NO. 2211. GENERA DESMOQNATHUS AND LEUROGNATHUS—DUNN. 403 



white in the young, soon begins to blacken in an area laterally very 

 definitely defined by a line between the ventral insertion of the 

 legs. Medium-sized specimens have the back dark, the sides light, 

 and the belly dark. The dorsal surface is never lighter than the 

 sides as it is in all the other forms of the genus. Old specimens 

 become uniformly black save for the head and feet which remain 

 brownish and whitish. 



In view of the lack of definite color pattern, question may arise 

 as to the appUcability of Holbrook's name to this species. The 

 proportions and ventral coloration of Holbrook's plate are those of 

 this form. The dorsal spots are more distinct in Holbrook's figure 

 than in any specimen I have seen. But, as Stejneger says, some 

 specimens approach it and the other characters agree exactly with 

 this form and with no other. 



I can find no sexual differences in this form. The testes are not 

 pigmented, as are those oifusca and ochrophaea. 



Habits. — According to my own experience, this is by far the most 

 aquatic species of the genus. I have found them swimming about 

 in streams as large as the Linville River. They are extremely 

 active and hard to catch or hold. It is a surprising sight to see 

 one of these big black salamanders run down a sloping bank into 

 the water for all the world Uke a water snake.. 



On several occasions captured specimens attempted biting. 



Around Brevard, North Carolina, they came dowTi to 2,100 feet 

 in large streams, to which they seemed practically confined. At 

 Linville, North Carolina, where the valley was 3,800 feet, they inhab- 

 ited large and small streams, irrespectively. 



The explanation seems to be that D. monticola inhabits small 

 streams which rise at or below 4,000 feet, while D. quadramaculata 

 follows down streams which rise above that Une. This is not an 

 absolute rule, but it holds in most cases, and a stream in which 

 quadramaculata is common usually contains few monticola, and the 

 opposite. 



The food seems not definitely known. 



Nothing has been published on the breeding habits. The condi- 

 tion of the ovaries in females would indicate September as the 

 beginning of the egg-laying period. Specimens have been seen 

 which were taken in May, July, and September. In May the eggs 

 were not developed. In July the eggs were in all cases well devel- 

 oped. In September the eggs were developed in some cases and 

 not in others. This would show that some of the September females 

 had spawned and others had not. 



The larvae reach a much larger size than in any of the other 

 species. The smallest transformed example seen was 52 mm. long, 

 and the largest larva was 79 mm. long. Females appear to be 



