54 }- H. Pozi'crs 



era, like Cryptobranclius and Necturus, display somewhat similar 

 outlines. 



It was partially this resemblance of the profile, and indeed oi 

 the whole head, to some of the lower types, together with such 

 peculiarities as extreme resistance to metamorphosis, which led 

 me in my article on metamorphosis into the convenient and classic 

 error of ascribing this variation to reversion. Such larvae seemed 

 to me to represent a more primitive and perhaps perenibranchiate 

 ancestor. Had I, at that time, worked out the causes of such 

 other morphological developments as I have recorded in this 

 paper, I might soon perhaps have gathered a clue to the secret of 

 this special form. As it was I found it slow work to do so. 



Gradually, however, I accumulated a few data. First, the 

 variation was decidedly uncommon. Second, when })resent, it 

 was not by itself, but only in small numbers among a great many 

 other ordinary individuals. So few were the specimens that it 

 seemed the form could hardly be strongly hereditary. A single 

 female deposits from six or seven hundred to at least fourteen 

 hundred eggs. The whole stock of larvae in an ordinary pond 

 represents but a few such depositions. If hereditary, therefore, 

 one should find every third, fifth, or tenth individual, say, of this 

 type. Third, the type existed not only in its extreme develop- 

 ment, but showed combination with certain other characters as 

 well as transition forms. Most of the scattered specimens se- 

 cured through several years were of this nature, one of which is 

 shown in figure i, plate IX. Fourth, the individuals showed very 

 variable, yet always singular habits. Many absolutely refused to 

 eat liver or raw meat, something which no healthy and hungry 

 larva had persisted in before. Others would swallow meat in 

 pieces of prodigious size. Nearly all of the specimens secured 

 were very indolent and tame, yet one foiled my best attempts at 

 measurement, because no sooner did I approach its head with 

 compass point or other instrument, than it snapped at the same 

 most viciously, while out of hundreds of common larvae so ap- 

 proached not one ever offered such a response. Fifth, so far as I 

 could judge, the larvae of this type were only present in rather 

 clear water, not in the nuiddy ponds which were the commoner 



250 



