Population Genetics in Forestry

Genetic management of forest tree species requires a long-term perspective on the development of breeding populations. To achieve any 1 recurrent selection objectives, we must understand gene interactions, the distinctions between mating demes and selection demes, and the stability of these interacting systems. The present understanding of gene actions in forest trees and the interaction of selection and mating systems are discussed, and some implications are drawn for breeding in partially managed forests.

The complex cytology of European birches is considered from three points of view namely hybridisation in controlled and natural conditions, somatic chromosomal variation and the role of the dwarf birch in speciation and variation of arborescent birch.

European studies, including recent work at Aberdeen, have indicated that it is possible to produce viable seed and hybrid progeny from all combinations of B. pubescens, B. pendula and B. nana crosses. The success of hybridisation is subject to environmental influences and cytological factors such as relative ploidy levels of forests.

Although fertility of hybrids is low and there is no evidence of their ability to survive under natural conditions studies of Scottish birch populations appear to indicate that hybridisation and backcrossing contribute to the variation within and between populations. The situation is complicated by the occurrence of aneuploidy and somatic chromosome variation.

Isozyme markers have stimulated efforts to measure the mating system in natural and planted populations of forest trees. When open-pollinated progenies have been surveyed for allozyme variation, the arrays have usually been analyzed in terms of the model of mixed self-fertilization and random outcrossing. This procedure has encountered several problems of which the more difficult are due to heterogeneity in outcrossing rates or in pollen allele frequencies. Recent models which deal with these problems and enrich the study of plant mating systems include multilocus estimation, measurement of ’effective’ selfing and of differential male fertility. More reliable estimates of outcrossing, which these methods encourage, will be needed to predict selection gains, management practices and optimal conservation strategies in forest trees.

The genetic structure of natural plant populations results from the interaction of selection, gene flow and genetic drift. Reviews of the plant allozyme literature demonstrate that the distribution of allozyme variation within and among plant populations is closely associated with the species’ mating system, pollination ecology and seed dispersal mechanism. Yet, there are relatively few species for which dependable estimates of the mating system or of pollen or seed dispersal are available.

Estimates of plant mating systems based on the mixed mating model have provided insights into the breeding structure of a few species. There are, however, relatively few data available on temporal and spatial variation in the mating system. Also, the assumptions of the mixed mating model are often violated in natural populations. Finally, the mixed mating model is limited in its ability to provide information about the breeding structure of populations.

Our understanding of gene flow via pollen or seed in natural populations is poor at best. Estimation procedures based on pollinator movements have underestimated pollen movement by not adequately dealing with pollen carryover and do not measure the effective movement of genes. Procedures using genetic markers, although giving a more accurate measure of pollen flow in natural or artificial populations typically produce results which are limited in scope and generalizations are difficult.

The use of paternity analysis to identify the father of individual seeds or seedlings removes many of the problems inherent to estimates of gene flow or the mating system. Although requiring considerable effort, paternity analysis can determine several genetic parameters that have previously been difficult or impossible to measure. From such analyses a detailed picture of the mechanisms which interact to produce the genetic structure of plant populations can be obtained.

Partial inbreeding due to sibmating can reduce heterozygosity beyond that resulting from partial self-fertilization alone. A negative autocorrelation in inbreeding over time, results in higher heterozygosity than when there is no autocorrelation. High selfing can result in a greater probability of polymorphism when there are selective differences over space. When two loci are considered simultaneously, inbreeding may increase or decrease the likelihood of polymorphism. Genetic hitchhiking may be basis of genetic change in highly selfed plants.

From previous investigations on genotypic equilibria existence, we know that a diallelic polymorphic overdominance equilibrium cannot simultaneously show a high asymmetry in homozygotic fitness disadvantage and a strong deficit in heterozygote frequency relative to Hardy-Weinberg proportions. Since on the other hand sexually asymmetrical selection during reproduction is able to compensate a reduction in heterozygote frequency originating from inbreeding effects, the question arises as to whether the above property becomes meaningless in the presence of sexually asymmetrical selection. To answer this, polymorphic equilibria are calculated and also characterized for a multiallelic model of fertility resource allocation and partial self-fertilization where the fertilities are allowed to show extreme sexual asymmetry.

