ON THE SPECIFIC HEAT OF WATER. 241 



of the third degree, which is, however, of no importance for the present purpose. 

 To get the mean value for any period we have 



J,, = 1 I Jrle = 4 > 2146-0-0013450 + 0-0000170 a -0-0000000660 il , 



Jo 



from which we get from BARNES' figures for the mean calorie from 13" C. to 55 C. 

 the result 



J, 3 15 = 4-175. 

 Our own result is 



Ji3 M = 4-182, 



which is higher than the Callendar-Barnes value by 0"007. On the other hand, our 

 value for the 15 calorie is less than that of BARNES' by 0'005, whilst the values for 

 the 26 calorie are identical, and the values for the 80 calorie differ by nearly 2 per 

 cent. Our value for the 30 calorie is identical with that obtained by ROWLAND for 

 30 with the mechanical method. 



It is difficult to explain these discrepancies, which clearly result from differences of 

 method involving defects which are more or less concealed. We are inclined to 

 attribute them to some extent to the thermoid effect described above in Section 6. 



Another point which might lead to discrepancies between different observers is one 

 necessarily incidental to electric calorimetry, and which has been already discussed in 

 Section 6, i.e. the difficulty of exactly defining where the leads to the heatitig 

 resistance end and the heating resistance begins. Owing to the large scale of our 

 apparatus, it was possible to reduce the debatable portion to an amount of not more 

 than 1/10,000 of the heating resistance. We do not find sufficient data to enable us 

 to say how this matter was dealt with by former observers. 



It is obvious that both the above matters- might affect not only the absolute values, 

 but also the relative values obtained at different parts of the temperature scale. 

 Whether these causes in fact account for the differences in the values obtained for 

 the 15 calorie, and in the still larger differences beyond 35 C., it is impossible to say 

 with certainty. 



It is particularly unfortunate that such differences should exist with reference to 

 so widely accepted a standard as the 15 calorie. But it is clear that, apart from the 

 points above mentioned, there is a general consensus of observers that the value of J 

 for 15 C. comes on a very sloping portion of the curve and may therefore be affected 

 by small temperature differences. It is curious that the result of all observers who 

 have given a range of absolute values of J (i.e., ROWLAND, GRIFFITHS, BARNES, and 

 ourselves) is to give the value of the 30 calorie as 4'174 within 0"001. This may 

 be seen by reference to fig. 11, and may be connected with the fact that at 30 C. 

 the value of J comes on what is generally recognised to be the flat portion of the 

 curve near the minimum value. But it would be much better were the standard 

 value chosen over a longer range on the flat portion of the curve, and in such a way 



VOL. CCXI, A. 2 I 



