the Diffusion of Gases. 245 



Experiment. Thermometer 64. To a diffusion-tube over 

 water, with 5 measures air, 80 chlorine gas were added, making 

 together 85 measures, which, diffusing into damp air, expanded 

 8 measures in the first eight minutes, 18 measures in eighty-two 

 minutes, and, finally, 19 measures in one hundred and six mi- 

 nutes ; but the same gas, in a close standard tube of the same dia- 

 meter, contracted, owing to absorption of the gas by water, 5 

 measures in eight minutes, 15 measures in thirty-three minutes, 

 and 18 measures in thirty-nine minutes, the rate of absorption 

 diminishing evidently from the water in the tube becoming satu- 

 rated and abiding in it. But the absorption of gas by water in 

 the two experiments cannot be well compared, for, in the diffu- 

 sion experiment, the chlorine is rapidly diluted with return-air, 

 which protects it from absorption, and, indeed, before the end of 

 the experiment, must occasion a portion of the dissolved chlorine 

 gas to reassume the gaseous form, vaporizing away from the water 

 which held it in solution, and rising into the upper part of the 

 tube. The absorption in the diffusion-case would certainly be 

 overrated at one-half of what occurred in the comparative expe- 

 riment in the same time. At the outset, however, we may pre- 

 sume that the same absorption took place in both cases. Hence 

 the expansion in the diffusion experiment would be 3 -f- 5, or 8 

 measures to the first eight minutes. The absorption, however, 

 would tell two ways in lessening the expansion ; first, so much 

 gas has disappeared by absorption, the quantity to be added to 

 the expansion ; second, so much less chlorine has really been sub- 

 mittedte diffusion : 80 parts have not been diffused, but 80 di- 

 minished by this quantity. 



Merely adding the observed absorption in the first thirty- 

 nine minutes, namely 18 measures to the expansion observed of 

 19 measures, we have an expansion from diffusion of 37 measures, 

 which approaches, as near as we can expect from the method, to 

 45 measures, the theoretical expansion on 78 measures dry chlo- 



VOL. XII. PART I. I 1 



