecological conse-
quences of flowering asynchrony in monoecious figs: a simulation study. Ecology
71:21452156
References 271
Brown AHD, Hardner CM (2000) Sampling the gene pools of forest trees for ex situ
conservation. In: Young A, Boshier D, Boyle T (eds) Forest conservation genetics.
Principles and practice. CSIRO, Collingwood, and CABI, Oxford, pp 185196
Brown AHD (1978) Isozymes, plant population genetic structure and genetic conser-
vation. Theor Appl Genet 52:145157
Bruenig EF (1996) Conservation and management of tropical rainforests. CABI,
Wallingford
Brunner AM, Busov VB, Strauss SH (2004) Poplar genome sequence: functional
genomics in an ecologically dominant plant species. Trends Plant Sci 9:4956
Bryndum K, Hedegart T (1969) Pollination of teak (Tectona grandis L.). Silvae Genet
47:155158
Buchert GP, Rajora OP, Hood JV, Dancik BP (1997) Effects of harvesting on genetic diver-
sity in old-growth eastern white pine in Ontorio, Canada. Conserv Biol 11:747758
Burdon RD (1989) When is cloning on an operational scale appropriate? In: Gibson
GI, Griffin AR, Matheson, AC (eds), Breeding tropical trees: population structure
and genetic improvement strategies in clonal and seedling forestry. Proceedings of
the IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry
Institute, Oxford, and Winrock International, Arlington, pp 927
Burgess IP, Williams ER, Bell JC, Harwood CE, Owen JV (1996) The effect of out-
crossing rate on the growth of selected families of Eucalyptus grandis. Silvae Genet
45:97100
Burgman MA, Possingham H, Lynch AJJ, Keith DA, McCarthy MA, Hopper SD, Drury
WL, Passioura JA, Devries RJ (2001) A method for setting the size of plant conser-
vation target areas. Conserv Biol 15:603616
Burley J (1973) Sources and distribution of seedlots in the CFI international prove-
nance trial of Cedrela odorata (including C. mexicana and C. tubiflora). In: Burley J,
Nikles DG (eds) Tropical provenance and progeny research and international coop-
eration. Commonwealth Forestry Institute, Oxford, pp 234237
Burley J, Nikles DG (eds) (1972) Selection and breeding to improve some tropical conifers,
vol 1. Commonwealth Forestry Institute, Oxford, and Department of Forestry, Brisbane
Burley J, Nikles DG (eds) (1973) Selection and breeding to improve some tropical
conifers, vol 2. Commonwealth Forestry Institute, Oxford, and Department of
Forestry, Brisbane
Butcher PA (2004) Genetic mapping in acacias. In: Kumar S, Fladung M (eds)
Molecular genetics and breeding of forest trees. Haworth, New York, pp 411427
Butcher PA, Moran GF (2000) Genetic linkage mapping in Acacia mangium.
2. Development of an integrated map from two outbred pedigrees using RFLP and
microsatellite loci. Theor Appl Genet 101:594605
Butcher PA, Moran GF, Perkins HD (1998) RFLP diversity in the nuclear genome of
Acacia mangium. Heredity 81:205213
Butcher PA, Glaubitz JC, Moran GF (1999) Applications for microsatellite markerks in
the domestication and conservation of forest trees. For Genet Resour 27:3442
Butcher PA, Moran GF, Bell R (2000) Genetic linkage mapping in Acacia mangium.
1. Evaluation of restriction endonucleases, inheritance of RFLP loci and their
conservation across species. Theor Appl Genet 100:576583
Byrne M (2000) Disease threats and the conservation genetics of forest trees. In: Young A,
Boshier D, Boyle T (eds) Forest conservation genetics. Principles and practice. CSIRO,
Collingwood, and CABI, Oxford, pp 159166
272 References
Guries RP (1990) Forest genetics and forest tree breeding. In: Young RA, Giese RL (eds)
Introduction to forest science. 2nd edn. Wiley, New York, pp 99118
Gyi MKK (1993) Teak in Myanmar. In: Wood H (ed) Teak in Asia. FORSPA publica-
tion no 4. Technical document GCP/RAS/134/ASB. FAO, Bangkok, pp 5162
Haffer J, Prance GT (2001) Climatic forcing of evolution in Amazonia during the
Cenozoic: On the refuge theory of biotic differentiation. Amazoniana 16:579605
Hall BG (2001) Phylogenetic trees made easy: a how- to manual for molecular biolo-
gists. Sinauer, Sunderland
Hall P, Chase MR, Bawa KS (1994a) Low genetic variation but high population differ-
entiation in a common tropical forest tree species. Conserv Biol 8:471482
Hall P, Orrel LC, Bawa KS (1994b) Genetic diversity and mating system in a tropical
tree, Carapa guianensis (Melicaceae). Am J Bot 81:11041111
Hall P, Walker S, Bawa K (1996) Effect of forest fragmentation on genetic diversity and
mating system in a tropical tree, Pithecellobium elegans. Conserv Biol 10:757768
Hamilton MB (1999) Tropical tree gene flow and seed dispersal. Nature 401:129
Hamrick JL, Godt MJW (1989) Allozyme diversity in plant species. In: Brown AHD,
Clegg MT, Kahler AL, Weir BS (eds) Plant population genetics, breeding, and genetic
resources. Sinauer, Sunderland, pp 4363
Hamrick JL, Godt MJW (1996) Effects of life history traits on genetic diversity in plant
species. Philos Trans R Soc Lond Ser B 351:12911298
Hamrick JL, Godt MJW (1997) Allozyme diversity in cultivated crops. Crop Sci 37:2630
Hamrick JL, Murawski DA (1991) Levels of allozyme diversity in populations of
uncommon Neotropical tree species. J Trop Ecol 7:395399
Hamrick JL, Godt MJW, Sherman-Broyles SL (1992) Factors influencing levels of
genetic diversity in woody plant species. New For 6:95124
Hanski I, Gilpin M (1991) Metapopulation dynamics: brief history and conceptual
domain. Biol J Linn Soc Lond 42:316
Hart TB (1990) Monospecific dominance in tropical rain forests. Trends Ecol Evol
5:611
Hart TB, Hart JA, Murphy PG (1989) Monodominant and species-rich forests of the
humid tropics: causes for their co-occurrence. Am Nat 133:613633
Hartmann HT, Kester DE (1983) Plant propagation. Principles and practices. Prentice
Hall, Englewood Cliffs
Hattemer HH (1969) Versuche zur geographischen Variation bei der japanischen
Lrche. Silvae Genet 18:123
Hattemer HH (1987) Are the EEC directives on forest reproductive material genetically
adequate? Silvae Genet 36:94102
Hattemer HH (1991) Genetic analysis and population genetics. In: Fineschi S, Malvolti
ME, Cannata F, Hattemer HH (eds) Biochemical markers in the population genetics
of forest trees. SPB, The Hague, pp 522
Hattemer HH (1995) Concepts and requirements in the conservation of forest genetic
resources. For Genet 2:125134
Hattemer HH (1997) Gene conservation in tropical forests. Nat Resour Dev 45/46:4253
Hattemer HH (1999) Wald aus der Retorte? G Aug Nachr Univ Gtt 2128
Hattemer HH (2006) On the appropriate size of forest genetic resources. In: Geburek T,
Turok J (eds), Conservation and management of forest genetic resources in Europe.
