Fig. 4: Clustering analysis of native and introduced populations of Mimulus guttatus.

K-means clustering analysis of native and introduced populations of Mimulus guttatus. The analysis is based on the first 300 principal components. A Bayesian Information Criterion values for models ranging from 2 to 15 clusters. B Group membership of each geographic group for the optimal number of clusters (K = 8). C Principal component analysis depicted in Fig. 3A but coloured by the groups identified in the K-means cluster analysis (K = 8). To improve clarity a polygon was drawn around each of the eight genetic groups. The genetic groups of native populations used for the ABC analysis are named in parenthesis: AKBC = North Pacific Clade.