345 
would be equal to one another, and equal to those in the third column. 
The differences between the actual numbers are surprisingly small, but 
are so consistent that they cannot be attributed to errors of observa- 
tion. It is possible they may be due to a want of perfect agreement 
in the values of a degree on the different thermometric scales ; but 
it seems more probable that they represent true discrepancies from 
theory, and are therefore excessively interesting, and possibly of 
high importance with a view to estimating the effects of inequalities 
of surface and of interior conductivity. The final means of the num- 
bers in the second and third columns are 
Calton Hill . . . *11702 
Experimental Garden . '11061 
Craigleith Quarry . . *06988 
The thermal capacities of specimens of the trap rock, the sand, and 
the sandstone of the three localities were, at the request of Professor 
Forbes, measured by Regnault and found to be respectively 
*5283, *3006, and -4623. 
Hence, according to proposition (3), stated above, the thermal con- 
ductivities are as follows : — 
Trap rock of Calton Hill, . . . 121*2 
Sand of Experimental Garden, • . 77*19 
Sandstone of Craigleith Quarry, . . 273*6 
These numbers do not differ much from those given by Professor 
Forbes, who for the first time derived determinations of thermal 
conductivity in absolute measure from observations of terrestrial 
temperature. In consequence of the peculiar mode of reduction, 
followed in the present investigation, it may be assumed that the 
estimates of conductivity now given are closer approximations to the 
truth. To reduce to the English foot as unit of length, we must 
multiply by the square of 1*06575; to reduce, further, to the 
quantity of heat required to raise 1 lb of water by 1° as unit of heat, 
we must multiply by 66*447 ,* and lastly, to reduce to a day as unit of 
time, we must divide by 365J. We thus find the following results : — • 
Trap rock of Calton Hill, 
23*5 
Sand of Experimental Garden, 
15*0 
Sandstone of Craigleith Quarry, 
53*5 
These numbers show the quantities of heat per square foot con- 
ducted in a day through a layer of the material one foot thick, kept 
with its two surfaces ata difference of temperature of one degree, — the 