Arbora, Zvolen, pp 413434
282 References
Hattemer HH, Gregorius H-R (1996) Bedeutung der biologischen Vielfalt fr die
Stabilitt von Waldkosystemen. In: Mller-Starck G (ed) Biodiversitt und nach-
haltige Forstwirtschaft. Ecomed, Landsberg, pp 110
Hattemer HH, Melchior GH (1993) Genetics and its application to tropical forestry. In:
Pancel L (ed) Tropical forestry handbook. Springer, Berlin Heidelberg New York,
pp 333380
Hattemer HH, Mller-Starck G (1990) Evolution von Baumpopulationen als Folge des
Waldsterbens? In: Hattemer HH (ed) Erhaltung forstlicher Genressourcen.
Sauerlnder, Frankfurt, pp 3445
Hattemer HH, Ziehe M (1997) Genetic control of phenotypic traits with relevance to
gene conservation in trees a survey of methods. In: Mtys C (ed) Perspectives of
forest genetics and tree breeding in a changing world. IUFRO world series, vol 6.
IUFRO, Sopron, pp 135148
Hattemer HH, Gregorius H-R, Ziehe M, Mller-Starck G (1982) Klonanzahl forstlicher
Samenplantagen und genetische Vielfalt. Allg Forst-Jagdz 153:183191
Hattemer HH, Ziehe M, Bergmann F (1993) Einfhrung in die Genetik fr
Studierende der Forstwissenschaft. Sauerlnder, Frankfurt
Hedegart T (1976) Breeding system, variation and genetic improvement of teak
(Tectona grandis L.f.). In: Burley J, Styles BT (eds) Tropical trees; variation, breeding
and conservation. Academic, London, pp 109121
Hedegart T, Lauridsen EB, Keiding H (1975) Broadleaved seed orchards. Part D.
Teak. In: Faulkner R (ed) Seed orchards. Her Majestys Stationery Office, London,
pp 139142
Hedrick PW (2000) Genetics of populations. Jones and Barlett, Sudbury
Heithaus ER, Opler PA, Baker HG (1974) Bat activity and pollination of Bauhinia
pauletia: plant-pollinator coevolution. Ecology 55:412419
Herre EA (1996) An overview of studies on a community of Panamanian figs.
J Biogeogr 23:593607
Herre EA, Machado CA, Bermingham E, Nason JD, Windsor DM, McCaferty SS, Van
Houten W, Bachmann K (1996) Molecular phylogenies of figs and their pollinator
wasps. J Biogeogr 23:521530
Hewitt G (2000) The genetic legacy of the Quaternary ice ages. Nature 405:907913
Hewitt GM (2001) Speciation, hybrid zones and phylogeography or seeing genes in
space and time. Mol Ecol 10:537549
Hill DS (1967a) Figs (Ficus spp.) of Hong Kong. Hong Kong University Press, Hong Kong
Hill DS (1967b) Figs (Ficus spp.) and fig wasps (Chalcidoidea). J Nat Hist 1:413-434
Hill JL, Hill RA (2001) Why are tropical rain forests so species rich? Classifying, review-
ing and evaluating theories. Prog Phys Geogr 25:326354
Hodgson LM (1976a) Some aspects of flowering and reproductive behaviour in
Eucalyptus grandis (Hill) Maiden at J.D.M. Keet Forest Research Station. 2. The fruit,
seed, seedlings, self fertility, selfing and inbreeding effects. S Afr For J 98:3243
Hodgson LM (1976b) Some aspects of flowering and reproductive behaviour in
Eucalyptus grandis (Hill) Maiden at J.D.M. Keet Forest Research Station. 1. Flowering,
controlled pollination method, pollination, receptivity. S Afr For J 97:1828
Hodgson LM (1977) Methods of seed orchard management for seed production and
ease of reaping in Eucalyptus grandis. S Afr For J 100:3842
References 283
Hollingsworth PM, Dawson IK, Goodall-Copestake P, Richardson JE, Weber JC, Sotelo
Montes C, Pennington RT (2005) Do farmers reduce genetic diversity when they
domesticate tropical trees? A case study from Amazonia. Mol Ecol 14:497501
Holsinger KE (2000) Demography and extinction in small populations. In: Young AG,
Clarke GM (eds) Genetics, demography and viability of fragmented populations.
Cambridge University Press, Cambridge, pp 5574
Honadle G (1993) Institutional constraints on sustainable use: lessons from the tropic
showing that resource overexploitation is not just an attitude problem and conser-
vation education is not enough. In: Aplet GH, Johnson N, Olson JT, Sample VA
(eds), Defining sustainable forestry. Island, Washington , pp 90126
Hong SO (1975) Vegetative propagation of plant material for seed orchards with spe-
cial reference to graft-incompatibility problems. In: Faulkner R (ed) Seed orchards.
Her Majestys Stationery Office, London, pp 3848
Hosius B (1993) Wird die genetische Struktur eines Fichtenbestandes von
Durchforstungseingriffen beeinflut? Forst Holz 48: 306308
Howe HF, Smallwood J (1982) Ecology of seed dispersal. Annu Rev Ecol Syst 13: 201228
Hubbell SP, Foster RB (1992) Short-term dynamics of a neotropical forest: why eco-
logical research matters to tropical conservation and management. Oikos 63:4861
Hufford KM, Hamrick JL (2003) Viability selection at three early life stages of the trop-
ical tree Platypodium elegans (Fabaceae, Papilionoideae). Evolution 57:518526
Hughes CE (1989) New opportunities in Leucaena genetic improvement. In: Gibson
GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: population structure and
genetic improvement strategies in clonal and seedling forestry. Proceedings of the
IUFRO conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute,
Oxford, and Winrock International, Arlington, pp 218226
Hhn M (1986) Theoretical studies on the necessary number of components in mixtures.
3. Number of components and risk considerations. Theor Appl Genet 72:211218
Huxley J (1938) Clines: an auxiliary taxonomic principle. Nature 143:219220
Indrioko S, Gailing O, Finkeldey R (2006) Molecular phylogeny of Dipterocarpaceae in
Indonesia based on chloroplast DNA. Plant Syst Evol (in press). DOI
10.1007/s00606-006-0435-8
Ishiyama H, Kado T, Iwasaki M, Matsuoka M, Shukor N, Szmidt A, Yamazaki T (2003)
Nucleotide variation in the GapC region of four species of Shorea and their putative
hybrids. Tropics 13:8999
Iwata H, Konuma A, Tsumura Y (2000) Development of microsatellite markers in the
tropical tree Neobalanocarpus heimii (Dipterocarpaceae). Mol Ecol 9:16611686
James T, Vege S, Aldrich P, Hamrick JL (1998) Mating systems of three tropical dry for-
est tree species. Biotropica 30:587594
Janson CH, Terborgh J, Emmons LH (1981) Non-flying mammals as pollinating agents
in the Amazonian forest. Biotropica 13(Suppl):16
Janzen DH (1970) Herbivores and numbers of tree species in tropical forests. Am Nat
104:501528
Janzen DH (1976) Why bamboos wait so long to flower. Annu Rev Ecol Syst 7:347391
Janzen DH (1979a) How to be a fig. Annu Rev Ecol Syst 10:1351
Janzen DH (1979b) How many parents do the wasps from a fig have? Biotropica
11:127129
284 References
Laurance WF, Bierregaard RO (1997a) A crisis in the making. In: Laurance WF,
Bierregaard RO (eds) Tropical forest remnants: ecology, management, and conser-
vation of fragmented communities. University of Chicago Press, Chicago, pp XIXV
Laurance WF, Bierregaard RO (eds) (1997b) Tropical forest remnants: ecology, manage-
ment, and conservation of fragmented communities. University of Chicago Press,
Chicago
Laurance WF, Lovejoy TE, Vasconcelos HL, Bruna EM, Didham RK, Stouffer PC,
Gascon C, Bierregaard RO, Laurance SG, Sampaio E (2002) Ecosystem decay of
Amazonian forest fragments: a 22-year investigation. Conserv Biol 16:605618
Law BS, Lean M (1999) Common blossom bats (Syconycteris australis) as pollinators in
fragmented Australian tropical rainforest. Biol Conserv 91:201212
Lawson GJ (1994) Indigenous trees in West African forest plantation: the need for
domestication by clonal techniques. In: Leakey RRB, Newton AC (eds) Tropical
trees: the potential for domestication and the rebuilding of forest resources. Her
Majestys Stationery Office, London, pp 112123
Leakey RRB, Newton AC (eds) (1994a) Tropical trees: the potential for domestication
and the rebuilding of forest resources. Her Majestys Stationery Office, London
Leakey RRB, Newton AC (1994b) Domestication of Cinderella species as the start of a
woody-plant revolution. In: Leakey RRB, Newton AC (eds) Tropical trees: the poten-
tial for domestication and the rebuilding of forest resources. Her Majestys
Stationery Office, London, pp 36
Leakey RRB, Last FT, Longman KA (1982) Domestication of tropical trees: an approach
securing future productivity and diversity in managed ecosystems. Commonw For Rev
61:3342
Leakey RRB, Newton AC, McDick J (1994) Capture of genetic variation by vegetative
propagation: processes determining success. In: Leakey RRB, Newton AC (eds)
Tropical trees: the potential for domestication and the rebuilding of forest resources.
