148 PROCEEDINGS OF THE AMERICAN ACADEMY. 



Thus 96,610 coulombs correspond to one gram equivalent of an 

 electrolyte. 



These numbers are of interest from a theoretical standpoint, but their 

 practical value as means of determining current strength must depend 

 upon tlie ease of their application. The value for copper is a mean be- 

 tween two limits, and is not easy to reproduce with exactness ; but for 

 most purposes the weight deposited by current densities between 0.008 

 and 0.012 amperes per square centimeter, from ten per cent solutions of 

 cupric sulphate cooled below zero and protected by an inert atmosphere, 

 is sufficiently close to the real value. Nevertheless, empirical tables 

 like those of Gray will probably remain the most convenient method of 

 computing current strength from the deposition of copper. 



On the other hand, the intricacy added by the porous cup in the silver 

 determination is not serious, and the results obtained seem trustworthy. 

 Obviously any method capable of being repeated with constant results, when 

 standardized with accuracy by means of known currents, would serve all 

 practical purposes ; and a method which is constant because the sources 

 of eri-or have been at least partially removed is certainly worthy of 

 attention. 



This point suggests a discussion which has recently ariseu concern- 

 ing the electrical determination of Joule's equivalent, which depends upon 

 measuring the heat corresponding to a known amount of electrical 

 energy. Griffiths, in an appendix* to an elaborate paper f upon this 

 subject, calls attention to the fact that an error in the electrochemical 

 equivalent of silver of 0.1 per cent would explain the difference between 

 his value of the mechanical equivalent of heat and Rowland's. The 

 equation of Griffiths for the calculation of his value is 



in which i? signifies the constant potential difference, R:= the resistance 

 converting electricity into heat, M the heat-capacity which is heated, 8t 

 the interval of time, and SO the change of temperature. The electro- 

 chemical equivalent of silver enters into this expression in the determina- 

 tion of the value of £ in reference to H. Griffiths's value for J is larger 

 than Rowland's, hence his value for E is greater than that which Row- 

 land would have observed if he had used an electrical method; or, in 

 other words, too small an amount of silver may have been taken as the 



* Nature, 56, 258 (1897). i Proc. Roy. Soc, 53, 6 (1S03J. 



