476 proceedings: philosophical society 



The effusion method which has been widely used for this purpose 

 has been shown to be unreliable, errors of over 10 per cent not being 

 unusual in practice. A study of the theory of the effusion process has 

 revealed the main sources of error in this method. The effusion method 

 under the most favorable conditions can probably not be relied on to 

 better than 1 or 2 per cent. 



In order to secure precise results a specific gravity balance employ- 

 ing an indirect weighing method was developed; this enables one to se- 

 cure results accurate to about 0.2 per cent. This apparatus is con- 

 venient to use, may be made portable, and requires no preliminary 

 calibration. 



Discussion. The paper was discussed by Messrs. Priest, Hersey, 

 Sweet, Buckingham, and Humphreys. 



Informal communications: Mr. Buckingham outlined a theory ac- 

 counting for the errors in the effusion method of determining gas 

 densities. 



Mr. C. A. Briggs exhibited two combination time and percussion 

 caps for shrapnel shells of types now in use on the Russian battle front. 



Mr. H. F. Stimson exhibited a 2-stage mercury vapor condensation 

 pump for operating between about 4 cm. and 0.0004 mm. of mercury. 

 The low pressure stage is built with a long nozzle of the De Laval type 

 and the high pressure with a short nozzle. In each nozzle the ratio of 

 the higher to the lower pressure is maintained above the critical value 

 of about 2, and the velocity of the vapor in the nozzle is greater than 

 the velocity of sound. 



Donald H. Sweet, Secretary. 



