Introduction
Seed orchards constitute an important component in most tree improvement
programs, and seeds from seed orchards are superior to stand seeds. Nevertheless,
it does not seem easy to achieve the optimal function of seed orchards because of
many factors. Such factors are described, and general principles and practices of
seed orchards are reviewed in this introduction.
Overview of seed orchards
1. Seed orchard - general
Reforestation can be achieved naturally or artificially. For artificial regeneration,
seeds are collected from natural stands or from seed orchards. Seeds are
important in both natural and managed populations as a resource of reproductive
material and also for their commodity value in the production of forest products.
To meet immediate seed needs, one can collect seeds from individually good
phenotypes, good stands, seed production areas (seed stands) and proven seed
sources (Zobel and Talbert, 1984). Long-term seed needs are often met by
establishing seed orchards; greatly improved seeds are obtained from seed
orchards. Seed orchards are the most common and cost-effective means of
making available a stable supply of genetically improved seed (Varghese et al.,
2000).
A seed orchard is defined as an area where seeds are mass-produced to obtain the
greatest genetic gain, as quickly and inexpensively as possible (Zobel et al.,
1958). Also, it is defined as a plantation of selected clones or progenies which is
isolated or managed to avoid or reduce pollination from outside sources, and
managed to produce frequent, abundant, and easily harvested crops of seed
(OECD, 1974; Feilberg and Soegaard, 1975). In study VII, seed orchards are
defined as the production populations where genetic gain from tree breeding is
transferred into planting materials and to commercial forest crops.
The objectives of a seed orchard and methodology for its establishment may be
modified when seeds are not needed for immediate use, but where there is a
perceived future need for seeds. Clonal propagation techniques have been
improving in effectiveness, both biologically and economically, for many years.
These may influence the nature of future output systems and modify the present
role of seed orchards (see review by Sweet, 1995). In the future, mass
propagation techniques, like tissue culture, somatic embryogenesis and artificial
seeds, may make it possible to get seeds or plants directly from in vitro
conditions. In the foreseeable future, however, the role of seed orchards as a
primary production system for genetic improvement programs is likely to
continue.
8
2. Seed orchard establishment and management
Seed orchards may be established vegetatively with grafts, cuttings, rooted
cuttings or tissue culture plantlets propagated from a selected tree, or with
seedlings produced from seeds of the selected parent (El-Kassaby and Askew,
1998). In either case, a careful testing and evaluation program can be used to
separate the environmental and genetic factors associated with the trees’
superiority.
Seedling seed orchards (SSO) have a broad genetic base because of the large
number of parents involved, but the selection differential is less than in the
vegetative seed orchard. When the seedlings are grown from open-pollinated
seeds collected from selected trees, the cost for raising material and time required
for establishment will usually be low in comparison with costs of establishing
clonal seed orchards (Toda, 1964). Seedling seed orchards are suitable for species
like most Eucalyptus spp., Picea mariana, the early flowering pines, and many
hardwoods that produce seed at a young age. In seedling seed orchards, some
progenies may outgrow others that may increase inbreeding.
The breeding seedling orchard (BSO), a derivative of the SSO, is a flexible
strategy that lies somewhere between a SSO and a progeny test (Barnes, 1995). In
the BSO, the conventional hierarchy of sequential testing, selection and seed
production populations is combined in a single planting. This may be a cheaper
and less complicated way to run a tree improvement program. It can provide the
potential to respond to new materials and new demands, and allows the breeder
freedom to be adventurous without taking unacceptable risks.
In clonal seed orchards (CSO), there may be some level of graft incompatibility,
but the problem has simply been accepted. On the other hand, grafting is
particularly useful for species where flowering is delayed, as grafts retain the
physiological age of the parents. Root deformities and low productive output may
also be problems in seed orchards established with cuttings or by tissue culture.
The possibility of selfing is considerably greater in vegetative orchards, but this
can be minimized by keeping ramets of the same clone well separated. The
decision whether to use clonal, seedling or polycross seed orchards should be
made on the basis of economic and genetic considerations (Sweet, 1995). The
choice of seedling versus clonal seed orchards depends primarily on the earliness
of flowering of grafted trees and of seedlings, the relative difficulty, cost and
speed of establishing grafted trees and seedlings, and selection differential
possible with clones and families.
If trees are moderately self-sterile and have overlapping flowering periods (e.g.,
in Acacia and Populus species), it should be possible to produce significant
quantities of hybrid seeds in seed orchards (Griffin et al., 1992). Some trees
reveal high specific combining ability (SCA) in certain combinations. If they are
9
outstanding for certain characteristics, such as volume production, tree form and
disease resistance, they can be used for the establishment of bi-clonal seed
orchards. There is considerable interest in taking advantage of these
combinations by establishing two-clone seed orchards, the progeny from which
will be planted locally or on specific problem areas (Kellison, 1971; Sedgley and
Griffin, 1989).
Determination of the number of families or clones to be included in a seed
orchard depends on the short- and long-term goals of the breeding program (ElKassaby and Askew, 1998). Breeding programs that use seed orchard genotypes
as parents for the next successive generation will reduce the effective population
size just a few generations if a small number of parents is used. Also, the gene
diversity may soon become low (VIII; Bila, 2000). If breeding (e.g., clonal
archives) and production populations (e.g., seed orchards) are maintained
separately, the breeding population must reflect the diversity of the original
population and be sufficiently large to maintain gene diversity for many
generations ahead, whereas the production population can be managed to meet
specific needs, e.g., high gain can be combined with an acceptable gene diversity
in a production forest.
To guarantee pollination within an orchard and the success of seed orchard
material in reforestation, seed orchards should be established with superior
genotypes originating from geographical distribution areas (Sarvas, 1970, Koski
1980b). In other words, the location of a seed orchard should be geographically
within the natural range of the species, but transfer outside this area to a warmer
climate in the south can be advantageous for seed maturation, earlier flowering
and physical isolation from pollen contamination (Sarvas, 1970). Moreover,
transfer from high to lower elevations should be used with similar effects in cases
where there are appreciable differences in altitude (Hellebergshaugem, 1970).
Moving orchards should be done only after consideration that it seems likely
normal fruit and seed are produced in the new environment. It has also been
observed that maternal environment might influence growth and growth rhythm
of the offspring, the so-called “after-effect” (Bjornstad, 1981; Dormling and
Johnsen, 1992; Lindgren and Wei, 1994). This may be a factor that deserves
consideration in seed orchard localization.
Frost, drought and wind may all have adverse effects on the establishment of
orchards, and/or flower crops and seed set. Of key importance is whether the
environment of an area favors the production of seed crops. An orchard site
should be intermediate in fertility. Fertile sites often involve problems with heavy
vegetative and poor reproductive growth. Sloping ground creates problems with
management and cone harvest. Abandoned agricultural field, flat terrain and
gently sloping land can thus be good (Zobel and Talbert, 1984). Damage by
animals is also a problem, and could be considered in choosing the spot where to
10
place a seed orchard. Cold air pockets may be a problem, especially where the
land looks flat that deserves attention.
Seed orchards must be protected or isolated from contamination by outside
pollen. Pollen dilution zones are most critical for advanced-generation seed
orchards because of the greater potential loss of genetic gain. Conversely,
contamination from foreign can increase seed set as a young orchard first comes
into production, when pollen production is typically low, and can increase
genetic diversity of seed crops. Concerning protection, it is good to orient the
long axis of the orchard with the direction of the prevailing wind at the time of
pollen dispersal if a rectangular configuration is used. Most of the pollen from a
single tree is dispersed within a very short distance. Even so, considerable
amounts may migrate from several hundred or thousand meters away during a
year of very heavy flowering (Silen, 1962). For most transects, there is a very
rapid decrease in pollen density within 400-600m from the edge of the pollen
source, beyond which the pollen frequency is fairly constant (Andersson, 1955).
However, efforts to isolate seed orchards have not been very successful and
pollen contamination is often large (Lindgren, 1991).
Collecting seed crops is an expensive process. To facilitate seed collection, the
crowns of orchard trees should be pruned to encourage the development of a
short, wide and bushy habit except for those species where fallen seeds are
collected from the ground, such as oak and beech. Pruning must be done with
care, as it can reduce cone production or increase male flowering. The loss of
seed production to insects, disease, mammals and birds should also be minimized.
How do we determine when cones are ripe and ready for harvest? Key to insuring
efficient cone harvest is recognition of cone maturity. Harvesting mature cones
will do much to promote proper cone opening and extraction of high quality seed.
If cones are harvested while immature, they will fail to open properly, seed yield
per cone will be reduced, and seed viability will be diminished. Sometimes, a
remedy may be found in the proper treatment of the cones after cone collection,
so the seeds can mature after collection. As seeds of conifers become mature, the
specific gravity of the cones decreases because of water loss. This change in
specific gravity provides one of the most reliable tests for determining ripeness of
conifer cones (Hartmann et al., 1997). It has been found that seeds from cones
harvested 3 weeks prior to cone maturity are physiological mature in a loblolly
pine, but premature harvesting of the cones results in case-hardening (Matthews,
1963). Cells on the top and bottom of the cone scales dry at different rates,
causing the scale to flex outward thereby releasing the seed.
In some cases, accidental mislabeling of grafts or missing identification of
rootstocks that developed into trees may cause variation in the number of ramets
among clones (e.g., Kossuth et al., 1988; Paule, 1991) and in the estimates of
genetic composition of seed orchard crop (VII). Even if these errors become
11
apparent, the mislabeled materials must be identified or discarded. Harvesting
seeds from alien genotypes and even leaving alien ones in the seed orchard leads
to a substantial contamination of the gene pool of the orchard crop (Gömöry et
al., 2000). This may increase gene diversity, but the effect would be marginal
compared to the loss in genetic gain.
For establishment clonal seed orchards, there are many possibilities for errors,
starting from collecting and labeling the scions, over the manipulation during the
transport, grafting, handling the material in the nursery, up to planting (Gömöry
et al., 2000). Scions may also fail, leaving only the rootstock to develop a seedproducing crown. These mislabeled or alien ramets represent genotypes of
unknown genetic quality. So, the presence of any alien material must be
considered as in internal contamination of the orchard gene pool. A rapid and
non-destructive means of clone certification is needed, using biochemical
markers such as monoterpenes or isozymes that are not realistic in practical
operations (Kossuth et al., 1988). DNA fingerprinting for confirmation of clone
identity may be readily available soon, but it may not be cheap.
3. Seed production and quality in seed orchards
The progress of a seed orchard program depends on a plentiful delivery of viable
seed. The final cone and seed yields may be influenced by breakdown of any one
of the processes of pollination, pollen grain germination, pollen tube growth,
fertilization and embryo development (Brown, 1971). Seed yield is also reduced
if insufficient pollen is supplied to the female strobilus. Unpollinated ovules
degenerate during the first growing season, and only the wings continue to
develop in most conifers (Sarvas, 1962). For maximum yield of seed, the female
strobili must be pollinated when they are fully receptive to pollen, i.e., when the
axis of the strobilus is fully elongated and the ovuliferous scales are separated
from each other. Empty seed and seed of low viability may also result from a
breakdown in embryogeny. A further loss of seeds is occasioned by premature
abscission of the cones from the tree.
There has long been interest in the flowering behavior of forest trees. The
development of seed orchards for production of improved seeds has intensified
this interest and focused attention on individual clones. There are many desires
for a well functioning seed orchard and still more to get it to work as an ideal
object (II, V, XI; Sarvas, 1970; Eriksson et al., 1973; Jonsson et al., 1976; Koski,
1980b; El-Kassaby and Reynolds, 1990; Wheeler and Jech, 1992; Harju, 1995).
Most existing orchards have been established based on the assumption that each
clone and ramet (or family-plot and seedling tree) in the orchard would


12
flower during the same period (synchronization, random mating),
have the same cycle of periodic heavy flower production,




