20 BULLETIN OF THE 



those which I employed, that it seems justifiable to make use of his 

 conclusions in the present case. The ingenious device employed by 

 Baumert ('53, p. 46), in verifying the accuracy of his previous results, 

 to suppress the branchial respiration, and thus secure the effects of 

 intestinal respiration alone, could not, from the nature of the difference 

 between the modes of respiration, be made available in the case of 

 Lepidosteus. 



The means (viz. protracted boiling) which Baumert employed for 

 extracting the absorbed gases contained in the water of experimenta- 

 tion were as complete as the chemical methods at his time permitted. 

 Since then, however, Grehant ('69) has shown that simply boiling gas- 

 impregnated water, although sufficient to eliminate all the oxygen and 

 nitrogen, does not remove all the carbon dioxide, but that a combina- 

 tion of the mercurial pump with the method of boiling is capable of 

 accomplishing this result. By this process he has demonstrated that 



renewal of the water, gave in 100 volumes of gas respectively 1.77, 0.47, and 0.13 

 volumes of carbon dioxide, and 10.46, 13.71, and 11.92 of oxygen. It will be seen, 

 therefore, that not only the conditions under which the collections were made have 

 been in both cases very similar, but also that the composition of the gas as deter- 

 mined by analysis was nearly the same in Lepidosteus that it was in Cobitis. Since 

 the ratio of the gases in the two cases was practically the same, there is no reason 

 to suppose that the water absorbed a greater proportion of gas in one instance than 

 in the other. 



As regards the temperature of the water at which the experiments were made, 

 although not definitely stated by Baumert, it is safe to infer, from the temperature 

 at which his numerous experiments on branchial respiration were made, that it was 

 considerably lower than the temperature of the water in which the Lepidostei were 

 placed, so that the tendency to an absorption of carbon dioxide in the latter case 

 would certainly have been less than in Baumert's experiments. 



One of the plans adopted by Baumert to test the influence of the water and its 

 contained gases upon the emitted gas — whether the latter were either deprived of 

 any of its oxygen or carbon dioxide, or received accessions while exposed to the 

 water — was (11 to prepare artijicial mixtures of atmospheric air and carbon diox- 

 ide in different proportions (4 : 1 and 6:1) and cause them to bubble through water 

 in which fishes were living, and after collection to allow these gases to remain 

 exposed to the water, as in the experiment with the natural gas ; (2) in a similar 

 way, to cause atmospheric air to pass through water in which the fishes had been 

 living for several days. In the second of these experiments, the atmosphere 

 remained absolutely unaffected, showing not only that it did not lose oxygen, but 

 that it also did not acquire carbon dioxide. In the first of these methods, however, 

 while the proportion of nitrogen and oxygen remained practically the same, the 

 carbon dioxide was diminished by absorption to the extent of about one half its 

 original volume. To this evidence of considerable absorption, it seems to me, 

 Baumert does not allow proper weight, when, by a more satisfactory method, he 

 subsequently finds what he believes to be the true composition of the emitted gas. 



