MEMOIR XXVI.] CAPILLAltY ATTRACTION, ETC. 359 



By this apparatus it was found that one thousand 

 measures of atmospheric air exposed to atmospheric air 

 underwent no change either in volume or composition. 

 The exposure in some cases lasted an hour. 



One thousand measures of hydrogen, in the bubble, 

 were exposed to atmospheric air in the bell. In five 

 minutes there remained only four hundred and seventy- 

 two. (It will be understood that the numbers quoted 

 in this and other succeeding experiments are for the pur- 

 pose of illustrating the general principle. They change 

 w r ith the relative proportion of gases inside and outside 

 of the bubble.) 



A reverse action ensues when nitrogen is substituted 

 for hydrogen. The bubble swells instead of diminish- 

 ing. Thus one hundred measures of nitrogen in half an 

 hour became one hundred and seven and a half. 



Oxygen decreases in bulk. Two hundred and fifty 

 measures in ten minutes became one hundred and fifty- 

 three. This gas passes more rapidly through the bub- 

 ble than nitrogen. 



Carbonic acid passes through the bubble very rapidly. 

 When five hundred measures were used, the bubble col- 

 lapsed almost as fast as it had expanded. Under a 

 water roof half an inch thick and two inches in diame- 

 ter, five thousand measures escaped into the air in forty- 

 eight hours. In its place there were found two hundred 

 measures of atmospheric air, which had passed in the op- 

 posite direction through the water roof. 



By this apparatus it was proved that these motions 

 through soap-bubbles continue until the gases on both 

 sides of the bubble have the same chemical composition. 



It has long been known that liquids and gases pass 

 through porous structures though resisted by consider- 

 able force. Thus, if over the mouth of a cylindrical jar 

 Z>, Fig. 8-4, a thin sheet of India-rubber be tied, and the 