Her Majestys Stationery Office, London, pp 7283
Le Corre V, Kremer A (2003) Genetic variability at neutral markers, quantitative trait
loci and trait in a subdivided population under selection. Genetics 164:12051219
Ledig FT (1973) An analysis of methods for the selection of trees from wild stands. For
Sci 20:216
Ledig FT (1986a) Conservation strategies for forest gene resources. For Ecol Manage
14:7790
Ledig FT (1986b) Heterozygosity, heterosis, and fitness in outbreeding plants. In: Soul
ME (ed) Conservation biology. The science of scarcity and diversity. Sinauer,
Sunderland, pp 77104
Ledig FT (1992) Human impacts on genetic diversity in forest ecosystems. Oikos
63:87108
Ledig FT, Kitzmiller JH (1992) Genetic strategies for reforestration in the face of global
change. For Ecol Manage 50:153169
Lee SL (2000) Mating system parameters of Dryobalanops aromatica Gaertn.
f. (Dipterocarpaceae) in three different forest types and a seed orchard. Heredity
85:338345
Lee SL, Wickneswari R, Mahani MC, Zakri AH (2000) Genetic diversity of a tropical
tree species, Shorea leprosula Miq. (Dipterocarpaceae), in Malaysia: implications for
conservation of genetic resources and tree improvement. Biotropica 32:213224
288 References
Lee SS (ed) (1995) A guide book to Pasoh. FRIM technical information handbook no 3.
Forest Research Institute Malaysia, Kepong
Lefvre W (1984) Die Entstehung der biologischen Evolutionstheorie. Ullstein,
Frankfurt
Leighton M, Leighton DR (1983) Vertebrate responses to fruiting seasonality within a
Bornean rain forest. In: Sutton SL, Whitmore TC, Chadwick AC (eds) Tropical rain
forest: ecology and management. Blackwell, Oxford, pp 181196
Leinemann L, Hattemer HH (2006) Genetic variation and mating pattern in a stand of
yew (Taxus baccata L.). Allg Forst-Jagdz (in press)
Lemes MR, Gribel R, Proctor J, Grattapaglia D (2003) Population genetic structure of
mahogany (Swietenia macrophylla King, Meliaceae) across the Brazilian Amazon,
based on variation at microsatellite loci: implications for conservation. Mol Ecol
12:28752883
Lesica P, Allendorf FW (1995) When are peripheral populations valuable for conserva-
tion? Conserv Biol 9:753760
Lessa EP, Cook JA, Patton JL (2003) Genetic footprints of demographic expansion in
North America, but not Amazonia, during the late Quaternary. Proc Natl Acad Sci
USA 100:1033110334
Levin DA, Kerster HW (1974) Gene flow in seed plants. Evol Biol 7:139220
Lewontin RC (1984) Detecting population differences in quantitative character as
opposed to gene frequencies. Am Nat 123:115124
Li Q, Xu Z, He T (2002) Ex situ genetic conservation of endangered Vatica guangxien-
sis (Dipterocarpaceae) in China. Biol Conserv 106:151156
Libby WJ (1973) Domestication strategies for forest trees. Can J For Res 3: 265276
Libby WJ (1982) What is a safe number of clones per plantation? In: Heybroek HM,
Stephan BR, von Weissenberg K (eds) Resistance to diseases and pests in forest trees.
Pudoc, Wageningen, pp 342360
Liengsiri C, Yeh FC, Boyle TJB (1995) Isozyme analysis of a tropical forest tree,
Pterocarpus macrocarpus Kurz. in Thailand. For Ecol Manage 74:1322
Liengsiri C, Boyle TJB, Yeh FC (1998) Mating system in Pterocarpus indicus Kurz in
Thailand. J Hered 89:216221
Lindgren D, Lindgren K (1996) Long distance pollen transfer may make gene conser-
vation difficult. In: Kurm M, Tamm Y (eds) Conservation of forest genetic resources.
Nordic group of forest genetics and tree breeding meeting, Estonia, June 1996.
Estonian Agricultural University, Tartu, pp 5162
Lindgren D, Paule L, Shen X, Yazdani R, Segerstrm U, Wallin J-E, Lejdebro ML (1995)
Can viable pollen carry Scots pine genes over long distances? Grana 34:6469
Lindquist B (1946) Den skogliga rasforskningen och praktiken. Svenska Skogsvrds-
freningen, Stockholm
Lira CF, Cardoso SRS, Ferreira PCG, Cardoso MA, Provan J (2003) Long-term popu-
lation isolation in the endangered tropical tree species Caesalpinia echinata Lam.
