THE A M ERICA N NA TURALIST. [Vol. XXXVII. 



eating, followed by indigestion and bloating, frequently upsets 

 the metabolic equilibrium sufficiently to induce the change, and 

 this even in quite small specimens ; the accumulation of gas in 

 the digestive tract causes the animal to float for several days, no 

 food is taken, and by the time recovery has occurred the irre- 

 versible gill shrinkage, etc., has set in. A careless observer 

 might possibly interpret this floating at the surface, followed by 

 metamorphosis, as an instance where aerial respiration ushered 

 in the change. There is no occasion for such interpretation, 

 however: the animal floats on one side, with the mouth and 

 entire head under water. 



Quite in accord with the explanation here given, and con- 

 stituting a minor continuation of it, is the fact observed by 

 myself, and independently by an assistant in my laboratory, that 

 in any lot of larvae of approximately one age it is not the largest 

 specimens that metamorphose first ; but an intermediate size. 

 The smallest larvae will not have finished certain developmental 

 changes which constitute the preparation for the metamorphosis ; 

 while the very largest are usually fat, phlegmatic, anabolic larvae, 

 which are less easily stimulated to sudden katabolic change. 

 Thus in seven experiments, in each of which three larvae were 

 taken varying somewhat in length, but all of them between 

 eleven and fourteen centimetres, there was but one instance in 

 which the largest of the three began metamorphosis noticeably 



mediate or the smallest specimen was notably in advance. 

 Much more striking are the facts noticeable in aquaria containing 

 many larva?, where a few extra large, anabolic individuals fre- 

 quently postpone metamorphosis for weeks after many of the 

 smaller specimens have become adult. 



In close connection with such facts as those just stated and 

 subject to the same explanation is the fact, which I have noted 

 again and again, that the tendency to sudden metamorphosis is 

 directly correlated with the disposition or temperament of these 

 larva. Variable as these animals are, in no way do they show 

 greater differences than in their suceptibility to excitement. 

 One of the first ponds well stocked with Amblystoma that I 

 discovered was filled with larvae of the usual type, but which had 



