MUSEUM OF COMPAKATIVE ZOOLOGY. 23 



tubing, as indicated in the figure. After traversing these tubes, the air 

 was compelled to bubble through two sets of potash bulbs (k 7 ), and 

 thence by a glass tube was led through the stopper of the experimental 

 jar (j). The glass tube was carried to within about 10 cm. of the bottom 

 of the jar, and, after describing a U-shaped bend, terminated in a plati- 

 num "rose" (r). Thence bubbling through the water, the air was led 

 from the air-chamber of the jar by a glass tube to the baryta bulb (b). 

 To guard against the effect of an accidental diminution of pressure in 

 the hydrant water, and a consequent reflow of water into the baryta 

 bulb, the latter was not directly connected with the pump, but an ordi- 

 nary Wolffs bottle (w) was interposed. 



The preliminary test imposed upon the apparatus was, that it should 

 run twenty-four hours with no fishes in the water, but otherwise under 

 the same conditions that would be required when the fishes were intro- 

 duced, without giving a perceptible precipitate in the baryta bulb. It 

 required some time and considerable attention to details before this was 

 attained. All the connections were made by means of close-fitting rub- 

 ber tubing, which was made as limited as possible, narrow glass tubing 

 being used wherever practicable, and all the joints were in addition 

 sealed with melted paraffine. While the pump was in operation, the 

 water in the jar was of course under diminished pressure, 1 the diminu- 

 tion being equivalent to a column of water equalling the distance be- 

 tween the surface of the water in the jar and the lower end of the bent 

 glass tube to which the "rose" was attached. Evidently the removal 

 of the stopper from the neck of the jar (through which it was" proposed 

 to introduce the fishes) under such circumstances would allow the en- 

 trance of considerable impure air, which might give a precipitate as 

 soon as the pumping was resumed. To obviate this source of error, the 

 tubing of the apparatus was clamped at x, and by means of a hand 

 bellows air was injected through the potash system until equilibrium 

 was restored in the chamber of the jar, this being indicated when the 

 water rose inside the bent tube to the same height as outside. In the 

 preliminary test, this was done in the same manner as subsequently 

 when the fishes were introduced, and the stopper was also removed for 

 the length of time which experience had shown would be necessary for 

 the introduction of the fishes. Thus, as nearly as possible the same 

 conditions were observed in the preliminary and the final experiments. 



1 It is to be observed that, so far as this diminution of pressure affects the 

 absorption of gases by the water, it can only act favorably in preventing such 

 absorption. 



