A unifying study of phenotypic and molecular genetic variability in natural populations of Anadenanthera colubrina var. cebil from Yungas and Paranaense biogeographic provinces in Argentina.

Anadenanthera colubrina var. cebil is a discontinuously distributed native tree species in South American subtropical forests. Thirteen quantitative traits and eight nuclear microsatellite loci were examined in individuals from two biogeographic provinces of Argentina to determine the number and composition of genetically distinguishable groups of individuals and explore possible spatial patterns of the phenotypic and genetic variability. Means of reproductive traits were higher in the Yungas than in the Paranaense biogeographic province, whereas five out of eight nonreproductive quantitative traits showed higher mean values in the latter. Variance coefficients were moderate, and there were significant differences between and within provinces. Three clusters were defined based on spatial model for cluster membership for quantitative traits. One cluster grouped the individuals from the Paranaense biogeographic province whereas the individuals from the Yungas biogeographic province grouped regarding its population of origin. Parameters of molecular genetic variability showed higher values in the Yungas than in the Paranaense biogeographic province. Observed heterozygosity was lower than expected heterozygosity in both biogeographic provinces, indicating an excess of homozygosity. The homozygosity test by Watterson and the exact test by Slatkin suggested diversifying selection for locus Ac41.1. Bayesian clustering spatial model for microsatellites loci data were performed for both all loci and for all loci excluding locus Ac41.1. In both analyses two clusters were inferred. Analysis of molecular variance revealed similar results for all genotypes and for all genotypes defined excluding locus Ac41.1. Most of the total variance is attributable to genetic variation within clusters. The presence of homogeneous clusters was detected for both the phenotypic and molecular genetic variability. Two Bayesian clustering analyses were performed according to molecular genetic data, and two clusters were inferred. Individuals were assigned to their provinces of origin. Genetic molecular variation was higher in the populations of the Yungas biogeographic province which translates into highly qualified populations for conservation. Populations from the Paranaense biogeographic province showed the highest mean value of number of seeds per fruit making them valuable as well with regard to the exploitation of management strategies as a means to recover the impacted areas where these populations are located.