240 



MESSRS. W. R. BOUSFIELD AND W. ERIC BOUSFIELD 



Henoe the preponderance of the results by various methods is in favour of the lower 

 minimum. We are confirmed in the view that the minimum is in fact at about 25 C. 

 by an entirely different set of experiments which were made by us by heating with a 

 constant power of about 250 watts regulated by a watt-balance, so that determina- 

 tions of current and resistance were unnecessary. This set of experiments also gave 

 the minimum value at about 25 C. We have not here set out the details of these 

 experiments, as they could not be relied on for absolute values, and would have 

 h'ligthfiH'd this communication without corresponding advantages. But for relative 

 values, which alone are required for determining the minimum, such measurements are 

 better adapted than any others for determining the comparative rise of temperature 

 at different points of the scale for the same heating, and they strengthen our view as 

 to the probability that the minimum is in fact in the neighbourhood of 25 C. 



On the other hand, as regards the absolute value of the 15 calorie which we place 

 at 4'179, the preponderance of the results is against this low figure and in favour of 

 the value of 4'184, at which it is placed by BARNES. JOULE'S determination, which 

 works out at 4'173 at 16'5 C., or about 4'174 at 15 C., is the only one which is 

 lower than our own ( JOULE, 'Phil. Trans.,' vol. 169, p. 3G5, 1878). All the electrical 

 determinations give higher values than ours. 



For the purpose of further comparing our figures with the most recent figures 

 given by BARNES, we have worked out a formula for BARNES' figures and have found 

 that BARNES' values from C. to 80 C. may be represented with sufficient accuracy 

 by the equation 



J = 4-2146-0 > 002G90 + 0-0000510 2 -0-0000002650 3 . 



In the subjoined Table XIII. BARNES' values for J are given and the values 

 calculated from the above expression, and also the differences. 



TABLE XIII. 



It will be seen that from C. to 80 C. the agreement is well within the limits of 

 Beyond this point there is a serious divergence for any curve 



