of dark Heat-rays by Gases and Vapours. 15 



nometer had remained in the circuity the cooling of the thermo- 

 pile by the rarefaction of the air in A would have produced a 

 powerful thermo-current injurious to so delicate a galvano- 

 meter. After A had been filled with the gas or vapour to be 

 investigated an interval of two hours was allowed, so that its 

 temperature, which had been raised by the gas rushing into it, 

 might have sunk again to that of the surrounding water. 

 Then the separate experiments were made again, in the same 

 manner as already described. 



As some of the more extended series of experiments re- 

 quired 12 to 14 hours, and so long a time was seldom at our 

 disposal during the day, we almost always made our experi- 

 ments at night — a plan which had other advantages also, in 

 consequence of the very unfavourable position of the In- 

 stitute. The vibration from the crowded street visibly affected 

 the galvanometer in the daytime ; and its position of equili- 

 brium was much more constant at night. 



Results of Experiments. 

 I. Pure dry Air. 



The absorption of heat by this mixture of gases is of the 

 highest physical interest, especially in questions of radiant 

 heat. The results we obtained at first differed very much 

 among themselves ; and it was not until we had learned to 

 take all the precautions which have been described that we ob- 

 tained agreeing results. Of the large number of our experi- 

 ments, in part of which snow was employed as cooling-agent, 

 we give two series. 



In all the following Tables, E denotes the position of rest 

 of the galvanometer before the heating took place, H x the 

 calculated position of rest after steam had been allowed to 

 stream against the little plate a ; J gives the radiation which 

 has passed through the medium, the intensity of the original 

 radiation being put equal to 100 ; h denotes the ratio of two 

 consecutive amplitudes, and t the temperature of the water- 

 bath in centigrade degrees. 



It must be remarked, with reference to the arrangements 

 of this and following Tables, that, in order to save space, two 

 quite different series of experiments are often separated only 

 by a line. 



