Figure 1.
Phylogenetic relationships of Wallabia and the three Macropus subgenera, M. (Macropus), M. (Osphranter) and M. (Notamacropus).
(A) The supertree of Cardillo et al. [6] summarizing previous molecular and morphological phylogenies and (B) Meredith et al.'s [7] evolutionary timescale (ave. of four BEAST analyses), showing the 2â2.4 Ma duration divergence cluster. Both trees are modified to include only the taxa sampled in the present study. Dendrolagini was not recovered by Cardillo et al. [6], however its inclusion in the summary tree is warranted on subsequent strong evidence from morphology [2] and all recent molecular analyses. Photos include (from the top) W. bicolor, M. rufogriseus (left), M. irma (right), M. rufus and M. giganteus. Photo credits â Matt Phillips, except M. irma (Ric Dawson) and M. rufus (Daniel Hoops).
Figure 2.
Phylogenetic analysis of kangaroos and wallabies.
Maximum likelihood phylogenies inferred from the (A) mitochondrial (Mt16) and (B) nuclear (Nuc17) concatenated datasets, with RAxML bootstrap values (BPML) above branches and MrBayes Bayesian posterior probabilities (BPP) below branches. The mt placement of M. dorsalis is derived from the reduced-length Mt17 and the mt placements of M. antilopinus and W. bicolor (NSW, New South Wales) are derived from the Cytb18 alignment. Support for grouping M. eugenii and M. agilis increases (BPML = 88; BPP = 0.98) for Mt16, which excludes M. dorsalis, but increases sequence length. Asterisks indicate full support. Clades including members of Macropus are shaded.
Figure 3.
Macropodid clade support from datasets simulated under coalescence.
(A) Simulation workflow. (B) Mean number of the five nuclear genes supporting each clade in maximum likelihood analyses of 200 simulations of the combined data *BEAST species tree for Ne values of 1,000 (triangle), 10,000 (open circle), 100,000 (square) and 1,000,000 (filled circle). For comparison, the grey bars show the number of genes supporting each clade on the observed data. (C) Percentage of ML analyses supporting each clade among 200 mtDNA simulations on the nuclear-only *BEAST species tree for Ne values set to mitochondrial equivalency for the same populations (one quarter of the corresponding nuclear values). Abbreviations: Lagor.; Lagorchestes, Wall.; Wallabia, M. (Notamac.); M. (Notamacropus), M. (Osphran.); M. (Osphranter).
Figure 4.
Macropodid species tree estimates from the combined mitochondrial and nuclear sequences (MtNuc16).
(A) concatenated sequences, showing BPP/BPML (@ = 0.89/59), (B) *BEAST partitioned between the mtDNA and five nuclear genes, showing BPP values. (C) both MDC and BUCKy, which recovered the same tree from the mt and five nuclear gene trees. (D) Meredith et al. (2008) with BPML values included for comparison. Several supraspecific clades that were identical across all reconstructions were collapsed for visualization convenience. Relationships within each of the collapsed clades were as inferred in Figure 2. Asterisks indicate full BPP or BPML support.
Table 1.
Approximately unbiased (AU) test results.
Table 2.
Individual nuclear gene âlnL differences and SH test results.
Table 3.
Macropodoid divergence time estimates in millions of years before present.