432 
favor of the view that the Australian Marsu- 
pials have been derived from opossum-like 
ancestors. Since then the opportunity has been 
afforded of examining the case in greater detail, 
and it may accordingly be of interest to notice 
certain of the results obtained. 
As nearly every writer who has dealt with the 
group has observed, the Australian Marsupials 
exhibit extensive resemblances to Placentals 
in respect to certain features of their outer 
organization, notably the adaptive modifications 
of the teeth and feet. By means of a very re- 
liable, though roundabout method, this fact may 
be made effective use of in determining the 
character of the Marsupial ancestors. Thus, 
by reference to the paleontological history of 
Placentals, it is possible to recognize the se- 
quence of events in the development of their 
structural modifications, and from this to infer 
a similar sequence in the similar modifications 
of Marsupials. Then, by carefully excluding 
those forms which show obvious signs of de- 
generation, it is possible to select from the 
others those modifications which are primitive 
or Original from those which are secondary or 
derived. 
Such an analysis is easily effected because 
the Australian Marsupials, notwithstanding 
their great diversity of form, constitute an ex- 
tremely simple and homogeneous group. They 
show no indications of having been derived 
from more than one ancestral type. 
The stem-form is found to have possessed the 
following characters, or more primitive ones: 
Dental formula 
gO 2 
3 TPs b 
al 
upper molar teeth triangular, each with three 
main cusps and an outer row of incipient styles ; 
lower molar teeth each with an anterior triangle 
of three main cusps, an antero-external shelf, 
and a posterior heel with three terminal cusps. 
Foot with the great toe opposable, but with the 
other digits normal. 
These characters are not found together in 
any one of the Australian forms, but, with one 
minor exception,* are exactly reproduced in the 
American opossums. 
* The opossums have four lower incisor teeth. 
SCIENCE. 
[N.S. Vou. XIII. No. 324. 
In addition, it appears entirely probable that 
other structural modifications of the Marsupial 
organization are also departures from an opos- 
sum type. The anteriorly opening pouch of 
the kangaroos, wallabies, and phalangers, and 
the posteriorly opening pouch of the bandicoots 
appear to be modifications of a type like that 
met with in the opossums, where the opening of 
the pouch is directed vertically downwards. 
This condition has been preserved in the Aus- 
tralian dasyures. Again, it is probable that a 
sealy prehensile tail, like that of the opossums, 
has to be ascribed to the Marsupial stem-form, 
since various grades of an apparently disap- 
pearing prehensilism are to be met with in the 
phalangers. 
There is also the significant fact that during 
the Oligocene period opossums were widely dis- 
tributed throughout the northern hemisphere 
and that at a slightly later period opossum-like 
forms (Microbiotheriidz) existed in South Amer- 
ica. To this may now be added the probability 
that it was at about the middle of the Tertiary 
that the Marsupial radiation began. Of the latter 
there is the following evidence: The most ad- 
vanced modifications of the teeth and feet in Mar- 
supials (kangaroos) are decidedly primitive as 
compared with the most advanced modifications 
of the same structures in Placentals (Ungulates, 
especially the horses). Furthermore, although 
the Marsupials possess the representative char- 
acters of Placental orders, it is barely possible to 
separate them, in some cases, into families. 
Both of these facts indicate that the Marsupial 
radiation has been of much shorter duration 
than that of the Placentals, and therefore, if the 
Placental radiation began in the Cretaceous or 
early Hocene, the Marsupial radiation, as indi- 
cated above, could not have begun before the 
middle of the Tertiary. - 
If these indications are reliable, it would ap- 
pear that the entire Marsupial fauna of Aus- 
tralia has been derived by a comparatively re- 
cent adaptive radiation from a single ancestral 
type, of which the American opossums are the 
existing representatives. 
Naturally the greatest objection to such a 
view is the fact that there are no opossums in 
Australia at the present time. This difficulty 
is, however, quite a superficial one. If we 
