I 30 THE AMERICAN NATURALIST. [VOL. XXXV. 
placental type, but that, as far as present evidence goes, the 
alternative view is just as plausible 
The question now to be decided is whether or not the latter 
explains the facts of the case equally well. 
Admitting the reduced allantois to be secondary, and recog- 
nizing the importance of the yolk-sac stage in marsupials, is it 
not just as probable that the reduced allantois has proceeded 
from one of sauropsidan type as that it has proceeded from 
one of placental type? A failure, due to whatever cause, of 
the allantois to form a placental connection furnishes a more 
plausible explanation of the premature birth of marsupials than 
that of Dollo (99, p. 203), which attributes it to the loss of 
placenta in the violent gymnastics of an arboreal life. And it 
thus affords a simple explanation of the reduction of the allan- 
tois, inasmuch as premature birth, instituted through the failure 
of the allantois to form a placental connection, provides not only 
for the nutrition of the embryo but also for its respiration. 
This condition being perfected, the allantois, which is a provision 
for respiration during the later embryonic stages, having its 
function usurped, has naturally become reduced. 
Why the allantois should have failed to form a placental 
connection cannot at present be explained. The question 
naturally suggests itself, Is not this also attributable to arbo- 
real habit? It is not improbable that there may have been 
something in the assumption by metatherian animals of an 
arboreal habit which, for mechanical reasons, may have made 
a placental formation impossible. 
While still concerned with the former relationships of mar- 
supials, it is of interest to notice certain facts concerning their 
ancestry which have been derived from paleontology. 
In a recent paper on the origin of mammals, Osborn (9) 
has dealt with this subject and has represented in the form of 
a chart the probable geological and phylogenetic relationships 
of the mammalian subclasses. In this the Prototheria and - 
Eutheria! are depicted as primary groups in the Triassic, the 
! Following Gill's division of 1872 (cf. Gill,'72). The Eutheria as defined by 
Gill are not assigned a placental or non-placental character. Hence they may be 
; interpreted as equivalent either to the Metatheria of Huxley or to his Eutheria 
as interpreted by Wilson, Hill, and Dollo. 
