sl8 



Professor Sir James Dewar 



as before ; and in every case straight line graplis are thus obtained. 

 These straight lines all show a marked deviation at 0° C, which 

 luiturallv suggests that water in some condition is the associated 

 cause. In the case of carbonic acid, another break was obtained at 

 about - :^w ' C. The gases used were dried through a suitably cooled 

 U-tube before entering the box round the membrane shown in 

 Fig. 1 (a) ; but there is the possibility that water in some form of 

 colloidal association is a constituent of the complicated structure 

 of the rubber membrane. The second break in the carbonic acid 



250 



a. 200 



5 



u 



a. 150 



Q. 



100 



50 



10 



20 



Pressures in Atmospheres 



Fig. 19. 



curve at - 37° C. suggests a possible combination of the membrane 

 with carbonic acid below this temperature. 



These results were obtained with gases at the atmospheri pressure 

 passing into the vacuous space of known volume behind the mem- 

 brane. Working at higher pressures — up to twenty atmospheres — 

 this curious break in the logarithmic graph is much less evident. 

 The conditions of experimenting were not quite comparable, as the 

 gases passing through the membrane were collected at atmospheric 

 pressure and measured either over brine or heavy mineral oil in the 

 case of carbonic acid, or over water when hydrogen was employed. 



