BETWEEN THE FREEZING AND BOILING-POINTS. 
171 
Arbor, Michigan University (‘ Physical Review,’ vol. 9, p. 288. 1899), with a type of 
dynamometer designed by themselves. The results of these measurements show a 
wide divergence. The values found are 
Glazebrook and Skinner . . . 1-4342 volt at 15°. 
Kahle . .. 1-43285 ,, ,, ,, 
Carhart and Guthe . 1-4333 ,, ,, ,, 
The large discrepancy in the value of the mechanical equivalent of heat obtained 
by the electrical methods used by Professor Griffiths and Professors Schuster and 
Gannon, as compared to the value given by the direct mechanical method, has so far 
hinged on the value to be assigned to the Clark cell. The older, and for so long a 
time accepted, value, 1*4342, there is every reason now to think is too high. The 
value given by Dr. Kahle, i.e., 1 '43285 volt, is at the same time probably a little too 
low. The value found by Carhart and Guthe depends on the mean of three 
determinations differing in the extreme by "5 mvt. These three determinations were 
made for two Clark cells in series, one of which was afterwards compared to the 
Reichsanstalt cells and found to be in good agreement. 
At present there is a grave uncertainty in the absolute value of this fundamental 
constant, which requires immediate attention. It has been pointed out that the 
value of the mechanical equivalent of heat found by Griffiths would be brought 
into harmony with the values found by Rowland by the direct mechanical method, 
by assuming the Clark cell 2 mv. lower than the value found by Glazebrook and 
Skinner. The value found by Schuster and Gannon requires a somewhat smaller 
correction in the same direction. 
In the face of these uncertainties in the value to be assigned to e, I have adopted 
the older value, 1*4342 international volt, as the basis of my calculations of the 
absolute value of the mechanical equivalent, in order to bring my results into 
comparison with those of Griffiths and of Schuster and Gannon. On this basis 
I have had the temerity to combine the mean value of the mechanical equivalent 
obtained by integrating the curve of absolute values between 0° and 100° with the 
determination of the mean value obtained by Reynolds and Moorby, and have 
obtained by that way an absolute measure of the Clark cell in terms of the 
mechanical units, which is probably as accurate a value as has yet been obtained, 
provided the values assumed for my resistance standards are correct. 
The discrepancy in the two values of the mean mechanical equivalent, the one 
obtained by integrating the variation curve, and the other obtained as a direct 
determination, is "132 per cent. As I have used the value e of the Clark cell in my 
measurements squared, this reduces to "066 per cent, on 1 "43420, and shows that 
the value assumed for my cells is too high by this amount. The true value of the 
Clark cell I have assumed for calculation is 1"4342, which would give for the two 
z 2 