revealed by chloroplast microsatellites. Mol Ecol 12:32193225
Listabarth C (1996) Pollination of Bactris by Phyllotrox and Epurea. Implications of the
palm breeding beetles on pollination at the community level. Biotropica 28:6981
Lobo JA, Quesada M, Stoner KE (2005) Effects of pollination by bats on the mating sys-
tem of Ceiba pentandra (Bombacaceae) populations in two tropical life zones in
Costa Rica. Am J Bot 92:370376
References 289
Loiselle BA, Ribbens E, Vargas O (1996) Spatial and temporal variation of seed rain in
a tropical lowland wet forest. Biotropica 28:8295
Longman KA (1976) Conservation and utilization of gene resources by vegetative mul-
tiplication of tropical trees. In: Burley J, Styles BT (eds) Tropical trees; variation,
breeding and conservation. Academic, London, pp 1924
Loveless MD (1992) Isozyme variation in tropical trees: patterns of genetic organiza-
tion. New For 6:6794
Lowenfeld R, Klekowski EJ Jr (1992) Mangrove genetics. I. Mating system and muta-
tion rates of Rhizophora mangle in Florida and San Salvador Island, Bahamas. Int J
Plant Sci 153:394399
Loyn RH, MacAlpine C (2001) Spatial patterns and fragmentation: indicators for conserv-
ing biodiversity in forest landscapes. In: Raison RJ, Brown AG, Flinn DW (eds) Criteria
and indicators for sustainable forest management. CABI, Wallington, pp 391422
Lu Z, Johnson WE, Menotti-Raymond M, Yuhki N, Martenson JS, Mainka S, Shi-Qiang
H, Zhihe Z, Li G, Pan W, Mao X, OBrien SJ (2001) Patterns of genetic diversity in
remaining giant panda populations. Conserv Biol 15:15961607
Lugo AE (1988) Estimating reductions in the diversity of tropical forest species. In: Wilson
EO, Peter FM (eds) Biodiversity. National Academy Press, Washington, pp 5870
Luu HT (2005) Genetic variation and the reproductive system of Dipterocarpus cf. con-
dorensis Pierre in Vietnam. Cuvillier, Gttingen
Machado CA, Herre EA, McCaferty SS, Bermingham E (1996) Molecular phylogenies
of fig pollinating and non-pollinating wasps and the implications for the origin and
evolution of the fig-fig wasp mutualism. J Biogeogr 23:531542
Mandal AK, Gibson GL (ed) (1998). Forest genetics and tree breeding. CBS, New Delhi
Maghembe JA, Kwesiga F, Ngulube M, Prins H, Malaya FM (1994) Domestication
potential of indigenous fruit trees of the miombo woodlands of southern Africa. In:
Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the
rebuilding of forest resources. Her Majestys Stationery Office, London, pp 220229
Malcolm JR, Ray JC (2000) Influence of timber extraction routes on Central African
small-mammal communities, forest structure, and tree diversity. Conserv Biol
14:16231638
Manel S, Schwartz MK, Luikart G, Taberlet P (2003) Landscape genetics: combining
landscape ecology and population genetics. Trends Ecol Evol 18:189197
Margraf J, Milan PP (2004) Rainforestation farming. A farmers guide to biodiversity
management for the Philippines. Institute of Tropical Ecology, Leyte State University,
Baybay
Marsh EK (1970) Selecting adapted races of introduced species. In: Proceedings of
the 2nd world conference on forest genetics and forest tree breeding. FAO, Rome,
pp 12491265
Marshall DR, Brown AHD (1975) Optimum sampling strategies in genetic conserva-
tion. In: Frankel OH, Hawkes JG (eds) Crop genetic resources for today and
tomorrow. Cambridge University Press, Cambridge, pp 5380
Mtys C, Yeatman CW (1992) Effect of geographical transfer on growth and survival
of jack pine (Pinus banksiana Lamb.) populations. Silvae Genet 41:370375
Maynard Smith J (1998) Evolutionary genetics. Oxford University Press, Oxford
McCauley DE (1991) Genetic consequences of local population extinction and recolo-
nization. Trends Ecol Evol 6:58
290 References
Moncur MV, Mitchell A, Fripp Y, Kleinschmidt GJ (1995) The role of honeybees (Apis mel-
lifica) in eucalypt and acacia seed-production areas. Commonw For Rev 64:350-354
Montagnini F, Jordan CF (2005) Tropical forest ecology. The basis for conservation and
management. Springer, Berlin Heidelberg New York
Moran GF, Brown AHD (1980) Temporal heterogeneity of outcrossing rates in Alpine
ash (Eucalyptus delegatensis R.T. Bak.). Theor Appl Genet 57:101105
Moran GF, Bell JC (1983) Eucalyptus. In: Tanksley SD, Orton TJ (eds) Isozymes in plant
genetics and breeding, part B. Elsevier, Amsterdam, pp 423441
Moran GF, Hopper SD (1983) Genetic diversity and the insular population structure of
the rare granite rock species, Eucalyptus caesia Benth. Aust J Bot 31:161172
Moran GF, Muona O, Bell JC (1989a) Acacia mangium: a tropical forest tree of the
coastal lowlands with low genetic diversity. Evolution 43:231235
Moran GF, Bell JC, Griffin AR (1989b) Reduction in levels of inbreeding in a seed
orchard of Eucalyptus regnans F. Muell. compared with natural populations. Silvae
Genet 38:3236
Morgenstern EK (1996) Geographic variation in forest trees. UBC Press, Vancouver
Morin PA, Luikart G, Wayne RK, SNP workshop group (2004) SNPs in ecology, evolu-
tion and conservation. Trends Ecol Evol 19:208216
Moritz A, Ldders P (1993) Pollenkeimung, Pollenschlauchwachstum und Befruch-
tungsverhalten verschiedener Klone der Paranu (Bertholetia excelsa Humb. &
Bonpl.). Angew Bot 67:107112
Mottura MC, Finkeldey R, Verga AR, Gailing O (2005) Development and characteriza-
tion of microsatellite markers for Prosopis chilensis and Prosopis flexuosa and cross-
species amplification. Mol Ecol Notes 5:487489
Mller G (1976) A simple method of estimating rates of self-fertilization by analysing
isozymes in tree seeds. Silvae Genet 25:1517
Mller G (1977) Cross-fertilization in a conifer stand inferred from enzyme gene-
markers in seeds. Silvae Genet 26:223226
Mller-Starck G (1982) Reproductive systems in conifer seed orchards. I. Mating prob-
abilities in a seed orchard of Pinus sylvestris L. Silvae Genet 31:188197
Mller-Starck G (1985) Genetic differences between tolerant and sensitive beeches
(Fagus sylvatica L.) in an environmentally stressed adult forest stand. Silvae Genet
34:241247
Mller-Starck G (1991) Genetic processes in seed orchards. In: Giertych M, Mtys C
(eds) Genetics of Scots pine. Akadmiai Kiad, Budapest, pp 147162
Mller-Starck G (1996) Beitrge der Forstgenetik zur nachhaltigen Waldbewirtschaf-
tung. In: Mller-Starck G (ed) Biodiversitt und Nachhaltige Forstwirtschaft.
Ecomed, Landsberg, pp 259283
Mller-Starck G, Ziehe M (1984) Reproductive systems in conifer seed orchards.
3. Female and male fitnesses of individual clones realized in seeds of Pinus sylvestris L.
Theor Appl Genet 69:173177
Muluvi GM, Sprent JI, Soranzo N, Provan J, Odee D, Folkard G, McNicol JW, Powell W
(1999) Amplified fragment length polymorphism (AFLP) analysis of genetic varia-
tion in Moringa oleifera Lam. Mol Ecol 8:463470
Muona O (1989) Population genetics in forest tree improvement. In: Brown AHD,
Clegg MT, Kahler AL, Weir BS (eds) Plant population genetics, breeding, and genetic
resources. Sinauer, Sunderland, pp 282298
292 References
Muona O, Harju A (1989) Effective population size, genetic variability, and mating sys-
tem in natural stands and seed orchards of Pinus sylvestris. Silvae Genet 38:221228
Murawski DA, Bawa KS (1994) Genetic structure and mating system of Stemonoporus
oblongifolius (Dipterocarpaceae) in Sri Lanka. Am J Bot 81:155160
Murawski DA, Hamrick JL (1991) The effect of the density of flowering individuals on
the mating systems of nine tropical tree species. Heredity 67:167174
Murawski DA, Hamrick JL (1992) The mating system of Cavanillesia platanifolia under
extremes of flowering-tree density: a test of predictions. Biotropica 24:99101
Murawski DA, Dayanandan B, Bawa KS (1994a) Outcrossing rates of two endemic
Shorea species from Sri Lankan tropical rain forests. Biotropica 26:2329
Murawski DA, Gunatilleke IAUN, Bawa KS (1994b) The effects of selective logging on
inbreeding in Shorea megistophylla (Dipterocarpaceae) from Sri Lanka. Conserv Biol
8: 9971002
Murillo Gamboa O (1997) Genetische Untersuchungen an natrlichen Populationen
von Alnus acuminata spp arguta (Schlectendal) Furlow in Costa Rica und Panam.