It turns out that for this model two types of equilibria can occur. For equilibria of the first type the above mentioned property can serve to investigate their existence. For equilibria of the second type all genotypic fitnesses are always identical, so that the above property is trivially fulfilled and does not provide any further insights. Moreover, the latter type of equilibrium directly reflects the influence of sexual asymmetry.

As a consequence, it is confirmed that sexual asymmetry represents an efficient mechanism for the establishment of additional polymorphic equilibria even in the presence of strong inbreeding effects, and it is suggested that protectedness of polymorphisms can be classified according to the underlying mechanism of either symmetry in homozygotic disadvantage or sexual asymmetry.

Isozymes from viable embryos were used to study the distribution of the effectively incorporated pollen pool in a Scots pine seed orchard. Although selfing was not a significant component of the mating system in this orchard, several forms of non-random mating were observed. These included possible non-random mating patterns marked by one locus, heterogeneous pollen pool distributions, and possible asynchronous sexual phenology schedules among the various clones. Implications of these, distortions on the genetic structure and effective population size of the filial generation are discussed.

To quantify the reproductive success of individual genotypes in different environments-, the realized female and male fitness values were estimated for single clones in each of two spatially and two temporally different environments. Each of the studied clones carries a unique multilocus genotype enabling the determination of its contribution to the offspring. The results are based on 1o-locus genotyping of the parental clones and the four seed orchard offspring populations, represented by 500 seeds each.

The reproductive success of the clones differ substantially from one another, regardless of whether the female, the male or both fitness components of a clone are concerned. Strong sexual asymmetries are observed which also occur in sexually opposite directions. The results do not support the assumption of genetically controlled reproductive efficiencies. However, in some cases it cannot be ruled out that this is due to environmentally dependent selection with respect to the sex function of single clones. Some consequences of the observed phenomena are outlined briefly.

Genotypes at 18 allozyme loci were determined for all clones and their half-sib embryo families in a clonal seed orchard of Scots pine. The proportion of pollen gametes with genotypes that could not have been produced by parental clones was determined and used as a minimum estimate of external pollen contamination. We found that 37.8% of analysed embryos were fertilized by external pollen.

The spatial distribution of unique allozymes among 785 naturally regenerated plants (aged 10 -20 years) established under seed trees was determined using electrophoretic techniques. Genes from four seed trees carrying marker alleles were traced in seedlings in the vicinity of those trees. The contribution of a seed tree to plant regeneration close to that seed tree was about 5 %; thus plants are usually not the progeny of the closest seed tree. Groups consisting mostly of half-sibs were not evident. As the half sibs seemed to be distributed rather widely, half-sib mating is probably not a common source of inbreeding in a naturally regenerated stand.

A model of gene dispersal was developed. According to this model the proportion of all genes dispersed at certain distances from the source are calculated. The approximate percentages of genes travelling different distances are 0 – 5 m, 4 percent; 5 – 10 m, 8 percent; 10 –15 m, 11 percent; 15 –50 m, 40 –75 percent.

Examination of embryos from seeds produced by seed trees has revealed a large excess of homozygotes. This excess of homozygotes probably resulted from inbreeding caused by selfing. Estimation of selfing by rare and unique isozyme markers from thirteen trees gave an average value of 11.8 percent at the embryo stage. Selfing rates appear to be dependent on seed-tree density, climate conditions at time of pollen flight and size of pollen and seed crop.

Processes involved in the evolution of plant reproductive systems are reviewed. It is shown that constraints imposed by the rules of genetic transmission, and by the properties of available genetic variation in the breeding system, may significantly affect the outcome of natural selection on reproductive systems. Theoretical predictions of selection models are compared with the evidence from genetic and comparative studies of breeding systems. Some practical implications for the plant breeder are discussed.

