ss 
Steam Engine, 
of hot water and a pint of cold, is, afto* mixture, very" 
nearly half way between that of the two extremes. But 
this is not the case, when equal quantities of different bo- 
dies, at diSerent temperatures, are employed. 
(a) If a pint of quicksilver at 100^ Fahrenheit, be mix- 
ed with a pint of water at 40^, the resulting temperature 
will not be 70^ (the arithmetical mean), but only 60^. 
Hence the quicksilver loses 40^ of heat, which neverthe- 
less raise the temperature of the water only 20^ ; in other 
words, a larger quantity of caloric is required to raise the 
temperature of a pint of water, than that of a pint of mer- 
cury, through the same number of degrees. Hence it is 
inferred, that water has a greater capacity for caloric than 
is inherent in quicksilver. 
(b) The experiment may be reversed, by heating the 
water to a greater degree than the quicksilver. If the 
water be at 100®, and the mercuryat 40^, the resulting 
temperature will be nearly f 0® ; because the pint of hot 
water contains more caloric, than is necessary to raise the 
quicksilver to the arithmetical mean. 
(c) Lastly if we take two measures of quicksilver to 
one of water, it is of no consequence which is the hotter ; 
for the resulting temperature is always the mean between 
the two extremes ; for example, 70-, if the extremes be 
fOO® and 40®. Here, it is manifest, that the same quan- 
tity of caloric, which makes one measure of water warmer 
by 30®, is sufficient for making two measures of quick- 
silver w^armer by the same number. Quicksilver has,, 
therefore, a less capacity than water for caloric, in the 
proportion, when equal measures are taken, of one to two. 
if, instead of equal bulks of quicksilver and water, we 
had taken equal weights, the disparity between the spe- 
cihc caloric of the mercury and water tvould have been 
still greater. Thus a pound of water at 100®, mixed with 
a pound of mercury at 40®, gives a temperature of 97i, 