Cuvillier, Gttingen
Myers N (1986) Tropical deforestation and a mega-extinction spasm. In: Soul ME (ed)
Conservation biology. The science of scarcity and diversity. Sinauer, Sunderland,
pp 394409
Nagamitsu T, Inoue T (1997) Cockroach pollination and breeding system of Uvaria
elmeri (Annonaceae) in a lowland mixed-dipterocarp forest in Sarawak. Am J Bot
84:208213
Namkoong G (1969) Non-optimality of the local races. In: Proceedings of the 10th
southern conference on forest tree improvement, Raleigh, Southern Forest Tree
Improvement Committee, pp 149153
Namkoong G (1981) Introduction to quantitative genetics in forestry. Castle House,
Tunbridge Wells
Namkoong G (1984) A control concept of gene conservation. Silvae Genet 33:160163
Namkoong G (1985) The population genetic basis of breeding theory. In: Gregorius
H R (ed) Population genetics in forestry. Springer, Berlin Heidelberg New York, pp 215
Namkoong G (1986) Genetics and the forests of the future. Unasylva 38:218
Namkoong G (1989) Systems of gene management. In: Gibson GL, Griffin AR,
Matheson AC (eds) Breeding tropical trees: population structure and genetic
improvement strategies in clonal and seedling forestry. Proceedings of the IUFRO
conference, Pattaya, Thailand, November 1988. Oxford Forestry Institute, Oxford,
and Winrock International, Arlington, pp 18
Namkoong G (1991) Maintaining genetic diversity in breeding for resistance in forest
trees. Annu Rev Phytopathol 29: 325342
Namkoong G (1999) Plantations vs. naturally regenerated forests: risks and stability in
using bred, cloned, or biotech products. In: Mtys C (ed) Forest genetics and sus-
tainability. Kluwer, Dordrecht, pp 147154
Namkoong G, Barnes RD, Burley J (1980) A philosophy of breeding strategy for trop-
ical forest trees. Commonwealth Forestry Institute, Oxford
Namkoong G, Kang HC, Brouard JS (1988) Tree breeding: principles and strategies.
Springer, Berlin Heidelberg New York
Namkoong G, Koshy MP, Aitken S (2000) Selection. In: Young A, Boshier D, Boyle T
(eds) Forest conservation genetics. Principles and practice. CSIRO, Collingwood,
and CABI, Oxford,, pp 101111
References 293
Poke FS, Vaillancourt RE, Potts BM, Reid JB (2005) Genomic research in Eucalyptus.
Genetica 125:79101
Prabhu R, Ruitenbeek HJ, Boyle TJB, Colfer CJP (2001) Between voodoo science and
adaptive management: the role and research needs for indicators of sustainable for-
est management. In: Raison RJ, Brown AG, Flinn DW (eds) Criteria and indicators
of sustainable forest management. CABI, Wallington, pp 3966
Prance GT (1985) The pollination of Amazonian plants. In: Prance GT, Lovejoy TE
(eds) Key environments. Amazonia. Pergamon, Oxford, pp 166191
Prance GT (1994) The resource of usful tree species: identification of priorities for
domestication. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential for
domestication and the rebuilding of forest resources. Her Majestys Stationery
Office, London, pp 715
Primack R, Corlett R (2005) Tropical rain forests. Blackwell, Oxford
Prober SM, Brown AHD (1994) Conservation of grassy white box woodlands: popula-
tion genetics and fragmentation of Eucalyptus albens. Conserv Biol 8:10031013
Proctor M, Yeo P, Lack A (1996) The natural history of pollination. HarperCollins,
London
Pryor LD, Byrne GR (1969) Variation and taxonomy in Eucalyptus camaldulensis. Silvae
Genet 18:6471
Putz FE, Romero C (2001) Biologists and timber certification. Conserv Biol 15:313314
Putz FE, Blate GM, Redford KH, Fimble R, Robinson J (2001) Tropical forest manage-
ment and conservation of biodiversity: an overview. Conserv Biol 15:720
Quaile DR, Mullin LJ (1984) Provenance and progeny testing in Eucalyptus camaldu-
lensis in Zimbabwe. In: Barnes RD, Gibson GL (eds) Provenance and genetic
improvement strategies in tropical forest genetics. Commonwealth Forestry
Institute, Oxford, and Forest Research Centre, Harare, pp 438450
Rabouam C, Comes AM, Bretagnolle V, Humbert J-F, Periquets G, Bigot Y (1999)
Features of DNA fragments obtained by random amplified polymorphic DNA
(RAPD) assays. Mol Ecol 8:493503
Raison RJ, Flinn DW, Brown AC (2001) Application of criteria and indicators to
support sustainable forest management: some key issues. In: Raison RJ, Brown AG,
Flinn DW (eds) Criteria and indicators for sustainable forest management. CABI,
Wallington, pp 518
Ramirez N, Sobrevila C, de Enrech NX, Ruiz-Zapata T (1984) Floral biology and breed-
ing system of Bauhinia benthamiana Taub. (Leguminosae), a bat-pollinated tree in
Venezuelan Llanos. Am J Bot 71:273280
Rautner M (2001) Designer trees. Biotechnol Dev Moni 44:27
Raven PH (1988) Our diminishing tropical forests. In: Wilson EO, Peter FM (eds)
Biodiversity. National Academy Press, Washington, pp 119122
Raymond CA, Lindgren D (1990) Genetic flexibility a model for determining the
range of suitable environments for a seed source. Silvae Genet 39:112120
Reddy KV, Rockwood DL (1989) Breeding strategies for coppice production in a
Eucalyptus grandis base population with four generations of selection. Silvae Genet
38: 148151
Reed DH, Frankham R (2001) How closely correlated are molecular and quantitative
measures of genetic variation? A meta-analysis. Evolution 55:10951101
Reed DH, Frankham R (2003) Correlation between fitness and genetic diversity.
Conserv Biol 17:230237
References 297
Regal PJ (1982) Pollination by wind and animals: ecology of geographic patterns. Annu
Rev Ecol Syst 13:497524
Remington DL, Purugganan MD (2003) Candidate genes, quantitative trait loci, and
functional trait evolution in plants. Int J Plant Sci 164:720
Richards C (2000) Genetic and demographic influences on population persistence:
gene flow and genetic rescue in Silene alba. In: Young AG, Clarke GM (eds) Genetics,
demography and viability of populations. Cambridge University Press, Cambridge,
pp 271291
Richards PW (1981) The tropical rain forest. Cambridge University Press, Cambridge
Richardson JE, Pennington RT, Pennington TD, Hollingsworth PM (2001) Rapid diversi-
fication of a species-rich genus of neotropical rain forest trees. Science 293:22422245
Ritland K (1996a) Inferring the genetic basis of inbreeding depression in plants.
Genome 39:18
Ritland K (1996b) A marker-based method for inferences about quantitative inheri-
tance in natural populations. Evolution 50:10621073
Ritland K, El-Kassaby YA (1985) The nature of inbreeding in a seed orchard of Douglas
fir as shown by an efficient multilocus model. Theor Appl Genet 71:375384
Ritland K, Jain S (1981) A model for the estimation of outcrossing rate and gene fre-
quencies using n independent loci. Heredity 47:3552
Roberds JH, Namkoong G, Skrppa T (1990) Genetic analysis of risk in clonal popu-
lations of forest trees. Theor Appl Genet 79:841848
Roche LR (ed) (1975) The methodology of conservation of forest genetic resources.
