of Heat by Gaseous and Liquid Matter, 93 



The conclusion that the sum of the absorptions is greater than 

 the absorption of the sum is here amply verified. The Tables 

 also show that the ratio of the sum of the absorptions to the 

 absorption of the sum is practically constant for all the gases. 

 Dividing the first mean by the second in the respective cases, we 

 have the following quotients : — 



Carbonic oxide" 1*70 



Carbonic acid 1*72 



Nitrous oxide 1*61 



defiant gas (mean of both) . . 1*68 



The sum of the absorptions ought to be a maximum when 

 the two chambers are of equal length. Supposing them to be 

 unequal, one being in excess of half the length of the tube, let 

 us consider the action of this excess singly. Placed after the 

 half-leu gth, it receives the rays which have already traversed 

 that half; placed after the shorter length, it receives the rays 

 which have traversed the shorter length. In the former case, 

 therefore, the excess will absorb less than in the latter, because 

 the rays in the former case have been more thoroughly sifted 

 before the heat reaches the excess. From this it is clear that, 

 as regards absorption, more is gained by attaching the excess to 

 the short length of the tube than to the half-length ; in other 

 words, the sum of the absorptions, when the tube is divided into 

 two equal parts, is a maximum. This reasoning is approximately 

 verified by the experiments. Supposing, moreover, one of the 

 lengths constantly to diminish, we thus constantly approach the 

 limit when the sum of the absorptions and the absorption of the 

 sum are equal to each other, the former being then a minimum. 

 The effect of proximity to this limit is exhibited in the first expe- 

 riment in each of the series; here the lengths of the compart- 

 ments are very unequal, and the sum of the absorptions is, in 

 general, a minimum. 



After the absorption by the permanent gases had been in this 

 way examined, I passed on to the examination of vapours. They 

 were all used at a common pressure of 05 of an inch of mer- 

 cury, or about -g^th of an atmosphere. The liquid which yielded 

 the vapour was enclosed in the flasks described in my previous 

 memoirs, and the pure vapour was allowed to enter the respec- 

 tive compartments of the experimental tube without the slightest 

 ebullition. The following series of Tables contains the results 

 thus obtained. 



