Morphological J^ariation and Its Causes /u A. tigrinum 53 



external appearance. Thus the width of the head is not due, as 

 in most broad-headed specimens, to the development of the paro- 

 tid glands ; on the contrary, in this type of larva these structures 

 are non-developed, the width being due to an enormous extension 

 of the whole buccal and branchial apparatus. The width of the 

 gape essentially equals the width of the head. How singular this 

 distortion of proportions is becomes clear when we estimate what 

 the width of head in figure 2 would be were the parotids and pos- 

 terior portion of the head developed laterally until they exceeded 

 the gape of the mouth proportionately as much as they do in cer- 

 tain small-mouthed but wide-headed individuals of the tvpe of 

 figure I, plate II. The width of the head in this hypothetical case 

 would equal one-fourth the entire length of the animal. Not that 

 I would suggest any such carrying out of proportion as possible; 

 it is not. as the analysis of the development of the form will show. 

 But such comparisons serve to show, better than simple dimen- 

 sions, how extremely aberrant the t}pe is. ]\Iost obvious and 

 striking of all the peculiar external characters, at- least to one fa- 

 miliar with the species or even the family, is the singular profile. 

 It is not exaggerated nor distorted in figure i, and it is also 

 neither accidental nor pathological. Yet one may handle hun- 

 dreds or even thousands of the larvae of A. tigrinum without 

 even the suggestion of such a type. Comparison should be made 

 •with the normal profiles on plate i. also figure 3, plate II, and 

 figure 5, plate III. The profiles of the adults on plate III are also 

 instructive, figure 3 only being related strongly to the larval type 

 we are discussing. A comparison with the outline profiles given 

 by Cope for nearly the whole genus is also instructive. Only in 

 the case of the one-specimen species. A. copeannm (which for 

 one who understands tigrinum is almost certainly a variety, and 

 may obviously be related to the type which I am describing) 

 shows a profile resembling, in some degree, the one I figure. In 

 the closely related genus. Chondrotus, only the one species, C. ten- 

 ebrosus (like tigrinum, very large) seems, according to Cope's 

 figures of both larva and adult, to possess a dished profile. Oth- 

 erwise such profiles are very uncommon throughout the entire 

 Urodela. Only some of the lower and permanently aquatic gen- 



249 



