GAS DISEASE IN FISHES. 351 



well filled with gas the condition modifies considerablj^ the macroscopic 

 appearance of the gill, and in fishes of some size the individual enil)oli 

 may be seen on careful inspection by the naked eye. 



The gas has not been observed in the capillaries of the bod}-, but is 

 confined to the larger vessels of the systemic circulation and the gills. 

 It does not distend the veins, though bul)bles may be seen in them. 

 In sculpins in full roe the arteries ramifying over the surface of the 

 ovar}" attract immediate attention by their appearance as pale blood- 

 less streaks in contrast with the green background of the ovar}- and 

 the dark red of the veins which accompany them. Gas bubbles may 

 be seen in the pyloric cceca, in the walls of the intestine, and also 

 within the intestine itself, though these latter ma}' be due to other 

 causes. 



CAUSE OF THE GAS DISEASE IN FISHES. 

 POSSIBILITY OF INFECTIOX BY GAS-FKODUCING BACTERIA. 



The inference to which all the gas symptoms at first give rise, of 

 infection with gas-producing species of bacteria, has been negatived 

 b}' repeated attempts to obtain cultures from the blood and tissues of 

 affected fishes, among both the Woods Hole marine forms and those of 

 fresh water. The microscope gives no evidence of infection, and inocu- 

 lated culture media remain sterile. The Woods Hole sea water suf- 

 fered no unusual pollution and the bacterial count at the intake in 

 January and Februar}' averaged only 191 per cul^ic centimeter. The 

 rapidity of the pathologic process, furthermore, contraindicates 

 infection. 



ABNORMAL GAS CONTENT OF WATER IN WHICH THE DISEASE OCCURS. 



The sea water in which fishes die with these described lesions always 

 has an extraordinar}' gaseous content. At the Woods Hole station it 

 had passed through a pumping plant which elevated it to storage tanks 

 to provide a gravity flow for aquarium and hatching purposes. Steam 

 pumps took the water from the sea through a long suction pipe and 

 forced it to a height of about 18 feet into tanks, from w^hich it flovred 

 to the aquaria and hatching boxes. At the point of intake the sea 

 water was of normal qualit}' and fishes lived in it without unusual 

 S3;mptoms. The suction pipe was of wood, had been long in use, and 

 by deterioration had developed areas of porosit}" or open leaks, so 

 that air continually gained access to the pipe and could readily be 

 dtemonstrated at the pump, which forced a mixture of water and large 

 quantities of air bubbles instead of a solid body of water. Immedi- 

 ateh' upon passing the pump this air and water came into a region of 

 about 8 pounds hydrostatic pressure in addition to that of the atmos- 

 phere, and continued under this pressure through a long stretch of 

 level water main. As the sea water was approximatel}' saturated with 