be completely inter-fertile with all its neighbors and yield identical numbers
of viable seed per tree,
have the same degree of resistance to self-incompatibility, and
have a similar rate of growth and crown shape as all other trees.
Non- or minimum relatedness and pollen contamination are also expected.
It is almost certain that these assumptions are virtually never satisfied. On the
other hands, there are not likely limited deviations from these assumptions, which
lead to serious consequences. But it is difficult to formulate the consequences
numerically. Thus, it is important to evaluate their impact and consider them
properly when establishing and managing seed orchards. The biological impact of
forest tree reproductive processes and the extent of management practice of seed
orchards and nurseries on the genetic representation of new forests were
reviewed by Edwards and El-Kassaby (1996).
The first problem related to seed production for seed orchard programs is to
estimate the amount of seed needed in the future. A buffering amount should be
calculated for loss or failure of seed production because good seed crops will not
be obtained every year. Also, the time in seed storage (at least 2-3 years) should
be considered if storage is possible. Zobel and Talbert (1984) suggested that 30%
greater amount than currently predicted requirement would be for a sufficient
buffer in planning the target amount of seed from seed orchard programs. The
buffering may also be obtained by keeping older seed orchards for a longer time
or creating options to produce more seeds in cheaper ways if needed.
Trees may grow more quickly (fast DBH growth) and produce more cones under
CO2 enrichment because developing seeds are strong carbon sinks (Lawlor, 1991).
It is recently reported by LaDeau and Clark (2001) that trees are twice as likely to
be reproductively mature and produce three times as many cones and seeds as
trees at ambient CO2 concentration, as a result of 3 years’ CO2 fumigation in a
Pinus taeda half-sib plantation. This implies that trees in seed orchards can
produce plenty of seeds at early ages and smaller sizes if the application of this
technique is sustainable and cost-effective.
Seed orchards have been established with selected superior trees to produce seeds
that are physiologically and genetically better than those obtainable from natural
stands (Harju, 1995). The degree of superiority is in general determined by
comparing the performance of the progeny of the orchard trees with the
performance of a checklot family in a carefully designed test (Li et al., 1999).
Due to the deviation from the ideal situations, the improvement of seed quality
cannot be accurately calculated and does not reach at the maximum. The
minimum requirements can thus be set and allowances made for deviations from
the ideal situations (Koski, 1980b).
13
Forest trees are generally still in early stages of domestication. Most tree
breeding programs were initiated in the late 50’s or 60’s. The first-generation
seed orchards were established using ramets or seedlings of selected plus-trees
without further genetic information (Hodge and White, 1993). Usually, the
intention was to use as an equal number of ramets, but there were often
deviations for practical difficulties (VII). Improved seed orchards (1.5 or 2.5
generation SO) consist of the very best genotypes, based on testing of their
progeny. No matter what kind of seed orchard is established, clone or family
records must be kept to ensure use of the best genotypes and to minimize
deleterious related matings, especially in the advanced-generation seed orchards
where genetic information from progeny tests is available and parents will be
more closely related and perhaps inbred.
Orchard managers want seed orchards with well-pruned trees as in apple orchards,
using topping and branch pruning, but trees easily grow beyond reach for
convenient cone collection. Several methods of cone collection are currently
practiced in seed orchards; they include the use of ladders, mechanical lifts,
netting, vacuum machines, tree shakers and helicopters (Ho and Schooley, 1995).
Well-trained small animals (e.g., monkeys) can also be used. Among them,
ladders and ground netting are generally considered as the cheapest and most
practical methods. Tree climbers may be employed, but this method is expensive,
and the branch apices that bear cones are often difficult to reach safely. Thus, it
seems increasingly more unrealistic to rely on human climbers.
The use of special stocks, often of a different species or variety from the scion,
has been practiced in fruit and forest orchards to reduce the stature of the grafts
and to encourage early fruiting and flowering (Gardner et al., 1922). The ability
to produce broomy, dwarf trees may be of substantial value in grafted seed
orchards (Buckland and Kuijt, 1957). The development of dwarfing rootstocks is
therefore an important concern in seed orchard technology, Christmas tree culture
and horticulture (Duffield and Wheat, 1963). Use of root suckers as stocks may
also be of benefit to reduce the height of seed trees.
Some grafted scions do not develop satisfactorily due to some reasons such as
failure in graft development and inadequate vascular transport of water and
nutrient through xylem tissues (Cameron and Thomson, 1971), so-called graft
incompatibility or uncongeniality (Slee and Spidy, 1970). In less severe cases,
incompatible grafts cannot be detected for some time after establishment, and
their occurrence modifies irrevocably an orchard layout, disturbs clone
representation (VII) and reduces seed production. Also, there is a large variation
in incompatibility among clones and among species (Ahlgren, 1972). Dyson
(1975) reported that a large proportion of the ramets on Pinus radiata stocks had
a dwarf, bushy form caused by graft incompatibility, but appeared to be bearing
more ripening cones than grafts made on Pinus patula stocks in P. patula clonal
seed orchards. Heteroplastic grafts in a black pine (P. nigra) seed orchard
14
displayed more and better seeds than homoplastic ones (Climent et al., 1997).
Therefore, if grafts are made with genetically similar materials, the problems
associated with graft incompatibility may not be serious in clonal seed orchards.
Seeds are often transferred between countries or between areas within countries.
For the trade or transfer, adequate information about source and history should
accompany the shipment (Jones and Burley, 1973; OECD, 1974). Certification is
defined as the authoritative attestation of facts or statements and it usually
implies documentation by a formal written certificate. Seed certification is an
official statement that a seed lot conforms to certain standards, which may
include specific identity, origin, genetic characters and seed purity (Muhs, 1986;
Portlock, 1988). In seed orchards, seed certification implies, in general, a uniform,
high genetic quality that may be described by genetic gain, gene diversity and
germination ability. An effective number, which gives better genetic information,
is recommended to describe gene diversity of orchard populations and orchard
crops (VII).
Genetic gain
Genetic improvement includes selection, testing and breeding from the species,
through the population and family, to the clonal level (see Fig. 3; Barnes, 1995).
The amount of improvement that can be made by plus-tree selection among a
number of initial parents depends on the amount of genetic variation among the
parents. In general, there is great variation found within and among populations
in natural forests. By selecting plus-trees and using them to establish seed
orchards, one can get genetic gain in seed crops. In the orchards, the gain made
by the plus-tree selection is realized from the general combining ability (GCA) of
the selected trees, which comes from the additive variance in the reference
population.
For gain from specific combining ability, parents are chosen for crossing and the
best combinations are selected (cf. a bi-clonal seed orchard). But, the specific
combining ability is a reflection of dominance and epistatic gene actions. So, its
value cannot be predicted from phenotypes before the cross is made (Zobel and
Talbert, 1984). To get high intensity of selection and to maintain high diversity
require a large number of parents, therefore more parents than needed are used in
the first-generation seed orchards (Hodge and White, 1993). This also facilitates
genetic thinning on later stages.
In advanced-generation orchards, breeding values and/or fertility variation of
orchard parents will be known from genetic testing or fertility assessment. With
knowledge of breeding values and other genetic information, it would be possible
to establish new seed orchards using intentionally different numbers of ramets
per clone or progenies per family to attain the optimal genetic gain with desirable
diversity (VII; ‘linear deployment’ by Lindgren and Matheson, 1986; ‘step-wise’
15
Selected plus
trees
Selection from natural stands or
unimproved plantations
Genetic gain by selection (7% in volume)
Progeny
population
Progeny
test
Crossing between superior
genotypes
1st generation
seed orchard
Genetic information
Genetic gain by roguing inferior genotypes
or by collection best ones (12%)
Information
Additional gain by
repeated roguing
Improved SO
(1.5 gen.)
Selection and crossing; Genetic gain
Information
Genetic gain
nd
2 generation seed
orchard
Infusion
(new genetic materials)
(13-21%.)
Genetic gain by roguing
to keep the best 30% (26-35%)
Improved SO
(2.5 gen.)
Figure 1. Flow chart of the generation of seed orchards. Values for genetic gain are
from Li et al. (1999).
16
relationship by Hodge and White, 1993). These optimal and step-wise solutions
could result in more gain than truncation selection, while minimizing the
disadvantage connected to reduction in diversity.
Establishing improved seed orchards by genetic thinning with high intensity may
be a more effective way to get high genetic gain rather than establishing of new
seed orchards, but some loss of gene diversity is unavoidable after roguing (I). Li
et al. (1999) reported that loblolly pine in the southern United States grown from
seeds of first-generation seed orchards produced 7-12% more volume per unit
area at harvest than trees grown from seeds of wild forests (Fig. 1), and that of
second-generation seed orchards showed 13-21% gain in rotation volume. The
estimated gain from rogued second-generation seed orchards varied from 2635%, depending on the breeding region considered.
The genetic gain estimated by Li et al. (1999) could be over estimated because of
the unequal contribution and relatedness among parents. In the Western Gulf
Forest Tree Improvement program, roguing improved the performance of loblolly
pine seed orchards by an average of 2.9% in breeding value for volume growth
(Byram et al., 1998). On the other hand, Matheson et al. (1986) showed that
Pinus radiata trees from first-generation seed orchards had considerably better
growth (15-30%) than control seed lots. Sluder (1994) also reported that high
gain in rust resistance and height growth could be obtained in a secondgeneration slash pine seedling seed orchard.
Genetic gain can be related to fertility of orchard parents (Griffin, 1982). In
practice, orchards are culled on the basis of breeding value of the parents, rather
than their fertility, but one should predict to what extent is reduction of total seed
production likely as a result of clonal selection on the basis of breeding value. It
may thus be necessary to give some weight to the fertility or collection cost as
well. If general combining ability (GCA) and fertility are known, the expected
genetic gain can be expressed as
G   GCAi  pi 
N
i 1
[1]
where N is the census number and pi is the fertility of ith parent in the seed
orchard. If parents contribute equally to the seed orchard crop, genetic value and
gene diversity in bulked orchard seeds are expected to be the same as that of the
orchard parents. Due to fluctuations in fertility, however, genetic gain and gene
diversity are difficult to predict accurately (Griffin, 1982).
Gene diversity and group coancestry
The accumulated loss of gene diversity is the group coancestry (Lindgren and
Kang, 1997). The gene diversity can be estimated relative to the reference
17
population (II and VII). The reference population is defined as having an infinite
number of unrelated individuals, and thus a group coancestry of 0 and gene
diversity of 1. No genes of individuals in the reference population are identical by
descent. The inbreeding in the reference population is also considered to be 0
(III; Yeh, 2000).
Loss of gene diversity may occur in two stages: (1) when seed orchards are
established with a limited number of parents; and (2) when there is unequal
contribution of gametes by the orchard parents (Harju, 1995). On the other hand,
pollen contamination increases the gene diversity of orchard crop (Lindgren and
Mullin, 1998).
The term group coancestry was introduced by Cockerham (1967), who defined
the group coancestry coefficient as referring to the probability of identity of a
random pair of genes sampled, with replacement, among the 2N genes of the N
individuals in a population. By definition, therefore, group coancestry is the
average coancestry of all pairs of population members including individuals with
themselves (self-coancestry). Group coancestry is equivalent to “average
coancestry” or “average kinship”, as used in some studies (e.g., Askew and
Burrows, 1983; Fries, 1994; Lacy, 1995), but other studies (or even in the same
studies) also use the terms “average coancestry” or “mean kinship” referring to
averages of other sets of values. As the term is used with different meanings by
various investigators, without being intuitively clear as to which values are to be
averaged, the term easily causes misunderstandings. Lindgren et al. (1996) coined
“status number” calculated from the group coancestry, which gives clear intuition
on the concept of effective number.
Objectives
The main aim of this thesis is to give a theoretical foundation for genetic aspects
of seed orchards and their management. Several factors, such as selection
intensity, genetic value, group coancestry, relatedness, fertility, selfing and gene
migration, are considered in estimating and predicting genetic quality of seed
orchard crops. Most emphases are placed on evaluating effects of fertility
variation. Much of the theory heads at evaluating methods for efficient orchard
management.
The following questions are addressed:
1. How can seed orchard managers predict and improve the genetic quality of
seed orchard crops and modify the balance between genetic gain and gene
diversity? (I)
2. If parents do not contribute equally, what are the effects and consequences of
unequal contribution in transferring genes to the planted forest? Is it possible
18
to predict genetic composition of the seed crop by the assessment of female
and male strobilus production? (II, III and V)
3. What are the implications of fertility variation in the management of seed
orchards? If the parental fertility is different between genders, what is the
consequence of correlation on the gene diversity of orchard crops and on the
management of seed orchards? (III and VI)
4. How large is the fertility variation in the real populations of forest trees or
clones, and what kinds of factors can have importance in the real
populations? (IV)
5. Why do fertility variation and ramet variation occur? (VII) and what are the
consequences of relatedness, selfing, parental number and pollen
contamination in transmitting genes from orchard parents to their offspring?
(VIII and IX)
6. How can breeders or conservationists decide the optimum tradeoff between
gene diversity (status number) and seed production by balancing and/or
controlling maternal fertility in seed collections? (X)
7. When there is relatedness and fertility variation among orchard parents, how
do these variables affect the gene diversity of seed crops? What is the effect
of gene migration from extraneous pollen sources? (IX and XI)
19
Materials and methods
The main purpose of this thesis was to develop breeding theory in seed orchards.
New theoretical methods were developed and documented in the included
publications. The methods and models were applied for various cases, using
empirical data and information on flowering and ramets from different species,
years and countries.
Flowering and ramet data
Numbers of female and male strobili were counted in Pinus densiflora and Pinus
thunbergii clonal seed orchards that are located on Anmyon island (36o30’N,
126o20’E, elevation 35 m), which is the western part of Korea. Flowering data
from the clonal archive of Pinus koraiensis were also collected for consecutive
five years from 1991 through 1995, which was established in Suwon (37o70’N,
127o20’E, elevation 100 m). All grafts originated from plus trees selected
throughout Korea. Strobilus production was assessed by counting all female
strobili, and by estimating the number of male strobili.
Flowering assessment was then used to estimate fertility and fertility variation
(II, III, V, VI and XI). Data collection has been done almost every year in
Korean seed orchards. Four-year observations were used in paper V, and fiveand six-year data in papers XI and VI, respectively. Fertility variation among
species was estimated and compared using a single year observation from three
seed orchards in different species in study II.
The term ‘fertility’ is defined as the capacity of an organism to produce living
offspring (Krebs, 1985). The notion of fertility is the actual level of performance
in the population based on the number of successful progeny, which is a measure
of physiological or genetic capacity. The similar term ‘fecundity’ is the potential
level of performance of the population, or physical capacity. In the present study,
the female and male fertilities are measured as the number of female and male
gametes produced by an individual relative to others in the orchard population
(Gregorius, 1989).
In study VII, we used orchard records from a large number of seed orchards in
Finland, Korea and Sweden, in order to estimate ramet variation and to predict its
effect on the gene diversity of orchard crops. All registered seed orchards of
Pinus koraiensis and Pinus densiflora in Korea, and Pinus sylvestris and Picea
abies in Finland were used to study variation in ramet numbers among parents.
Not all, but many Swedish seed orchards were also included. Unlike in Korea and
Finland, orchard records in Sweden are not kept centrally and orchards may have
several owners. Newer seed orchards must be registered at the Swedish National
Board of Forestry and this documentation includes ramet numbers. Many records
20
were available, either directly from the orchard owners, or via the Swedish Forest
Research Institute (SkogForsk).
Theoretical methods
1. Relatedness among seeds from seed orchards
Pedigree and relatedness among orchard crops from a seed orchard are illustrated
in Figure 2. Coancestry is a quantitative description of relatedness. AB means the
coancestry between individuals A and B, and it is 0 between unrelated individuals
(Fig. 2-b). In the case of self-coancestry, i.e., an individual’s coancestry with
itself, AA= 0.5(1+FA), where FA is the coefficient of inbreeding for the parent A.
When the parent is not inbred, the value of FA is zero.
Self-coancestry can also be understood as the group coancestry for a population
with a single member (Fig. 2-a), thus an extrapolation of the ordinary coancestry
concept. So, the group coancestry of a non-inbred individual (FA= 0) is 0.5.
A
a) self-coancestry
(A,A)=
0.5(1+FA)
(Y1,Y2)=0.25(AA+2AB+BB)
A
A
B C
D
Y
K
b) different parents
A
B
X1
X2
X1
Y
d) between self-sibs
e) self and outcross sibs
(X1,X2)= (A,A)= 0.5(1+FA)
(Y,Z)=0.25(AA+AC+AB+BC)
A
B
A
Y1
Y2
c) full-sibs
(Y,K)=0.25(AC+AD+BC+BD)
B
A
C
Y
Z
f) half-sibs
(X1,Y)=0.5(AA+AB)
Figure 2. Pedigree and relatedness (coancestry, ) among orchard seeds: a) selfcoancestry; b) unrelated seeds; c) full-sibs; d) self-sibs; e) double pairing of a parent;
and f) half-sibs. FA represents the coefficient of inbreeding for parent A.
In figure 2-d), it looks like self-sibs. If the seed parent A were not inbred,
relatedness between selfed seeds would be 0.5. It will be higher when the parent
is inbred. In some papers, it was assumed that the seed parents were unrelated
and non-inbred (II, III and VI), but this was not so in papers IX and XI.
21
Under the idealized situation of non-relatedness and equal contribution among
parents, the gene diversity of orchard crop will be maximized. Gene diversity is
reduced if there is relatedness or unequal fertility among orchard parents due to
the occurrence of sib relationships (IX and XI). This can be compensated for by
having more parents in the seed orchard, but then the gain may be reduced (III).
2. Fertility variation estimation
Each orchard genotype has a fertility value (pi), the ability to produce successful
gametes. Fertility can be regarded as the progeny mothered or fathered by an
individual relative to others in the population (Devlin and Ellstrand, 1990).
The “sibling coefficient” () is defined here and estimated directly from
observations as the sum of squared parental contributions multiplied by the
census number (see also Table 1). Expected fertility of an individual is an
average of maternal and paternal fertilities as follows (II and III):
 f  mi 
Ψ  N  p  N  i

i 1
i 1 
2 
N
2
i
N
2
2
where N is the census number, fi is the fertility as female and mi is the fertility as
male in a genotype i. The sibling coefficient () was denoted as A in papers I, II,
III and VII. Note that pi2 is the probability that two genes in the offspring come
from the same parent i, summed over all parents, i.e., the likelihood that two
gametes in the gamete pool originate from the same individual.
The sibling coefficient expresses how fertility varies among parents as the
increase in the probability that sibs occur compared to the situation where parents
have equal fertility (III). The sibling coefficient cannot be smaller than one. If 
= 1, all individuals have the same fertility. If  = 2, for instance, it means that the
probability that two individuals share a parent is twice as high, compared to when
the parental fertility is equal across the population.
Fertility differences among population members can also be described by the
coefficient of variation (CV) in fertility, so that
CV 2 ( N  1)
Ψ
1
N
3
where N is the size of sample (II and IV). When fertility is measured from all
population members (III and IX), then N disappears from formula [3] and the
sibling coefficient will hence be  = CV2 + 1. The sibling coefficient carries the
same information as the coefficient of variation, but  is based on a probabilistic
22
aspect, while CV is based on a variance aspect. Also,  relates to the orchard
crop or offspring, while CV refers to the parental population. Therefore,  may
have the potential to give genetic information for describing gene diversity of
orchard crops that is closer to the seed crop than is the CV.
Fertility variation occurs, not only among seed parents, but also among pollen
parents. Thus, the sibling coefficient for parental fertility can be divided into
female (f) and male (m) components as
N
 f  N  f i 2 CV f2  1 and
i 1
[4]
N
 m  N  mi2 CVm2  1
i 1
where CVf and CVm are the coefficients of variation in female and male fertilities,
respectively.
Fertility can also be described by a probability density function f(x), where x is
the percentile of trees ranked by fertility and f(x)dx= 1. For a large population of
trees, p= f(x)/N is the predicted fertility of the parent with rank corresponding to x
(II). An example of a density function is the power function, F(x)= xa for a≥1.
The derivative of the power function is f(x)= axa-1. For a given population, an
individual fertility (pi) can also be estimated in a relative sense, based on the sum
of observations. Parents are then ranked for the fertility and transformed to
cumulative contributions summing up to one. Cumulative contributions are
plotted against the parents’ proportions, and then the a value that gives the bestfitting curve is chosen (Bila, 2000).
The exponent of the fitted power function, a can also be expressed as a function
of  (II and X), thus there may not be direct need of the parameter (a) fitted to
empirical data because  is enough to specify a well-fitted power function.  is
related to the exponent a as

1 
a  Ψ 1  1  
Ψ 

[5]
3. Group coancestry and status number
Status number (Ns) is defined as half the inverse of group coancestry ()
(Lindgren et al., 1996).
23
Ns 
0.5

[6]
An attractive property of status number is that it is equivalent to the census
number for a population of unrelated, non-inbred individuals. The status number
describes an equivalent population of unrelated, non-inbred individuals, where
the probability to draw two genes identical by descent is the same as for the
population under study. It is thus regarded as an effective number (II).
Status number calculations were based, not only on relatedness among parents
and pollen contamination from extraneous undesirable sources (IX and XI), but
also on ramet variation (VII) and fertility variation (II and III). Relative status
number (Nr), which is the relationship between Ns and N, was calculated as Nr =
Ns/N = 1/ (II, III, VII and X). The relative status number relates to how the
seed orchard is functioning in terms of relatedness or fertility variation; thus, it is
a convenient way of describing the orchard situation from the management point
of view. When the relative status number is low, orchard managers should
consider some action to improve the efficiency of seed orchard.
The effective number of parents (Np), as a function of fertility variation ( ), was
calculated (III and VI) as
Np 
N
Ψ
[7]
By knowing the magnitude of fertility variation among genotypes that will consist
of a seed orchard, the census number can be chosen to achieve satisfactory gene
diversity (IV).
The effective number of clones, Nc (c.f., status number) was also calculated,
based on the variation of ramet numbers among clones that are unrelated and noninbred (VII), as
Nc 
ntotal
N
 ni
i 1

1
1

N
N n
 ri
 total
i 1
i 1
ni
[8]
where ntotal is the total number ramets in the seed orchard, and ni is the number of
ramets of clone i. When relative values (ri) are used, the effective number of
clones becomes the status number of orchard crops where parents are unrelated,
non-inbred and parental fertility is proportional to the ramet number (VII;
Lindgren and Mullin, 1998; Nikkanen and Ruotsalainen, 2000).
24
4. Symbols and definitions
The following list of symbols is used throughout the thesis (Table 1). More
complete definitions and rules of operation for seed orchard populations are
given in the individual chapters.
Table 1. List of symbols, terms and definitions of main parameters used in the study
Symbol
Term
Definition
G
genetic gain
i
selection
intensity
M
gene migration
Rate of immigrant genes from outside the seed
orchard. It is half the pollen contamination
N
census number
Total number of genotypes (in a seed orchard)
Ns
status number
Half the group coancestry (=0.5/; see Table 5)
Nr
relative status
number
Ns / N; proportion of the status number to the
census number
f
inbreeding
coefficient
F
(group)
inbreeding
coefficient
coancestry
Probability of alleles being identical by descent in
an individual; the chance that both homologous
genes in the same zygote are identical by descent
Average inbreeding coefficient among individuals
in a population

Change in the average breeding value (additive
genetic value) achieved by artificial selection
Number of standard deviations by which the
mean of the selected individuals exceeds the
mean of the base population
Probability that genes, taken at random from each
of the concerned individuals, are identical by
descent (= coefficient of kinship).

group
coancestry
Average of all possible coancestries between
individuals in a group, including self-coancestry

sibling
coefficient
Probability of two random alleles being identical
by descent in a set of gametes from the same
group. It is designated in some papers as A.
25
t
26
generation
Time, measured in generations. Initial members of
the source population (initial generation; t= 0) are
always assumed to be unrelated and non-inbred.
Summarized results and discussion
All publications included in this thesis have their unique results and
consequences on genetic gain and gene diversity of orchard crops, and on the
management of seed orchards. In this chapter, therefore, results from these papers
are summarized and discussed from various genetic and management viewpoints.
Genetic gain
Genetic gain is always connected to gene diversity, as artificial selection is to
genetic diversity. There are both short-term and long-term considerations in
practical tree breeding programs (Fig. 3). In the short term, forestry practices
should result in productive stands that are able to tolerate changing
environmental pressures for the duration of the rotation. The long-term concern is
the maintenance of reservoirs for genetic variability, which is needed for current
breeding populations, but still more important to meet future needs (Savolainen
and Kärkkäinen, 1992). In short-term breeding, genetic gain is usually maximized
in clonal forestry, family forestry and seed orchards. For long-term breeding
strategy, breeding and base populations should be managed for sustainable
genetic diversity.
Higher genetic gain
CF
Short-term
SO
Breeding pop.
Long-term
Base population
population
Base
Broader genetic diversity
Figure 3. Relationship between genetic gain and diversity on the pyramid of
improvement in any tree-breeding program. CF represents clonal forestry with a few
superior clones and SO represents seed orchards.
Genetic improvement can be defined as a process whereby genetic value is
improved while joint consideration is given to the genetic diversity of deployed
material. As shown in Fig. 1, genetic gain in seed orchards can be achieved at
different stages from plus-tree selection through roguing inferior parents and/or
crossing superior parents. Plus-tree selection itself gives some extra gain (maybe
27
1-2%) compared to natural stands, due to the avoidance of relatedness as the plus
trees are usually selected in different stands. In seed orchard management, there
are different techniques that may be used to increase genetic gain while retaining
gene diversity, including selective cone harvest, genetic thinning and/or a
combination of these techniques.
In study I, genetic gain and gene diversity of orchard crops from clonal seed
orchards were formulated considering selection intensity, genetic value, fertility
variation, pollen contamination, and inferiority of the alien pollen. Formulae
were derived to calculate genetic gain (also gene diversity) and applied to five
management strategies. The formulae and results of the study I can be used for
identifying favorable alternatives in the management of seed orchards. The main
results and discussion follow.
By truncation selection, the expected breeding value of the orchard crop (G) after
selection can be predicted as
G  ie A  MC


