Fig. 2.
Detection of the uppermost hierarchical level of genetic structure in the complete microsatellite dataset. Ln P(D) and convergence between replicates (open circles with error bars) as well as Delta K (filled circles) were used for the detection of the most likely number of K. The graph is modified from the output derived from Structure Harvester [36]. According to these results, K = 3 is adequate to describe the structure in the dataset