1921.] 
Marsden.—Some Tropical Heat-insulators. 
189 
were available locally, and were sun-dried. Elaborate laboratory methods 
were impossible, and the apparatus shown in fig. 1 was used. 
Tee was prepared from the purest water available, broken up, arid 
placed in the inner cylinder C 1 . This cylinder was surrounded by a 
concentric cylinder C 2 , the space between the two being filled with the 
material under test, and the ends closed by thick wooden plugs. The 
heat conducted through the material to the ice in the inner cylinder was 
measured by the amount of ice melted, since each gramme of ice requires 
79*6 calories to melt it (143*3 B.Th.U. per pound). The expression con¬ 
necting the coefficient of conductivity with the amount of heat conducted 
through the material may be obtained as follows :— 
Let the length of cylinder be l, the radius of the inner cylinder r 1; and 
of the outer cylinder r a . Take a cylindrical ring in the material of radius r 
and thickness dr; suppose dT to be the difference of temperature between 
the surfaces of this ring. Then, if K be 
the conductivity, the amount of heat, Q, 
winch crosses this ring (when conditions 
are steady) is given bv 
Q = K2 -rl 
dT 
dr 
i.e., 
dT 
Q 
ZttTK 
X 
dr 
Integrating between limits T L and T 2 
for temperature, and r i and for radii, 
we obtain 
Q log. r 2 /ri 
K = 
T 2 — T, 
Q foge r 2 /rt 
2t rl (T 2 — Ti) 
2ttIK 
= 2*3026 
or 
Q log i o — 
r i 
m (t 2 - t \y 
Fig. 2. 
This formula holds only for a steady radial flow. To ensure this 
condition measurements were based only on the central portion of the 
cylinder. Two partitions, A and A. (fig. 1), separated off a length l from 
