The object of our investigation has been to study the conduction of heat through a light powder and to find how it depends upon the pressure and thermal conductivity of the gas in which the powder is immersed. A solution of this question is part of the solution of the problem of the conduction of heat through a certain class of “solid” heat insulators—a class which includes those of lowest thermal conductivity. The class of insulator referred to are solids dispersed in gases, or gases dispersed in solids, and consists of three kinds of substances, (1) fibrous substances ( e.g., wool, eiderdown, asbestos), (2) cellular substances (e. g., cork, pumice stone) and (3) powders (e. g., lamp-black, powdered cork, silox or monox). It might be expected that substances so different as those mentioned would have very different thermal conductivities. Actually their conductivities range from about 8 to 11 times 10
-5
cal. cm.
-1
deg.
-1
sec.
-1
. As there is nothing common to the solid part of these substances, their conductivities, it would seem probable, are determined mainly by the factor which is common to them all, that is, the gaseous part, which is air. Our experiments have been made with a very light powder known as monox or silox, and the conductivity of this powder when immersed in air, in carbon dioxide, and in hydrogen at various pressures has been determined. We find that there is a linear relation between the conductivity of the powder and the logarithm of the pressure of the gas in which it is immersed, so that if
k
is the measure of the conductivity of the powder,
k
0
that of the gas in which it is immersed, and
p
the measure of the gas pressure, then
k
= ½
k
0
log
10
p/n
approximately, where
n
is a constant for a given gas.
The thermal conductivity of some gases at 0° C
