■-sb 



Professor Sir James Dewar 



[Jan. 21, 



sa])sequently increasing the diameter of the large circnlar fihn 

 produced. 



For studying the transit of gases fihus about 5 mm. in diameter 

 were used. Thev were mounted on a thick-walled glass tul)e ground 

 flat and polished at one end. The tube was fixed vertically and 

 centrally under the water surface before making the film. The water 

 is afterwards allowed to flow slowly out, carrying the film down, 

 thereby folding it securely and neatly over the ground end of the 

 tube. The water level inside and outside the tube must i;»e kept 

 the same by a suitable air inlet, otherwise the suction may break the 

 film. By this means tw^o, three or more black films can be overlaid, 

 and difl:'eVent coloured films obtained. 



In order to investigate the passage of air through such films a 

 small pressure was introduced into the tube supporting the film, and 

 the diminution with time measured on a sensitive alcohol displace- 

 ment manometer. A soap bubble was used to supply a small regulated 

 pressure. 



Minutes. 



Fig. 2. 



The above graph (Fig. 2; shows the observed fall of pressure 

 ip) as ordinates, with time {t) as abscissas, got from observations 

 from three separate films freshly dried in a warm oven before measure- 

 ment. 



Curve No. 1 is the graph of one film : Xos. 2 and :> of a double 

 and triple film. In all cases the fall of pressure was logarithmic. 

 That is to say, when log ^; is plotted Avith t a straight line graph is 



