GAS DISEASE IN FISHES. 371 



only after it has flowed a considerable distance from its source in a 

 shalloAV open stream. Failing this it may be carried through a circuit 

 of a long and wide trough, to pass finally through perforated deaerat- 

 ing pans. In general a complete exposure to the atmosphere is neces- 

 sary and the means for accomplishing this will vary with the conditions 

 of each individual case. The deficit of oxygen is more readily supplied 

 than the excess of nitrogen removed. The water eagerly takes up the 

 oxygen it lacks, but the last traces of excess of nitrogen come awa}^ 

 with difficulty. 



When water rises as springs or wells in the bottom of the fish ponds 

 themselves, it is trtill more difficult of correction, and quite impossi- 

 ble unless the head is strong enough to lift the water above the level 

 of the surface of the pond, and so permit the adoption of the above 

 measures. 



It is a fact of significance and importance, to be considered from 

 the standpoint of fish culture, that spring waters may vary consider- 

 ably from time to time in the amount of dissolved air the}^ contain. 

 An instance of this, recently observed, concerned the oxygen alone, a 

 marked deficiency being f ollow^ed after several days and subsequent to 

 a heavy rain, by a feJrly abundant supply. It is inferred that nitro- 

 gen variations may likewise occur, and presumably changes in the 

 solids in solution. Weather and seasonal conditions probably are con- 

 tributing causes of this variability, but not manj^ observations have 

 been made and little is known beyond the fact, which makes it neces- 

 sary not to place entire reliance on one examination of a given water. 



In three instances of gas disease at government fish-cultural stations 

 the excess of air has been actually determined b}^ analysis. In others 

 similar symptoms make a presumption of a similar cause. Meager 

 information of other cases of disease or mortality among fishes with 

 gas symptoms indicate with more or less probability the presence of 

 supersaturation. A spring at an abandoned private trout cultural 

 establishment in Vermont was found to be constantly giving up large 

 bubbles of air (Table I, page 372, sample 5). Trout culture was not 

 successful in this water, and the former superintendent gave a history 

 of bulging e3"es. Analyses were not made, but it seems extremely 

 probable that this water was supersaturated. 



In 1902, at the exhibit of the United States Fish Commission at the 

 Charleston Exposition, a sudden and severe loss occurred among the 

 marine fishes of the aquaria. The water precipitated quantities of 

 gas, and the fishes were described as showing external bubbles and 

 blisters of gas. The water supply was obtained b}" pumps with a long- 

 suction. The presumption is strong that the mortality was from 

 excess of air, and that its sudden disappearance was caused b}^ a change 

 in the suction pipe, which, though unwittingl}', corrected undetected 

 leaks. The trouble was not explainable on other grounds than these. 