 0.5 i( N f , N )  (1  2M )i( Nm , N )  A  MC
[9]
where ie is the effective selection intensity, the average selection intensity applied
to female and male parents; A is the standard deviation of breeding values for
orchard clones; M is the gene migration, that is half the pollen contamination; C
is the contamination inferiority; and Nf and Nm represent the selected number of
female and male parents, respectively. If selection for general combining ability
is to be effective, clone or family differences must exist or, synonymously, the
additive variance (A2) must be greater than zero (Namkoong et al., 1966).
The predicted genetic gain is not affected by fertility variation if the breeding
value is not correlated with fertility. While Schimidtling (1981) reported that a
negative relationship between growth and flowering had a genetic basis in
loblolly pine, Byram et al. (1986) found that there was no significant correlation
between volume growth and cone production in loblolly pine seed orchards. A
positive correlation was observed between vegetative and flowering
characteristics in tamarind (Varghese et al., 2000). It seems that there is a trend of
positive correlation between crown size and flower production, and negative
correlation between tree growth and cone production (Nikkanen and
Ruotsalainen, 2000).
Plus-tree selection is mass selection; so seed orchard establishment itself gives
genetic gain. Here, the scale for breeding value is defined so that the average
breeding value of plus trees (thus initial seed orchard parents) is set to zero (I).
Therefore, the genetic gain for orchard crops from the initial seed orchard will
only be affected by gene migration (i.e., pollen contamination and its inferiority),
28
G = M*C. The influence of pollen contamination on the genetic value of the crop
is usually negative, thus the contaminating pollen has a negative breeding value.
1. Selective harvest
Seeds can be collected selectively, based generally on genetic values of the seed
parents. In the case of open-pollinated orchards, no selection can be applied to
the pollen parent. Genetic gain from the selective harvest will be
G  0.5i( N m , N ) A  MC
[10]
Higher gain can be achieved by block-planting seedlings from families with high
genetic values, rather than mixing all seeds from a given seed orchard (IX; Li et
al., 1999). The family block system redistributes the offspring of the orchard, but
it does not much reduce the genetic diversity found in the orchard because the
paternal contribution is uncontrolled (Duzan and Williams, 1988). Genes from all
parents can thus be represented by the use of half-sib families. Even if seeds were
collected from the single mother, the effective number (status number) of the
seeds would be close to four (IX). Risk of cataclysmic disaster from pests may be
slight because genetic variability within a given half-sib family block is still quite
large (Bishir and Roberds, 1999).
A variation on selective harvest is to collect equal numbers of seed from each
parent (I, III, X; Bila, 2000). When each genotype contributes equally to the
progeny as a mother parent, gene diversity will improve depending on the levels
of fertility variation and correlation between gender fertilities (VI). For the
complete equal seed harvest, however, the size of the collection will be
determined by the parent having the lowest fertility, resulting in a great loss of
seed production.
A tradeoff among genetic gain, seed production and gene diversity (status
number) can be achieved when constraints are made on the selected fraction of
ranked parents (II and X). By controlling maternal contribution, high status
numbers could be obtained, although accompanied by some loss of seed
production (X). For a tradeoff between seed production and gene diversity,
expected seed production (y) is based on the probability density power function
when the genotypes are ranked according to fertility, F(x)= xa as
xj
a 1
a 1
y   ax a1dx  ax j (1  x j )  x j  x0a  ax j (1  x j )
a
[11]
x0
where a is a parameter; x0 is the minimum contribution (lower bound), meaning
that seeds are not collected from parents with very low fertility because it is
29
regarded to be expensive; and xj is the maximum contribution of individuals on
ranked parental fertility (upper bound; fertility balancing). This results in an
equal amount of seed collected from parents with fertility higher than the upper
bound, to avoid over-representation of the most fertile ones (see Fig. 6).
If there is a strong negative correlation between breeding value and fertility of
individuals, this procedure is not attractive, as gain will be lost. With a positive
correlation, however, the balancing may give desirable genetic quality in the seed
crop.
2. Genetic thinning
In many seed orchard programs, orchard parents are evaluated in genetic tests and
those that prove to be genetically inferior are removed, i.e., genetic thinning or
roguing. In genetic thinning, selection is applied simultaneously to both seed and
pollen parents with the same intensity. Genetic gain by roguing only will then
give
G  (1  M )i( Nm , N ) A  MC
[12]
There is no practical way in which an orchard can be designed to provide for
roguing, as one does not know ahead-of-time when clones are to be removed. As
a consequence, true roguing tends to result in a somewhat uneven distribution of
trees in the orchard (VII). If the number of clones to be removed is small, in
comparison to the total number of trees that need to be thinned to provide
adequate spacing, the problem is not too serious. Usually, some holes and clumps
are unavoidable. Roguing is also a management technique that can improve the
genetic quality of seed and increase seed yield due to better crown development.
Most concern about the genetic thinning is the loss of genetic variability in
orchard crops after roguing, because roguing is a kind of family selection.
However, simulated genetic thinning in the orchard did not drastically reduce
genetic diversity (heterozygosity), even when as many as 50% of orchard clones
were rogued (I; Stoehr and El-Kassaby, 1997). Savolainen and Kärkkäinen
(1992) also reported that the effect of random genetic drift on depleting
heterozygosity in seed orchards was low. Most parents with high fertility can
produce well for successive years, but some parents do not (III, V; Goddard,
1964; Eriksson et al, 1973). It could thus be possible to control fertility during
roguing (Hodge and White, 1993) or seed collection (X), so that sufficient
genetic variability would be maintained.
The overriding factor in roguing a seed orchard is economics. The potential loss
in seed production needs to be balanced against the expected improvement in
quality (X). Roguing gives an initial seed loss, but in the long run the remaining
30
trees will produce more seeds and the accumulated seed loss may not be so large.
The best roguing solution may be obtained by weighting each clone according to
its seed production and the percent of superiority of its seed. For example, a clone
producing 100 kg of seed annually that is 10% better than nursery-run seed is
equal to a clone producing 200 kg of seed that is only 5% better. Both are inferior
to a clone producing 75 kg of seed that is 20% better. It varies with the particular
circumstances whether genetic quality of the seed is a more important goal than
its quantity, or vise versa.
3. Combined method
The former alternatives can be combined, i.e., genetic thinning first, followed by
harvesting the best clones. In this alternative, parents that serve as pollen source
are the total number of parents remaining after thinning. Cones are then harvested
from more restrictively selected parents with higher breeding values from the
remaining part. Selection thus occurs twice with different intensity. The first is
simultaneous selection against pollen and seed parents at the time of thinning, so
that the selection intensity is derived from the initial number of parents. The
second is more intensive selection against seed parents for the selective cone
harvest. The genetic gain is thus calculated from the average of selection as
G
i
( Nm ,N )

 (1  2M )i( N f , N )  A
2
 MC
[13]
Calculations of relative total gain from orchard management (or breeding
methods) may change greatly with the many factors (Namkoong et al., 1966). Not
all factors can be kept comparable for a fair comparison of the management
alternatives. Many factors, such as heritability, gender fertility correlation, cost
and time, should also be considered to decide which tactics would be more
effective in the management of seed orchards (e.g., Danbury, 1971). The
foregoing rates of genetic gain are based on the gain obtained at the end of an
operation or cycle, which are the figures most commonly of interest to tree
breeders. On the other hand, seed orchard managers may wish to determine the
value of the orchard on the basis of realized gain over the lifetime of the seed
orchard, as in Figure 1.
4. Genetic parameters
Genetic gain is, in a statistical sense, calculated from simple linear regression, y =
bx. The regression coefficient, b (e.g., heritability) is the ratio of the covariance
between the predictor variable x (e.g., selection differential) and the predicted
gain (y) to the variance of the predictor x as b = cov(x,y)/var(x). The covariance is
composed of two parts: genetic differences and sampling error (Namkoong et al.,
1966). It is the genetic covariance between the breeding value of parents and the
31
expected growth of trees grown from seed orchard seeds. The covariance between
them is, in general, controlled by purely genetic and G x E interaction effects.
Genetic models are often complicated; so the composition of cov(x,y) or var(x) is
not clear and hence the heritability is not interpretable accurately (see Namkoong
et al., 1966 for summarizing). The other errors in genetic testing are faulty or
inaccurate procedures.
In publications V and XI, the analyses of variance indicated that there were
statistically significant differences among orchard clones in flower production.
Broad-sense heritability estimates (H2) also implied that the reproductive output
attributes were under genetic control (Matziris, 1994; Schmidtling, 1983). The
correlations between female and male strobilus production were all weakly
positive for the four years studied in a Pinus densiflora seed orchard, but they
were positive or negative for five successive years in a Pinus thunbergii orchard.
It has been reported that there were no significant correlations between volume
growth and cone production (Skroppa and Tutturen, 1985; Byram et al., 1986).
The relationship between fertility and progeny growth rate may thus be too weak
to use as a criterion for genetic thinning (I). Parents with desirable genetic values
might be penalized. Schmidtling (1981) and Nikkanen and Velling (1987)
reported negative genetic correlations between flowering and growth. In their
studies, correlations between mid-parent fruitfulness and progeny growth rates
were negative, with the exception of a positive correlation with diameter. It is
generally suggested that taller trees allocate much of their energy to vegetative
growth rather than to reproductive growth, so the most fruitful parents may
produce slower-growing progeny. If there were indeed negative correlations,
roguing based on flower production would result in a loss of volume growth.
Sexual characters (e.g., fertility, fruitfulness) are peculiar traits in comparison to
other traits (e.g., growth rate, gum yield) that are usually included in a breeding
program. Seedlings with high seed production potential may not be needed for
commercial forest planting, because it is foreseen that the vast majority of
planted stands are to be replanted artificially (Varnell et al., 1967). Desirable
orchard trees should still be heavy seed producers. It may be speculated that seed
orchard environments and stand environments are different, so that trees
flowering well in seed orchard environments may not necessarily produce
progeny that flower well in the field. But, investigations are difficult to conduct
because of the long lifespan of trees. The best might be to have trees which are
fertile only in seed orchard environments and not otherwise, and this might be
possible with a particular treatment regime (e.g., chemical floral induction) in
seed orchards.
Recently, new selection methods have been proposed in which both breeding
values and genetic relationships among members in the selected population are
considered. Brisbane and Gibson (1995) regarded the balance between genetic
gain and relatedness (gene diversity) as important to achieve high genetic gain
32
with minimal inbreeding. Similarly, low group coancestry combined with
optimized genetic gain was proposed by Lindgren and Mullin (1997), and by
Olsson et al. (2000).
Gene diversity (status number)
Concerns over the reductionist nature of the domestication of forest trees focus
on the possibility of genetic erosion. There are several steps in the forest-tree
domestication process in which genetic variation could be reduced if sufficient
safeguards are not considered (Stoehr and El-Kassaby, 1997). The steps are
phenotypic selection, breeding (tree improvement cycle), seed and seedling
production. Systematic monitoring of the various steps of the domestication
process should be identified (X; El-Kassaby and Namkoong, 1995).
Phenotypic selection of plus trees has a little impact on the reduction of gene
diversity in seed orchard populations. Seed orchards harbor levels of genetic
variation that are similar to, or even greater than, those found in their naturalpopulation counterparts (Knowles, 1985; Cheliak et al., 1988; Bergman and
Ruetz, 1991; Chaisurisri and El-Kassaby, 1994; Williams et al., 1995). Other less
obvious sources of genetic erosion could occur during the seed and seedling
production phases. El-Kassaby and Thomson (1996) reported that the genetic
variation of seedling crops was reduced due to the unintentional, directional
selection through nursery management practices.
The status number of the orchard parental genotypes is the same as their census
number, if they are assumed to be unrelated and non-inbred. Status number and
group coancestry are used as measures of what happened since tree improvement
started. Of interest is a possible relationship among trees from various geographic
origins, and between the origin and the seed production of clones (XI; Goddard,
1964). The main criteria for plus-tree selection are to select one, or a few widely
separated, vigorous, healthy and well-adapted trees within a stand. The criteria
are hard to fulfill in practice, so that there could be a few relatives and some
relatedness among selected trees. However, such deviation of criteria could be
ignored because a large number of parents have generally been used in firstgeneration seed orchards.
In seed orchards, status number is best described as the status number of orchard
crops (Lindgren and Mullin, 1998). Status number of the seed orchard itself is
equivalent to the status number of the orchard crops only when parents are
unrelated, non-inbred and have equal fertility, and when there is no
contamination by outside pollen sources.
1. Relatedness and group coancestry
When orchard parents are unrelated and non-inbred, group coancestry of orchard
crops is calculated only from self-coancestry (i.e.,  = 0.5/N). Expected status
33
number of seed crops will thus be the same as the census number (N). Under such
conditions, group coancestry of orchard crops can also be calculated from the
contribution of parents (pi) as
N
  0.5 pi2
[14]
i 1
If all parents are assumed to be non-inbred, indicating that all self-coancestry
equals 0.5, but are related each other, ij, group coancestry becomes
 0.5  ( N  1) N N  Ψ
  ij 

i 1 i  j



N
[15]
Group coancestry (status number) depends on the size of the orchard population
(Fig. 4), since group coancestry includes the repeated sampling of the same gene
(Rosvall, 1999). Status number decreases as relatedness among parents increases.
Relative status number
1.00
N = 10
N = 20
N = 100
0.75
0.50
0.25
0.00
0.00
0.05
0.10
0.15
0.20
0.25
Relatedness between parents
Figure 4. The relationship between relatedness and relative status number depending
on the size of orchard population. Equal fertility and no pollen contamination are
assumed.
The decrease of relative status number is profound in orchards with many
parental genotypes (Fig. 4), reflecting that status number cannot be very large if
population members are related. At a given level of relatedness, orchards with
fewer genotypes will retain higher relative status number. Gea et al. (1997)
34
reported, using stochastic simulations, that breeding schemes with small,
disconnected groups were slightly more efficient in preserving status number
through a large number of generations than those with large groups. However,
when breeding programs are started with very small groups, or with status
number of less than 10, there will be lower expected genetic gain and potential
problems with inbreeding (VII and IX).
Group coancestry for the gamete gene pool of the seed crop is calculated based
on the contribution and relatedness of orchard genotypes, and pollen
contamination as (I; cf. Lindgren and Mullin, 1998)
N
N
i 1
j 1
   ( f i  (1  2M )mi )  ( f j  (1  2M )m j ) ij
[16]
where fi and mi are the female and male contributions of genotype i (sums up to
0.5, respectively); fj and mj are those contributions of genotype j; and ij is the
coancestry between genotypes i and j. Here, mi and mj refer to the male
contributions in the absence of gene migration (M).
Theoretically, the group coancestry () can vary from a minimum of one-eighth
the census number to a maximum of one. And thus, the status number (Ns) has
inversely the range from the maximum of 4N to the minimum of 0.5 as
0.125N <  < 1 and inversely 4N > Ns > 0.5
When there is one completely inbred genotype and only selfed seeds are collected
from it, the coefficient of inbreeding will be one. The group coancestry of selfed
seeds from the pure line will also be one and Ns = 0.5 (IX). If a tree mates equally
with unrelated and non-inbred pollen sources, the seeds are all half-sibs and the
group coancestry of a large amount of seeds from the same seed parent
approaches 0.125 and, hence, Ns approaches four (I, IX; Lindgren and Mullin,
1998).
2. Fertility and ramet variation
Maximum gene diversity of seed orchard crops can only be attained when all
parents contribute equally to the gamete pool (see Fig. 5 and Table 2). This
assumption is virtually never fulfilled and it is commonly observed that a small
portion of orchard parents contributes a disproportionately large amount to the
orchard crop (El-Kassaby et al., 1989a; El-Kassaby and Cook, 1994). The
unequal contribution leads to an accumulation of genetic relatedness and the loss
of gene diversity in seed crops (II, III, IV and XI).
35
Balanced and unbalanced contributions from parents in different orchard
populations are illustrated in Figure 5. This might also represent different
flowering years. The results are presented in Table 2. In population A, all parents
contribute equally, so there is no fertility variation and the census number of the
orchard is exactly the same as the status number of orchard crop. Due to unequal
gamete contribution, the status numbers of orchard crops from B and C decrease
29% and 47%, respectively, when compared to the parental census number
(Table 2). Population C shows the typical unequal contribution that was found in
study IV ( = 2: CV = 100% in fertility).
Orchard population A
(equal contribution)
Orchard population B
(unequal contribution)
Orchard population C
(typical contribution)
Figure 5. An illustration of the parental gamete contribution in seed orchard
populations with five unrelated individuals
Table 2. Fertility variation (CV, ), status number (Ns) and relative status number (Nr)
for the populations illustrated in Figure 5 where parents are unrelated and non-inbred
Orchard pop.
No. of trees
CV

Ns
Nr
A
5
0
1
5
1.00
B
5
0.63
1.4
3.6
0.71
C
5
0.95
1.9
2.6
0.53
In paper III, an effective number of parents is proposed, based solely on the
fertility variation. Effective numbers of seed and pollen parents can be estimated
separately (VI; Roberds et al., 1991). Inherent variation in male flowering is
large, but the effects are seldom quantified. Male strobilus production is more
difficult to assess than female strobilus production, because catkins are numerous
and usually do not persist after pollen shedding (Schmidtling, 1983).
If there is no correlation between maternal and paternal fertilities, the sibling
coefficient () in an orchard is calculated with pollen contamination (M) as (III,
VIII; cf. Nikkanen and Ruotsalainen, 2000)
36
 f  (1  2M )mi 
Ψ  N  p N   i

