204 



SCIENCE. 



[N. S. Vol. XIV. No. 345. 



nion, filled with fluid and surrounding the 

 embryo, is clearly a protective apparatus. 

 Its origin can in no way be explained by the 

 presence of the large yolk or of an egg-shell, 

 for in the sharks which have both no amnion 

 appears. As a protective apparatus it, of 

 course, would have the greatest value in an 

 embryo developing inside the mother, pro- 

 tecting it, like a water cushion, from the 

 peristaltic and other movements of the 

 mother. Hence, it is more reasonable to 

 look for the origin of the amnion in vivi- 

 parous rather than in oviparous forms. 

 Now, an amnion formed after the type found 

 in the hedgehog can act as a protection from 

 the very first, while one arising as in reptiles 

 cannot have any protective value until the 

 folds have completely closed over the em- 

 bryo, and so it is difficult to account for 

 the incipient stages in the reptilian type. 

 Again, it is easy to derive the reptilian con- 

 dition from that described, while the am- 

 nion of the hedgehog or of man is not easily 

 explained on an hypothesis of descent from 

 the reptilian condition. Hence it follows 

 that there is no inherent improbability, and 

 that there is much plausibility, in the view 

 that the amnion formed by splitting is the 

 primitive, that 'by overgrowth the derived 

 type. 



Now, where are there features that could 

 have given rise to such structures ? The an- 

 cestral form must have been viviparous, and 

 it must have had a two-layered ectoderm. 

 Now, the amphibia fulfill the conditions in 

 both respects, for there are salamanders 

 which bring forth living young, and no- 

 where in vertebrates, except in the amphi- 

 bia, do we find a two-layered ectoderm, 

 while here a two-layered condition occurs 

 throughout the group ; and further the outer 

 layer does not participate in the formation 

 of the embryo. 



In conclusion, it may be said that at 

 present the weight of evidence is in favor 

 of an amphibian ancestry for the mammals. 



but when the known forms of amphibia are 

 examined none is found which will meet 

 exactly the requirements of the case. The 

 limbless csecilians are, of course, out of the 

 line ; the anura, with their reduced verte- 

 bral columns and reptilian ear structures, 

 are equally out of the question. The uro- 

 deles approach more nearly to the ancestral 

 form, but their skull is so degenerate that it 

 cannot give rise to the zygomatic arch so 

 characteristic of the mammals. There re- 

 mains only the group of stegocephala. 

 These are extinct forms, the earliest fossils 

 of which appear in the Carboniferous, the 

 subclass dying out in the Triassic. In every 

 known feature these are closer to what the 

 ancestors of the mammal must have been 

 than are any of the other groups, and yet 

 not a single form of stegocephalan is known 

 which can be said to meet the demands re- 

 quired for the ancestor of the mammals. 

 This ancestor must be some form closely 

 allied to, but yet more primitive than, any 

 known stegocephalan. Further, it may be 

 said that we cannot derive urodeles, csecil- 

 ians or anura from any stegocephala as yet 

 discovered. 



The earliest known stegocephala are well 

 differentiated and widely distributed, and 

 they have a structure greatly different from 

 that of the crossopterygian ganoids from 

 which they have in all probability de- 

 scended. The ancestor of the mammals 

 partakes of characters intermediate be- 

 tween those of the crossopterygians and 

 those of the most primitive stegocephalan 

 known, and yet one in which the amphibian 

 characters predominate over the ganoid 

 features. 



We must, however, remember that the geo- 

 logical record is as yet imperfectly known. 

 We have as yet found no form which serves 

 to bridge the gap between the finned and 

 the limbed vertebrates. Footprints have 

 been found in Devonian rocks in Pennsyl- 

 vania which, in the light of our present 



