RECAPITULATION. 487 



retained in the uterus during the branchiate condition, and 

 not be born until the metamorphosis has been accompHshed. 

 We have good reason to beheve, however, that the transi- 

 tion to the Amniota from Amphibian-like ancestors was not 

 effected in this manner, since viviparity even in Mammals was 

 evidently acquired at a later stage, and the change came 

 about by means of the development of an egg-shell in which 

 the larva was contained and protected during its development. 

 None of the existing Amphibia, though they exhibit various 

 different methods by which the eggs are kept out of water 

 during development, show any tendency to form an egg- 

 shell. But it may reasonably be suggested that the 

 formation of this structure was originally due in some way 

 to the stimulation of the epithelium of the oviduct by the 

 retention in It of a laro^e-yolked ovum. Mr. Macbrlde 

 distinguishes two classes of stimuli to evolution, change of 

 environment, and increase or decrease of function, but it 

 appears to me that all modifications may be described as 

 the effects of stimulation, of which the Increased or 

 decreased activities of organs are particular cases. To my 

 mind it Is not possible to conceive the origin of the shelled 

 eofor as due to the selection of individuals In which the shell 

 appeared as a variation, unless we have some idea of the 

 cause of Its appearance. 



Judging from the structure of the aquatic larvae of 

 existing Amphibia, and from the general requirements 

 of aquatic life, we may suppose that in the actual an- 

 cestors of the Amniota the free aquatic larvae possessed 

 a median membranous fin -fold. We do not, however, 

 find that this or any trace of it Is preserved In the em- 

 bryo of the Amniota. Lateral fin-folds do not exist in 

 the tadpole of the Batrachia, and we need not expect, there- 

 fore, to find them In the embryos of Amniota. But the 

 disappearance of the median fin-fold Indicates that the 

 ancestral larval condition may undergo considerable modi- 

 fication when It becomes embryonic. The mere fact that 

 they were functional originally In a free larval condition is 

 probably not the whole explanation of the temporary de- 

 velopment of a complete system of gill-arches and gill-slits 



