Within M. megacephalus, network and population genetic
analyses of the eastern clade find that the most
common haplotypes are from the eastern subclade (i.e.,
localities of Beryl and Minersville; Figs. 1 and 3a). This
indicates that kangaroo mice from southwestern Utah
may represent the ancestral population for the eastern
clade, a possible source for recent radiations and expansions
to the north and west (Table 3; Fig. 4a). Indeed,
the location of this ancestral population (immediately
south of the Great Salt Lake) suggests that the demographic
history of the eastern clade may be linked
intimately with the lacustral intervals of the ancient
Pleistocene Lake Bonneville.
Within the central clade of M. megacephalus, kangaroo
mice from multiple localities in the central subclade appear
to represent the ancestral population (localities Gold Reed,
SE Tonopah, Sunnyside, Warm Springs, NE Warm Springs,
and SE Warm Springs; Figs. 1 and 3b) and population
genetic analyses support a strong signal of population expansion
(Table 3; Fig. 4b). Evidence of population expansion
in the central clade also was supported in directional
analyses of phylogeographic patterns (DAPP; Hafner et al.
2008) performed in Hafner and Upham (2011). These DAPP
analyses uncovered a web orientation pattern of haplotype
sharing for the central clade of M. megacephalus, indicating
the presence of a southern refugium during cooler climatic
periods followed by subsequent expansions to the north
during warmer times (possibly during the last 100,000
years; Fig. 4b). Interestingly, kangaroo mice isolated in theMono Basin of the central clade’s western subunit (localities
Benton and Fletcher; Figs. 1 and 3b) consist of unique
haplotypes that are not shared with other M. megacephalus.In the M. megacephalus western clade, the ancestral population
seems to be represented in northwestern Nevada