Report on a pilot study. FAO, Rome
Rossetto M, McNally J, Henry RJ, Hunter J, Matthes M (2000) Conservation genetics of
an endangered rainforest tree (Fontainea oraria Euphorbiaceae) and implications
for closely related species. Conserv Genet 1: 217229
Rotach P (2000) Zur Gefhrdung seltener Baumarten in der Schweiz: Grundstzliche
berlegungen, Situationsanalyse und zwei Fallbeispiele. For Snow Landscape Res
75:267284
Rothe GM (1994) Electrophoresis of enzymes. Laboratory methods. Springer, Berlin
Heidelberg New York
Russel JR, Weber JC, Booth A, Powell W, Sotelo-Montes C, Dawson IK (1999) Genetic
variation of Calycophyllum spruceanum in the Peruvian Amazon Basin, revealed by
amplified fragment length polymorphism (AFLP) analysis. Mol Ecol 8:199204
Sa-Ardavut P, Chairuangsirikui T, Chanpaisaeng S (1989) In situ gene conservation in
northeast Thailand. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropi-
cal trees: population structure and genetic improvement strategies in clonal and
seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand,
November 1988. Oxford Forestry Institute, Oxford, and Winrock International,
Arlington, pp 291297
Saatchi S, Agosti D, Alger K, Delabie J, Musinsky J (2001) Examining fragmentation
and loss of primary forest in the southern Bahian Atlantic Forest of Brarazil with
radar imagery. Conserv Biol 15:867875
Sakai KI (1985) Studies on breeding structure in two tropical tree species. In: Gregorius
H-R (ed) Population genetics in forestry. Springer, Berlin Heidelberg New York, pp
212225
Salwasser H, MacCleery DW, Snellgrove TA (1993) An ecosystem perspective on sus-
tainable forestry and new directions for the US National Forest system. In: Aplet GH,
298 References
Sherwin WB, Moritz C (2000) Managing and monitoring genetic erosion. In: Young
AG, Clarke GM (eds) Genetics, demography and viability of fragmented popula-
tions. Cambridge University Press, Cambridge, pp 934
Sim BL (1984) The genetic base of Acacia mangium Willd. in Sabah. In: Barnes RD,
Gibson, GL (eds) Provenance and genetic improvement strategies in tropical forest
genetics. Commonwealth Forestry Institute, Oxford, and Forest Research Centre,
Harare, pp 597603
Simons AJ, MacQueen, Stewart JL (1994) Strategic concepts in the domestication of
non-industrial trees. In: Leakey RRB, Newton AC (eds) Tropical trees: the potential
for domestication and the rebuilding of forest resources. Her Majestys Stationery
Office, London, pp 91102
Simpson D (1998) Selection of superior trees. In: Puri S (ed) Tree improvement.
Applied research and technology transfer. Science, Enfield, pp 107124
Sinclair FL, Verinumbe I, Hall JB (1994) The role of tree domestication in agroforestry. In:
Leakey RRB, Newton AC (eds) Tropical trees: the potential for domestication and the
rebuilding of forest resources. Her Majestys Stationery Office, London, pp 124136
Siregar IZ, Hattemer HH (2001) Gene flow and mating system in a seedling seed
orchard and a natural stand of Pinus merkusii Jung. et de Vriese in Indonesia. In:
Mller-Starck G, Schubert R (eds) Genetic response of forest systems to changing
environmental conditions. Kluwer, Dordrecht, pp 281292
Skrppa T, Johnsen (1999) Patterns of adaptive genetic variation in forest tree
species; the reproductive environment as an evolutionary force in Picea abies. In:
Mtys C (ed) Forest genetics and sustainability. Kluwer, Dordrecht, pp 4958
Smith NJH, Williams JT, Plucknett DL, Talbot JP (1992) Tropical forests and their
crops. Cornell University Press, New York
Smits WTM (1993) Mass propagation of dipterocarps by vegetative methods in
Indonesia. In: Davidson J (ed) Proceedings of the regional symposium on recent
advances in mass clonal multiplication of forest trees for plantation programmes.
RAS/91/004 field document no 4. UNDP/FAO Forest Tree Improvement Project, Los
Baos, pp 3845
Smits WTM, de Fraiture AC, Yasman I (1994) Production of dipterocap planting stock
by cuttings in Indonesia. In: Leakey RRB, Newton AC (eds) Tropical trees: the poten-
tial for domestication and the rebuilding of forest resources. Her Majestys
Stationery Office, London, pp 267272
Smouse PE, Dyer RJ, Westfall RD, Sork VL (2001) Two-generation analysis of pollen
flow across a landscape. I. Male gamete heterogeneity among females. Evolution
55:260271
Sneath PHA, Sokal RR (1973) Numerical taxonomy. Freeman, San Francisco
Sokal RR, Rohlf FJ (1998) Biometry, 3rd edn. Freeman, New York
Soonhuae P, Boyle T, Yeh F (1995) The population genetics of Dalbergia cochinenesis
Pierre and implications for genetic conservation. In: Baradat P, Adams WT, Mller-
Starck G (eds) Population genetics and genetic conservation of forest trees. SPB,
Amsterdam, pp 371385
Sork VL, Campbell D, Dyer R, Fernandez J, Nason J, Petit R, Smouse P, Steinberg E (1998)
Proceedings from a workshop on gene flow in fragmented, managed, and continuous
populations. Research paper no 3. National Center for Ecological Analysis and
300 References
van Wyk G (1977) Pollen handling, controlled pollination and grafting of Eucalyptus
grandis. S Afr For J 101:4753
van Wyk G (1981) Inbreeding effects in Eucalyptus grandis in relation to degree of relat-
edness. S Afr For J 116:6063
van Wyk G, Schnau APG, Schn PP (1989) Growth potential and adaptability of young
eucalypt hybrids in South Africa. In: Gibson GL, Griffin AR, Matheson AC (eds)
Breeding tropical trees: population structure and genetic improvement strategies in
clonal and seedling forestry. Proceedings of the IUFRO conference, Pattaya, Thailand,
November 1988. Oxford Forestry Institute, Oxford, and Winrock International,
Arlington, pp 325333
Vavilov NI (1992) Origin and geography of cultivated plants. Cambridge University
Press, Cambridge
Vavilov NI (1996) Five continents. IPGRI, Rome
Vellend M (2005) Species diversity and genetic diversity: parallel processes and corre-
lated patterns. Am Nat 166:199215
Vendramin GG, Lelli L, Rossi P, Morgante M (1996) A set of primers for the amplifica-
tion of 20 choloroplast microsatellites in Pinaceae. Mol Ecol 5:585598
Verkerke W (1989) Structure and function of the fig. Experientia 45:612622
Viana VM, Tabanez AAJ, Batista LF (1997) Dynamics and restoration of forest frag-
ments in the Brazilian Atlantic moist forest. In: Laurance WF, Bierregaard RO (eds)
Tropical forest remnants: ecology, management, and conservation of fragmented
communities. University of Chicago Press, Chicago, pp 351365
Viera MF, de Carvalho-Okano RM (1996) Pollination biology of Mabea fistulifera
(Euphorbiaceae) in Southeastern Brazil. Biotropica 28:6168
Villaincourt RE, Potts BM, Watson M, Volker PW, Hodge GR, Reid JB, West AK (1995)
Detection and prediction of heterosis in Eucalyptus globulus. For Genet 2:1119
von Gadow K (2005) Forsteinrichtung. Analyse und Entwurf der Waldentwicklung.
