96 
PKOFESSOE KOPP ON THE SPECIFIC HEAT OF SOLID BODIES. 
perature of the contents of the glass at the time the maximum temperature has been 
attained in the calorimeter, it is unnecessary to give the indications of T' in hundredths 
of a degree; and since the temperature T, to which the glass with its contents was 
heated in the mercury-bath, only serves to deduce the difference T — T', it is unimportant 
in giving this temperature to do so in hundredths of a degree. The accuracy of the 
determinations of specific heat, in so far as it depends on determinations of temperature, 
is limited by the accuracy with which the difference of T— T' and t ! —t are determined 
(where t is the original temperature of the water in the calorimeter, and the other 
letters have the meanings previously assigned to them). To have one of these differences 
very accurately, while the other is much less accurately determined, avails nothing for 
the accuracy of the final results. It is at once seen that in my experiments, and especially 
in those of Neumann and Regnault, the hundredths of a degree have a greater signifi- 
cance for the small difference tf— t, than the tenths of a degree for the great difference 
T-T'. 
The correction for educing the value of T', which I have just discussed, is of course 
more important the smaller the difference T— T' ; for most of my experiments in which 
this difference is about 30°, the significance of this correction is inconsiderable, if the con- 
tents of the glass be a good conductor. I give a few numbers. The experiments given 
in § 25 on the specific heat of mercury, which, by using this correction, give it at 0‘ 0335 
in the mean, give it = 0-0331 if this correction is neglected, that is, T' made=^. 
The fourth series of experiments, given in § 27, for determining the specific heat of coal- 
tar naphtha A, give it at 0-425 when this correction is made, and at 0-420 when it is 
omitted. The first series of experiments in § 33, for determining the specific heat of 
sulphur, give it at 0-159 when this correction is used, and at 0-152 when it is neglected. 
Whether in all such cases T' is put —t\ or=£ ,o +0°-3, is of inconsiderable importance. 
The correction in question is inadequate if the substance in the glass is a bad conductor ; 
for example, when the solid in the glass is a pulverulent or porous mass, in which the 
moistening liquid stagnates (compare § 18). That, under such circumstances, the numbers 
obtained for the specific heat are found somewhat too small must be remembered in 
§ 41 in the case of chromium, and in § 52 in the case of chloride of chromium. Too 
small numbers are also obtained, if in the experiments the maximum temperature of the 
cooling water exceeds that of the air by much more than 2°. Such experiments are not 
comparable with the others, for example, with those made for the purpose of ascer- 
taining the ancillary magnitudes occurring in the calculation of the results ; for them this 
correction is inadequate, and the loss of heat which the contents of the calorimeter ex- 
periences between the time which elapsed between immersing the glass and the establish- 
ment of the maximum temperature is too great. By individual examples in § 25 in the 
case of water, in § 39 in the case of copper, and § 41 in the case of iron, I shall call to 
mind how this source of error may give somewhat too small numbers for the specific 
heat ; but I have always tried to avoid this error, since I saw its importance in my first 
preliminary experiments. 
