148 
PROFESSOR HUGH L. CALLENDAR ON ELECTRIC CALORIMETRY. 
proportion of dissolved ice molecules, which are also required to account for the 
anomalous expansion. But here there is no obvious guide to the proportion required, 
or to the latent heat of solution to he assumed. The basis adopted by Sutherland 
appears to he too empirical, and the results to which it leads are too vague and 
inexact to permit of a profitable comparison with experiment. In order to place any 
such hypothesis as that represented by formula (16) on a sound experimental basis, 
it would be necessary to determine the variation of specific heat of several typical 
liquids with great accuracy over a wide range. The required data are not at present 
obtainable, but the suggestion appears to be worth recording as a possible physical 
interpretation of the- variation of specific heat. 
