394 



THE AMERICAN NATURALIST. [Vol. XXXVI J. 



point, these two larvae were, one after the other, striken by the 

 sudden impulse to change, the great gills withering almost in a 

 single night ; they underwent metamorphosis at about fourteen 

 centimeters in length, in the very jar of water in which they had 

 been flourishing as larvae so surpassingly well. My aquaria have 

 furnished any number of essentially similar cases. Whether or 

 not it be true that oxygenated water develops the large gill, it 

 certainly cannot be too strongly emphasized that the large gill, 

 when developed, is no impediment to metamorphosis. I once 

 investigated a pond well stocked with thriving larvae which 

 struck me by the meagreness of gill development ; the rami were 

 short, and very thin, the fringes likewise, irregular, few and 

 short. A few weeks later the gills had considerably increased 

 in size and general development. Yet it was following this 

 period of gill expansion that rapid metamorphosis began. The 

 expansion of the gills was here doubtless in no sense the cause 

 of the metamorphosis. They may have been essentially unre- 

 lated phenomena. It is possible however that a connection 

 existed as follows : The small-gilled larvae, although having 

 reached considerable size, had still the delicate larval skin well 

 adapted in itself to subserve the purpose of aquatic respiration. 

 Metamorphosis seldom begins until this larval skin has given 

 place to the more or less thickened integument which is to 

 finally characterize the adult. This change in integument is not 

 here considered as a part of the metamorphosis proper; for the 

 larva may change its integument and yet long retain its aquatic 

 form. Yet this change is a necessary preparation for metamor- 

 phosis ; and it is quite possible that it should render the skin 

 less effective for respiratory purposes. The cutaneous circula- 

 lation might even be checked, and a corresponding increase of 

 blood flow to the gills, causing their growth. Certain it is that 

 larvae with the thickened dermis usually have at least well 

 developed gills, and such larvae, if the right stimulus comes, 

 are ripe for metamorphosis. I will mention in this connection 

 the surprising fact, developed in the course of several experi- 

 ments, that the adaptation of these larvae to entire aquatic res- 

 piration stands in no constant relation whatever to branchial 

 development. Many experiments were made to induce metamor- 



