Genetic analysis using cellulose acetate electrophoresis was performed on 12 populations of C. v. nodulosus throughout most of its distributional range, and including those used for mark-recapture studies above. A total of 308 specimens were included in the analysis for a mean sample size of 25 individuals and 16 allozyme loci (Matern et al. 2009; Fig. 2). Despite small population sizes (inferred from small areas of suitable habitats at most sampling sites), we found no evidence for inbreeding, as populations did not deviate significantly from Hardy-Weinberg equilibrium (GENEPOP 3.4, Raymond and Rousset 1995). However, allele numbers and the amount of polymorphism per population were low, with only eight of the investigated loci found to be polymorphic and a maximum of seven polymorphic loci per population, indicating genetic impoverishment.
Population differentiation and correlation between genetic and geographical distances between samples were quantified with the Arlequin package (Excoffier et al. 2005). Despite the low number of private alleles, populations were highly differentiated (overall FST > 0.45) owing to locally differentiated allele frequencies among populations, as exemplified by Fig. 2. This finding supports the possibility that populations were affected by genetic drift due to small population sizes. Neither clinal variation nor a directional loss of alleles that could hint at any postglacial recolonization scenario according to the allele elimination hypothesis (Reinig 1938; Hewitt 1999) could be detected. Generally, geographic distance only had a marginal influence (< 10%) on the isolation of populations (Mantel test). A neighbor-joining tree based on genetic distances among all C. v. nodulosus populations confirms this result, showing poor agreement between genetic and geographical distance across the entire range (not shown). Even neighboring populations at distances of only 2-3 km in the same forest and drainage system were found to be virtually independent of each other as shown, for example, by significant pairwise FST values of 0.27 and 0.14 between populations 1 and 2 and populations 5 and 6, respectively.
Strongly separated were two populations from Southern Slovenia (sites 11 and 12) which showed deviating allele frequencies and some private alleles compared to the other investigated populations. This is confirmed by Bayesian structure analysis (STRUCTURE, Pritchard et al. 2000) which generally assigns the individuals of these two populations to one distinct genetic cluster if no a priori assumptions of geographical origin are made (Matern et al. 2009). This pattern
J 200 km \ 12 \
Fig. 2 Allele frequencies of (a) alcohol dehydrogenase, (b) glucose-6-phosphate isomerase, (c) aspartate aminotransferase, and (d) mannose-6-phosphate isomerase in the investigated populations of C. variolosus nodulosus. Different colours indicate various alleles at the respective locus
Fig. 2 Allele frequencies of (a) alcohol dehydrogenase, (b) glucose-6-phosphate isomerase, (c) aspartate aminotransferase, and (d) mannose-6-phosphate isomerase in the investigated populations of C. variolosus nodulosus. Different colours indicate various alleles at the respective locus may be caused by genetic drift and an efficient present isolation between the Southern Slovenian samples and the remaining Central European populations that might have existed since the survival of these populations in different glacial refugia (cf. Chaps. 4 and 5.1, this contribution).
Was this article helpful?