FUNCTIONS OF SWIMBLADDER OF FISHES. 93 



into a bicarbonate, called "half-bound" or "bicarbonate" carbon 

 dioxide. The half-bound is of course not in such stable union 

 as the fixed. Although plants can make use of the bicarbonate 

 for photosynthesis it can be of very little importance to fishes. 

 It is only the free carbon dioxide that is considered in the fol- 

 lowing experiments. 



The determination of carbon dioxide is not especially accurate 

 when done by titration methods and can not always be relied 

 upon for absolute results. Probably the most accurate method 

 is that of Pettenkofer's described and modified by Birge and 

 Juday (1911). Such a method was used in these experiments. 



Carbon dioxide has a marked effect upon fishes. As in other 

 animals the physiological effect is that of a narcotic, stimulating 

 in small quantities and intoxicating in larger quantities. Carbon 

 dioxide and acids apparently produce similar effects when 

 judged by the behavior of fishes. Reuss (1910) working with 

 varying amounts of carbon dioxide upon rainbow trout found 

 that the effects were similar to those upon birds and mammals. 

 Shelford (1918) also showed that when individuals of several 

 species where dropped into water containing 168 cubic centi- 

 meters of carbon dioxide per liter, the first effect was one of 

 stimulation but in a few minutes fishes lost their power to per- 

 form correlated movements and died. Shelford also showed that 

 fish display a preference for low carbon dioxide and that they 

 react very definitely to amounts naturally found in ponds and 

 lakes. He concludes that the carbon dioxide content of the 

 water is the best single index of the suitability of the water for 

 supporting fishes. 



A study was made of the effects of carbon dioxide in the sur- 

 rounding water on the gases in the swimbladders of perch. The 

 experiments were carried on in the laboratory. Carbon dioxide 

 gas was bubbled through a large carboy of water. A constant 

 flow of water was kept through the carboy to an aquarium con- 

 taining the experimental fishes. Individual fishes were removed 

 at intervals and analyses made of the swimbladder gases. Water 

 samples were taken at the same time. All the experiments were 

 begun by having normal lake water in the aquarium and gradual- 

 ly increasing the amount of treated water. The results are 

 shown in Table VI. In every case the oxygen and carbon 



