356 EEPOET OF THE COMMISSIONER OF FISHERIES. 



the drop and in the splash at the surface, all mortality and symptoms 

 of gas could be prevented. From water standing without flow in 

 ordinary containers the excess of course finally disappeared, but in the 

 large Woods Hole aquaria signs of excess were still evident after seven 

 days. A cylindrical glass hatching jar of about 2i gallons capacity, 

 after filling with supersaturated water, required to stand two or three 

 days before this water failed to produce an external precipitation on the 

 body of a toracod immersed in it as a test. 



r6lES of nitrogen and oxygen in CAUSATION OF THE DISEASE. 



Some consideration may now be given to the separate roles of the 

 two gases nitrogen and oxj^gen in the disease. A reference to Table 

 II, page 373, shows that the gas from the fixed gas lesions, that is, from 

 the exophthalmia, from the fin blebs, and particularly from the cham- 

 bers of the heart, is very high in nitrogen. 



The sample from the sacs of rainbow-trout fry was taken from 

 specimens preserved in formalin and some oxygen may have been lost 

 on this account. All the others were from fresh material. 



The samples upon which these figures are based were very small, and 

 in obtaining them it was impossible to exclude with certainty all con- 

 tamination from atmospheric sources. In each case a part of the small 

 percentage of oxygen found certainly came directly from the air. 

 The sample from the eyes of scup was most liable to this error. That 

 from the hearts of various fishes indicates that the gas which causes 

 the fatal embolism in the vessels is almost pure nitrogen, and samples 

 from this source more accurately represent the gas as released from 

 the blood than those from the external blebs or the tissues about the 

 eyes. The one sample of the latter sort obtained was largely from 

 scup in which gas had inflated the conjunctiva so that this gas was 

 separated from the water only b}^ a very thin transparent membrane, 

 through which oxygen from the water may have diffused. Likewise 

 all the fin blebs have but a similar osmotic membrane protecting the 

 contained gas from changes in its original composition. The heart 

 gas, however, doubtless represents solel}^ a direct precipitation from 

 the blood. It would appear, then, that it is the nitrogen gas chiefl}^, 

 if not solely, which plays the essential part in the disease. The 

 separation of gas from the supersaturated blood is certainly not in pro- 

 portions analogous to that of the separation of nitrogen and oxygen 

 from water supersaturated with air. In air-saturated w^ater the 0x3^- 

 gen is about 33 per cent of the total oxygen and nitrogen dissolved. 

 In water air-supersaturated under the mechanical conditions here 

 described the percentage of oxygen dissolved is slightly less, for the 

 excess is not taken up in the same proportions that it is from the atmos- 

 phere. When unsaturated water is shaken with air at ordinary pres- 

 sure, the residue of undissolved gas is richer in nitrogen than the 



