and D lather mancy of Air and Hydrogen. 



421 



ments the reverse was the case. Moreover the distance was 

 less than 45 inillims., which gave greater deflections. For 

 the sake of comparison, the results of experiments repeated 

 with dry air and the vacuum are also given. 



Table XI. 

 (Each experiment lasted six minutes.) 





T. 



T'. 



t. 



T. 



millims. 



100 r 

 T ■ 



Dry air 



(630 



jeoo 



; 59-0 

 158-0 



600 



51-0 



48-0 

 470 



42-2 



43-6 



42-5 



40-5 



28-0 

 (304) 

 15-2 



17-8 

 162 



5 



738-5 



742-6 

 6560 



7203 



34-86 



419 

 400 



Dry defiant gas... 



From these experiments the diathermancy of olefiant gas 

 appears somewhat higher than that of air. True, I cannot 

 warrant the purity of the gas ; but I believe that any possible 

 admixtures must have tended rather to increase than to dimi- 

 nish the absorption. 



The diathermancy of a hydrogen-vacuum of 1*5 millim. ten- 

 sion may be considered as very closely approximating to that 

 of an absolute vacuum. The comparable average value of the 



numbers for a hydrogen-vacuum gave — - — = 45 '4. Dividing 



6 



the values of the other gases by this number, and multiplying 

 the quotient by 100, the following ratios are obtained : — 



Name of the gas. 



Hydrogen, dry \ 



Air, dry 



>» 



>> 



>» 



>> 



Air-vacuum, moist at 1 2° C 



Air, saturated with vapour at 12° C... 



Carbonic acid, dry 



Olefiant gas .• 



Tension, 



Diather- 



Absorp- 



in millims. 



mancy. 



tion. 



15 



1000 







750-760 



102 







750-760 



45-6 



54-4 



520-6 



54-5 



45-5 



414-5 



600 



400 



254-5 



63-0 



37-0 



108-0 



804 



196 



121 



87-5 



12-5 



1-5 



95-6 



4-6 



12-9 



74-2 



25-8 



756-6 



432 



56-8 



7500 



42-1 



579 



7500 



53-0 



47-4 



Hydrogen at ordinary pressure always gave a somewhat 

 higher number than a vacuum ; so that I cannot suppose that the 

 ditference is due merely to errors of observation. Nevertheless 

 I do not regard this as a proof of conduction, which would re- 



