442 MR. W. J. M. RANKINE ON THE COMPUTATION 
The first term of this expression corresponds to the formula employed by 
M. Reenautr. To correct the results given in his table of experiments, we must, 
therefore, subtract from each of them the quantity 
(P—@)v 
K,,,(T,—T,) 
As T, and T, were always low temperatures, I have treated K,,, as a con- 
stant quantity in computing the corrections, its value being 
Number. ieee 
ogarithm. 
In feet per degree of Fahrenheit, . : } 5 772 2°8876173 
In feet per centigrade degree, - : 5 ‘i 1389°6 3:1428898 
In métres per centigrade degree, . ; : : 423:54 2-6268944 
In the following table, the numbers in the first column refer to certain groups 
of experiments in M. Rrecnavtt’s table, the mean results of which are given in 
the succeeding columns. 
The correction, which is scarcely appreciable for temperatures near the ordi- 
nary boiling point, increases rapidly as the temperature in the boiler rises. 
The temperatures are all stated according to the scale of a centigrade air ther- 
mometer. 
TABLE I. 
Reference to ‘ K 
Ki,2 Correction to = 
M. REGNAULT’S 2 ~5 
Sea ere as computed by| be subtracted.| , S42 
MSR N EN M. REGNAULT. Corrected. 
1, 2, 3. 1: | 1:00384 0:00009 1:00375 
A 10> ihe : 100665 0-00010 1:00655 
26, 27, 28, 29. 1:00871 0-00092 1:00779 
30, 31, 32, 33. ; 101140 0-00121 1:01019 







36, 37, 38. 1:01581 0-00162 1:01419 
Empirical Formule. 
The results of experiment, as thus corrected, agree very nearly with those of 
the following empirical formula, in which K is the apparent specific heat of liquid 
water at the temperature T, and K,, its apparent specific heat at the tempera- 
ture T,, which is that of the maximum density of water; viz., 4°1 centigrade, 
or 39°°4 Fahr. 
a is a constant coefficient, whose value is, 
For the centigrade scale, ; : d e . 0-000001 
For Fahrenheit’s scale, . ; 2 : F ; 0:000000309 

