Figure 1.
Localities for all Galaxiella samples examined.
Refer to Table 1 for corresponding locality details. Each species is represented by a different symbol, with populations from both lineages of G. pusilla color coded to match Figs. 2, 3 and 5. Shaded areas refer to the general distribution of each Galaxiella species. Low sea level drainage patterns are shown to the minus 135 m bathymetric contour. Bathymetry predicts a large depression that we refer to as Lake Bass. The small inset of Australia shows the relevant biogeographic provinces, and the black central southern region represents the extent of the Eucla Basin.
Table 1.
Locality data for all individuals examined.
Figure 2.
Phylogenetic results for all Galaxiella samples examined.
Maximum likelihood trees for Galaxiella based on analysis of cytochrome b (A) and S7 (B) sequences and a neighbour joining tree for allozymes (C). Panel A shows the complete tree with operational taxonomic units (OTUs) for each Galaxiella species collapsed. Expanded trees within each species are shown below, with the lower scale bar applying to each of the four subtrees. Bootstrap values are based on 1000 pseudoreplicates. Trees are rooted with Brachygalaxias. Each OTU code is based on the sampling location number and name in Table 1 and Fig. 1, while the coloured symbols match Figs. 1, 3 and 5.
Figure 3.
Haplotype networks for each Galaxiella species examined based on the cytochrome b gene.
Each haplotype within G. pusilla (A) is color-coded relative to Figs. 1, 2 and 5. Circle size represents haplotype abundance; the key to circle size is in the center of panel A. The ancestral haplotype in each network is indicated by a box. Haplotype counts are given in parentheses when multiple populations share the same haplotype. Haplotype labels consist of the population number and name from Table 1 and Fig. 1. Unsampled haplotypes are represented by small filled circles.
Table 2.
Results from BEAST dating analyses based upon different combinations of calibrations.
Figure 4.
Historical demography of Galaxiella pusilla based on Bayesian skyline plots of female effective population size (Ne) changes through time.
Mean posterior Ne estimates for each species (darker lines) are bounded by upper and lower 95% highest posterior densities (G. pusilla west, pink shading; G. pusilla east, blue shading). The x-axis represents units of time in thousands of years ago (ka), scaled according to posterior mutation rates estimated in BEAST. The y-axis shows estimated population size in hundreds of thousands, calculated assuming a G. pusilla generation time equal to 1.0. The fish shown is a male G. pusilla west (by MPH).
Figure 5.
Principal coordinates analyses of Galaxiella pusilla.
PCO for all 100 G. pusilla east (A) and all 82 G. pusilla west (B) individuals. Relative PCO scores are plotted for the first and second dimensions, which individually explain 41% and 10% (A) and 25% and 17% (B) respectively of the total multivariate variation present. Individuals are colour-coded relative to Figs. 1, 2 and 3.