GAS DISEASE IN FISHES. 361 



the quantit}^ of air dissolved is great enough. Supposing- the temper- 

 ature constant, it is the interrehition of the dissolved air factor and 

 the pressure factor which determines the fact of the excess, and since 

 the condition of excess of air is to be defined only as a preponderance 

 of the dissolved air factor over the pressure factor, the cause of the 

 gas disease maj" be defined broadly as due to an excess of air; and 

 more narrowly, since there is much evidence that nitrogen alone is 

 essentially concerned, as due to an excess of nitrogen. 



SUPERSATURATION OF NATURAL WATERS. . 



The symptoms and fatalit}- at Woods Hole were the result of 

 artificial conditions. A modification by the hand of man of the condi- 

 tions under which air is usually taken up by the water resulted in an 

 excess of the air so taken up. The pertinent question will immediately 

 suggest itself whether natural waters ever acquire a similar excess, or 

 any excess at all, of air or of the constituents of air. Such excesses 

 are found to occur. Natural springs of water and flowing wells are 

 known to emit a gas, sometimes in considerable quantities, which has 

 approximately the composition of air. These are not ver}^ common. 



Conditions at Erwin, Tenn. — Such a spring occurs on the reserva- 

 tion at the Fisheries Station at Erwin, Tenn., in a limestone region 

 near the foot of a considerable mountain ridge. This spring has a 

 superficial area of about 600 square feet and its maximum depth is 

 about 4 feet. The bottom is partly of mud, partly of gravel and the 

 outcropping of the limestone strata. The water wells up chiefly from 

 the gravel, and from each wellspring a quantity- of gas in large bub- 

 bles is evolved at intervals of a few moments. The gravelly bottom 

 about the sources of water holds mechanicall}' large amounts of gas, 

 frfr, upon tapping it gently with a stick, an unusuall}" large quantity is 

 liberated and comes bubbling up through the water. The evolution 

 of gas then ceases for a longer period than usual, but begins again 

 spontaneously within a few minutes. This periodical deliver}^ of gas 

 continues day and night at all seasons. Evidently there is a constant 

 flow of gas accompanying the flow of water and at all times in the 

 earth or gravel beneath the spring and through which the water rises 

 are entangled large quantities of gas, a small fraction of which is 

 evolved every few minutes as the pressure beneath determines. 



This gas is air with the nitrogen and carbon dioxid considerably 

 increased. (Table I, p. 372, sample 1.) As springs do not usually 

 discharge both water and free air, the original access of air is of more 

 than passing interest. It is evident that it must be mainly derived 

 from the atmosphere. 



The region about this spring is mountainous and largely of a lime- 

 stone formation, in which caverns have been formed by the usual 

 process of solution of the limestone by water containing carbon dioxid. 



