G. Jounstong Sronzy on the Penetration of Heat across Layers of Gas. 17 
B by the process here described, and for which I would suggest the name 
penetration, by forming an expression which aims at roughly representing the 
quantity of heat absorbed by the gas per second froma square centimetre of A. 
One such expression is approximately— 
GQ eg a NET eb bal Wes ONY (2) 
in which V is the velocity with which the adjustment is made, « the heat which 
would raise a gramme of the gas one degree in temperature, and p, the density 
(referred to water) of the gas, where it is in contact with A. 
To get the loss by penetration per second from the whole surface of the cooling 
i : 
body, we have to find the value of the integral hh S dA, dA being an element of 
dQ 
the surface of the cooling body, and °F having the value assigned to it above. If 
the surface is everywhere equally exposed, a condition easily secured in making 
experiments with thermometer bulbs, this becomes simply A a , or— 
AO 
AVooi—— CARAT Ck A eA a fe (e), 
] 
where A is the area of the surface of the cooling body. 
7. It will be instructive to compare the loss of heat by penetration with the quantity 
which is carried off by convection. ‘To estimate the latter, let @ be the section of 
the convection current, » its average density, Ae the average excess of its temper- 
ature, and v its velocity. Then the total quantity of heat which will be removed 
per second by convection will be— 
Ad 
Qvop SY aa ies POMCL CS) 
This to be compared with (e) the expression for the total loss of heat per second 
by penetration. 
Now, in the cases that occur in laboratory experiments, Acp, is seldom many 
times larger or many times smaller than Qs, but V is always very much 
larger than v, whence (2) may have a value comparable with (4), while ao, 
is very much less than ag; in other words when the processions between 
the opposed surfaces have but slightly different velocities. We learn from 
this that the escape by penetration may be expected to manifest itself as soon as the 
Crookes’s layer has become in a moderate degree compressed.* It is also evident 
that the co-existence of a convection current will not much affect the escape of heat 
by penetration, inasmuch as convection currents are sluggish when compared with 
the promptness with which re-adjustments are made in Crookes’s layers. It is 
therefore worth while to examine the numerous records of experiments upon the 
velocity with which bodies cool in gases, with a view to finding whether instances of 
the escape of heat by penetration can be found among them. 
* Hence, also, thermal experiments may be expected to explore Crookes’s layers with more sensitiveness 
than contrivances for manifesting the mechanical force which is also present. 
