63 
of Edinburgh, Session 1872-73. 
but the early determinations were rejected for obvious reasons, and 
the ultimate determination was made from a mean of fourteen 
readings, the experiments having been performed in this case with 
extraordinary care, and with all the experience derived from 
previous trials. 
Let the heat required to raise 1 gramme of water 1° C. be 
taken as our unit of heat. Assuming that in the formation of the 
ice a statical state of temperature has been reached, we have 
x 
when F = the flux of heat, h = the coefficient of conductivity, 
SO = — (the temperature of the freezing mixture), 
x = the thickness of ice. 
Jc is assumed to be the quantity of heat which crosses an area of ice 
1 square centimetre section in 1 minute, the thickness of ice being 
1 centimetre, and the difference of temperature of the two sides of 
the ice being 1° 0. 
But F = the quantity of water raised 1° Cent., in 1 minute, 
over a surface of 1 sq. centimetre. 
= (volume of ice formed in centimetres) x (latent 
heat of water) x (specific gravity of ice). 
dt 
S.L. 
S.L. • ~ 
n dt 
kse.t = s.l f 
Employing this formula in the series of experiments alluded to 
above as being worthy of the greatest confidence, fourteen values 
of h were found corresponding to different values of x. Now, for 
small values of x an error in the value of x will introduce an error 
into the value of h greater than for large values of x. The value 
VOL. VIII. 