This paper gives a personal view of the evolution of outbreedlng systems, emphasizing the many features that hermaphrodite and monoecious populations, especially the more outbred ones, have in common with sex-polymorphic populations. Such features include frequency-dependency for fitness, functional sex, and combined gamete selfing rate. This last selfing rate differs from that for the ovules, since the pollen selfing rate differs from the ovule rate and is frequency dependent, because of variation in ovule and pollen fertilities. Five topics are emphasized. These are: (1) the concept of successful gametes, defined as gametes (of both sexes) that take part in fertilization; (2) sexual asymmetry, defined as non-constant pollen to ovule ratios among individuals (or genotypes or phenotypes); (3) allocation of resources to male or female reproduction, where such allocation may differ among individuals; (4) individual selection; and (5) genetic control. The concept of successful gametes is used to define fitness (the total number of successful gametes per individual), functional sex (the number of successful ovules as a proportion of all successful gametes, or fitness value), and combined gamete selfing rate (the number of ovules and pollen grains of an individual that take part in selfing, as a proportion of all successful gametes of that individual). Sexual asymmetry is he-Id to be of fundamental importance not only for sex-polymorphic, but also for the more outbred hermaphrodite populations, and evidence for asymmetry is presented for hermaphrodite and monoecious populations. Asymmetry is responsible for maintenance of polymorphisms, and also for frequency-dependent selection, differential ovule and pollen outcrossing rates, and variation in extents of male and female functioning, so that these last three characters are common to asymmetric hermaphrodite and to sex-polymorphic populations. A simple model of male/female resource allocation ensures that any variation in such allocation results in asymmetry with frequency-dependent selection. Numerical examples show that intermediate or incompletely dominant gene action for resource allocation may result in overdominance or in equal genotypìc fitnesses in equilibrium populations, for hermaphrodite, gynodioecious, or subdioecious populations. There is a continuum among these population types. It is held that individual selection is sufficient to bring about the evolution of sex polymorphisms, and such evolution is not always accompanied by the evolution of outcrossing. The importance of the mode of genetic control of a sex phenotype is emphasized, since differences in such control may entirely determine whether a polymorphism can be maintained or not.

The change of the genetic structure of three plant species of woodland clearings, Senecto sylvattcus, Dtgttalts purpurea and ScrophuVarta nodosa has been studied in relation to the management of these clearings.

All populations of the biennial Digitalis purpurea consisted of plants with glabrous and hairy flower stalks, the latter being the recessive homozygote. Mowing of herbs and gresses, but not of shrubs and young trees in woodland clearings extended the survival of the populations for some more generations, and gave rise to a selection in favour of the glabrous genotype of D. purpurea.

Populations of the perennial Scrophularia nodosa contained iron efficient and inefficient plants. The ability to change the pH around the root seems to be coded by a single gene. In addition to this acidification Fe-efficiency was further determined by the Fe-translocation from the root to the shoot. Genotypes with a slow Fe~translocation were inferior in flower and seed production and unable to produce seeds if ploughing of woodland clearings before afforestation had diminished the soil organic matter of the A_o-horizon.

Two tropical forest tree species, Altingia excelsa Noronha and Agathis borneensis Warb. were investigated for their breeding structure in natural forests. Assuming that inbreeding would produce higher similarity than random mating in the pattern of isoperoxidase bands in individual trees, the degree of disagreement between trees in isoperoxidase patterns was calculated for estimating occurrence of inbreeding. By finding that trees alike in the isoperoxidase patterns were forming separate groups in the forest in Altingia, five family clumps were presumed (Figure 3). It was found that those different families were dissimilar in respect of isoperoxidase constitution as well as in several leaf characters. The distance between two trees at which they could mate was estimated to be 16 to 18 meters and the area one family occupies 200 to 250 m² on the assumption that inbreeding in dioecious Altingia should have occurred by consanguineous mating within each family. In spite of growing together of two families within the mating distance, it was found that they showed apparent genetic differentiation between them suggesting that they had been sexually isolated from each other. This sexual isolation, if it happened, is supposed to be due to genetic difference in flowering time.