Universittsverlag Gttingen, Gttingen
von Gadow K, Kleinn C (2005) Managing forest ecosystems. In: Peterson CE, Maguire
DA (eds) Balancing ecosystem values innovative experiments for sustainable
forestry. USDA Forest Service general technical report PNW-GTR-635, pp 1524
von Humboldt A (1808) Ansichten der Natur. Neuausgabe 2004. Eichborn, Frankfurt
Vornam B, Kuchma O, Kuchma N, Arkhipov A, Finkeldey R (2004) SSR markers as
tools to reveal mutation events in Scots pine (Pinus sylvestris L.) from Chernobyl.
Eur J For Res 123:245248
Vornam B, Bldner C, Schubert R, Baumann R, Mller-Starck G, Hattemer HH, Polle A
(2003) Preliminary results of expression profiling of spruce (Picea abies (L.)
Karst.) in response to drought stress. In: Galris D (ed) Free radicals and oxidative
stress: chemistry, biochemistry and pathophysiological implications. Proceedings of
the meeting of the society for free radical research European section, Ioannina,
Greece, 2629 June 2003. Medimont, Bologna, pp 109113
Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Hornes M, Frijters A, Pot J,
Peleman J, Kuiper M, Zabeau M (1995) AFLP: a new technique for DNA finger-
printing. Nucleic Acids Res 23:44074414
Wang K (2001) Gene flow and mating system in European beech (Fagus sylvatica L.).
Cuvillier, Gttingen
Warren JM, Kalai, Misir S (1995) An unnatural breeding system polymorphism in
cacao (Theobroma cacao, Sterculiaceae) in Trinidad. Am J Bot 82:11261130
References 303
Webb CJ, Bawa KS (1983) Pollen dispersal by hummingbirds and butterflies: a com-
parative study of two lowland tropical forest plants. Evolution 37:12581270
Weising K, Gardner RC (1999) A set of conserved PCR primers for the analysis of sim-
ple sequence repeat polymorphisms in chloroplast genomes of dicotyledonous
angiosperms. Genome 42:919
Weising K, Nybom H, Wolff K, Kahl G (2005) DNA fingerprinting in plants. Principles,
methods, and applications. Taylor and Francis, Boca Raton
Wellendorf H, Kaosa-Ard A (1988) Teak improvement strategy in Thailand. DSR,
Copenhagen
Wells OO, Wakeley PC (1966) Geographic variation in survival, growth and fusiform-
rust infection of planted loblolly pine. For Sci Monogr 11:140
Wells OO, Wakeley PC (1970) Variation in longleaf pine from several geographic
sources. For Sci 16: 2842
Westfall RD (1992) Developing seed transfer zones. In: Fins L, Friedman ST, Brotschol
JV (eds) Handbook of quantitative forest genetics. Kluwer, Dordrecht, pp 313398
Wheeler NC, Jech KS (1992) The use of electrophoretic markers in seed orchard
research. New For 6:311328
Whelan RJ, Ayre DJ, England PR, Llorens T, Beynon F (2000) Ecology and genetics of
Grevillea (Proteaceae): implications for conservation of fragmented populations. In:
Young AG, Clarke GM (eds) Genetics, demography and viability of fragmented pop-
ulations. Cambridge University Press, Cambridge, pp 253269
White G, Powell W (1997) Isolation and characterization of microsatellite loci in
Swietenia humilis (Meliaceae): an endangered tropical hardwood species. Mol Ecol
6:851860
White GM, Boshier DH (2000) Fragmentation in Central American dry forests: genetic
impacts on Swietenia humilis (Meliaceae). In: Young AG, Clarke GM (eds) Genetics,
demography and viability of fragmented populations. Cambridge University Press,
Cambridge, pp 293311
White GM, Boshier DH, Powell W (1999) Genetic variation within a fragmented pop-
ulation of Swietenia humilis Zucc. Mol Ecol 8:18991909
White TL (1987) A conceptual framework for tree improvement programs. New For
4: 325342
Whitmore TC (1975) Tropical rain forest of the far east. Clarendon, Oxford
Whitmore TC (1998) An introduction to tropical rain forests. Oxford University Press,
Oxford
Wickneswari R, Boyle TJ (2000) Effects of logging and other forms of harvesting on
genetic diversity in humid tropical forests. In: Young A, Boshier D, Boyle T (eds)
Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and
CABI, Oxford, pp 115122
Wickneswari R, Norwati M (1993) Genetic diversity of natural populations of Acacia
auriculiformis. Aust J Bot 41:6577
Wickneswari R, Norwati M (1995) Spatial heterogeneity of outcrossing rates in Acacia
auriculiformis A. Cunn. ex Benth. in Australia and Papua New Guinea. In: Baradat P,
Adams WT, Mller-Starck G (eds) Population genetics and genetic conservation of
forest trees. SPB, Amsterdam, pp 329337
Wickneswari R, Lee SL, Din M (1995) Utility of RAPD markers in detecting genetic
variation in multiple seedlings of Hopea odorata Roxb. (Dipterocarpaceae). Malay
For 58:4250
304 References
Wickneswari R, Lee CT, Muhammad N, Boyle TJB (1999) Impact of logging on genetic
diversity in humid tropical forests. In: Mtys C (ed) Forest genetics and sustain-
ablility. Kluwer, Dordrecht, pp 171181
Wiebes JT (1979) Co-evolution of figs and their insect pollinators. Annu Rev Ecol Syst
10:112
Wilcox BA, Murphy DD (1985) Conservation strategy: the effect of fragmentation on
extinction. Am Nat 125:879887
Williams CG, Savolainen O (1996) Inbreeding depression in conifers: implications for
breeding strategy. For Sci 42:102117
Williams ER, Matheson AC, Harwood CE (2002) Experimental design and analysis for
tree improvement. CSIRO, Melbourne
Wilson EO (1988) The current state of biological diversity. In: Wilson EO, Peter FM
(eds) Biodiversity. National Academy Press, Washington, pp 318
World Commission of the Environment and Development (1998) http://www.sustain-
ablemeasures. com/Training/Indicators/index.html
Wright JW (1976) Introduction to forest genetics. Academic, London
Wright S (1931) Evolution in Mendelian populations. Genetics 16:97159
Yang RC, Yeh FC, Yanchuk AD (1996) A comparison of isozyme and quantitative
genetic variation in Pinus contorta ssp. latifolia by FST. Genetics 142:10451052
Yates CJ, Broadhurst LM (2002) Assessing limitations on population growth in two
critically endangered Acacia taxa. Biol Conserv 108:1326
Yeh F (1989) Isozyme analysis for revealing population structure for use in breeding
strategies. In: Gibson GL, Griffin AR, Matheson AC (eds) Breeding tropical trees: pop-
ulation structure and genetic improvement strategies in clonal and seedling forestry.