i 1
i 1 
2

 0.25 f  (1  2M ) 2 m   0.5(1  2M )
N
2
i
2
N
[17]
where N is the number of genotypes contributing to the gamete pool; pi is the
contribution of genotype i; f and m are the fertility variations for female and
male parents, respectively; and M is the gene migration due to pollen
contamination (i.e., half the pollen contamination).
When there is a correlation between female and male fertilities, the correlation
coefficient (r) should be considered for estimating total fertility variation in the
seed orchard (VI). The sibling coefficient () can be calculated as
Ψ
 f  (1  2M ) 2 m  2(1  2M )  2r (1  2M ) ( f  1)( m  1)
4(1  M ) 2
[18]
If there is a perfect positive correlation (r = 1) between gender fertilities, the
sibling coefficients for maternal, paternal and total fertility all are the same (  =
f = m). On the other hand, if there is a perfect negative correlation (r = -1), the
sibling coefficient for total fertility equals one ( = 1) and f = m.
Most studies have shown that male and female flowering are weakly, positively
correlated (II, V; O’Reilly et al., 1982; Schoen et al., 1986; Kjær, 1996;
Nikkanen and Velling, 1987; Nikkanen and Ruotsalainen, 2000), or not related
(III, XI; Savolainen et al., 1993). Negative genetic correlations between female
and male flowering have been found in slash pine (Schultz, 1971) and in Scots
pine (Savolainen et al., 1993). This supports the central assumption of the theory
of sexual allocation: given a fixed total quantity of resources available, there
should be a negative correlation between female and male allocation. This
negative relationship could also be an important consideration in seed orchard
management; if the parents are rogued with strong intensity, some of the best
pollen sources may be eliminated (Danbury, 1971; O’Reilly et al., 1982;
Schmidtling, 1983). This in turn would narrow the genetic base and perhaps
reduce the contribution of some of the better genotypes.
One way to reduce the loss of gene diversity due to fertility variation is to restrict
the parental contribution to the next generation (VIII; Wei, 1995; Bila, 2000).
This restriction is more likely to be applied to the maternal contribution. By
keeping the contribution equal among genotypes, genetic relatedness is
minimized and gene diversity is maximized in the orchard crop (Lindgren et al.,
1996). When flower phenology is a concern, removing parents with unusually
early or late flowering from orchards may also be an effective way to maintain
37
high genetic gain and gene diversity, and a means to reduce pollen contamination
and selfing (Xie and Knowles, 1994).
Equal seed harvest is often proposed to mitigate the effect of unbalanced
contribution among parents in stands and seed orchards (III, VI, X; Schmidtling,
1983; Bila, 2000). Equal seed harvest means that female fertility is kept constant
and thus no correlation between female and male fertilities. When an equal
amount of seed is collected from each parent, the sibling coefficient is a function
of male fertility variation (m) and pollen contamination (M) as
Ψ 
(1  2M ) 2 m  4M  3
4(1  M ) 2
[19]
In paper VI, I evaluated how much better was equal seed harvest, compared to
seed collection where both maternal and paternal fertilities vary. The magnitude
of equal seed harvest depended on fertility variation and correlation strength
between genders. In general, it was concluded that collecting the same amount of
seeds across the population reduced considerably relatedness of seed crops, and
this option was an effective way of maintaining high gene diversity. Equal seed
collection among parents could also be considered to apply in germplasm
collection and when establishing stands for ex situ gene conservation (X).
Sexual asymmetry was investigated in study V by means of a maleness index, Mi
= mi/(fi+mi) (cf. functional gender in Devlin and Ellstrand, 1990). A maleness
value near 0.5 indicates that female and male fertilities are nearly equal, while a
value near one means that the individual transmits genes to the seed crop
predominantly through pollen. Femaleness equals 1-Mi. It was shown in study V
that orchard parents commonly differed in their genetic contributions to seed
crops through pollen and ovules, so that fertility estimates based solely on one
gender might be misleading.
As mentioned earlier, harvesting or mixing equal amounts of seeds from each
parent is suggested as a means to mitigate the imbalance of gamete contributions
among parents (III, VI, X; Bila, 2000). Completely equal seed harvest will
discard most of the crop, because equalizing of female fertility will be set to the
least-fertile parent, and translate to a substantial loss of seed production. It is, of
course, impossible to apply when there is an individual with zero fertility. In
study X, a tradeoff between seed production and gene diversity was achieved by
equalizing maternal contributions from over-represented parents and discarding
seed from the least-fertile parents. By controlling female fertility, high gene
diversity could be obtained while an acceptable amount of seed was produced.
Revised results from study X are shown in Figure 6. When all seeds were
collected, the status number was 149. By balancing with constraints of x0= 0.1
38
and xj= 0.9, the status number increased as 183 and the loss of seed production
was only 13% (i.e., 87% of the total crop was collected). This balancing between
gene diversity (measured by status number) and seed collection will be more
important in genetic resource conservation programs.
0.025
Observation
Balancing
Individual fertility
0.020
All seed production
x 0=0 & x j =1
0.015
Balancing
x 0=0.1 & x j =0.9
0.010
0.005
0.000
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
Fraction of ranked parents (x 0 & x j )
Figure 6. Tradeoff between gene diversity and seed production by controlling female
fertility in the collection of Corylus avelana L. from native Danish populations.
Balancing improved fertility variation and thus increased status number.
Discarding seed by constraints will increase Ns, but may decrease the likelihood
of retaining rare alleles present in the population. Loss of gene diversity caused
by selection is quantified by heterozygosity or reduction in allele numbers
(Allendorf, 1986). However, the number of alleles remaining after selection is
important for the long-term response, while heterozygosity provides a good
measure of the response in diversity to selection immediately (see formulae 23
and 24).
Fertility variation is higher in stands than in seed orchards (IV). Differences in
fertility were usually higher during poor flowering years and in young
populations. Forecasts of the effect of tree breeding and conservation operations
require information on fertility variation. In most cases, the information does not
exist or is highly unreliable. For predictions concerning objects that are neither
juvenile nor characterized by poor flowering, it is suggested in paper IV as a
rough generalized heuristic rule that  equals 2 (CV in fertility = 100%) for seed
orchards and  = 3 (CV = 140%) for stands. These values are a little higher than
values actually indicated by observations, but after truncation they come out
39
about right. That they are a little high may mean an extra safeguard for gene
diversity, which seems prudent. They are recommended for use in calculations
where other information does not exist (probably in most cases). Generally
speaking, therefore, the effective number is about 150% larger in seed orchards
than in stands of the same size.
In seed orchards, trees are widely spaced with well-developed crowns and limited
height. Genetic thinning is done to eliminate families or clones with low genetic
values and those with low flowering and fruiting abilities (Varghese et al., 2000).
Topping and pruning are performed to maintain a short, wide crown, encouraging
the growth of lateral branches and thus increasing the number of potential flower
and seed production sites (Ho and Schooley, 1995). Irrigation, fertilizers and
plant growth regulators may be used to induce and enhance flowering (Setiawati
and Sweet, 1995; Owens, 1995). The effects of these practices are likely to make
trees more similar, which correspond well with the lower fertility variation in
seed orchards compared to stands (IV).
The sibling coefficient is increased with unbalanced parental contribution to the
progeny. When the sibling coefficient equals 2, it means that relatedness and
inbreeding of seed crops will build up over generations twice as fast as the
reference population (VIII). The “20:80 rule” is often quoted (Anon., 1976) to
describe the imbalance, asserting that 20% of trees in a seed orchard often
produce 80% of the cone crop. This rule has been proven in many orchard species
(Eriksson et al., 1973; Griffin, 1982; Schmidtling, 1983; El-Kassaby et al., 1989a;
El-Kassaby and Reynolds, 1990). Understanding such imbalance has implications
for selective treatment, prediction of realized genetic gain, expected gene
diversity of orchard crops, and crop prediction models (Byram et al., 1986).
Parental balance curves have also been used widely (V; Griffin, 1982; ElKassaby and Reynolds, 1990; Adams and Kunze, 1996). They are informative in
determining the present contribution of the various clones to either female or
male strobilus production, and in identifying the high and low producers (ElKassaby and Askew 1991).
Besides fertility variation, the variation in the number of ramets among parents
will also affect the genetic composition of orchard seeds. In study VII, a survey
of 255 seed orchards indicated that there was a considerable ramet variation
among clones. The coefficient of variation (CV) in ramet number was about 60%.
The average size of orchards was about 12 ha and the average total number of
ramets was 3,798 per orchard (Table 3). Thus, orchards had an average density of
approximately 320 ramets per ha. On average, the surveyed orchards were 23
years old.
Table 3. Average clone number (N), total numbers of ramets, CV for ramet number
and relative effective number (Nr= 1/) in first-generation clonal seed orchards
40
Orchard
location
Species
Number
of SO
N
Finland
P. sylvestris
176
139.0
4,137
0.469
0.81
Picea abies
25
75.4
3,182
0.602
0.74
P. koraiensis
13
70.4
2,689
0.852
0.60
P. densiflora
8
94.0
5,454
0.752
0.65
P. sylvestris
22
79.8
4,614
0.503
0.80
Picea abies
11
71.2
2,713
0.382
0.86
88.3
3,798
0.593
0.74
Korea
Sweden
Average
Number of CV for ramet
ramets/SO
number
Nr
Ramet variation found in study VII was mainly caused by factors such as graft
availability, incompatibility, graft replacement, and biotic or abiotic injuries.
Unusual weather can significantly delay a seed orchard program and also cause
variation in numbers of ramets (Byram et al., 1998). Moreover, intentional
variation often occurs in later stages. As seed orchards get older and their
associated progeny test data become available, further reduction in numbers of
parents is expected through the genetic thinning of poor parents. There were large
differences in the numbers of clones and ramets between unthinned and thinned
orchards, and also between young and mature orchards (VII).
The census number of clones (N) in surveyed seed orchards was generally high
and could give information on gene diversity. The effective number (Ns) will be
more informative in advanced-generation seed orchards where parents will be
more closely related. Also, the effective number estimated from ramet and/or
fertility variation can probably give information for the deployment of genetically
improved stock for capturing high gene diversity (maybe also gain) and the
development of seed-trade guidelines. It is thus recommended in study VII to use
the concept of effective number of parents for the certification of seed orchards
and their crops following OECD scheme.
When estimates of orchard gain are based on tests of open-pollinated progeny, it
is often assumed that parental fertilities are equal and that pair-wise mating
probabilities are identical. However, the violation of such assumptions,
particularly where a small number of parents contribute most of the gametes to
the seed crop may lead to inaccurate estimates of gain (Schoen and Stewart,
1987). Under such a system of mating, the actual response to selection may also
be higher or lower than expected, even when accurate estimates of GCA have
been obtained (Griffin, 1982).
41
In mature orchards, seed production is more evenly distributed among parents in
good or intermediate crop years (V; Eriksson et al., 1973; Byram et al., 1986;
Kjær, 1996). Maximum panmixia is thus more readily achieved in good or
moderate flowering years than in poor years (O’Reilly et al., 1982; El-Kasaby et
al., 1989a; Chaisurisri and El-Kassaby, 1993). It was also found for the Pinus
densiflora seed orchard reported in study V that parents contributed more equally
to the seed crop and that status number was higher in moderate flowering years.
Parental ranks also varied year-to-year, mainly among parents with intermediate
fertility.
Variation in fertility has important implications in breeding (Griffin, 1982; Xie
and Knowles, 1992; El-Kassaby, 1995) and conservation programs (Sedgley and
Griffin, 1989). Differences in gamete contributions among orchard parents
influence the genetic composition of offspring by over-representing the most
productive genotypes (V; Kjær, 1996). Fertility variation represents fertility
selection that changes gene frequency and reduces effective number and viability.
Genetic drift and inbreeding will occur more rapidly than expected from the
census number. Special attention should thus be given to fertility variation in
order to avoid rapid genetic erosion in gene conservation or breeding programs.
Fertility in some studies was estimated by counting the number of reproductive
structures; however, there may be other factors, such as separation distance,
phenology, pollen dispersal, genetic incompatibility and local weather conditions,
playing an important role, especially for the male fertility. The effect of the
localization of individual clones on the selfing and genetic composition of
orchard crop is generally neglected (Gömöry et al., 2000). In seed orchards,
repetition of parents with several times (often 30-40 times in clonal seed
orchards, VII) and differences in floral phenology are supposed to overlay the
effect of distance among mates as a determination of mating success (El-Kassaby
et al., 1984; Askew, 1988; Adams and Birkes, 1991; Matziris, 1994). Separation
distance may significantly affect male mating success in Douglas-fir seed
orchards (Burczyk and Prat, 1997). If there is high proportion of mislabeled or
incorrectly planted individuals, the distance might become an even more
important factor (Gömöry et al., 2000).
There have been many reports that there is a positive correlation between the
number of seed cones and the proportion of filled seeds (Reynolds and ElKassaby, 1990; Chaisurisri and El-Kassaby, 1993; El-Kassaby and Cook, 1994;
Kjær and Wellendorf, 1997), and that pollination success is a positive function of
flowering production (Schoen and Stewart, 1987; Allison, 1990). Schmidtling
(1983) reported that the genetic correlation between female flowers and sound
seeds were strongly positive, indicating that clones with produced more female
flowers were the ones that produced more seeds. The simple counting of numbers
of strobili can thus give a reasonable indication of the unequal contribution
among parents in seed orchards (II, III, IV and XI), notwithstanding that heavy
42
flowering may not always result in a heavy seed crop, especially in natural stands
(Koski, 1980a).
Synchronization of pollen shedding and flower receptivity is important for
gamete pairing and influences the impact of foreign pollen, especially in seed
orchards with large numbers of parents (Askew, 1988; El-Kassaby et al., 1984).
On the other hand, the effect of reproductive phenology on the genetic
composition was found to be negligible in black spruce (O’Reilly et al., 1982)
and radiata pine seed orchards (Griffin, 1982). Variation in fertility within all
investigated orchard populations could be attributed to genetic and environment
effects, since both are confounded, although reproductive traits (e.g., flowering,
fruiting, seed production) are generally moderately strong genetic control (V, XI;
Schmidtling, 1983; Byram et al., 1986; El-Kassaby and Cook, 1994; Fries, 1994;
Kjær, 1996).
Fertility variation can be studied by marker genes, but such techniques have not
been sufficiently accurate, nor have data sets been large enough for quantitative
analysis. If a clone at one locus carries one or two alleles unique or rare to the
orchard, it may then be possible to trace the contribution of that particular clone
to the seed crop (Rudin and Lindgren, 1977). In the absence of such unique
alleles, the use of open-pollinated seeds collected from a clone and counting the
plants arising from these can document cross-pollination (Kormut'ák and
Lindgren, 1996). In addition, information on the occurrence of selfing can be
obtained.
Hormonal treatments (e.g., gibberellins) are recommended when reproductive
induction and synchrony are fundamental prerequisites for the production of
seeds reflecting the gene diversity represented in the orchard (Ross, 1978; Pharis
et al., 1980; Ross et al., 1985; Wheeler et al., 1985). Some hormones (e.g., GA 3)
have a negative or no affect on male flowering (Chalupka, 1980; Eriksson, 1996),
depending on environmental conditions and hormone concentration. Physical
treatments (e.g., root pruning, girdling) are sometimes recommended, but these
would often harm the treated trees for long-term seed production (Eriksson,
1996).
Supplemental mass pollination (SMP) is defined as the broadcast application of
pollen to non-isolated female strobili (Bridgwater and Trew, 1981). SMP has
been proven to be a vital tool for improving reproductive success, on average of
20%, due to the reproductive phenology displacement between male and female
strobili, and more uniform parental contribution in Sitka spruce (Franklin, 1971a;
El-Kassaby et al., 1989b; Chaisurisri and El-Kassaby, 1993) and Scots pine seed
orchards (Yazdani et al., 1986; Eriksson, 1996). SMP can produce more and
better-filled seeds compared to those produced by normal wind pollination (ElKassaby and Reynolds, 1990). An overhead cooling system with SMP on earlyand late-reproductive phenology classes in moderate and good flowering years is
43
effective in disrupting the synchrony, in minimizing pollen contamination and
inbreeding, and in reducing seed loss due to insect damage in Douglas-fir seed
orchards (Fashler and Devitt, 1980; El-Kassaby and Ritland, 1986b; Fashler and
El-Kassaby, 1987; El-Kassaby et al., 1990; El-Kassaby and Davidson, 1990).
3. Pollen contamination
Pollen contamination is a main source of gene migration into seed orchards,
which is probably the most serous impediment to the genetic quality of orchard
crops (Adams and Birkes, 1989). It is hard to avoid pollen contamination in
practical orchard management, even though wide isolation strips are often used to
reduce background pollen (Wheeler and Jech, 1986). Gene dispersal through
pollen influences levels of inbreeding and effective population size (Adams et al.,
1992). Numerous studies have been concluded that considerable pollen
contamination occurs in seed orchards (Table 4). It seems that many of the seeds
from seed orchards will have unknown fathers from outside the orchard.
Although contamination levels are high, the accuracy of many estimates is in
doubt (Stewart, 1994). For estimating background pollination, it is evident that
orchard size, shape, pollen production and phenology must be important. For
example, Yazdani and Lindgren (1991) reported that pollen contamination was
higher in the windward part of the orchard and lower in the center. Understanding
the underlying process of pollen movement from outside to seed orchards is also
important (Di-Giovanni and Kevan, 1991).
Levels of pollen contamination are of concern because seed resulting from
fertilization by alien pollen is expected to have lower genetic gain than seed
fertilized by orchard pollen (I). The contamination affects only the male side, and
thus considerably more than half of the gene in the orchard crop will have a seed
orchard origin. Low levels of pollen contamination could actually increase gene
diversity of orchard crops (IX; Lindgren and Mullin, 1998), but high levels could
either increase or decrease gene diversity, depending on the source of
contamination (Adams and Kunze, 1996). Reduction in genetic gain may even be
worse if contaminating pollen comes from trees that are poorly adapted to the
intended planting sites. Orchard managers should have a better understanding of
how various parameters such as orchard size, age, composition and degree of the
isolation influence levels of contamination, and the degree to which pollen
management tactics are effective in limiting pollen contamination (Wheeler and
Jech, 1986).
Table 4. Estimates of background pollination (%) in seed orchards
Species
Larix decidua
44
Background
pollination
5
References
Gömöry and Paule, 1992
Larix europaea x L.leptolepis
1.0-6.7
Häcker and Bergmann, 1991
Pinus elliotii
83.5
Squillace and Long, 1981
Pinus sylvestris
2
38
21-36
26-33
30
6-20
69-74
11
48
26
36-60
Müller-Starck, 1982
Nagasaka and Szmidt, 1985
El-Kassaby et al., 1989b
Harju, 1995
Wang et al., 1991
Pakkanen and Pulkkinen, 1991
Yazdani and Lindgren, 1991
Paule and Gömöry, 1992
Harju and Nikkanen, 1996
Friedman and Adams, 1981
Friedman and Adams, 1985b
Picea abies
10-17
Paule et al., 1993
Picea mariana
31.4
Caron, 1987
Pseudotsuga menziesii
15
5-15
29-52
0.2
12
22-56
50
49
34 a
Fashler and Devitt, 1980
Clare, 1982
Smith and Adams, 1983
El-Kassaby and Ritland, 1986a
El-Kassaby and Ritland, 1986b
Wheeler and Jech, 1986
Adams and Birkes, 1989
Adams et al., 1997
Xie et al., 1991
Pinus taeda
a
) clone bank. Note that the author has not corrected all observations to make them
comparable and accurate.
The main factor that detracts from the usefulness of seed from orchards is
background pollination (Pulkkinen, 1994). When migrating pollen contributes
effectively to pollination and fertilization, it may implicate some genetic
consequences in the adaptation process of forest trees (Koski, 1970; Lindgren,
1975). If the seeds are hybrids pollinated by the surrounding stands, the resulting
seedlings may have poor chances for survival in the areas from which the parents
originate because adaptation depends on the genotypes of both parents (Mikola,
1993). Stoehr et al. (1994) point out that if pollen contamination is not prevented,
the faster-growing seedling sired by background pollen may be preferentially
selected during culling in the nursery and out-planted on sites to which they are
maladapted. Pollen contamination in seed orchards involves a long-distance
transport of pollen (Chalupka, 1998). The outside pollen may also influence
genetic variation in the seed-orchard progeny (I; Müller-Stark, 1982).
45
In general, some female flowers are receptive before the tree sheds pollen.
Female receptivity often occurs before pollen shedding in young seed orchards.
In this case, most pollen comes from outside the seed orchard. On the other hand,
Yazdani et al. (1995) reported a situation where pollen production in the orchard
commenced a little earlier than in surrounding stands, because more sunshine
reaches the ground in a seed orchard as the canopy is more open and the trees
shorter in the seed orchard than in stands. There may be insufficient pollen at
early dates and female strobili may remain receptive until the orchard pollen
production commences. At later dates, the surrounding pollen cloud from local
sources may be more intense relative to the seed orchard pollen cloud. Harju and
Nikkanen (1996) reported that background pollination varied during the
pollination season.
The genetic quality of seeds from orchards is increased by genetic thinning or by
establishing advanced-generation seed orchards, but the full genetic potential is
never reached because of pollen contamination (Bridgwater et al., 1998). DiGiovanni and Kevan (1991) estimated that a genetic gain of 8% from selection
was reduced to 6% in the next generation if 50% of the pollen originated outside
the orchard. Due to pollination from foreign pollen, the number of alleles at
allozyme loci in orchard crops is clearly higher than in the orchard parents
(Harju, 1995).
Numerous reports, for a variety of conifer species, indicate that even in mature
seed orchards with heavy pollen production, levels of pollen contamination could
exceed 30-40% (Table2). Furthermore, the relationship between levels of pollen
contamination and within-orchard pollen production is not clear. Smith and
Adams (1983) predicted that as the orchard matured and pollen production within
the orchard increased, pollen contamination would diminish in a Douglas-fir seed
orchard. This is not the case in Scots pine seed orchards (Harju, 1995; Pakkanen
and Pulkkinen, 1991).
Contamination by extraneous pollen should be avoided. Possible methods used to
reduce the contamination include:



