[willey] BRANCHIODERMA AND BRANCHIOTREMA 99 



The movements of the external gills of Necturus are very ir- 

 regular and those of one side can be waved independently. Whether 

 they move together or separately, with a single movement or with 

 consecutive rhythmic action, the flapping of the gills is the sole res- 

 piratory movement of Necturus. Neither in the abdominal wall, 

 the floor of the mouth, nor in the external nares is there any res- 

 piratory twitch. Necturus, although an Amphibian in grade, is 

 not in any sense amphibious in habit, but is purely aquatic. It has 

 been captured on hooks and in nets, but there is no record of its having 

 been seen swimming near the surface in the open. It seems highly 

 probable that it does not normally make periodical visits to the surface 

 for the express purpose of breathing air. Necturus does not hiber- 

 nate but remains active through the winter; Eycleshymer has re- 

 peatedly taken them through the ice on set lines during the months 

 of January and February. 



The arrangement of the vascular system points to the same con- 

 clusion. The pulmonary arteries of Necturus arise from the last 

 pair of efferent branchial arteries, whilst all the blood which leaves 

 the heart passes through the afferent branchial arteries to the gills, 

 except that the first afferent artery on each side gives oft" an external 

 carotid artery. Boas (1882) was able to inject the pulmonary artery 

 from the radix aortae. Perhaps it may be presumed that the lungs 

 are filled with air by gaseous diffusion from the blood as with the air- 

 bladder of fishes. Apparently when this air accumulates in excess, 

 it can be got rid of through the minute glottis which lies far back and 

 has no relation to the internal nares. 



Hunger may drive Necturus to acts of cannibalism. This fact 

 alone shows how necessary it is to discriminate between different 

 kinds of behaviour in captivity. After having been kept through 

 the winter months almost entirely without food, a Necturus of mod- 

 erate size was found in the act of swallowing a smaller one, the 

 hind-legs and tail of the latter projecting out of its mouth. Twenty- 

 four hours later, the hind-legs had disappeared down the throat and, 

 after a few hours more, only the posterior moiety of the tail was still 

 protruding to a length of 1| inches. All this time the mouth of the 

 larger individual was necessarily kept permanently open under water. 



Very different is the behaviour of the aquatic salamander Die- 

 myctylus. According to Inez L. Whipple (1906), when retained in an 

 aquarium, Diemyctylus requires frequently to take air into the lungs 

 during the brief period when, by a rapid swimming to the surface, 

 sufficient momentum has been acquired to force the head for an in- 

 stant out of water. The nostrils are then useless as air-passages, 

 being full of water. The method employed is a quick, gulping mo- 



