THE FCETAL MEMBRANES OF THE VEHTEBRATES. 185 



If we do not accept the starting-point in the phicentation- 

 process to be represented in the ung-uhite arrangement, a 

 proposal whicli the systematic position of the Ungulata would 

 in itself render doubtful, we must tlienlook for another phylo- 

 genetic sequence whicli will help us to rightly interpret that 

 momentous process of placentation. And here the important 

 results of Hill's investigation of very intense placental pheno- 

 mena in some marsupials, such as Perameles, have great 

 weight. 



We may fairly conclude that kangaroos, phalangers, oppos- 

 sums and other marsupials have only gradually become 

 aplacentary, parallel to those other formidable changes which 

 must have accompanied the elaboration of that peculiar type 

 which we call our recent Didelphia, in which the dentition, 

 the lactation, and those adaptations of the new-born animals 

 for nutrition during their life inside the marsupium form such 

 distinctive characters. 



And so if the Didelphia are in reality eri'atic Monodelphia 

 secondarily modified and with an allantois that has been thrown 

 out of the line of its normal development, with the exception 

 of Perameles, Dasyurus, and in part Phascolarctos, then we 

 have again to look, not amongst them, but amongst the 

 Monodelphia, for such forms that can give ns an indication as 

 to what may have been the primitive stage of placentation. 



And I may here state that my own researches on the 

 placentation of both primates and of insectivores have led me 

 to the conclusion that we should look in quite another direction 

 than the one alluded to above, which starts from diffuse 

 placentation. In the earlier part of this address 1 have con- 

 sidered those early phylogenetic stages when, in viviparous, 

 air-breathing tetrapods, the larval layer, the trophoblast, 

 found the most diverse possibilities open to it. 



I believe that those forms of which the embryonic tropho- 

 blast actually attacked the maternal uterine mucosa phago- 

 cytically were the pioneers towards the formation of what has 

 later become the discoid placenta. In some forms, even 

 among our recent mammals, that phagocytic attack is com- 