46
geographical isolation by establishing orchards at the margin or slightly
outside range of species (Sarvas, 1970; Hadders, 1972) or at a different
elevation (Silen, 1963);
reducing the frequency of outside pollen by the use of SMP (Denison, 1973;
Bridgwater and Trew, 1981; El-Kassaby and Ritland, 1986b), establishing
larger orchards (Wright, 1953), and/or surrounding orchards by isolation
zones (Wright, 1953; Squillace, 1967); and
physiological isolation through phenological manipulation, e.g., by the use of
water-spray cooling treatments (Silen and Keane, 1969; Fashler and Devitt,
1980; El-Kassaby and Davidson, 1990).
There are two main methods of estimating levels of background pollination. In
the “migration model”, migration changes the allelic frequency of the recipient
population towards that of the source population. If there are appreciable
differences between the source and recipient population, the migration can be
estimated based on neutral genetic markers or progeny tests. In the “finite
population model”, the finite orchard population produces only a limited number
of all the possible gamete genotypes, while the infinite surrounding populations
are expected to produce all possible combinations of alleles present (Savolainen,
1991). Pollen genotypes that cannot be produced in the seed orchard are detected
as background pollination.
For contamination levels in seedling seed orchards, estimation based on
differences among gene frequency estimates derived from single-locus
outcrossing pollen, ovule and outside seed is suggested by El-Kassaby and
Davidson (1990), and Wheeler and Jech (1992). In clonal seed orchards where
the number of within-orchard genotypes is detected by the number of clones, the
method that relies on detecting unique multilocus gametes is recommended by
Smith and Adams (1983). Stoehr et al. (1998) evaluated pollen contamination and
SMP efficacies using a genetic marker on the paternally inherited chloroplast
genome, and proposed to use DNA makers for assessing orchard mating
dynamics and management efficacies in Douglas-fir seed orchards.
If significant differences among maternal (ovule), paternal (pollen), and progeny
pools are observed, it indicates that the pollen pool includes some pollen from
outside the orchard. On the other hand, if there are no significant differences
among the three gene pools, it implies similarity between the allelic frequencies
within the orchard population and in the outside pollen source. This similarity
makes the detection of contamination based on unique alleles or differences in
allele frequencies very difficult (El-Kassaby et al., 1989b).
4. Different sex ratio
While most animal species reproduce by outcrossing and have separate sexes,
plants display a wide variety of sexual systems ranging from hermaphroditism
and complete self-fertilization to dioecism and complete outcrossing
(Kärkkäinen, 1994). Avoidance of self-fertilization is the key factor shaping the
evolution of reproductive structures in plants (Darwin, 1877). Darwin believed
that separate sexes helped to avoid self-fertilization and sexual allocation might
also be an important factor. That is evidently true in forest tree populations and
maintains high gene diversity. Most economically important tree species have a
mixed mating system, predominantly outcrossing with some selfing, which makes
the assay of gene diversity difficult.
In papers I and IX, formulae were derived to calculate gene diversity (status
number) based on sex ratio, pollen contamination and fertility variation in seed
47
orchard populations. The status number (Ns) for the seed crop is described,
considering the number of female parents (Nf), male parents (Nm) and parents that
are functioning as both female and male at the same time (Nfm) by
NS 
4N f Nm
N m  (1  2M ) 2 N f  2(1  2M ) N fm
[20]
This is equivalent to the effective number given by Frankel and Soulé (1981) and
by Falconer and Mackay (1996) when there is no gene migration, i.e., the status
number becomes Ns = 4*Nf*Nm/(Nf + Nm) in dioecious species. If the sexes are not
equal, Ns is less than N. On the other hand, Ns is the same as N when the sexes are
all equal (Nf = Nm = Nfm), as under an idealized situation. Note that Nfm cannot be
larger than the smaller of Nf and Nm.
An especially important case is that in which the types of parental numbers are
divided as seed parents (Nf) and alien fathers, i.e., Nfm and Nm are disregarded,
meaning that orchard crops are pollinated entirely by contaminating pollen. For
this situation, the status number becomes Ns = 4, assuming that contaminating
pollen is related neither to the orchard pollen nor to itself, and contributes equally
to the seed parents (IX; Lindgren and Mullin, 1998).
The amount of genetic drift in a seed crop where the sex ratio of the parents is
skewed is higher than that for a crop where the parental sex ratio is balanced (I).
From the sex-ratio standpoint, Ns is the size of an ideal population having a 1:1
sex ratio, which is subject to the same degree of genetic drift as a real orchard
population. In other words, the amount of random genetic drift in the orchard
population size of N is equal to that in an ideal population of size Ns, where
genders are represented equally and mated randomly. It should be noted that
inbreeding or genetic drift depends mainly on the numbers of the less-numerous
sex (Caballero, 1994). For instance, if an orchard were maintained with an
infinitely large number of pollen parents (Nm  ) but only one seed parent (Nf =
1), the status effective number would be only four (I and IX).
Concerning the number of alleles, allelic diversity is less sensitive to unequal
number of sexes. Let’s recall the case of one seed parent (Nf = 1) and 100 pollen
parents (Nm = 100). The loss of gene diversity in this case is estimated to be
12.5% relative to the reference, which is the same as an ideal population
consisting of four individuals (formula 23). What proportion of allelic diversity
will be lost then? This will depend on the number of seed and pollen produced
(Allendorf, 1986). A large number of seed crops are produced from seed
orchards, so each of the father parents has the opportunity to contribute to the
offspring. Thus, there may be a little or no loss of allelic diversity in the progeny.
Low frequency alleles may not contribute much to the tree improvement (genetic
gain and diversity) in seed orchard programs.
48
5. Over generations
Seed orchard programs follow progressive development as each generation
produces new parents for the next generation (El-Kassaby and Askew, 1998). In
some breeding programs, the seed orchards can function as the sources of bred
materials. In such breeding programs, care must be more taken to minimize the
problems associated with small effective populations while continuing to achieve
an increase in genetic gain. Many factors such as selection criteria of plus trees,
arrangement of trees, number of parents, reproductive success and management
practices contribute to the realized levels of genetic gain and gene diversity that
are achieved in the seed crop (I) and subsequent reforestation stock.
An increasing in inbreeding within the population during successive generations
of recurrent selection is potentially a major problem in long-term breeding
programs (Gea et al., 1997). Inbreeding appears one (if selfing is allowed) or two
(if not) generations later than genetic drift caused by the initial selection
(Caballero, 1994). Breeding and conservation programs that use seed orchard
populations with a low status effective number may lead to a loss of gene
diversity in the plantations (VIII; Lindgren et al., 1996).
Coancestry of the parental generation becomes inbreeding in the offspring
generation if they mate at random (Falconer and Mackay, 1996). In other words,
the group coancestry of orchard parents will be the expected group inbreeding of
orchard crops following random mating as (III, VIII; Bila, 2000)
 parents  F̂crops
[21]
In study VIII, an accumulation of group coancestry was monitored over
generations. Predictions over five generations showed that group coancestry and
inbreeding accumulated quickly at early generation shifts (VIII; Bila, 2000). In
the study, the link between the generations was the gamete; hence, the gene pool
of the offspring (t) was identical to that of the successful gametes of the parents
(t-1). The group coancestry of the gamete gene pool can be described as a
function of parent inbreeding (F), fertility variation () and census number (N)
as
 gametes 
0.5(1  Ft 1 )Ψ t 1
N t 1
 Ψ  N   0.5(1  Ft 1 )
 1  t 1  t 1 t 1
N t 1  1
 N t 1 
[22]
Relative status number declined very quickly over generations. The increase in
inbreeding and group coancestry was accelerated by fertility variation, and the
accumulation was faster and greater when fertility of both genders varied than
49
when maternal fertility was kept constant. Gene diversity was decreased faster
when fertility variation was large, but was maintained higher when the orchard
population size was reasonably large (VIII). Thus, breeding programs that use
small status numbers will lead to the accumulation of inbreeding and group
coancestry, and do not provide for a sustainable long-term breeding strategy
(Yeh, 2000).
Group coancestry increased and gene diversity decreased at an almost constant
rate over generations, even if there was high fertility variation. Status numbers
dropped remarkably at the first-generation shift (VIII; Bila, 2000), so Ns seems to
be less suitable for describing the loss of diversity over generations, especially in
long-term breeding populations (Rosvall, 1999). The most appropriate application
could be for situations (e.g., seed orchards) concerned with an actual state rather
than the rate of change (Lindgren and Mullin, 1998), and in situations where the
loss of rare alleles should be emphasized as in small-endangered populations
(Ballou and Lacy, 1995).
Interpretation of status number
Status number is a convenient way of expressing group coancestry in terms of an
effective population size. Status number can be seen as the number of genotypes
sampled from the reference population of non-inbred, unrelated individuals
(Lindgren and Kang, 1997), which has the same group coancestry as the
population in question. The status number of an orchard crop shares the
following characteristics with the same size of sample from the reference
population:







the same group coancestry;
the same degree of average relatedness;
the same probability that genes taken at random from the gamete gene
pool are identical by descent;
the same expected deviation of gene frequencies from the reference
population;
the same gene diversity and expected heterozygosity;
a similar number of rare alleles; and
the same expected inbreeding in the progeny following random mating.
Status number can also be interpreted as the number of individuals that are
unrelated and non-inbred (II, VII), the number of genotypes having equal fertility
(III), or the number of population members having the same sex ratio (I). Status
number expresses the accumulated genetic drift from the reference population to
which the concepts of inbreeding and coancestry refer (Lindgren et al., 1996).
Status number is compared with the traditional concepts of effective population
size in Table 5. The variance effective population size (Ne(v)) is the size of an
50
ideal population with the same sampling variance of allelic frequencies per
generation (i.e., the same random genetic drift) as the seed orchard (Crow and
Kimura, 1970). Ne(v) refers to the variance of change in group coancestry  (i.e.,
status number) from the reference population with all unique alleles (II; Rosvall,
1999).
Table 5. Effective numbers studied in seed orchards
Type
Symbol
Definition
Reference
Status
effective
number
Ns
The number of unrelated and
non-inbred genotypes, which
has the same group
coancestry. Sometimes, it can
be called the effective
number of parents (III).
Lindgren et al., 1996;
Lindgren & Mullin
1998;
Roberds et al., 1991;
Skroppa, 1994;
Lacy, 1995
Inbreeding
effective
population size
Ne(i)
The size of an idealize
population that gives the
same increase in inbreeding
coefficient (in average
homozygosity)
Wright, 1931;
Robertson, 1961;
Kimura & Crow,
1963;
Pollak, 1977;
Harju, 1995
Variance
effective
population size
Ne(v)
The size of an idealized
population that gives rise to
the same expected change
(random genetic drift) in gene
frequency
Crow & Kimura,1970;
Crow & Denniston,
1988;
Burczyk, 1996;
Crossa & Vencovsky,
1997
Eigenvalue
effective
population size
Ne(e)
The ideal population size that
gives the same expected
change in gene frequency and
variance in the transition of
genes (cf. diversity effective
size, random extinction)
Haldane, 1939;
Crow & Kimura,
1970;
Ewens, 1982;
Gregorius, 1991;
Caballero, 1994
The other conventional effective population size is the inbreeding effective
population size (Ne(i)), which is the size of an ideal population that accumulates
inbreeding F at the same rate as the orchard population (Kjær and Wellendorf,
1997). If the seed orchard is under Hardy-Weinberg equilibrium (random
mating), both Ne(v) and Ne(i) are expected to be identical. While Ne(i) depends
primarily on the size of the parental generation, Ne(v) is more dependent on the
number of offspring. In contrast to the conventional effective population sizes
(Ne(v), Ne(i)), status number (Ns) refers to the accumulated relatedness in the gene
51
pool. Ns is thus an effective number describing at a particular state of
development.
From Table 5, it is apparent that there are various kinds of effective number, and
these can sometimes lead to quite different expressions even though they are
usually very similar (Crow and Kimura, 1970; Kjær and Wellendorf, 1997).
Wright’s procedure is based on the probability of homozygosity arising from
common ancestry, and hence it is called an inbreeding effective number
(Caballero, 1994). The amount of allele frequency drift per generation, as
measured by its variance, can be used to define a variance effective number
(Crow and Denniston, 1988). One can also consider the asymptotic rate of decay
of segregating loci, which is determined by the largest non-unit eigenvalue of the
transition. Ewens (1982) pointed out that the eignvalue effective number could be
estimated by tracing the progress of a quantity such as heterozygosity.
The inbreeding effective population size does not make much sense in firstgeneration clonal seed orchards (II). Ne(i) describes how fast a population
accumulates inbreeding over generations. If the clones are assumed to be
unrelated, inbreeding arises only from selfing. Thus, it is more relevant to
describe selfing explicitly, rather than try to base an effective number on it (II;
Lindgren and Mullin, 1998). Gregorius (1991) also proposed to use ‘diversity
effective size’ that was a derivative of Ne(e). Ns is not informative about selfing,
therefore Ne(v) can be a supplement for managing seed orchards and describing
their crops. Ns and Ne(v) are also more important than Ne(i) in germplasm collection
and regeneration (X; Crossa and Vencovsky, 1997).
Gene diversity and heterozygosity
Loss of gene diversity by drift in a small population (e.g., seed orchards) results
in a loss of allelic variants and lower heterozygosity, while inbreeding due to
non-random mating only changes the level of heterozygosity (Rosvall, 1999).
Nevertheless, empirical results on variability in natural stands and orchard crops
confirm that expected heterozygosities are similar in both populations
(Savolainen and Kärkkäinen, 1992). If orchards have a low effective number,
some of the important genetic variability may be in the form of rare alleles, e.g.,
disease resistance. Genetic drift and inbreeding affect both target traits and
neutral alleles, while selection (e.g., genetic thinning) will only affect the target
and linked genes.
Expected gene diversity of seed crops (GD) from first-generation seed orchards
can measured relative to the group coancestry of a reference population. GD is a
function of group coancestry (Lacy, 1995) and inversely proportional to the status
number (II) as
52
GD  1  
1
1
2Ns
[23]
Gene diversity can be defined as the probability that genes are different by
descent, and thus the loss of gene diversity compared to an infinite reference
population is the accumulation of group coancestry (II; see also Gregorius,
1991). When group coancestry is monitored in seed orchards, one can say how
much gene diversity has been lost compared to the reference population, i.e., the
wild-forest (Yeh, 2000). Group coancestry also measures gene dispersion
(Lindgren and Mullin, 1998).
The loss of gene diversity assessed by GD is approximately linear for several
generations, reflecting that the amount of additive genetic variance decreases at
the same rate as the gene diversity or expected heterozygosity (VIII; Verrier et
al., 1991; Savolainen and Kuittinen, 2000). Conversely, the decrease in Ns over
generations is non-linear and reflects the sampling process of founder genes,
being proportional to the probability that low-frequency genes will be lost (VIII;
Lindgren et al., 1996; Rosvall, 1999; Bila, 2000).
Gene diversity (GD) is equivalent to expected average heterozygosity (He). The
expected average heterozygosity (Het) in orchard crops can thus be demonstrated
under drift theory on the same basis as gene diversity (Stoehr and El-Kassaby,
1997; Savolainen and Kuittinen, 2000) as

