Allele frequencies, genetic relationships and heterozygosity in Polygala vulgaris populations from contrasting habitats in southern Britain

Polygala vulgaris is a widespread perennial plant species of basic to neutral grassland throughout Britain. It is variable in flower colour and some vegetative characters in most populations. Gene flow is restricted in all populations since the plants are normally self-pollinating (up to 20% outcrossing, calculated from figures of heterozygosity) and seeds are normally dispersed only up to 2 m from the parent plant. Examination of several enzyme systems, using starch gel electrophoresis, showed that all enzyme loci appeared to have been duplicated, probably owing to the fact that P. vulgaris is a tctraploid in a genus which has primarily diploid members in Europe. Polymorphisms were found at both of the duplicated loci in four enzyme systems and at one locus in each of two further systems. Fifty-one populations from a diversity of habitats were sampled in southern Britain. In most populations from the typical habitat of lowland calcareous grassland, polymorphisms occurred at five or more loci, and the alleles were distributed throughout the populations. In all these sites there was a diversity of microhabitat and, in one area, marked differences in microhabitat distribution between different enzyme morphs. It is likely that differential selection is an important factor in the maintenance of the polymorphisms in all these areas. These populations were divided, genetically, into two main groups, those from the Gower peninsula, South Wales, and those from the English chalk. Within each group the populations were quite closely related genetically, though some large differences at particular loci occurred even between adjacent populations. Diffuse clumps of alleles occurred within all these populations, and it is likely that limited gene flow has affected the genetic structure of the populations, particularly since cycles of local extinction and recolonization are likely to be recurring. The genetic structure of each population, at the allozyme loci examined, is likely to be the consequence of a balance between differential selection and restricted gene flow. The number of heterozygotes was small, though higher than expected from an analysis of experimental progeny, suggesting a hcterozygotc advantage. Few strong multilocus associations were found despite the predominance of self-pollination. In the populations from sand dunes, roadsides, submontane grassland and water meadows there was, in general, much less polymorphism than in those from lowland calcareous grassland and some populations were entirely monomorphic; in most others there were large, entirely monomorphic groups of plants. Genetic distances between these populations were variable, even within one habitat and area, and they did not appear to show any particular relationship to each other. It is likely that the genetic structure of these populations, at the loci studied, has arisen largely through chance colonizations and genetic drift.