No. 410] ` THE MAMMALIAN SUBCLASSES 121 
hallux. The greatest departure from this condition is seen in 
Peragale leucura (Fig. 1, 4) and in Cheropus castanotis, in both 
of which the hallux is entirely obliterated, the fourth digit 
greatly enlarged, and the remaining digits much reduced. 
The exclusive predominance of the fourth digit in Choeropus 
Dollo regards as significant evidence of former arboreal habit, 
since in other cases in which there is a tendency towards mono- 
dactylism, as, for example, in the horse, it is not the fourth 
digit which predominates, but the third. The fourth digit must 
have been enlarged, and the third reduced, in the Peramelidze, 
in connection with arboreal habit, before the tendency towards 
monodactylism appeared in connection with terrestrial, curso- 
rial habit ; so that the already enlarged fourth digit became still 
more predominant, while the third became still more reduced. 
The kangaroos and wallabies also have a functionally mono- 
dactyl foot, resembling closely that of the bandicoots, and 
showing a similar range of modification. The musk-kangaroo 
(Hypsiprymnodon moschatus) (Fig. 1, 2), for example, which is 
the most generalized type of the family, has a foot which, like 
that of Perameles doreyana, is but little modified away from 
the phalangerine condition. The true kangaroos, on the other 
hand, have an extreme monodactyl foot, like that of the higher 
bandicoots. 
Dollo considers separately the South American form Coeno- 
lestes (Epanorthidae) and the Australian burrowing form Noto- 
ryctes (Notoryctida). With regard to the former he quotes 
the description of the foot given by Thomas, as “ hallux short, 
clawless, and not properly opposable ; other digits subequal, 
the fourth slightly the longest.” He recognizes this case as 
doubtful, but thinks it entirely probable that Coenolestes also 
is of arboreal derivation. 
In Notoryctes, finally, he finds indicated predominance of 
the fourth digit, and reduction and syndactylism of the second 
and third. These characters, he states, have not been adap- 
tively developed in connection with a burrowing habit because 
they are not found in other burrowing animals, such as the 
moles (Talpa and Chrysochloris). . They point rather to à 
former arboreal habit. 