1
Het  1 
2 N e( v )


 Het 1

[24]
where Ne(v) is the variance effective population size; and Het-1 is the
heterozygosity in the parental orchard population (Nei et al., 1975; Lacy, 1995).
By definition, the heterozygosity of first-generation orchard parents will be one
(Het-1= 1) because all parents are considered to be a sample from the reference
population (VII). The expected heterozygosity of orchard crops will thus be the
direct function of variance effective population size. If the Ne(v) is larger than 10,
the depletion of heterozygosity caused by random drift alone will be very small
(Fig. 7).
53
Expected heterzygosity (He )
1.00
0.75
0.50
0.25
0.00
1
10
20
30
40
50
Variance effective pop. size (Ne (v ))
Figure 7. Relationship between expected heterozygosity (He) and variance effective
population size (Ne(v)). When Ne(v) is larger than 10, the depletion of heterozygosity
caused by genetic drift will be small.
Variance effective population size (Ne(v)) is a function of fertility variation ( )
and census number (N) in a seed orchard consisting of unrelated and non-inbred
parents, Ne(v) = N/(-1). It describes the change of gene contributions between the
orchard parents and their offspring, and the change of status number over
generations (I, II; Bila, 2000). When the sibling coefficient is equal to two,  = 2
(CV = 100%), Ne(v) will be the same as the orchard census number (N) where
parents are unrelated and non-inbred.
Concerning gene diversity, the main question is how many parents we have to use
in seed orchards. Skroppa (1994) suggested that 20-30 individuals could
guarantee a sufficiently high effective population size. If there are too many
parents, one can keep much diversity, but selection intensity is low, resulting in
little genetic gain (Zobel and Talbert, 1984). It also causes some operational
difficulties such as high cost, complicated layout, unfavorable flexibility and so
forth. In advanced-generation seed orchards, it is highly recommended to balance
high genetic gain with desirable gene diversity.
Selfing and inbreeding
Selfing is expected in a seed orchard (Rudin and Lindgren, 1977) and will reduce
the genetic value of the orchard seeds. The proportion of seed due to selfing is
54
important because of the poor survival and growth of selfed offspring (Franklin,
1970). In clonal seed orchards, selfing can occur by mating among ramets of the
same clone as well as self-pollination within individual ramets (II). Frequencies
of selfed seedlings vary considerably among clones and years (Squillace and
Goddard, 1982). If individual clones have a high level of viable selfed offspring,
these clones need to be identified and either rogued from the orchard or treated
by pollen management techniques (e.g., SMP) to reduce the amount of selffertilization to acceptable levels (El-Kassaby and Ritland, 1986b).
The selfing rate can be estimated from mating probability (Gömöry et al., 2000)
or based on isozyme makers (Burczyk, 1991; El-kassaby et al., 1989b; Harju,
1995). In general, selfing rates estimated by isozyme analysis are 2-7% (less than
10%) in clonal seed orchards (Müller-Starck, 1982; Rudin et al., 1986; Adams
and Birkes, 1989), but may be higher in particular cases (Rudin and Ekberg,
1982; Lai and Wang, 1997). Selfing tends to be greater in the lower part of
crowns (Franklin, 1971b; Rudin and Lindgren, 1977; Shen et al., 1981; Philipson,
1997). Since male strobili occur mainly in the lower and mid-crown, female
strobili in the lower crown are more apt to be self-pollinated than those in the
upper crown (Squillace and Goddard, 1982), although El-Kassaby et al. (1986)
found no significant differences among crown levels and aspects in a Douglas-fir
seed orchard.
Inbreeding is not desirable in reforestation material (Kang and Namkoong, 1988;
Harju, 1995). With proper design in orchard composition and spatial arrangement
of parents, inbreeding can be minimized in orchard crops even though inbreeding
is substantial in the parent population (Lindgren and El-Kassaby, 1989; Xie and
Knowles, 1994). It may be further reduced by SMP and pollen contamination
(Rudin and Lindgren, 1977; El-Kassaby, 1995). Inbreeding also affects
differently the production of filled seeds, depending on the types of mating
(Griffin and Lindgren, 1985). Lindgren and Gregorius (1976) argued that
inbreeding might be a tool in a breeding strategy because it increase genetic
variance and makes it possible to keep a higher number of unrelated individuals
(i.e., higher effective number) in breeding populations.
It has been suggested that the genetic cause of embryo collapse producing empty
seeds is an inbreeding effect, caused by the action of homozygous lethal genes
(Sarvas, 1962). However, the presence of simple recessive embryonic lethals
does not satisfactorily explain why the second-generation of selfing appears to
yield a much lower percentage of filled seeds than the first-generation of selfing
(Andersson et al., 1974; Sniezko, 1984).
The study of outcrossing rate in seed orchards is useful for evaluating the rate of
inbreeding and its associated effects (El-Kassaby et al., 1989b). In most
published reports, high outcrossing rates are observed in orchards (t > 0.9;
Friedman and Adams, 1985a; Omi and Adams, 1985; Ritland and El-Kassaby,
55
1985; El-Kassaby et al., 1988). Estimates of outcrossing rate obtained from seed
orchards are generally higher, or at least not lower, than those reported for natural
stands (Shaw and Allard, 1982; Rudin et al., 1986). It may be due to the
population structure, i.e., the physical arrangement of related individuals within a
population. In conjunction with the phenomenon of polyembryony, the
differential viability of selfed and outcrossed embryos maintains a high effective
outcrossing rate in spite of considerable natural self-pollination (Lindgren, 1975).
Koski (1973) estimated that, if the level of self-pollination were 20%, there
would be less than 1% of trees derived from selfing in regenerated Scots pine
stands. Thus, selfing does not seem to be alarming in seed orchards.
Under the mixed random mating and selfing model, an effective population size
(Ne) reduces to 0.5(1+t)N, where t is the outcrossing rate. Ne is thus always
smaller than N in the model, unless the rate of outcrossing is unity. Consequently,
if a species exhibits high outcrossing, the effect of inbreeding on the decay of
gene diversity is small, relative to that due to genetic drift (Yeh, 2000). Kjær and
Welledorf (1997) formulated the inbreeding effective population size,
considering inbreeding and selection against self-pollination, as Ne(i)= 1/[(1-t) +
w*t(fimi)], where w is the fitness of selfed zygotes compared to outcrossed
zygotes (Gömöry et al., 2000).
The effect of inbreeding in forest trees is of interest from many points of view. A
major question is the evaluation of the hazards involved in selecting related
individuals for use in advanced-generation seed orchards. Another aspect is the
possibility of using inbreeding as a tool in tree breeding (Andersson et al., 1974;
Sniezko, 1984). Estimates of the inbreeding coefficients of orchard crops cannot
always be simply estimated, because they are functions of the amount of
inbreeding in the current and previous generations. A simple way to determine
the degree of expected inbreeding (F) in the seeds from a first-generation seed
orchard is (van Buijtenen, 1971)
F = (P1 * 0.5) + (P2 * 0.25) + (P3 * 0.125)
where P1 = the probability of obtaining seed from selfing
P2 =
"
"
"
from full-sib crosses
P3 =
"
"
"
from half-sib crosses
From studies II, III, X and XI, the expected inbreeding in orchard crops caused
by fertility variation and/or relatedness would be small and could be ignored in
first-generation seed orchards.
1. Inbreeding in clonal seed orchards
56
In a clonal seed orchard consisting of unrelated, non-inbred clones, the only
factor influencing the degree of inbreeding is the probability P1 of obtaining seed
from selfing as
P1 = (A1 + A2)  C  D
where A1 is the pollen contribution from the ramet itself; A2 is the pollen
contribution from other ramets of the same clone (A2<A1); C is the seed set from
selfs (relative value); and D is the proportion of sound seed from selfs (for selfsterile clones, D= 0).
The frequency of mutants (e.g., albino, chlorophyll deficient, etc.) appearing in
open-pollinated progenies carrying marker genes can give a good estimate of the
magnitude of the factor P1 (Squillace and Kraus, 1963) as
P1 = Y / X
where Y is the percent markers showing in open pollinated seeds; and X is the
percent marker showing in the selfed seeds.
For instance, 1% albinos are observed in open-pollinated seeds and 16% in selfed
seeds, the value of P1 is 0.0625 and F would be 0.031 (Franklin, 1971b; Squillace
and Goddard, 1982).
From a practical standpoint, orchard managers need not be very concerned about
the yield of selfed seedlings from seed orchards. Admittedly, since selfs are
usually inferior in growth rates and less vigorous than cross-pollinated seedlings,
they are likely to be culled in the nursery or often will be suppressed and die
when out-planted.
2. Inbreeding in seedling seed orchards
The situation in seedling seed orchards is much more complicated because
selfing, crossing among full sibs or crossing among half-sib may all occur. Thus,
P1 cannot be modified readily by the design of the orchard, but P 2 and P3 can be
modified if separation between half-sibs by one position is considered adequate,
which is still not easy.
Inbreeding reduces seed and wood production, although a higher genetic gain due
to a more intense and unrestricted selection may be obtained if related individuals
are used in the same seed orchard (Lindgren and Gregorius, 1976). Williams and
Savolainen (1996) advocated using experimental inbred populations for direct
utilization of inbreeding as a breeding method. Matheson et al. (1995) reported
that it was better to allow some proportion of related mating in the order of
parent-offspring pairs, half-sib and full-sib families. In advanced-generation seed
57
orchards, therefore, it is advisable to consider the possibility of using a limited
amount of related individuals in the same seed orchard (VII), based on detailed
calculations of the expected gain from different alternatives. It could be a kind of
linear deployment (Lindgren and Matheson, 1986) or step-wise arrangement
(Hodge and White, 1993). In general, it is better to include more genetic entries
with fewer individuals per entry rather than vise versa for less inbreeding in seed
crops. It may also appear better to use more ramets of the best relative and thus
fewer entries for high gain than to include several relatives if the inbreeding is
acceptable.
3. Inbreeding depression
Inbreeding depression can be defined as the reduction in seed production caused
by inbreeding compared with the seed production from outcrossing. It can be
caused by heterozygote superiority at specific loci (over-dominance hypothesis)
or arise from recessive or partly recessive deleterious genes (dominance
hypothesis) (Crow, 1952; Charlesworth and Charlesworth, 1987; see also review
by Williams and Savolainen, 1996). The latter one is suggested as the main
contributor to inbreeding depression, which is maintained by mutation-selection
balance.
Inbreeding depression expressed during seed development is especially severe in
conifers (Sorensen, 1969; Koski, 1971, 1973; Lindgren, 1975), although there
may be strong selection against deleterious genes or recessive mutations due to
rather frequent self-fertilization. Griffin and Lindgren (1985) suggested that gain
estimates should be adjusted downward at least in proportion to the expectation
of sib-mating (i.e., inbreeding depression), if breeders were to include multiple
selections from full- or half-sib families in the seed orchards. Surprisingly, there
was little inbreeding depression in germination rate and percent mortality through
two generations of selfing in loblolly pine (Sneizko, 1984). Inbreeding depression
on height growth in the nursery depends also on growth conditions in Norway
spruce (Andersson et al., 1974). On the other hand, inbreeding depression for
survival and growth vigor was consistent through three generations of selfing in
silver birch (Wang et al., 1999).
Inbreeding depression on height growth of S1 progeny has a large range from zero
to almost 50%, averaging 22%, for 20 conifer species (see Table 8.2 in Sedgley
and Griffin, 1989). It seems that less inbreeding depression is found among
families produced from parents with high GCA for growth (Mullin et al., 1978).
When the population size is kept constant, and each parental genotype is equally
represented, serious inbreeding and inbreeding depression can be avoided for
many cycles of selection (Barnes, 1995).
Inbreeding depression does tend to be linear with the respect to the coefficient of
inbreeding (F) if there is no epistatic interaction among loci (Sniezko, 1984). As
58
inbreeding accumulates, reproductive capacity will be deteriorate. Due to the
both factors, it soon becomes impossible to avoid the loss of some individuals or
of some lines. The survivors are then a selected group to which the theoretical
expectations no longer apply. The precise measurement of the rate of inbreeding
depression can thus generally be made only over the early stages, before the
inbreeding coefficient reaches high levels (Falconer and Mackay, 1996).
Inbreeding depression results in reduced yield of viable seeds, poor germination,
segregation of deformed or chlorophyll deficient seedlings, and varying degrees
of reduced growth (Franklin, 1970). As discussed above, however, the
outcrossing rate in most conifer species is very high, selfed embryos are killed by
lethal alleles, selfed seedlings are often out-competed in the nursery, and very
seldom do they remain in the mature stand. Thus, selfing in plantations derived
from orchard crops has probably more or less negligible effects.
59
Concluding remarks
Genetic gain and gene diversity of orchard crops were evaluated in seed orchard
populations. High genetic gain could be obtained using orchard management
alternatives (e.g., selective cone harvest, genetic thinning), while reasonable gene
diversity (measured by status number) was maintained. Effective management of
forest genetic resources is a key element in future forestry. In the future, seed
orchards must balance genetic gain and gene diversity, in order to ensure high
genetic quality of orchard crops.
There was a large variation in ramet number and flower production among
parents in seed orchards, although the sibling coefficient was generally in the
order of  = 2 (CV = 100%). Reproductive characteristics were under genetic
control and positively correlated with years. Fertility variation will accelerate the
accumulation of group coancestry and inbreeding, and hence reduce the expected
genetic gain and gene diversity of seed orchard crops. Relative status number (Nr)
was estimated to be 0.74 due to the ramet variation. By increasing seed quality
and quantity in seed orchards rather than by establishing larger area, seed orchard
genetics in combination with proper silvicultural management will provide better
opportunity for the use of seed orchards for breeding and conservation purposes.
If the status number (Ns) of an orchard crop was larger than 10, the depletion of
gene diversity in the following generation, due to genetic drift and fertility
variation, would be small. On the other hand, there would be lower expected
genetic gain and a potential inbreeding problem in seed orchards or breeding
populations with the status number of less than 10. A large number of parents are
generally used in first-generation seed orchards; thus, genetic loss and erosion
due to genetic drift do not seem to be alarming during the process of
domestication, because of the large number of orchard parents. Gene diversity
should, however, be monitored and controlled to maintain as much diversity as
possible in future-generation seed orchards.
Seed orchards are the link between breeding programs and reforestation through
the delivery of consistent, abundant yields of genetically improved seed. Seed
orchard crops represent the mechanism by which new forests are created and
genetic variation is transmitted from one generation to another. Status number
(Ns) is a useful tool for monitoring gene diversity of orchard crops. Status number
will also bring better information for seed certification in future seed orchards.
To describe the effect of random genetic drift, Ns and Ne(v) seem to be more
informative than Ne(i) in seed orchard populations. When the desirable level of
effective number is maintained, seed orchards can produce orchard crops with
high genetic gain and gene diversity.
An understanding of reproductive processes (e.g., random mating, phenological
synchrony, parental balance) and the impact of management practices (e.g., equal
60
seed harvest, genetic thinning, SMP, overhead cooling, and mixing seeds from
different years) is essential for maximizing genetic gain obtained from seed
orchard programs and for maintaining the gene diversity required to meet
environmental contingencies. Management options such as selective harvesting,
partial silvicultural thinning, genetic roguing, flower induction and SMP are
available for orchard managers to promote better reproductive balance in seed
orchards. Orchard managers should consider options that are cheaper and more
practical. Much closer attention also needs to be focused on reproductive
balancing when seed orchards are rogued in order to raise the overall average
breeding value of parents represented and thus maintain durable gene diversity of
reforestation stock.
Realized genetic gain and gene diversity of orchard crops usually deviate from
the expectation for an ideal case. For the successful functioning of seed orchards,
orchard managers, specialists and researchers should diligently observe,
assiduously collect and consistently study all essential information on parental
behavior, compatibility, combining ability and gene flow, and then translate this
information into practical terms so that it may be employed in future seed
orchards.
Genetic gain and gene diversity will vary considerably from orchard to orchard,
and from year to year. Cost and time for establishment and management will also
vary. Gene diversity can have importance for both the long-term stability and the
short-term productivity of forest plantations. Hence, there is need for information
on seed orchard genetics in the near future. Furthermore, tree improvement
through seed orchards is probably one of the most cost-effective alternatives
available to forest managers to improve forest productivity.
Modern seed orchard research is the science, business and art of managing and
conserving genetic resources for continuing economic, social and environmental
benefit. It involves the balanced utilization and optimum management of forest
resources for maximum yields of improved seed from seed orchards. Recent
orchard study also draws upon the knowledge and experience from many
disciplines and other professions, and plays a wide role in the development and
implementation of new techniques. Orchard seeds yield good forests and give
additional protection because they contain the offspring of carefully selected
superior parents. Plantations from seed orchard crops can also provide more
protection against disease or insect catastrophe, adapt well to varying site and soil
conditions, and enhance biological diversity.
61
Future perspectives related to this thesis
Relatedness among candidates for seed orchards will increase in advancedgeneration breeding programs, and relatives will likely be used in future seed
orchards. The effect of mating between relatives (among ramets of related clones
or among related trees) needs to be better evaluated, so the breeder knows better
in advanced-generation seed orchards whether relatives with high genetic values
can be effectively used or not. Orchard managers will want to get high genetic
gain rather than high diversity even if some inbreeding depression may result.
However, the question of how much gain and diversity are sacrificed when
relatives are allowed to mate in seed orchards is still not fully analyzed. The
outcome of a seed orchard is also dependant on selfing. Selfing is an important
consideration, particularly when the need for a higher effective number of clones
is considered. Thus, the effect of selfing and its relationship to effective number
in seed orchards ought to be better studied.
In previous studies, which mainly dealt with clonal seed orchards, the assumption
