130 Scientific Proceedings, Royal Dublin Society. 



bably be a very efficient measurer of small changes of pressure. A 

 film of this kind which had sufficiently little rigidity to be deflected 

 by a very small change of pressure would not in general be strong 

 enough to support high pressures, and if a column of liquid be 

 added on top of it to assist it the surface may not be in stable 

 equilibrium. For instance, if it is attempted to get over the diffi- 

 culty of the evolution of hydrocarbons by indiarubber, and the 

 facility it offers for diffusion by pouring a liquid on top of the 

 film and increasing the air pressure below, so as to keep the film 

 flat, it will be found that the film is no longer stable, but buckles 

 up in some places and down in others. If the pressure below be 

 not sufficient, the liquid gradually extends the indiarubber until 

 it bursts it. By choosing the right pressure for the air below it 

 would be possible to have the film just near the point of neutral 

 equilibrium, and then any very small change in the pressure would 

 make a considerable one in the position of the film. Stretched 

 films of indiarubber are, however, a nuisance to work with, they 

 tire so rapidly. 



A third method of measuring small pressures accurately is by a 

 plan which is practically the same as the balanced barometer tube. 

 If a bell glass be turned over into a liquid, any small change- 

 of pressure in the air inside produces a corresponding force on 

 the glass, and if it be attached to a balance this change in pres- 

 sure may be readily weighed. As this pressure varies as the area 

 of the surface of liquid enclosed, and the capillarity only varies 

 as its perimeter by making the vessel large, the relative effect of 

 capillarity may be made very small. The only difficulty in this 

 method arises from the very small range of motion of the sus- 

 pended vessel. In the case of water being the liquid, the motion 

 of the suspended vessel would only be the hundredth of a milli- 

 meter for the millionth of an atmosphere ; but this should be quite 

 easily detected by means of a microscope, or a movable mirror, 

 or Newton's rings. This method is liable to the objection that a 

 liquid with a possible, though perhaps small, vapour pressure is 

 present. Another objection to liquids which have necessarily 

 horizontal surfaces is the danger of their absorbing gases super- 

 ficially, and so changing in density. The only one of these methods 

 that is quite free from all these objections is the one depending on 

 the use of large, flat, thin glass or metal films. 



