No. 412.] THE AUSTRALIAN MARSUPIALIA. 261 
been a migration of these animals into both countries from 
a common center in Asia ; on the other hand, there may have 
been simply a passage of forms via Asia. In either case the 
former presence of opossums on the latter continent would 
be indicated. 
However this may be, the actual evidence at our disposal 
is very slight, and it is equally conceivable that ancestral 
Didelphyide may have entered by way of Antarctica. This 
possibility has been suggested by Osborn (99b), although no 
preference is indicated by him for either view. As already 
indicated, the time at which the ancestral marsupials gained 
access to the Australian region has been estimated by Wallace 
as *far back in the Secondary," by Spencer as Cretaceous, 
and by Lydekker as early Eocene. It is probable, however, 
that these estimates are much too great, and that the time of 
entry was considerably later. Comparing the marsupial and 
placental radiations with respect to the modifications of the 
teeth and feet, we notice the fact that while there is a general 
correspondence between the most primitive characters of each 
series, there is no such correspondence between the most 
specialized ones. The most specialized characters of mar- 
supials are decidedly primitive as compared with the most 
specialized characters of placentals. To illustrate the case, 
while the insectivorous members of the Dasyuride and the 
Didelphyidae approach closely the trituberculate pentadactyl 
ancestors of the placentals, the Macropodidæ, which are the 
most specialized of the marsupials, are much more primi- 
tive than the most specialized placentals (perissodactyl 
Ungulata) whose evolution they parallel. The lophiodont 
teeth of these forms are very similar to those of such per- 
sistent primitive types as the Tapiride and their imme- 
diate ancestors, having none of the complex characters of 
those of the progressive lophiodont forms (Equide and 
Rhinocerotidz). The functionally monodactyl, cursorial feet 
of the Macropodidz are primitive as compared with the simi- 
larly modified feet of the Equide. Further, there is the 
remarkable fact that while the marsupial families have the 
value of placental orders, it is barely possible to separate them 
