Morphological Variation and Its Causes in A. tigrinum 6i 



reached 18 cm. in length, a change began to appear. They con- 

 tinued to feed as before, and had, indeed, learned to haunt the 

 corners of the large box for the few remaining frog tadpoles, 

 now grown very large; both these and the other amblystomas 

 were taken. But from this point on, the whole development 

 changed and progressed again toward the normal type. The 

 body grew more and the head less ; the parotids expanded and 

 the heads became full-cheeked; and, strangest of all, though the 

 food had not changed, save that it was now much more scarce, 

 the animals soon became excessively full bodied, large gilled, and 

 fat. The results of early cannibalism were not lost entirely ; even 

 after metamorphosis, the adults were conspicuous for heavier 

 lower jaws and broader heads than any others of the season's 

 development. The loss of the early acquired characters was, 

 however, very rapid and very conspicuous. Watching the 

 process from day to day, too, the cause became perfectly clear. 

 The peculiar cannibal type developed progressively just so long 

 as the food animals were disproportionately large and great dif- 

 ficulty occurred in the swallowing. I had thought that possibly 

 the nature of the food (flesh of same species or of other Am- 

 phibia) might be a contributing cause. This experiment and 

 others still more conclusive showed that it is not. The cause of 

 the general leanness, too, of the reduction of the parotids, and, 

 as we shall see, of other parts of the body, seems solely due to 

 the struggle between the organs of the body, much nourishment 

 being drafted away to produce- the hypertrophy of buccal and 

 branchial apparatus as well as body length, digestive tract, etc., 

 that other organs are, of necessity, greatly reduced. So soon as 

 the excessive labor of swallowing ceases, assimilation rapidly 

 tends to a more normal distribution, and the ordinary type tends 

 to be partially restored. Such observations as these explain fully 

 such types of larvae as figure i on plate IX. This is a neotenic 

 larva, undoubtedly a cannibal during its first season. During its 

 second season, it fed upon small larvae or a mixed diet, and thus 

 lost a portion of its peculiar characters, acquiring its large gills, 

 etc. I have tested this explanation by examining stomachs of such 

 specimens. First-year larvae usually live on a remarkably uni- 



257 