was normally made that orchard parents were unrelated. This may have been a
reasonable attitude in the past and even for the present, for most attention has
been on clonal seed orchards. Still, there will be relatives in seedling seed
orchards (breeding seedling orchards) and in advanced-generation seed orchards.
Formulae for genetic gain and gene diversity in future-generation seed orchards
with complicated pedigree should be developed. More practical variables, such as
inbreeding, inbreeding depression and seed production, should also be considered
for optimizing and balancing genetic gain and gene diversity of orchard crops that
will give maximum orchard benefit.
For contaminating pollen, it was assumed in most studies that the contribution of
the foreign pollen was uniform and equal for all orchard parents, which may not
be true. Flower phenology and synchrony between the pollen producers and
female flowers were not considered, and observations that allow for
generalizations were difficult to obtain. By considering phenological aspects, the
impact of pollen viability, pollen density, pollen dispersal, seed size, germination
parameters, seed viability and seedling production on the genetic gain and gene
diversity of orchard crops can be better analyzed.
The study showed that controlling female fertility (i.e., equal seed harvest or
mixing seed equally among parents) increased gene diversity of seed orchard
crops. However, if an amount of seed production is an urgent matter in a seed
orchard, one could elaborate on the point with balancing on the male side rather
than the female side by removing ramets of a clone or by controlling family sizes.
This could be done without loss of seed if there is sufficient pollen, while it
increases gene diversity. The problem is that the female-male correlations
apparently vary among years. From theoretical and practical points of view, it
62
could be interesting to investigate the controlling male fertility and its effect on
genetic gain and diversity of orchard crops.
Linear deployment, using orchard parents in proportions linearly related to their
breeding values, gives an optimal compromise between genetic gain and gene
diversity of crops from unrelated, non-inbred clones. More appropriate orchard
deployment principles will be needed in advanced-generation seed orchards
where the parents will be related. Algorithms and software should also be
developed for designing of future seed orchards and for cheaper forest
improvement options including seed stands and seedling seed orchards.
Computer-based tools (by MS® Excel) in seed orchard management and some
useful programs in breeding theory can be found at the web-site managed by Prof.
Dag Lindgren; http://www.genfys.slu.se/staff/dagl/Breed_Home_Page/
63
References
Adams, W.T. and Birkes, D.S. 1989. Mating patterns in seed orchards. In: Proc. 20th
SFTIC conf. Charleston, South Carolina. p.75-86.
Adams, G.W. and Kunze, H.A. 1996. Clonal variation in cone and seed production in
black and white spruce seed orchards and management implications. For. Chron.
72: 475-480.
Adams, W.T., Griffin, A.R. and Moran, G.F. 1992. Using paternity analysis to
measure effective pollen dispersal in plant populations. American Naturalist 140:
762-780.
Adams, W.T., Hipkins, V.D., Burczyk, J. and Randall, W.K. 1997. Pollen
contamination trends in a maturing Douglas-fir seed orchard. Can. J. For. Res. 27:
131-134.
Ahlgrn, C.E. 1972. Some effects of inter- and intraspecific grafting on growth and
flowering of some five-needle pines. Silvae Genet. 21: 122-126.
Allendorf, F.W. 1986. Genetic drift and the loss of alleles versus heterozygosity. Zoo
Biology, 5: 181-190.
Allison, T. 1990. Pollen production and plant density affect pollination and seed
production in Taxus canadensis. Ecology 71: 516-522.
Andersson, E.W. 1999. Gain and diversity in multi-generation breeding programs.
Ph.D. thesis. Swedish University of Agricultural Sciences, Umeå, Sweden. Acta
Universitatis Agriculturae Sueciae, Silvestria 95, 1999.
Andersson, E. 1955. Pollen dispersal and the isolation by distance of seed orchards.
Norrlands SkogsvFörb. Tidsskr., Stockh. 1: 35-100.
Andersson, E., Jansson, R. and Lindgren. D. 1974. Some results from second
generation crossings involving inbreeding in Norway spruce (Picea abies). Silvae
Genet. 23: 34-43.
Askew, G.R. 1988. Estimation of gamete pool composition in clonal seed orchards.
Silvae Genet. 37: 227-232.
Askew, G.R. and Burrows, P.M. 1983. Minimum coancestry selection I. A Pinus
taeda population and its simulation. Silvae Genet. 32: 125-131.
Ballou, J.D. and Lacy, R.C. 1995. Identifying genetically important individuals for
management of genetic variation in pedigreed populations. In: Population
management for survival and recovery: Analytical methods and strategies in small
population conservation. Columbia Univ. Press. NY. p.76-111.
Barnes, R.D. 1995. The breeding seedling orchard in the mutiple population breeding
strategy. Silvae Genet. 44: 81-88.
Bergman, F. and Ruetz, W. 1991. Isozyme genetic variation and heterozygosity in
random tree samples and selected orchard clones from the same Norway spruce
populations. For. Ecol. Manage. 46: 39-47.
Bila, A.D. 2000. Fertility variation and its effects on gene diversity in forest tree
populations. Ph.D. thesis. Swedish University of Agricultural Sciences, Umeå,
Sweden. Acta Universitatis Agriculturae Sueciae, Silvestria 166, 2000.
Bishir, J. and Roberds, H. 1999. On number of clones needed for managing risks in
clonal forestry. For. Genet. 6: 149-155.
64
Bjornstad, A. 1981. Photoperiodical after-effect of parent plant environment in
Norway spruce (Picea abies (L.) Karst) seedlings. Medd. Nor. Inst. Skogforsk 36:
1-30.
Bridgwater, F.E. and Trew, I.F. 1981. Supplemental mass pollination. In: Pollen
management handbook 587. Ed) Dr. Franklin, E.C. Washington, D.C. p.15-19.
Bridgwater, F.E., Bramlett, D.L., Byram, T.D. and Lowe, W.J. 1998. Controlled mass
pollination in loblolly pine to increase genetic gains. For. Chron. 74: 185-189.
Brisbane, J.R. and Gibson, J.P. 1995. Balancing selection response and rate of
inbreeding by including genetic relationships in selection decisions. Theor. Appl.
Genet. 91: 421-431.
Brown, I. 1971. Flowering and seed production in grafted clones of Scots pine. Silvae
Genet. 20: 121-132.
Buckland, D.C. and Kuijt, J. 1957. Unexplained brooming of Douglas-fir and other
conifers in British Columbia and Alberta. For. Sci. 3: 236-242.
Burczyk, J. 1991. The mating system in a Scots pine clonal seed orchard in Poland.
Ann. Sci. For. 48: 443-451.
Burczyk, J. 1996. Variance effective population size base on multilocus gamete
frequencies in coniferous populations: an example of a Scots pine clonal seed
orchard. Heredity 77: 74-82.
Burczyk, J. and Prat, D. 1997. Male reproductive success in Pseudotsuga menziesii
(Mirb.) Franco: the effects of spatial structure and flowering characteristics.
Heredity 79: 638-647.
Byram, T.D., Bridgwater, F.E., Gooding, G.D. and Lowe, W.J. 1998. 46th progress
report of the cooperative forest tree improvement program. Texas Forest Service,
23p.
Byram, T.D., Lowe, W.J. and McGriff, J.A. 1986. Clonal and annual variation in
cone production in loblolly pine seed orchards. For. Sci. 32: 1067-1073.
Caballero, A. 1994. Developments in the prediction of effective population size.
Heredity 73: 657-679
Cameron, R.J. and Thomson, G.V. 1971. An instance of clonal incompatibility in
grafted Pinus radiata seedlings. Silvae Genet. 20: 195-199.
Caron, G.E. 1987. Pollen monitoring in two black spruce seedling seed orchards.
Fraser Inc., Edmonston, N.B.
Chaisurisri, K. and El-Kassaby, Y.A. 1993. Estimation of clonal contribution to cone
and seed crops in a Stika spruce seed orchard. Ann. Sci. For. 50: 461-467.
Chaisurisri, K. and El-Kassaby, Y.A. 1994. Genetic diversity in a seed production
population vs natural population of Sitka spruce. Biodiv. Conserv. 3: 512-523.
Chalupka, W. 1980. Regulation of flowering in Scots pine (Pinus sylvestris L.) grafts
by gibberellins. Silvae Genet. 29: 118-121.
Chalupka, W. 1998. Pollen formed under pollution affects some quantitative
characters of Scots pine (Pinus sylvestrys L.) seeds. For. Genet. 5: 133-136.
Charlesworth, D. and Charlesworth, B. 1987. Inbreeding depression and its
evolutionary consequences. Ann. Rev. Ecol. Syst. 18: 237-268.
Cheliak, W.M., Murray, G. and Pitel, J.A. 1988. Genetic effects of phenotypic
selection in white spruce. For. Ecol. Manage. 24: 139-149.
Clare, L.R. 1982. Assessment of pollen contamination in Pacific Forest Products’
high elevation seed orchard. BSF. thesis, Univ. of British Columbia, Vancouver.
65
Climent, J.M., Prada, M.A. and Pardos, J.A. 1997. Increase of flowering in Pinus
nigra Arn subsp salzmanni (Dunal) Franco by means of heteroplastic grafts. Ann.
Sci. For. 54: 145-153.
Cockerham, C. 1967. Group inbreeding and coancestry. Genetics 56: 89-104.
Crossa, J. and Vencovsky, R. 1997. Variance effective population size for two-stage
sampling of monoecious species. Crop Sci. 37: 14-26.
Crow, J.F. 1952. Dominance and overdominance. In: Heterosis, ed) Gowen, J.W.
Iowa State College Press, Ames, p.282-297.
Crow, J.F. and Denniston, C. 1988. Inbreeding and variance effective population
numbers. Evolution 42: 482-495.
Crow, J.F. and Kimura, M. 1970. An introduction to population genetics theory.
Harper and Row Publishers, London, 591pp.
Danbury, D.J. 1971. Economic implications of selection for seed production in
radiata pine seed orchards. Aust. For. Res. 5: 37-44.
Darwin, C. 1877. The differential forms of flowers on plants on the same species.
John Murray, London.
Denison, N.P. 1973. A review of aspects of tree breeding in the Republic of South
Africa. In: Selection and breeding to improve some tropical conifers. Eds) Drs.
Burley, J. and Nikles, D.G. Oxford. Comm. For. Inst. 2: 277-284.
Devlin, B. and Ellstrand, N.C. 1990. Male and female fertility variation in wild
radish, a hermaphrodite. Am. Nat. 136: 87-107.
Di-Giovanni, F. and Kevan, P.G. 1991. Factors affecting pollen dynamics and its
importance to pollen contamination: a review. Can. J. For. Res. 21: 1155-1170.
Dormling, I. and Johnsen, O. 1992. Effects of the parental environment on full-sib
families of Pinus sylvestris. Can. J. For. Res. 22: 88-100.
Duffield, J.W. and Wheat, J.G. 1963. Dwarf seedling from broomed Douglas-fir.
Silvae Genet. 12: 129-133.
Duzan, H.W. and Williams, C.G. 1988. Matching loblolly pine families to
regeneration sites. Southern J. of Applied Forestry 12: 166-169.
Dyson, W.G. 1975. A note on dwarfing of Pinus patula grafts. Silvae Genet. 24: 6061.
Edwards, D.G.W. and El-Kassaby, Y.A. 1996. The biology and management of
coniferous forest seeds: genetic perspectives. For. Chron. 72: 481-484.
El-Kassaby, Y.A. 1995. Evaluation of tree-improvement delivery system: factors
affecting genetic potential. Tree Physiology 15: 545-550.
El-Kassaby, Y.A. and Askew, G.R. 1991. The relation between reproductive
phenology and reproductive output in determining the gametic pool profile in a
Douglas-fir seed orchard. For. Sci. 37: 827-835.
El-Kassaby, Y.A. and Askew, G.R. 1998. Seed orchards and their genetics. In: Forest
Genetics and Tree Breeding, Chap. 6: p.103-111. A.K. Mandal and G.L. Gibson
(eds.). CBS Publishers and Distributors. Daryaganj, New Delhi.
El-Kassaby, Y.A. and Cook, C. 1994. Female reproductive energy and reproductive
success in a Douglas-fir seed orchard and its impact on genetic diversity. Silvae
Genet. 43: 243-246.
El-Kassaby, Y.A. and Davidson, 1990. Impact of crop management practices on the
seed crop genetic quality in a Douglas-fir seed orchard. Silvae Genet. 39: 230-237.
66
El-Kassaby, Y.A. and Namkoong, G. 1995. Genetic diversity of forest tree
plantations: consequences of domestication. In: Consequences of changes in
biodiversity. IUFRO World Congress, Tampere, Finland, Vol 2. p.218-228.
El-Kassaby, Y.A. and Reynolds, S. 1990. Reproductive phenology, parental balance,
and supplemental mass pollination in a Sitka-spruce seed orchard. For. Ecol.
Manage. 31: 45-54.
El-Kassaby, Y.A. and Ritland, K. 1986a. Low levels of pollen contamination in a
Douglas-fir seed orchard as detected by allozyme makers. Silvae Genet. 35: 224229.
El-Kassaby, Y.A. and Ritland, K. 1986b. The relation of outcrossing and
contamination to reproductive phenology and supplemental mass pollination in a
Douglas-fir seed orchard. Silvae Genet. 35: 240-244.
El-Kassaby, Y.A. and Thomson, A.J. 1996. Parental rank changes associated with
seed biology and nursery practices in Douglas-fir. For. Sci. 42: 228-235.
El-Kassaby, Y.A., Fashler, A.M.K. and Sziklai, O. 1984. Reproductive phenology
and its impact on genetically improved seed production in a Douglas-fir seed
orchard. Silvae Genet. 33: 120-125.
El-Kassaby, Y.A., Parkinson, J. and Devitt, W.J.B. 1986. The effect of crown
segment on the mating system in a Douglas-fir (Pseudotsuga menziesii (Mirb.)
Franco) seed orchard. Silvae Genet. 35: 149-155.
El-Kassaby, Y.A., Ritland, K. Fashler, A.M.K and Devitt, W. 1988. The effect of
reproductive phenology upon the mating structure of a Douglas-fir seed orchard.
Silvae Genet. 37: 76-82.
El-Kassaby, Y.A., Fashler, A.M.K. and Crown, M. 1989a. Variation in fruitfulness in
a Douglas-fir seed orchard and its effect on crop-management decisions. Silvae
Genet. 38: 113-121.
El-Kassaby, Y.A., Rudin, D. and Yazdani, R. 1989b. Levels of outcrossing and
contamination in two Pinus sylvestris L. seed orchards in Northern Sweden.
Scand. J. For. Res. 4: 41-49.
El-Kassaby, Y.A., Edwards, D.G. and Cook, C. 1990. Impact of crop management
practices on seed yield in a Douglas-fir seed orchard. Silvae Genet. 39: 226-230.
Eriksson, U. 1996. Enhancing production of high-quality seed in Swedish conifer
breeding. Ph.D. thesis. Swedish University of Agricultural Sciences, Uppsala,
Sweden.
Eriksson, G., Jonsson, A. and Lindgren, D. 1973. Flowering in a clone trial of Picea
abies Karst. Stud. For. Suec. 110. 45pp.
Ewens, W.J. 1982. On the concept of the effective population size. Theoretical
Population Biology 21: 373-378.
Falconer, D.S. and Mackay, T.F.C. 1996. Introduction to quantitative genetics. 4th
edition. Longman, Malaysia. 464pp.
Fashler, A.M.K. and Devitt, W.J.B. 1980. A practical solution to Douglas-fir seed
orchard pollen contamination. For. Chron. 56: 237-241.
Fashler, A.M.K. and El-Kassaby, Y.A. 1987. The effect of water spray cooling
treatment on reproductive phenology in a Douglas-fir seed orchard. Silvae Genet.
36: 245-249.
Feilberg, J. and Soegaard, B. 1975. ’Historical review of seed orchards’. Forestry
Comm. Bull. No.54. Her Majesty’s stationary office, London.
67
Frankel, O.H. and Soulé, M.E. 1981. Population genetics and conservation. In:
Conservation and evolution. Cambridge University Press.Cambridge,UK. p.31-77.
Franklin, E.C. 1970. Survey of mutant forms and inbreeding depression in species of
the family Pinaceae. USDA For. Serv. Res. Paper SE-61. USDA. Ashville, NC.
Franklin, E.L. 1971a. Pollen management in southern seed orchard. In: Proc. 11th
SFTIC, Atlanta, Georgia, USA. p.218-223.
Franklin, E.C. 1971b. Estimation of frequency of natural selfing and of inbreeding
coefficients in loblolly pine. Silvae Genet. 20: 194-195.
Friedman, S.T. and Adams, W.T. 1981. Genetic efficiency in loblolly pine seed
orchards. In: Proc. 16th SFTIC conf. Blacksburg, VA. p.213-224.
Friedman, S.T. and Adams, W.T. 1985a. Levels of outcrossing in two loblolly pine
seed orchards. Silvae Genet. 34: 157-162.
Friedman, S.T. and Adams, W.T. 1985b. Estimation of gene flow into two seed
orchards of loblolly pine (Pinus taeda L.). Theor. Appl. Genet. 69: 609-615.
Fries, A. 1994. Development of flowering and the effect of pruning in a clonal seed
orchard of lodgepole pine. Can. J. For. Res. 24: 71-76.
Gardner V.R., Bradford, F.C. and Hooker, H.D. 1922. Fundamentals of fruit
production. MacGraw-Hill, New York. p.551-579.
Gea, L.D., Lindgren, D., Shelbourne, C.J.A. and Mullin, T. 1997. Complementing
inbreeding coefficient information with status number: implications for structuring
breeding populations. NZ J. For. Sci. 27: 255-271.
Goddard, R.E. 1964. The frequency and abundance of flowering in a young slash pine
orchard. Silvae Genet. 13: 184-186.
Gömöry, D. and Paule, L. 1992. Inferences on mating system and genetic
composition of a seed orchard crop in the European larch (Larix decidua Mill.). J.
Genet. & Breed. 46: 309-314.
Gömöry, D., Bruchanik, R. and Paule, L. 2000. Effective population number
estimation of three Scots pine (Pinus sylvestris L.) seed orchards based on an
integrated assessment of flowering, floral phenology, and seed orchard design.
For. Genet. 7: 65-75.
Gregorius, H.R. 1989. Characterization and analysis of mating system. Ekopan
Verlag, Germany, 158pp.
Gregorius, H.R. 1991. On the concept of effective number. Theor. Pop. Biol. 40: 269283.
Griffin, A.R. 1982. Clonal variation in Radiata pine seed orchards. I. Some flowering,
cone and seed production traits. Aust. For. Res. 12: 295-302.
Griffin, A.R. and Lindgren, D. 1985. Effect of inbreeding on production of filled seed
in Pinus radiata experimental results and a model of gene action. Theor. Appl.
Genet. 71: 334-343.
Griffin, A.R., Wong, C.Y., Wickneswari, R. and Chia, E. 1992. Mass production of
hybrid seed of Acacia mangium x Acacia auriculiformis in biclonal seed orchards.
In: Breeding technologies for tropical acacias. ACIAR-Proceedings, No. 37. Proc.
of an international workshop, Tawau, Sabah, Malaysia, 1-4 July 1991. p.70-75.
Hadders, G. 1972. Kontroll av inkorsningen i en tallplantage. Foren.
Skogstradsforadlings Arsb. Uppsala. 1972.
Haldane, J.B.S. 1939. The equilibrium between mutation and random extinction.
Ann. Eugen. 9: 400-405.
68
Harju, A. 1995. Genetic functioning of Scots pine seed orchards. Ph.D. thesis.
University of Oulu, Finland. Acta Universitatis Ouluensis 271.
Harju, A.M. and Nikkanen, T. 1996. Reproductive success of orchard and nonorchard
pollens during different stages of pollen shedding in a Scots pine seed orchard.
Can. J. For. Res. 26: 1096-1102.
Hartmann, H.T., Kester, D.E., Davies, F.T. and Geneve, R.L. 1997. Techniques of
seed production and handling. In: Plant propagation – principles and practices.
Sixth edition. Prentice Hall International, Inc. p.147-172.
Häcker, M. and Bergmann, F. 1991. The proportion of hybrids in seed from a seed
orchard composed of two larch species (L. europaea and L. leptolepis). Ann. Sci.
For. 48: 631-640.
Hellebergshaugem, F. 1970. Seed orchards as seed sources in forestry. Norsk Skogbr.
16(20): 420-422.
Ho, R.H. and Schooley, H.O. 1995. A review of tree crown management in conifers
orchards. For. Chron. 71: 311-316.
Hodge, G.R. and White, T.L. 1993. Advanced-generation wind-pollination seed
orchard design. New For. 7: 213-236.
Jones, N. and Burley. J. 1973. Seed certification, provenance nomenclature and
genetic history in forestry. Silvae Genet. 22: 53-58.
Jonsson, A., Ekberg, I. and Eriksson, G. 1976. Flowering in a seed orchard of Pinus
sylvestris L. Stud. For. Suec. 165. 38pp.
Kang, H. and Namkoong, G. 1988. Inbreeding effective population size under some
artificial selection schemes. 1. Linear distribution of breeding values. Theor. Appl.
Genet. 75: 333-339.
Kärkkäinen, K. 1994. Deleterious mutations and sexual allocation in the evolution of
reproduction of Scots pine. Ph.D. thesis. Oulu University, Finland.
Kellison, R.C. 1971 Seed orchard management. 11th congress on SFTIC, Atlanta, Ga.,
p.160-172.
Kimura, M. and Crow, J.F. 1963. The measurement of effective population number.
Evolution 17: 279-288.
Kjær, E.D. 1996. Estimation of effective population number in a Picea abies (Karst.)
seed orchard based on flower assessment. Scand. J. For. Res. 11: 111-121.
Kjær, E.D. and Wellendorf, H. 1997. Variation in flowering and reproductive success
in a Danish Picea abies (Karst.) seed orchard. For. Genet. 4: 181-188.
Knowles, P. 1985. Comparison of isozyme variation among natural stands and
plantations: jack pine and black spruce. Can. J. For. Res. 15: 902-908.
Kormut'ák, A. and Lindgren, D. 1996. Mating system and empty seeds in silver fir
(Abies alba Mill.). For. Genet. 3: 231-235.
Koski, V. 1970. A study of pollen dispersal as a mechanism of gene flow in conifers.
Comm. Inst. For. Fenn. 70. Helsinki, 1970. 78pp.
Koski, V. 1971. Embryonic lethals of Picea abies and Pinus sylvestris. Commun. Inst.
For. Fenn. 75.3. Helsinki, 1973. 30pp.
Koski, V. 1973. On self-pollination, genetic load, and subsequent inbreeding in some
conifers. Commun. Inst. For. Fenn. 78.10. 40pp.
Koski, V. 1980a. On the variation of flowering and seed crop in nature stands of
Pinus sylvestris L. Silva Fenn. 14: 71-75.
69
