200 MESSRS. W. R BOUSFIELD AND W. KRIC BOUSFIELD 



of water combined with a solute at various temperatures and concentrations. It 

 seemed probable that similar series of observations upon the specific heat of solutions 

 over considerable ranges of concentration and temperature would throw further light 

 upon these matters, and the apparatus described in this paper was therefore primarily- 

 designed for the observation of the specific heat of solutions with the desired degree 

 of accuracy, and with the facility and ease of manipulation which are essential when 

 it is it-quired to amass a large body of data in a reasonable time. 



At an early stage it became apparent that the temperature-specific heat curve 

 of \\.-iter was entirely altered in character by the introduction of a small amount of 

 solute. With a half-normal solution of KC1 the more or less parabolic curve for water 

 U-c.iiin-s nearly a straight line, and even with fairly dilute solutions the water curve 

 is greatly modified. The appreciation of this modification necessarily involved as a 

 starting-point the consideration of the curve for pure water, as to the form of which 

 different observers have come to widely different conclusions. A reference to fig. 10 

 (Section 14 post), where the curves given by different observers are plotted, shows that 

 the latest form of the curve, which is the result of the researches of CALLENDAR and 

 BARNES, differs widely from the curves given by REGNAULT and by LUDIN.* At 80 C. 

 the values of the specific heat of water in terms of the 15 calorie are 



BARNES T0014, 



REGNAULT 1 '008.1, 



LUDIN T0113, 



showing a difference of 1 per cent., which was of the same order as the differences we 



were expecting to find between dilute solutions and water. We were therefore led 



first of all to use the apparatus to endeavour to verify the specific heat curve for 



water. Moreover, it seemed likely that differences in the value (in joules) of the 15 



calorie might underlie the large differences which appeared at the upper end of the 



temperature scale. The lowest value of the 15 calorie is that which is given by 



the determination by JOULE in 1878, t which works out at about 4174 joules, whilst 



the highest is that given by GRIFFITHS,} which is 4-198, BARNES' latest valueg being 



Hence it seemed desirable to endeavour to verify not only the relative values 



required for the specific heat curve, but also the absolute value of the mechanical 



equivalent. By slight modifications of our apparatus we were able to adapt it to this 



mrpose. The method of electrical heating being that adopted, we were able to use it 



tmuous-flow experiments of the general type of those of CALLENDAR and 



The investigation upon which we were thus launched proved sufficiently 



See the Landolt.Born.tein Tables,' ed. 1905, p. 393, where the result3 of previous observers are set 

 out in parallel columns. 



f JOUI.E, 'Phil. Trans.,' 1878, vol. 169, p. 365. 

 GRIFFITHS, 'Phil. Trans,' A, 1893, vol. 184, p. 361. 

 BARNES, < Koy. Soc. Proc.,' A, vol. 82, p. 394, 1909. 



