NO. 1443. CAVE SALAMANDERS— BANTA AND McATEE. 73 



minutes, in one instance fifteen, and often chose to stay under water 

 in daytime, remainino- under cover. But most of the day it remained 

 above water and at nig-ht seemed to do so entirely. 



The changes in the size of the gills preceding transformation are 

 the reverse of stages which are passed through during earl}" larval 

 life. These exclusively larval organs are very slightly developed in 

 the 3"oung below 18 mm. in length. In a specimen 17.5 mm. (March 

 20) and in one 18 mm. (February 16) they are very stubby, with short 

 filaments. At 21 mui. (February 16) the filaments are much length- 

 ened, and a corresponding increase is to be noted in larva? 36 mm. long 

 (November 12). Up to this stage the distal half of the gill is clear 

 and translucent, while the basal part is pigmented. At 48 mm. (Sep- 

 tember 30) some specimens show a reduction of the gills, but among 

 specimens having them most perfectly developed is one 50 mm. in 

 length (October 7). In these larvae the gills are pigmented to some 

 extent even on the delicate filaments. Absorption of the gills prob- 

 ably takes place in most cases when the larva? are between 50 and 55 

 mm. But we have one larva 56.5 mm. long (October 28) in which the 

 gills are perfect, although this specimen is longer than many of the 

 recently transformed adults. The state of the gills, however, shows 

 that it is a genuine larval form, and although we have observed one 

 longer specimen (67 mm.) with gills reduced and near transformation, 

 it seems certain that the growth of this individual was accomplished 

 under the most favorable conditions and that in it are realized the 

 utmost possibilities of larval development under natural conditions. 



The process of absorption of the gills is the reverse of that of growth, 

 in detail as well as in entirety. Whereas the basal stubs are present 

 in early life and from them are budded out the filaments, the latter 

 are the first parts absorbed, the main arches disappearing slowly. 

 After this is completed the place of joining of the edges of the gill-slit 

 is indicated for some time in the young adults b}^ a dark line, which 

 is visible on both the dorsal and ventral surfaces of the side of the 

 neck. Besides the changes in these particular organs of the trans- 

 forming larvaj the entire body seems to lose in bulk, becoming flat- 

 tened and undergoing a special reduction in vertical dimension in the 

 tail, which loses all traces of the keel. Further, the neck appears 

 longer and slimmer, making the head more conspicuous, and, finally, 

 the eyes become much more prominent. 



The evolution of form is accompanied by as great a change in color. 

 To the naked eye the young larvae appear uniformly gray, while the 

 adult is flaming orange with conspicuous black spots. The study of 

 the development of the color pattern is of so much unportance as an 

 aid in identifying the larval salamanders (a thing still impossible in the 

 case of ir any common species) that it has been separated from the 

 main account of the life history and treated in greater detail. 



