No. 410.] THE MAMMALIAN SUBCLASSES. ‘129 
placenta could not be assumed on a basis of retrogression in 
single structures. For example, if we assume retrogression of 
the allantois from a placental to a non-placental condition on a 
basis of the more retrogressive character of the teeth of Phas- 
colarctos as compared with those of Perameles, as suggested 
by Hill, we may also assume the reverse on the evidence of 
the feet, for, as Dollo has shown, the foot type of Perameles 
is a direct natural advance on that of Phascolarctos. 
The greatest objection, however, might be taken on the 
special grounds that the reduction and modification of the 
teeth which has taken place in Phascolarctos is not retrogres- 
sive (* decadent," Hill, p. 435) at all, but is, on the other hand, 
progressive and very comparable to that. which has taken place 
in ungulate placentals; and, further, that as far as the greater 
primitiveness of Perameles or Phascolarctos is concerned there 
is little to choose, for although the former is decidedly more 
archaic in general dentition, its individual upper molar teeth 
are well evolved (quadritubercular), and its foot is much more 
specialized. 
In all, it may be said that efforts to explain the reduction of 
the allantois, whether from a placental or non-placental condi- 
tion, by reference to retrogression of the general organization 
are foredoomed to failure; for whatever may be true of the 
Marsupialia in general, the Australian members of the group, 
far from being degenerate, show a progressive adaptive radia- 
tion (doubtless made possible through freedom from competi- 
tion) which parallels, and is in general very comparable to 
that which has taken place during the Tertiary period in 
placental mammals. 
And in view of this a parallelism of placenta, though not 
genetically related to them, need not be more surprising than 
the extensive parallelisms in otherstructures. In other words, 
there is no adequate reason why the placental allantois of 
Perameles should not be regarded as having arisen by progres- 
sive modification from one of non-placental (sauropsidan) type. 
These suggestions naturally do not imply that the placenta 
of Perameles is not primitive, or that in general the reduced 
allantois of the Marsupialia has not degraded from one of 
