CONTINUOUS ELECTRIC CALORIMETRY. 
145 
Dieterici (‘ Wiedemann’s Annalen,’ vol. 57, p. 333) has repeated the calculation, 
employing Rowland’s values of the specific heat at constant pressure. He seems to 
argue that the variation of the specific heat at constant pressure as discovered by 
Rowland is of the character to be expected from the large variation of the specific 
heat at constant volume. I do. not think, however, that we can fairly infer anything 
with regard to the variation of the specific heat at constant pressure from that at 
constant volume. The value of the latter can only be deduced by the aid of very 
uncertain data, and we have as yet no sure theoretical guide as to the way in which 
it ought to vary, apart from experiments on the specific heat at constant pressure. 
The most remarkable facts about the specific heat of water are its great constancy 
over a considerable range, and its high value as compared with that of the solid or 
the vapour. The specific heat of liquid mercury is nearly the same as that of the 
solid, and both are about double that of the vapour at constant volume. The 
specific heat of water is double that of ice, and nearly three times that of steam at 
constant volume. If we adopt Rankine’s hypothesis of a constant “ absolute ” 
specific heat for each kind of matter, we must admit that this absolute specific 
heat has a different value in different states. 
J. Macfarlane Gray ( £ Proc. Inst. Mech. Eng.,’ 1889), adopting Rankine’s idea 
of a constant specific heat for “ideal” water, gives, without proof, the following 
formula for the total heat h of water reckoned from 0° 0., 
h = P0106 (6 - 273) + 0v dp/dd .(13), 
in which the constant is the ideal specific heat, and dp/dd is the rate of increase of 
the steam-pressure with temperature. The term 6v dp/dd is evidently intended to 
represent the latent heat of expansion of the liquid from the ideal state against the 
steam-pressure. The value of v, however, is not the apparent expansion from 0°C., 
but is the “ actual volume of unit mass of water less its absolute matter-volume, the 
pressure during the heating being that due to the higher temperature. Absolute 
matter is no doubt much more dense than platinum; and the reduction from the 
apparent volume, being very small, may therefore be disregarded.” The value of v 
is therefore taken as the observed volume at the temperature 6 of unit mass of 
water. The justification of this argument is not very clear, but the values 
calculated by the formula on this assumption agree fairly with the observations of 
Regnault on the specific heat between 110° and 190° C. # 
W. Sutherland, in a recent paper “ On the Molecular Constitution of Water,” 
(‘Phil. Mag.,’ vol. 50, p. 460, 1900), has endeavoured to explain the properties of 
water on the assumption that it is a mixture of two kinds of molecules in varying 
proportions, “ trihydrol,” 3HoO, which is identical with -ice, and “ dihydrol,” 2H.,0, 
which constitutes the greater part of liquid water at higher temperatures. He 
* In a recent paper (‘Proc. Inst. Civil Engn.,’ vol. 147, 1902), Gkay gives a further elucidation of this 
formula, and a detailed comparison with our experimental results .— Added March 11 , 1902. 
VOL. CXCIX.—A. 
U 
