PHYSICAL CHARACTERS OE MIXTURES OE ALCOHOL AND WATER. 
599 
It will be seen that the specific heat of the 10 per cent, spirit, deduced from these 
experiments, is as much above that of water as in the former experiments. To explain 
the results obtained in both cases as an effect of evaporation, we must make the very 
improbable hypothesis that the amount of liquid which evaporates during the intro- 
duction of a heated metallic body into a liquid is exactly proportional to the tempera- 
ture of the body, whether it be heated to the boiling-point of the liquid or be consider- 
ably below it — an assumption obviously untenable. 
Apart, however, from these experimental data, a little consideration will show that the 
effect observed is not due to evaporation. 
The evaporation, if any, which affects the result must take place during the moment 
when the weight is in the act of entering the fluid ; when once fairly underneath the 
surface, the large mass of cold liquid prevents the superheating of any part of it, whilst 
the effect produced by any evaporation taking place from the surface of the liquid spon- 
taneously, is already provided for in the correction obtained by observing the fall in 
temperature for some time after the temperature of the calorimeter has reached its 
maximum. 
The amount of evaporation, then, not included in this correction will depend, in the 
first place, on the temperature of the heated metal, which, being nearly the same in all 
the experiments, affects all alike ; secondly, upon the time which the weight takes in 
actually entering the fluid, which is also alike in all the experiments, and is but a small 
fraction of a second ; and lastly, all other circumstances being the same, this evaporation 
will be proportional to the surface of the heated body submerged. From the dimen- 
sions of the two weights, which have been previously given, the total surface of each, 
including the fan-wheel, can be calculated. It amounts to 182 '4 square centimetres in 
the case of the copper ring, and 153T square centimetres in the case of the brass ring. 
The former weighs 614‘49 grins., the latter 246 - 49 grms. The proportion of surface to 
weight is thus twice and T as great in the brass weight as in the copper weight ; in 
other words, the evaporation from the surface of the brass weight (153T square centi- 
metres) will have twice and T as much effect on its temperature as the evaporation from 
the surface of the copper weight (182T square centimetres) has on its temperature; yet 
the results obtained with the two weights are almost identical. 
A series of experiments has nevertheless been undertaken, in which the specific heat 
of the 10 per cent, and 20 per cent, spirit was estimated, by observing the heating-effect of 
a known weight of either liquid, heated to a known temperature, upon a known weight of 
distilled water contained in a calorimeter. The liquid was at first enclosed in a thin glass 
bulb, which was heated with its contents to the desired point, by being immersed in a 
mercury-bath as recommended by Professor Kopp. The results, though generally con- 
firming the previous experiments, did not agree among themselves as well as could be 
wished. This seemed chiefly owing to the fact that a globular body, having the greatest 
cubical capacity for a given surface, is not well adapted for the rapid cooling of its con- 
tents. In consequence of this, when about 50 cub. centims. of liquid were employed, 
4 l 2 
