306 PHYSIOLOGY OF RESPIRATION OF FISHES 



further stimulated by the theory held by Roule (1915) that the 

 salmon {Salmo salar L.) does not respond to salinity or temperature 

 but that it always reacts in such a way as to bring it into water having 

 a higher oxygen content. 



Method. 



A very simple technique was employed. A fish was placed in a 

 2 quart Mason jar filled with. sea water and made air-tight with a 

 rubber stopper. The jar was then immersed in a water bath of 

 running sea water in which the temperature was almost constant. 

 The oxygen content of the water at the beginning of an experiment 

 was always sufficiently high so that the fish did not at first suffer from 

 oxygen want. The hydrogen ion concentration was determined by 

 the colorimetric method; and the oxygen, by the Winkler method 

 immediately after all movements of the fish had ceased. 



The hydrogen ion concentration of the sea water was varied by 

 aerating with carbon dioxide-free air or by the addition of a small 

 amount of sea water made alkaline by the addition of sodium hydrox- 

 ide and by the introduction of carbon dioxide. 



The fishes used in these experiments were kept in a small aquarium 

 of running sea water. They were allowed to rest from 6 to 12 hours, 

 so that they might recover from any shock suffered when collected. 

 All were rejected after being in the laboratory 2 or 3 days so as to 

 avoid as far a^ possible erratic results due to the ill effect of keeping 

 them under unnatural conditions. 



EXPERIMENTAL. 



The mean oxygen content of the sea water of all experiments 

 having approximately the same pH at the time of death of the fish 

 was calculated. That is, if the highest oxygen content of the sea 

 water at the time of death of the cunner at 7.22 to 7.25 pH was 0.30 cc. 

 per liter, and the lowest 0.20 cc. per liter, the cunner was taken to be 

 able to absorb the oxygen from the sea water down to 0.25 cc. per 

 liter at a pH from 7.22 to 7.25. This was taken to be more nearly 

 exact and to eliminate to a greater extent the factor of individual 

 variation than if individuals were taken alone. Loeb (1912) has 

 shown that the mortality curve of Fundulus embryos under the in- 



