Morphological J'ariation and Its Causes in A. tigrinnm 59 



Three weeks later it had grown from 11 to 15 cm. in length, and 

 had decidedly begun to lose its peculiar characteristics. It 

 changed in form more rapidly than in disposition, sluggishness 

 and violence characterizing it up to and even after metamor- 

 phosis. But after three weeks of liver feeding the head, though 

 still broad, was noticeably less disproportionate, and the sides 

 had begun to slope toward the muzzle. Sixteen days later still, 

 when metamorphosis began, at a length of 18.5 cm., the form 

 of the head, though broad and heavy, had nearly approached that 

 of other liver-fed, larvae grown from the ordinary type. In other 

 characters, posterior limbs, body, and tail, the specimen plainly 

 retained results of its early specialization. This instance, which 

 I observed with all possible care, and which I have since paral- 

 leled with a number of other cases, served to show the plasticity 

 of the lai-va. The characters had been lost in about the same 

 length of time which it had taken to develop them. 



The following two seasons I exerted myself to produce the 

 cannibalistic form experimentally. I succeeded but partially, al- 

 though my success was quite sufficient to add anything that was 

 lacking to the proof regarding the cause of the type. My most 

 extreme results resembled closely figure 3. plate VIII, although 

 body and tail were longer and the general cadaverous appear- 

 ance greater, owing to the fact that the animals had been in- 

 duced to swim much more. My experiments, however, proved 

 interesting, the failures as well as the partial successes throwing 

 much light on the process. The lack of success, partial or com- 

 plete, was always due to one of three causes — disease, failure in 

 inducing or maintaining cannibalism, or failure in keeping a sup- 

 ply of food animals of suitable size. 



I first placed several hundred young larvae within an enclosure, 

 twelve feet square, situated in the middle of a pond, with three to 

 five feet of water. The larvae when placed in the enclosure were 

 from 3 to 8 cm. in length. They were sparingly fed on daphnids 

 and Chironomous larvae. Unfortunately the water in the en- 

 closure was muddy and dark, and despite the close confinement 

 little cannibalism followed. Now and then, however, an animal 

 would acquire the habit, take a sudden leap forward in growth. 



255 



