No. 412. THE AUSTRALIAN MARSUPIALIA. 2 
55 
show some of these arboreal characters, indicating that up to 
a certain stage they have been arboreal in habit. It is thus a 
simple matter to arrange them in collateral groups according 
to the stage at which their terrestrial evolution began. Such 
an arrangement as this is represented diagrammatically in 
Pl. III, in the side lines passing off from the main arboreal 
line a—. The first series to be given off is that of the Dasyu- 
ride. In the least modified members of this family the animals 
are still arboreal, and the foot (Fig. 7) is like that of Didel- 
phys (Fig. a), except that the hallux is slightly reduced. As 
Dollo has pointed out, it is possible to recognize a complete 
series here, following which the hallux becomes gradually 
reduced and finally obliterated (e). 
Two other terrestrial offshoots, which parallel one another, 
and arise from the main arboreal line at the phalangerine stage, 
are represented by the Peramelidze and Macropodide. In 
both of these the evolution is cursorial as well as terrestrial, 
and is accompanied by a reduction of the hallux and an enor- 
mous enlargement of the fourth digit, so that the foot becomes 
functionally monodactyl. The extremes of structure in these 
lines are, as Dollo has indicated, in the Peramelidae Perameles 
doreyana (Fig. g) and Peragale leucura (Fig. h) and in the Mac- 
ropodida Hypsiprymnodon moschatus (Fig. 7) and Macropus 
(Fig. 7). 
Another terrestrial offshoot is represented by the Phasco- 
lomyide (Fig. f) although the relations of this branch are 
doubtful. The hallux is reduced in these animals as in other 
terrestrial forms. With respect to the remaining digits, the 
foot appears to be transitional between that of the Didelphy- 
idæ and that of the Phalangeride, the second and third digits 
not being so reduced as in the latter family. The difficulty arises 
here as to whether or not the second and third digits, from being 
fully reduced, as in the Phalangerida, have been secondarily 
enlarged. Such a condition would appear to be not improbable 
in slow-moving animals, such as the wombats; in the other 
animals which have passed through a phalangerine stage the 
tendency towards monodactylism has doubtless prevented the 
second arid third digits from becoming secondarily functional. 
