FUNCTIONS OF SWIMBLADDER OF FISHES. 83 



decomposition and yet a large percentage of the swimbladder 

 air is usually nitrogen. 



Woodland (1911) attempted to compromise the Jaeger- 

 Nusbaum controversy. He assumed that a lysin was secreted 

 by the gas gland cells. This he supposed to hemolyse the 

 erythrocytes and thus liberate oxygen which the gas gland cells 

 could hold and thus force into the swimbladder lumen. Wood- 

 land refuted his own hypothesis in the following year however' 

 (1912) and concluded "that hemolysis does not occur in the gas 

 gland," but he offered no suggestion as to how gas does enter the 

 swimbladder. It is obvious that a more or less continual hemo- 

 lysis or a disintegration of the erythrocytes would be a severe 

 drain on the blood, and it seems probable that some other 

 explanation, more economical for the organism, could be found. 

 The writer attempts to furnish such an explanation in the 



present paper. 



GENERAL METHODS. 



In the study of the functions of the swimbladder of fishes 

 reported in the following pages the attempt was made to employ 

 standard biochemical and physiological methods. In many 

 cases, however, special apparatus was needed. Such apparatus 

 was devised and is described in connection with the respective 

 experiments. 



In the analyses of the composition of the gases within the 

 swimbladder, the Haldane (1912) gas analysis apparatus was 

 used; the Newcomer (1919) modification being found to be the 

 most convenient form. A burette holding five cubic centimeters 

 was used instead of the one with the usual ten cubic centimeter 

 capacity, for the reason that the quantity of gas obtainable 

 from the swimbladder is, in most of the fishes used, less than ten 

 cubic centimeters. The sample was drawn from the fish into 

 a five cubic centimeter sampling tube (Fig. i.) The sampling 

 tube 6" connected with a levelling tube L and a hypodermic 

 needle H were filled with mercury. The hypodermic needle 

 was thrust through the body wall directly into the swimbladder. 

 The stopcocks were opened, the levelling tube lowered, and the 

 gas thus drawn into the sampling tube. The stopcocks were 

 then closed. The samples were introduced into the gas analysis 

 apparatus the same day and analyzed in the usual manner. 



