692 
Proceedings of the Royal Society 
r its rise of temperature corresponding to a fall t x of the temperature 
of the heater, then 
c' Tl = (sW + c)t .... (3) 
where s is the specific heat of the solution. Eliminating c between 
this equation and (2) we get 
v — y - — - 1 = (sW + c)t . 
h 
If we put Tx = t\ the last equation becomes 
(w + e)f = (sW + c)r 
or 
t' -t 
+c ~w ' 
Hence if, as the experiments prove to be the case, the specific heat 
per unit volume of the solution do not differ very much from unity, 
t' will be nearly equal to r, and the value of s will not be materially 
affected by a small error in the determination of c. 
The following is a description of the mode of performing an ex- 
periment. The volume of the solution is first roughly measured in 
a graduated flask, the liquid is then poured into the beaker and 
weighed. The beaker is then placed in the pad of cotton wool, 
and the thermometer put into position within it. The heater is 
placed on a hot plate and heated to about 95° C. ; it is then lifted 
off and shaken up to bring the whole of the mercury to one 
temperature, and placed in the solution, the temperatures of the 
solution and of the mercury in the heater being carefully read just 
before the heater is immersed. Thus all uncertainty as to what the 
temperature of the heater was when it was placed in the liquid was 
avoided. The temperature of the solution was noted four or five 
minutes after the heater was placed in it, the liquid having first 
been stirred to equalise its temperature. After this interval of time 
there was no appreciable difference between the temperatures of the 
heater and liquid. Another reading of the temperature of the 
solution was made after the lapse of an equal interval of time, and 
the difference of the two readings added to the first to allow for 
cooling. 
The following table gives the results of experiments on several 
solutions : — 
