208 AMPHIBIA 



living in tropical or subtropical countries. As such examples 

 we may mention nearly all African tree-frogs {Hylambates, Rap- 

 pia, etc.), the curious African genera Trichobatrachus and 

 Gampsosteonyx, the Solomon Island Frogs, Cornufer solomonis 

 and Ceratobatrachus guenthevi, etc. From the phylogenetic 

 point of view the question naturally arises, which type of egg 

 must be regarded as the most primitive. Some authors who 

 have dealt with the question have expressed the opinion that 

 the egg which is rapidly converted into the embryo, which after- 

 wards undergoes a lengthy larval period in the water, has given 

 rise by adaptation to the large egg, approaching the meroblastic 

 type, in which the young frog, subsisting on a large supply of 

 yolk, passes through part or the whole of the metamorphosis, 

 whilst others, on the contrary, have argued that modern Bat- 

 rachians are descended from land animals, and they have con- 

 sequently regarded the forms dispensing with the metamorphosis 

 as the most primitive. Most zoologists, however, hold that 

 Batrachians are probably derived from fishes allied to the Cros- 

 sopterygians and Dipnoans ; as these fishes have eggs which 

 are transitional between the extreme holoblastic and the mero- 

 blastic types, it seems rational to regard such eggs as the most 

 primitive, and the extremes exemplified at both ends of the 

 series by our common toad and by Hylodes martinicensis as 

 derived from them. 



From the evolutionary standpoint it is interesting to note 

 the various steps by which the most aberrant modes of nursing 

 are connected with the simple process of abandonment of the 

 offspring. The first step consists in protecting the eggs by de- 

 positing them in a hole or under some shelter on land ; further, 

 the parents, or one of the parents, watch over the eggs, or coil 

 themselves round them, this being followed by the actual trans- 

 port of the eggs, with or without structural modification in 

 the parent, such as brood-sacs. Again, we find the elaborate 

 dorsal pouch of certain tree-frogs developed at different degrees, 

 showing in what manner it may have been evolved. There is 

 also every form of transition between the extreme holoblastic 

 egg and one which is very nearly meroblastic. 



The examples quoted above of species within the same 

 genus, as in Rami, Hyla, Nototrema, differing so greatly from 

 each other in their early stages strikingly show how the neces- 



