362 EEPORT OF THE COMMISSIONEK OF FISHEEIE3. 



The surface water percolates tbrouo-li cavernous limestone. An aspi- 

 rating- effect is probably produced by the flow througb fissures and 

 narrow cbannels wliich bave access to air spaces, and tbe air is sucked 

 in and mingled with tbe down-flowing water, which it accompanies to 

 the mouth of the spring. During the journe}" a diminution of the 

 oxygen may occur from oxidation, which may reasonably explain the 

 modified proportions of these gases. Though the mountainous region 

 referred to abounds in springs, onl}" a single other bearing air was 

 found, and this a small one by the roadside. 



Air-bearing springs or wells of this character are to be distinguished 

 from the "breathing" or "blowing" wells abundant in some sections, 

 which alternately emit and suck in air from causes among which 

 variations in the barometer are important. In Nebraska many wells 

 having this remarkable peculiarit}' occur, and have been described b}^ 

 the United States Geological Survey (E. H. Barbour, 1899). The 

 springs of supersaturation which deliver bubbles of air constantly are 

 probably unrelated to breathing wells and, as far as known, pass the 

 air in one direction onl}-. 



The water of this Tennessee spring was apparently of excellent sani- 

 tary quality — clear, cold (about 12- C), slightly alkaline, and con- 

 tained an excess of nitrogen, but not of oxygen. It was slowl}" fatal 

 to fishes placed directly within the spring. Trout fry between 1 and 

 2 inches in length were killed by it sometimes' within a da}^ or two, 

 although some individuals would survive in it for weeks. On fish of 

 this small size no internal gas within the vessels was in any case 

 demonstrated with certainty. Neither were external symptoms usually 

 present, but in the hatcher}^ troughs supplied by the spring the}' were 

 more frequent and sometimes extremely conspicuous, consisting of 

 emphysema of the skin, either single cysts of gas, sometimes of rela- 

 tively great size, smaller multiple c^'sts, or small blisters of gas, which 

 usually had their seat upon the head or mucous membrane of the 

 mouth cavity. Apparently' the only inconvenience the fry experi- 

 enced from these was a mechanical one. The buoyanc}' of the gas 

 was often great enough to keep them constantly at the surface, and 

 its unequal lateral distribution gave them a list to one side or the other. 

 They did not appear to be materially Vvcakened. 



When older trout, yearling rainbows G to 8 inches in length, were 

 introduced into this spring, sjmiptoms more closely resembling those 

 at AVoods Hole resulted. Death occurred with moderate symptoms of 

 external gas, with gas free in the heart, though not abundant enough 

 to cause distention, and with emboli of gas in the gill filaments. The 

 susceptibility of species varied widely, and gold-fish were not atlccted 

 during a trial of sixteen days, while other cj'prinoids succumbed almost 

 as readily as the trout. These experiments with fishes in the spring 

 were made in live boxes and v\ere controlled by the same or similar 

 boxes in the spring water after it had passed from the spring and been 