Genetic variation of Pinus sylvestris was studied in three natural populations in northern Sweden (latitudes 66°40′N, 65°30′N, and 64°30′N). From seed harvests of each population the macrogametophyte and embryo of 133–134 seeds were studied with respect to 14 variable enzyme loci. The dègree of genetic differentiation between the populations was very low (G_ST=0.006). The average fixation indices in the populations were 0.135, 0.092, and 0.085, respectively. These positive fixation indices are probably mostly due to selfing. Garnetic disequilibrium between the loci was studied in zygotes and separately in the ovule and pollen pools. Disequilibrium was also divided into between and within individual components. Small, but significant disequilibria were found in the ovule pool of the northernmost population. The between individual component was also significant in this population. As the significantly associated locus pairs were unlinked, we concluded that the disequilibria are due to restricted effective population size. The decrease in the effective size may be due to uneven seed production, partial selfing and different male and female numbers, which are all known to occur in populations of Pinus sylvestris.

Genetic variation within and differentiation among nine Scots pine populations from the lowlands and the mountains in Poland were examined at nine gene loci coding for five enzyme systems involved in the energy and amino acid metabolism.

In general, the genetic variation in populations estimated as gene diversity appears to be higher than in other conifer species. However, lowland populations do not reveal a higher level of gene diversity than those from the mountains.

Based on estimates of genetic distance and a new measure of genetic differentiation, it was found that a considerable differentiation exists among lowland populations from different regions and between lowland and montane populations, whereas much lower differentiation was observed among populations from the mountains, thus roughly reflecting the different geographical distances. On the other hand, it was suspected that the low degree of differentiation found between the montane populations and the overall gene pool may indicate their relationship to former glacial refugia.

Air pollution is currently being shown to have many adverse effects on forest tree stands in Central Europe. However, ecological-genetic consequences of such a permanent stress factor were not considered up to now.

In order to obtain information about possible selection pressure caused by air pollution, numerous clones of Norway spruce were fumigated with SO₂ and subsequently ranked according to their visible damage. A comparison of a clone group of high sensitivity with one of low sensitivity revealed differences in genetic structure at four of eight enzyme gene loci studied. Surprisingly large difference in allele and genotype frequencies were found at one of these enzyme loci (G6PDH-A).

To demonstrate the possible effect of viability selection over the generations, the change in allele frequencies over several generations is predicted under certain simplifying assumptions concerning the fumigated clone material. It is shown that there is a very high probability of losing an allele at one or more of the four enzyme loci if air pollution stress persists.

It is argued that the studies of genetic differentiation within plant populations should explicitly take into account three major stages in the life cycle: adult plants, seed before dispersal, and seed after dispersal. The simultaneous consideration of these three stages is necessary in order to evaluate the effectiveness of the three main causes of differentiation, namely, non-random pollination and fertilization of the ovules, limited seed dispersal, and locally differential selection. In order to experimentally analyze genetic characteristics of adaptive strategies of plant populations, it is suggested that the association between genetic diversity and the combination of genetic differentiation at the three stages be considered.

The shortcomings of Wright’s F_ST measure for population differentiation are pointed out, and a new measure is proposed which consistently combines the levels of differentiation for the individual subpopulations with the total level of population differentiation. The use of this measure is demonstrated with the help of two experimental data sets, one referring to differentiation in the mating system of pines (differentiation within populations), and the other concerned with differentiation among populations of wild barley in Israel (differentiation between populations). Appropriate data sets alowing simultaneous anlysis of the three major life cycle stages do not yet seem to have been obtained.