OP ARTS AND SCIENCES. 187 



been hitherto regarded as among the most trustworthy data of science. 

 His determinations were all made by the method of counterpoise which 

 we adopted in our work, and he used balloons of twice the volume of 

 those we employed. When exhausted, the glass must have been con- 

 densed to an even greater extent than has been shown above ; but no 

 account whatever is taken of this shrinkage. As Regnault's constants 

 have been universally used, it is obvious that Lord Rayleigh's correc- 

 tion must be applied to all determinations of gas or vapor densities 

 hitherto made, and to all atomic weight determinations of any kind 

 which involve the calculation of the weight of a measured volume of 

 any gas or vapor. Except, however, in the case of hydrogen, the cor- 

 rection will be inconsiderable. 



J. P. C. Cambridge, March 15, 1888. 



