16 



BULLETIN OF THE 



of Harvard College were placed at my disposal, and all the subsequent 

 work of analysis was carried on in that Laboratory. I am also indebted 

 to Professor Cooke for valuable suggestions during the progress of the 

 work. To Prof. A. V. E. Young, then a private student in the Chem- 

 ical Laboratory, I am under especial obligation, since he carried on the 

 manipulations and measurements of the gases. From his skill and 

 previous experience in gas analysis the results are entitled to more 

 consideration than if the measurements had been conducted by one 

 less familiar with such manipulations. 



It was my aim in these experiments to ascertain, without sacrificing 

 the fishes, the composition of the gas in the air-bladder, or at least of 

 that which was emitted in the form of bubbles, presumably from the air- 

 bladder. No attempt was made to ascertain the results of branchial 

 respiration. 



The collection of the gas, even from fishes 20 cm. long, is a tedious 

 process. An inverted glass funnel large enough to allow the fish to 

 swim freely inside, was first employed ; but the inclined sides of the 

 funnel not allowing the fish to assume a horizontal position except 

 when deep in the water, appeared to interfere with the emission of 

 the gas. After ascending to the apex of the funnel, the fish would 

 make violent efforts, apparently for the purpose of attaining the surface 

 of the water, but only occasionally would it emit gas under these cir- 

 cumstances, the emission usually 

 taking place only after it had be- 

 gun to descend. If freed from 

 the restraint of the funnel, the fish 

 invariably came at once to the 

 surface outside the funnel, and 

 emitted the customary bubbles. 

 Afterwards a nearly flat-bot- 

 tomed glass dish, only a trifle 

 smaller than the one containing 

 the fish, was inverted over the 

 latter (see Figure 1) ; but this 

 apparatus was only slightly more 

 successful than the one at first employed. The fishes came to the under 

 surface of the inverted vessel, and after long intervals some of them 

 emitted small portions of gas, but this was too small an amount to 

 be satisfactorily analyzed. Another method — the one finally employed, 

 although a tedious process — will be mentioned further on. 



Figure 1. 



