208 PROCEEDINGS OF THE AMERICAN ACADEMY. 



from coming into contact with the molasses and glycerine by enclos- 

 ing the latter in an easily collapsible rubber tube, closed at the lower 

 end, and at the upper end tied over the mouth of the cylinder, as 

 shown in Figure 2. 



Molasses was the liquid used by Amagat in his manometer. A 

 heavy mineral oil, such as Barus used in a gauge of Amagat's type, 

 was found to be unsuitable for high pressure work, because it freezes 

 at room temperature under pressure. One grade of heavy oil tried in 

 this experiment froze at 20° under a pressure of 4500 kgm. Presum- 

 ably vaseline and such soft solids become unsuitable for the same 

 reason, although this point was not tested. For the same reason the 

 glycerine transmitting pressure from the pump had to be diluted with 

 water. The ease with which glycerine subcools, and the difficulty of 

 getting it pure, made any exact determinations impossible ; but it was 

 found that commercially pure glycerine was very apt to solidify at 

 G000 kgm. and 20°. 



Corrections to be applied to the Absolute Gauge 



In spite of the simplicity of this gauge, and the directness with 

 which it carries the measurement of pressure back to the fundamental 

 definition, there are two corrections which must be applied in prac- 

 tical use. These corrections are both so small, however, that neither 

 need be determined with much accuracy. 



The first correction is introduced by the slow leak, and is in amount 

 equal to the frictional force of the escaping liquid on the piston. The 

 equilibrating force must balance both the hydrostatic pressure on the 

 end of the piston and this frictional force. The effect of the correc- 

 tion, therefore, is to increase slightly the effective area of the piston. 

 If we assume that both cylinder and piston are perfectly cylindrical, 

 and that the crack between them is so narrow that the friction exerted 

 by the escaping liquid is equally divided between cylinder and piston, 

 then we easily see by writing down the equations of steady motion of 

 the escaping liquid that the friction increases the effective diameter 

 of the piston to the mean of the diameter of the piston and cylinder. 

 It appears from the equations that this correction is independent of 

 both the rapidity of leak and pressure. This is usually determined 

 by measuring the diameter of the piston directly, and the diameter of 

 the hole in the cylinder by some such indirect method as weighing the 

 quantity of mercury required to fill it. The dimensions of the gauge 

 used here were so small, however, that direct measurement of even 

 the piston could not be made with the desired percentage accuracy, 



