202 PROCEEDINGS OF THE AMERICAN ACADEMY 



XVIII. 



CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF 



HARVARD COLLEGE. 



ON A NEW METHOD OF DETERMINING 

 GAS DENSITIES. 



By Josiah Parsons Cooke. 



Presented June 12, 1889. 



In the well known method employed by Regnault for determining 

 the density or specific gravity of air, oxygen, nitrogen, hydrogen, and 

 carbonic acid, we deal primarily with tares, of which the weights to 

 be determined are the differences. The glass balloon which holds the 

 gas is tared by a similar balloon of exactly the same volume and of 

 nearly equal weight, suspended from the opposite pan of the balance. 

 The small difference of weight required to establish perfect equilibrium 

 is alone measured with standard brass or platinum weights. What- 

 ever may be the form of the subsequent calculation, the primary ob- 

 ject is to obtain the tare of the empty balloon when absolutely vacuous. 

 This known, the differences between such tare and the tare of the 

 balloon filled with various aeriform substances, gives the weights of 

 equal volumes of these substances under the temperatures and pres- 

 sures at which the balloon was filled. The volume of the counter- 

 poise is exactly adjusted to that of the balloon by the aid of a small 

 subsidiary glass bulb (Plate II.) ; and by sealing up more or less mer- 

 cury in this bulb it is easy to make the difference of weight such that 

 the standard weights required to complete the equilibrium will measure 

 the differences of tare to be determined, and no more. 



In the method of Regnault the tare of the empty balloon, or what 

 was equivalent to it, was found by exhausting the balloon with an 

 air-pump and weighing it after measuring the tension of the residual 

 gas while the glass was surrounded by ice. But it has been shown by 

 Agamennone * and Lord Rayleigh f that the results thus obtained are 



* Atti (Rendiconti) d. R. Acead. dei Lincei, 1885. 



t Proceedings of the Royal Society, vol. xliii. p. 362. 1888. 