Koski, V. 1980b. Minimum requirements for seed orchards of Scots pine in Finland.
Silva Fenn. 14: 136-149.
Kossuth, S.V., McCall, E. and Ledbetter, J. 1988. Clone certification by use of
cortical monoterpenes as biochemical markers. Silvae Genet. 37: 73-76.
Krebs, C.J. 1985. The experimental analysis of distribution and abundance. 3rd
edition. Harper and Row, NY.
Lacy, R.C. 1995. Clarification of genetic terms and their use in the management of
captive populations. Zoo Biology 14: 565-578.
LaDeau, S.L. and Clark, J.S. 2001. Rising CO2 levels and the fecundity of forest trees.
Science 292: 95-98.
Lai, H.-L. and Wang, Z.-R. 1997. Advance in research on mating system of forest
populations. World Forestry Research 10: 10-15.
Lawlor, W. 1991. In: Impact of global climatic changes on photosynthesis and plant
productivity. Eds) Y.P. Abrol et al., Oxford & IBB Publishing, New Delhi, India.
p.431.
Li, B., McKeand, S and Weir, R. 1999. Tree improvement and sustainable forestry impact of two cycles of loblolly pine breeding in the USA. For. Genet. 6: 229-234.
Lindgren, D. 1975. The relationship between self-fertilization, empty seeds and seeds
originating selfing as a consequence of polyembryony. Stud. For. Suec. 126. 24pp.
Lindgren, D. and El-Kassaby, Y.A. 1989. Genetic consequences of combining
selective cone harvesting and genetic thinning in clonal seed orchards. Silvae
Genet. 38: 65-70.
Lindgren, D. and Gregorius, H.R. 1976. Inbreeding and coancestry. In: Proc. IUFRO
working parties on population and ecological genetics, breeding theory,
biochemical genetics and progeny testing. 1976. Bordeaux, France. p.49-72.
Lindgren, D. and Kang, K.S. 1997. Status number - a useful tool for tree breeding.
Res. Rep., For. Gen. Res. Inst. Korea. 33: 154-165.
Lindgren, D. and Matheson, A.C. 1986. An algorithm for increasing the genetic
quality of seed from seed orchards by using the better clones in higher
proportions. Silvae Genet. 35: 173-177.
Lindgren, D. and Mullin, T.J. 1997. Balancing gain and relatedness in selection.
Silvae Genet. 46: 124-129.
Lindgren, D. and Mullin, T.J. 1998. Relatedness and status number in seed orchard
crops. Can. J. For. Res. 28: 276-283.
Lindgren, D. and Wei, R.-P. 1994. Effects of maternal environment on mortality and
growth in young Pinus sylvestris in field trials. Tree Physiology 14: 323-327.
Lindgren, D., Gea, L. and Jefferson, P. 1996. Loss of genetic diversity monitored by
status number. Silvae Genet. 45: 52-59.
Lindgren, D., Gea, L. and Jefferson, P. 1997. Status number for measuring genetic
diversity. For. Genet. 4: 69-76.
Matheson, A.C., Eldridge, K.G., Brown, A.G. and Spencer, D.J. 1986. Wood volume
gains from first-generation radiata pine seed orchards. CSIRO Div. For. Res. User
Ser No. 4, 13pp.
Matheson, A.C., White, T.L. and Powell, G.R. 1995. Effects of inbreeding on growth,
stem and rust resistance in Pinus elliottii. Silvae Genet. 44: 37-46.
Matthews, J.D. 1963. Factors affecting the production of seed by forest trees. Forestry
abstracts 24 (1), 13pp.
70
Matziris, D. 1994. Genetic variation in the phenology of flowering in Black pine.
Silvae Genet. 43: 321-328.
Mikola, J. 1993. Provenance and individual variation in climatic hardness of Scots
pine in northern Finland. In: Forest development in cold climates. Ed) Alden, J.
Plenum press, New York, p.333-342.
Muhs, H.J. 1986. A synopsis of some international and national rules for forest
reproductive material moving into trade-OECD scheme, EEC directives and
German law of forest reproductive material. Proc. IUFRO J. Mtg. of WP S2.04-05
and S2.03-14. Grosshansdorf, Germany. 25-28 June 1985. p.25-37.
Mullin, L.J., Barnes, R.D. and Prevost, M.J. 1978. A review of the Southern Pines in
Rhodesia. Rhod. Bull. For. Res. No. 7. Rhodesia Forestry Commission. 328pp.
Müller-Stark, G. 1982. Tracing external pollen contribution to the offspring of Scots
pine seed orchard. In: Proc. IUFRO Mtg. on breeding strategies including
multiclonal varieties. Lower Saxony FRI, Dep. of For.Tree Breeding, Escherode,
Germany. p.176.
Nagasaka, K. and Szmidt, A.E. 1985. Multilocus analysis of external pollen
contamination of a Scots pine (Pinus sylvestris L.) seed orchard. In: Population
genetics in forestry. Ed) Dr. Gregorius H.R. Lecture notes in biomathematics 60:
134-138.
Namkoong, G., Snyder, E.B. and Stonecypher, R.W. 1966. Heritability and gain
concepts for evaluating breeding systems such as seedling orchards. Silvae Genet.
15: 76-84.
Nei, M. Maruyama, T. and Chakraborty, R. 1975. The bottleneck effect and genetic
variability in populations. Evolution 29: 1-10.
Nikkanen, T. and Ruotsalainen, S. 2000. Variation in flowering abundance and its
impact on the genetic diversity of the seed crop in a Norway spruce seed orchard.
Silva Fenn. 34: 205-222.
Nikkanen, T. and Velling, P. 1987. Correlations between flowering and some
vegetative characteristics of grafts of Pinus sylvestris. For.Ecol.Manage.19: 35-40.
OECD. 1974. Establishing an OECD scheme for the control of forest reproductive
material moving in international trade. C(74)29(final). 23p.
Olsson, T., Lindgren, D. and Ericsson, T. 2000. A comparison of group merit
selection and restricted selection among full-sib progenies of Scots pine. For.
Genet. 7: 137-144.
Omi, S.R. and Adams, W.T. 1985. Variation in seed set and proportions of outcrossed
progeny with clones, crown position and top pruning in a Douglas-fir seed
orchard. Can. J. For. Res. 16: 502-507.
O'Reilly, C., Parker, W.H. and Barker, J.E. 1982. Effect of pollination period and
strobili number on random mating in a clonal seed orchard of Picea mariana.
Silvae Genet. 31: 90-94.
Owens, J.N. 1995. Constrains to seed production: temperate and tropical trees. Tree
Physiology 15: 477-484.
Pakkanen, A. and Pulkkinen, P. 1991. Pollen production and background pollination
levels in Scots pine seed orchards of northern Finnish origin. In: Pollen
contamination in seed orchards. Ed) Dr. Lindgren, D. Proc. Mtg Nordic Group for
Tree Breeding. 1991. Sweden. p. 14-21.
Paule, L. 1991. Clone identity and contamination in a Scots pine seed orchard. In:
Pollen contamination in seed orchards. Ed) Dr. Lindgren, D. Proc. Mtg Nordic
Group for Tree Breeding. 1991. Sweden. p.6-13.
71
Paule, L. and Gömöry, D. 1992. Genetic processes in seed orchards as illustrated by
European larch (Larix decidua Mill) and Scots pine (Pinus sylvestris L.). In: Proc.
international conference on phytotechnique and forest management in present
ecological conditions, Technical University, Zvolen. p.109-114.
Paule, L., Lindgren, D. and Yazdani, R. 1993. Allozyme frequencies, outcrossing rate
and pollen contamination in Picea abies seed orchards. Scan. J. For. Res. 8: 8-17.
Pharis, R.P., Ross, S.D. and McMullan, E.E. 1980. Promotion of flowering in the
Pinaceae by gibberellins. III. Seedling of Douglas.fir. Physiol. Plant. 50: 119-126.
Philipsson, J.J. 1997. Predicting cone crop potential in conifers by assessment of
developing cone buds and cones. Forestry 70: 87-96.
Pollak, E. 1977. Effective population numbers and their interrelations. In: Proc. WSU
conf. on Biomathematics and biostatistics, 1974. Washington State University.
Portlock, F.T. 1988. Forest tree seed inspectors' manual: OECD scheme for
certification of forest reproductive material moving in international trade.
Information Report DPC-X-22. Forestry Canada, Ottawa, 1988. Vol. 22. 45p.
Pulkkinen, P. 1994. Aerobiology of pine pollen: dispersal of pollen from non-uniform
sources and impact on Scots pine seed orchards. Reports from the foundation for
forest tree breeding 8. 23p.
Reynolds, S. and El-Kassaby, Y.A. 1990. Parental balance in Douglas-fir seed
orchards - cone crop vs. seed crop. Silvae Genet. 39: 40-42.
Ritland, K. and El-Kassaby, Y.A. 1985. The nature of inbreeding in a seed orchard of
Douglas fir as shown by an efficient multilocus model. Theor. Appl. Genet. 71:
375-384.
Roberds, J.H., Friedman, S.T. and El-Kassaby, Y.A. 1991. Effective number of pollen
parents in clonal seed orchards. Theor. Appl. Genet. 82: 313-320.
Robertson, A. 1961. Inbreeding in artificial selection programmes. Genet. Res. 2:
189-194.
Ross, S.D. 1978. Influences of gibberellins and cultural practices on early flowering
of Douglas-fir seedling and grafts. In: Proc. 3rd World consultation on forest tress
breeding, Vol. 2, Eds) Brown, A.G. & C.M. Palmberg. CSIRO. Canberra. p.9771007.
Ross, S.D., Webber, J.E., Pharis, R.P. and Owens, J.N. 1985. Interaction between
gibberellin A4/7 and root-pruning on the reproductive and vegetative process in
Douglas-fir. I. Effects on flowering. Can. J. For. Res. 15: 341-347.
Rosvall, O. 1999. Enhancing gain from long-term forest tree breeding while
conserving genetic diversity. Ph.D. thesis. Swedish University of Agricultural
Sciences, Umeå, Sweden. Acta Universitatis Agriculturae Sueciae, Silvestria 109.
Rudin, D. and Ekberg, I. 1982. Genetic structure of open-pollinated progenies from a
seed orchard of Pinus sylvestris. Silva Fenn. 16: 87-93.
Rudin, D. and Lindgren, D. 1977. Isozyme studies in seed orchards. Stud. For. Suec.
139. 23pp.
Rudin, D., Muona, O. and Yazdani, R. 1986. Comparison of the mating system of
Pinus sylvestris in natural stands and seed orchards. Hereditas 104: 15-19.
Sarvas, R. 1962. Investigations on the flowering and seed crop of Pinus sylvestris.
Commun. Inst. For. Fenn. 53. 198pp.
Sarvas, R. 1970. Establishment and registration of seed orchards. Folia Forestalia 89.
24p.
72
Savolainen, O. 1991. Background pollination in seed orchards. In: Pollen
contamination in seed orchards. Ed) Dr. Lindgren, D. Proc. Mtg Nordic Group for
Tree Breeding. 1991. Sweden. p.6-13.
Savolainen, O. and Kärkkäinen, K. 1992. Effect of forest management on gene pools.
New For. 6: 326-345.
Savolainen, O. and Kuittinen, H. 2000. Small population processes. In: Forest
conservation genetics: principles and practice. Eds) A. Young, D. Boshier and T.
Boyle. CABI publishing, Unite Kingdom. p.91-100.
Savolainen, O., Kärkkäinen, K., Harju, A., Nikkanen, T. and Rusanen, M. 1993.
Fertility variation in Pinus sylvestris: a test of sexual allocation theory. Amer. J.
Bot. 80: 1016-1020.
Schmidtling, R.C. 1981. The inheritance of precocity and its relationship with growth
in loblolly pine. Silvae Genet. 30: 188-192.
Schmidtling, R.C. 1983. Genetic variation in fruitfulness in a loblolly pine (Pinus
taeda L.). Silvae Genet. 32: 76-80.
Schoen, D.J. and Stewart, S.C. 1987. Variation in male fertilities and pairwise mating
probabilities in Picea glauca. Genetics 116: 141-152.
Schoen, D.J., Denti, D. and Steward, S.C. 1986. Strobilus production in a clonal
white spruce seed orchard: evidence for unbalanced mating. Silvae Genet. 35:
201-205.
Schultz, R.P. 1971. Stimulation of flower and seed production in a young slash pine
orchard. U.S. Southeastern For. Exp. Station. USDA Forest Service, SE-91, 10pp.
Sedgley, M. and Griffin, A.R. 1989. Sexual reproduction of tree crops. Academic
press, London.
Setiawati, Y.G.B. and Sweet, G.B. 1995. Increase of seed yield and physical quality
in a Pinus radiata control-pollinated meadow seed orchard. Silvae Genet. 45: 9196.
Shaw, D.V. and Allard, R.W. 1982. Estimation of outcrossing rates in Douglas fir
using isozyme markers. Theor. Appl. Genet. 62: 113-120.
Shen, H.-H., Rudin, D. and Lindgren, D. 1981. Study of the pollination pattern in a
Scots pine seed orchard by means of isozyme analysis. Silvae Genet. 30: 7-15.
Silen, R.R. 1962. Pollen disposal considerations for Douglas fir. J. For. 60(11): 790795.
Silen, R.R. 1963. Effect of altitude on factors of pollen contamination of Douglas-fir
seed orchards. J. For. 61: 281-283.
Silen, R.R and Keane, G. 1969. Cooling a Douglas-fir seed orchard to avoid pollen
contamination. USDA, For. Serv., Res. Note PNW-101. 10pp.
Skroppa, T. 1994. Impacts of tree improvement on genetic structure and diversity of
planted forests. Silva Fenn. 28: 265-274.
Skroppa, T. and Tutturen, R. 1985. Flowering in Norway spruce seed orchards. Silvae
Genet. 34: 90-95.
Slee, M.U. and Spidy, T. 1970. The incidence of graft incompatibility with related
stock in Pinus caribaea Mor. var. hondurensis E. et G. Silvae Genet. 19: 184-187.
Sluder, E.R. 1994. Gains in fusiform rust resistance and height growth in a second
generation slash pine seedling orchard. Silvae Genet. 43: 41-48.
Smith, D.B. and Adams, W.T. 1983. Measuring pollen contamination in clonal seed
orchards with the aid of genetic markers. In: Proc. 17th SFTIC. p.69-77.
73
Sniezko, R.A. 1984. Inbreeding and outcrossing in Loblolly pine. Ph.D. thesis. Dep.
of forestry, NCSU. 50pp.
Sorensen, F. 1969. Embryonic genetic load in coastal Douglas-fir, Pseudotsuga
menzieii var. menziesii. Am. Nat. 103: 389-398.
Squillace, A.E. 1967. Effectiveness of 400-foot isolation around a slash pine seed
orchard. J. For. 65: 823-824.
Squillace, A.E. and Goddard, R.E. 1982. Selfing in clonal seed orchards of slash pine.
For. Sci. 28: 71-78.
Squillace, A.E. and Kraus, J.F. 1963. The degree of natural selfing in Slash pine as
estimated from albino frequencies. Silvae Genet. 12: 46-50.
Squillace, A.E. and Long, E.M. 1981. Proportion of pollen from nonorchard clones.
In: Pollen managmenet handbook. Ed) E.C. Franklin. US. Dep. Agric. Agric.
Handb. No. 586. p.15-19.
Stewart, S.C. 1994. Simultaneous estimation of pollen contamination and pollen
fertilities of individual trees in conifer seed orchards using multilocus genetic data.
Theor. Appl. Genet. 88: 593-596.
Stoehr, M.U. and El-Kassaby, Y.A. 1997. Levels of genetic diversity at different
stages of the domestication cycle of interior spruce in British Columbia. Theor.
Appl. Genet. 94: 83-90.
Stoehr, M.U., Orvar, B.L., Vo, T.M., Gawley, J.R., Webber, J.E. and Newton, C.H.
1998. Application of a chloroplast DNA marker in seed orchard management
evaluations of Douglas-fir. Can. J. For. Res. 28: 187-195.
Stoehr, M.U., Webber, J.E. and Painter, R.A. 1994. Pollen contamination effects on
progeny from an off-site Douglas-fir seed orchard. Can. J. For. Res. 24: 21132117.
Sweet, G.B. 1995. Seed orchards in development. Tree Physiology 15: 527-530.
Toda, R. 1964. A brief review and conclusion of the discussion on seed orchards.
Silvae Genet. 13: 1-4.
van Buijtenen, J.P. 1971. Seed orchard design, theory and practice. Proc. 11th SFTIC,
Atlanta, Ga. 1971. p.197-206.
Varghese, M., Nicodemus, A., Nagarajan, B., Siddappa, K.R.S., Bennet, S.S.R. and
Subramanian, K. 2000. Seedling seed orchards for breeding tropical trees. Scroll
press, Coimbatore, India. 126pp. ISBN 81-900346-1.
Varnell, R.J., Squillace, A.E. and Bengston, G.W. 1967. Variation and heritability of
fruitfulness in slash pine. Silvae Genet. 16: 125-128.
Verrier, E., Colleau, J.J. and Foulley, J.L. 1991. Methods for predicting response to
selection in small populations under additive genetic models: a review. Livest.
Prod. Sci. 29: 93-114.
Wang, T., Hagqvist, R. and Tigerstedt, P.M.A. 1999. Inbreeding depression in three
generations of selfed families of silver birch (Betula pendula). Can. J. For. Res.
29: 662-668.
Wang, X.R., Lindgren, D., Szmidt, A.E. and Yazdani, R. 1991. Pollen migration into
a seed orchard of Pinus sylvestris L. and the methods of its estimation using
allozyme markers. Scand. J. For. Res. 6: 379-385.
Wei, R.-P. 1995. Predicting genetic diversity and optimising selection. Ph.D. thesis.
Swedish University of Agricultural Sciences, Umeå, Sweden.
74
Wheeler, N. and Jech, K. 1986. Pollen contamination in a mature Douglas-fir seed
orchard. In: Proc. IUFRO conf. on Breeding theory, progeny testing and seed
orchards. Williamsburg, Oct 13-17, Virginia. p.160-171.
Wheeler, N.C. and Jech, K.S. 1992. The use of electrophoretic markers in seed
orchard research. New For. 6: 311-328.
Wheeler, N.C., Masters, C.J., Cade, S.C., Ross, S.D., Keeley, J.W. and Hsin, L.Y.
1985. Girdling: an effective and practical treatment foe enhancing seed yield in
Douglas-fir seed orchards. Can. J. For. Res. 15: 505-510.
Williams, C.G. and Savolainen, O. 1996. Inbreeding depression in conifers:
implication for breeding strategy. For. Sci. 42: 102-117.
Williams, C.G., Hamrick, J.L. and Lewis, P.O. 1995. Multiple-population versus
hierarchical conifer breeding programs: a comparison of genetic diversity levels.
Theor. Appl. Genet. 90: 584-594.
Wright, S. 1931. Evolution in Mendelian populations. In: Genetics, Brooklyn Botanic
Garden, Brooklyn, New York. p.97-158.
Wright, J.W. 1953. Pollen-dispersion studies: some practical application. J. For. 51:
114-118.
Xie, C.Y. and Knowles, P. 1992. Male fertility variation in an open-pollinated
plantation of Norway spruce (Picea abies). Can. J. For. Res. 22: 1463-1468.
Xie, C.Y. and Knowles, P. 1994. Mating system and effective pollen immigration in a
Norway spruce (Picea abies (L.) Karst) plantation. Silvae Genet. 43: 48-52.
Xie, C.Y., Yeh, F.C., Dancik, B.P. and Strobeck, C. 1991. Joint estimation of
immigration and mating system parameters in gymnosperms using the EM
algorithm. Theor. Appl. Genet. 83: 137-140.
Yazdani, R. and Lindgren, D. 1991. Variation of pollen contamination in a Scots pine
seed orchard. Silvae Genet. 40: 243-246.
Yazdani, R., Hadders, G. and Szmidt, A.E. 1986. Supplemental mass pollination in a
seed orchard of Pinus sylvestris L. investigated by isozyme analysis. Scand. J. For.
Res. 1: 309-315.
Yazdani, R., Lindgren, D., Seyedyazdani, F., Pascual, L. and Eriksson, U. 1995.
Flowering, phenology, empty seeds and pollen contamination in a clonal seed
orchard of Pinus sylvestris in northern Sweden. In: Population genetics and
genetic conservation of forest trees. Ed) Baradat,, W.T. Adams & G. MullerStarck. SPB Academic Publishing, Amsterdam, The Netherlands. p.309-319.
Yeh, F.C. 2000. Population genetics. In: Forest conservation genetics - principles and
practice. Eds) A. Young, D. Boshier and T. Boyle. CABI publishing. London.
p.21-37.
Zobel, B. and Talbert, J. 1984. Applied forest tree improvement. John Wiley & Sons,
New York, USA. 505pp.
Zobel, B.J., Barber, J., Brown, C.L. and Perry, T.O. 1958. Seed orchards; their
concept and management. J. For. 56: 815-825.
75
Acknowledgements
I would greatly appreciate my supervisor, Prof. Dag Lindgren for accepting and
teaching me as a Ph.D. student at the Department of Forest Genetics and Plant
Physiology in the Swedish University of Agricultural Sciences (SLU) and also
for supervising my doctoral thesis. He has given me excellent guidance through
four years in my study in Sweden.
My thanks have to go to Drs. Adolfo Bila and Anni Harju for their collaboration
and valuable discussion. Also, many thanks to my colleagues, Dr. Derius
Danusevicius, Dr. Seog-Gu Son, Mrs. Thuy Olsson, Mr. Torgny Persson, Mr.
Daoqun Zang, and ex-colleagues, Dr. Erik Andersson, Dr. Huen-Lin Li, Dr.
Leopoldo S. Rodriguez, Dr. Yong-Qi Zheng and Mr. Seppo Ruotsalneinen. It
has been a pleasure working with such nice colleagues.
I wish to deeply thank Dr. Tim J. Mullin at NCSU in the USA, Dr. Erik D.
Kjær in Denmark, Profs. Jung Oh Hyun and Min Sup Chung in Korea, and Dr.
Yill Sung Park in Canada for their cooperation, comments and reviewing at the
stage of manuscripts. I also thank Dr. Teijo Nikkanen from the Finnish Forest
Research Institute and Mr. Curt Almqvist from the Swedish Forest Research
Institute (SkogForsk) for their co-operation in paper VII and fruitful discussion.
Also, many thanks to people from SkogForsk at Sävar. It was a particular
pleasure for me to have encouragement and congratulations from Dr. Gene
Namkoong on completing the piece of my work.
I would like to sincerely acknowledge all colleagues (especially Drs. Alfred E.
Szmidt, Anders Fries, Dag Rudin, Jan-Erik Nilsson and Katarina Lindgren)
at this department for providing me warm friendship and for sharing seminars and
discussion during my time in Umeå. I will keep in mind your cooperation,
Estelle, Laura, Maria, Wan, Xiao-Ru, Tongming and Vikram. Gertrud
Lestander, Maria Lindgren and Stefan Löfmark have helped and assisted me
like a mother, a sister and a brother, whenever and whatever I needed.
My employer, the Korea Forest Research Institute, the Swedish Kempe
Foundation and Föreningen Skogsträdsförädling have supported my study
grants. I thank my boss, Dr. Wan-Yong Choi, for encouraging me to be a Ph.D.
student in the field of seed orchard research.
I am especially grateful to my parents, brother and sisters for their
unconditional support and understanding. Finally, my infinite thank goes to my
wife, Eun-Joo Yang and to my son, Young-Jong (Swedish name, “Sam”) Kang
for their patience and tolerance of the long, dark, cold Swedish winter. How
could I live without you! Many kisses and hugs to both of you! I want to dedicate
this doctoral thesis to my grandmother and father-in-law who passed away before
this thesis is finished.
76