18790 Molecular Physics in High Vacua. 413 
Thus it happens that the extra energy of the impacts 
against the warm side of the diaphragm is exactly com- 
pensated by the increased number of impacts on the cool 
side. In spite therefore of the increased activity commu- 
nicated to a portion of the bees, the pressure on the two 
sides of the diaphragm will remain the same. This repre- 
sents what occurs when the extent of the box containing 
the bees is so great, compared with the mean free path, that 
the abrupt change in the velocities of those bees which 
rebound from the walls of the box produces only an insen- 
sible influence on the motions of bees at so great a distance 
as the diaphragm. 
I will next ask you to imagine that I am gradually re- 
moving bees from our box, still keeping the diaphragm warm 
on one side. The bees getting fewer the collisions will become 
less frequent, and the distance each bee can fly before 
striking its neighbour will get longer and longer, and the 
crowding in front of them will grow less and less. The 
compensation will also diminish, and the warmed side of 
the diaphragm will have a tendency to be beaten back. A 
point will at last be reached on the warm side, when the 
mean free path of the bees will be long enough to admit 
of their dashing right across from the diaphragm to the side 
of the box, without meeting more than a certain number 
of in-coming bees in their flight. In this case the bees will 
no longer fly quite in the same direction as before. They 
will now fly less sideways, and more forwards and backwards 
between the heated face of the diaphragm and the opposed 
wall of the box. Because of this preponderating motion, 
and also because they will thereby less effectually keep back 
bees crowding in from the sides, there will now be a greater 
proportionate pressure both on the hot face of the diaphragm 
and on that part of the box which is in front of it. Hence 
the pressure on the hot side will now exceed that on the 
cool side of the diaphragm, which will consequently have a 
backward movement communicated to it. 
I may diminish the size of the bees as much as I like, 
and by correspondingly increasing their number the mean 
free path will remain the same. Instead of bees let me call 
them molecules, and instead of having a few hundreds or 
thousands in the box let me have millions and billions and 
trillions ; and if we also diminish the mean free path 
to a considerable extent, we get a rough outline of 
the kinetic theory of gases. (I may just mention that the 
mean free path of the molecules in air, at the ordinary 
pressure, is the ten-thousandth of a millimetre.) 
2 E 2 
