VARIATION AND ADAPTATION 227 



the gills are long and branched. In the viviparous Salaman- 

 dra atra the long gills of the embryo not only serve for respi- 

 ration but also in the later stages of development absorb nour- 

 ishment from the walls of the uterus, thus performing another 

 function of the allantois which in the Mammalia is performed 

 by the allantois. Somewhat similar allantoic gills, as they 

 have been called, occur in the embryos of some species of Noto- 

 trema, in which the whole development is passed through in the 

 egg. In this case the gills envelope the embryo and are only 

 connected to the gill-arches by narrow stalks. In other cases 

 again, as in Hylodes martinicensis, in which also the develop- 

 ment takes place within the egg, the gills and clefts have been 

 lost altogether, and the organ of respiration of the embryo is 

 the large and vascular tail. It may be pointed out here that 

 when the tadpole is free and active in the water it is properly 

 called a larva, but when the tadpole develops in the egg with 

 more or less modification it becomes an embryo. All these 

 different embryonic adaptations in Amphibia may be regarded 

 as partially successful experiments as compared with the one 

 method which, originally adopted by one group, has given rise 

 to the reptiles, and has persisted with modifications in the birds 

 and mammals which arose from them. This method consisted 

 in the development of a large yolk, foreshadowed in some of 

 the Amphibia, the formation of a tough usually calcareous 

 egg-shell and the development of the amnion and allantois. 

 The amnion is a growth of the outer membrane covering 

 the yolk-sac, a fold of the membrane being formed which 

 grows up and encloses the embryo dorsally. It is perhaps pos- 

 sible that the large yolk was the original cause in the de- 

 velopment of both allantois and amnion, for the increase in 

 the size of the yolk-mass necessarily involves a distension of 

 the ventral region of the embryo, and an extension of the body- 

 cavity ventrally: into this enlarged body-cavity the allantois 

 is able to grow, and it may have been the growth of the allantois 

 which originally pushed the amniotic fold before it. 



The Amphibia present some interesting cases of adap- 

 tation in relation to the fertilisation of the eggs, and the 

 protection of the young, that is to say, in structures re- 

 lated to the sexual and parental instincts. Some of these 

 structures have been mentioned in the chapter on breeding 