Proceedings of the IUFRO Conference, Pattaya, Thailand, November 1988. Oxford
Forestry Institute, Oxford, and Winrock International, Arlington, pp 119131
Yeh F, Brune A, Cheliak WM, Chipman DC (1983) Mating system of Eucalyptus citri-
odora in a seed-production area. Can J For Res 13:10511055
Young A, Boshier D, Boyle T (eds) (2000) Forest conservation genetics. Principles and
practice. CSIRO, Collingwood, and CABI, Oxford
Young AG, Boyle TJ (2000) Forest fragmentation. In: Young A, Boshier D, Boyle T (eds)
Forest conservation genetics. Principles and practice. CSIRO, Collingwood, and
CABI, Oxford,, pp 123134
Young AG, Clarke GM (eds) (2000) Genetics, demography and viability of fragmented
populations. Cambridge University Press, Cambridge
Young AG, Merriam GH, Warwick SI (1993) The effects of forest fragmentation on
genetic variation in Acer saccharum Marsh (sugar maple) populations. Heredity
71:277289
Young AG, Boyle T, Brown T (1996) The population genetic consequences of habitat
fragmentation for plants. Trends Ecol Evol 11:413418
Young AG, Brown AHD, Zich FA (1999) Genetic structure of fragmented populations
of the endangered daisy Rutidosis leptorrynchoides. Conserv Biol 13:256265
Young AG, Brown AHD, Murray BG, Thrall PH, Miller CH (2000) Genetic erosion,
restricted mating and reduced viability in fragmented populations of the endan-
gered grassland herb Rutidosis leptorrhynchoides. In: Young AG, Clarke GM (eds)
Genetics, demography and viability of fragmented populations. Cambridge
University Press, Cambridge, pp 335359
References 305
A Allophyllus occidentalis 48
Abies alba 248 Altingia excelsa 134
Acacia albida 29 amplified fragment length
auriculiformis 31, 77, 193, 214, 246 polymorphism 9, 35, 189
crassicarpa 214 Ancistrocladus korupensis 79
mangium 11, 29, 193, 214, 217 Andira inermis 57
mearnsii 202, 215 androdioecy 44
nilotica 214 anemogamy see pollination by wind
retinodes 84 Anguria spp. 58
Acer saccharum 118 Anisoptera spp. 104
skutchii 10, 29 Annonaceae 58
Adansonia digitata 56 Apis spp. 62, 64
adaptability 3, 87, 142, 243 apomixis 51, 52
adaptation 21, 87, 113, 147, 163 Araucaria angustifolia 55, 157, 214
evolutionary 88 araucana 151, 248f.
physiological 87, 88 cunninghamii 194, 214
adaptedness 3, 20, 87, 155, 206 Ardisia escallonioides 73, 76, 77
adaptive genetic marker 248 Aristolochia spp. 58
adaptive potential 23, 87, 247 Astrocaryum mexicanum 77
adaptive trait expression 247ff. Austrocedrus chilensis 249
AFLP see amplified fragment length autochthonous 137, 148
polymorphism autogamy 44
agamospermy 51 Azadirachta indica 151, 165
Agathis spp. 131
Albertia edulis 48 B
allele 15, 17 Bambusa vulgaris 50
deleterious 80 barochory 66
dominant 15 base population 174, 208
recessive 15 basic material 221
allelic richness 25 basic rule of conservation 124
allelic structure 16, 17, 22 bats as fruit dispersers 117
allochthonous 137, 148f. as pollinators 56, 59, 121
allogamy 45 Bauhinia spp. 56
allopatric 149 pauletia 45
speciation 101, 104 ungulata 45, 46
308 Index
bees as pollinators 57, 59, 62, 64 Cedrela odorata 10, 11, 37, 68, 77, 246
beetles as pollinators 58, 64 Cedrelinga catenaeformis 222
Beilschmiedia pendula 72, 77 Ceiba pentandra 56, 72, 77
Bernardia nicaraguensis 47 Ceropegia 58
Bertholletia excelsa 77, 84 certification 142, 221
biparental inbreeding 133, 204, 225 Chlorophora tinctoria 47
birds as pollinators 57, 59, 117 chloroplast DNA 14, 18, 37, 68
Blastophaga quadraticeps 92 sequences of 103
Bombacaceae 66 cicadas as pollinators 59
Bombacopsis quinata 212, 224 cistron 6, 7
bottleneck 20, 122 Citrus spp. 52
breeding clonal
population 208 archive 234
products 215, 231ff. forestry 204f., 229, 232
progress 184, 186, 207 mixture 232ff.
strategy 195, 207ff. propagation 204f.
Brosimum alicastrum 72, 77 seed orchard 55, 60, 74, 196ff., 223f.
budding 199 variation 164, 167
bugs as pollinators 59 clone 50
bulbil 53 cluster analysis 27
Bursera simarouba 47 Coccoloba caracasana 48
tomentosa 47 floribunda 48
butterflies as pollinators 58, 59 coevolution 90, 91
Coffea arabica 215f.
C Coleoptera 58
Caesalpinia echinata 10, 68, 115 combining ability 185
Calamus spp. 45, 48 conservation 88, 90
Calliandra spp. 57 dynamic 251ff.
Calophyllum longifolium 63, 77 ex situ 251, 253ff.
Calycophyllum spruceanum 10, 29, 151 of figs 96
candidate gene 7, 89 in situ 251ff.
Carapa guianensis 77, 133 objectives 241ff.
procera 77 static 251, 255ff.
Carica papaya 44 strategy 100
Caryocar brasiliense 8 continuous-cover system 137
Castanea sativa 191 Cordia alliodora 50, 77, 84, 202
Castanopsis spp. 45 collococca 47, 84
Casuarina spp. 214, 220 dentata 84
equisetifolia 151, 165 inermis 84
nana 51 panamensis 47, 84
Castilla elastica 45 pringlei 84
cauliflory 56 correlated matings 133, 138, 156f., 204
Cavanillesia platanifolia 72, 76, 77, 85 Corythophora alta 67, 140
Cecropia spp. 55 cosexuality 42, 44
obtusifolia 73 Cotylelobium spp. 104
peltata 47 Couroupita guianensis 56
Index 309
homology 103 J
homoplasy 103 Jacaranda copaia 77
homozygous 15 Juglans neotropica 222
honey bees 114, 202, 225 regia 51
Hopea spp. 104
odorata 52 K
hummingbird 57 keystone species 91
as pollinators 117
hybrid seed orchard 198 L
hybrid vigor 192 Laelia rubescens 50
hybridization 193 land race 149f., 154, 165, 167, 194,
hydrochory 66 217, 230
Hymenoptera 43, 57 Lecythidaceae 57
legislation 221
I Leucaena spp. 213
imprinting 152 leucocephala 193, 216, 232
inbreeding 20, 73, 78, 96, 216 linkage map 189ff.
coefficient 33, 34, 79, 90, 117 Lovoa trichilioides 219
depression 51, 78, 80, 89, 115, 134,
192, 216f. M
-effective size of population 116 Mabea fistulifera 59
incompatibility 43, 58, 75, 80, 82 Macaranga spp. 103
indicators of sustainable forest mangrove 53
management 143ff. mating contact 115, 132
indigenous population 148 mating design 184ff.
individual effective pollen mating preference 42, 43, 96
cloud 156 mating, random 69
Inga spp. 84, 102, 104ff. mating system 20, 22, 69, 132, 138
inheritance 14 megagametophyte 14
biparental 14 Meliaceae 46
maternal 15, 18 Mendel, G. 13
paternal 15 metapopulation 125, 139, 254f.
study 12, 14 Metrosideros excelsa 114
uniparental 14, 18 microsatellite 8, 33, 62, 64, 78, 89
insect diversity 100 migration 20, 53, 65, 118
interaction 164, 179 minimum viable population 124ff.
internal transcribed spacer 103 mitochondrial DNA 14
Ionopsis utricularoides 91 mixed mating system 76, 80, 81, 85
Iryanthera macrophylla 44 molecular marker 7
isoenzyme see isozyme monoecy 44, 46, 84, 94
isolation 113, 115, 117, 197, Monotes kerstingii 104
200, 202 Moringa oleifera 10, 29
isozyme 12 moths as pollinators 58, 62
electrophoresis 12 mtDNA see mitochondrial DNA
gene locus 12, 30, 31, 33, 63, 73, multiple-population breeding 195,
76, 81, 85 206f., 240f., 253
312 Index
V
variance-effective size 116f.
variation parameters 246f.