G. F. Firzazratp—On the Mechameal Theory of Crookes’s Force. 67 
and K would be large if it were 100, so that even if a were 50 still 2aK would 
still be less than ~4th of X*» so that we may take approximately 
Q=—hVa. KX 
2 Kags 
From this we can calculate K ; for =1°6 in most gases, and if a—2°5, Va—1°5 and 
X=9700 as above -. k¥aX=22,310=2 x 10*g.p. Now at a distance of a fourth 
metre in air at atmospheric pressure, and with a difference of temperature of 10°C 
Q=10° ¢.p. so that in this case 
«=50 g.p. which is within the limits of the quantities obtained in the 
case of the spheroidal drops on liquids. 
That by this formula K varies nearly as Q and not as Q’is not to be wondered 
at because in the first place the formula only professes to represent an approx- 
mation to the true state of affairs, and in the second place it is only at distances 
and pressures at which the ordinary laws of conduction of heat cease to apply 
that it professes even approximately to represent it, 
The whole of these investigations are unsatisfactory to this extent, that I have 
been unable, from a consideration of the molecular encounters themselves, to dis- 
cover what is the actual distribution of velocities even in the simple case of two 
parallel surfaces. This is hardly to be wondered at for the problem is extremely 
complicated and evidently depends upon the undecided point in molecular physics, 
namely, the proportion of the molecules encountering in a given direction that are 
thrown off in the various other directions. We might very well assume with 
Maxwell that they are uniformly distributed in every direction after the encounter 
but even this does not simplify the question sufficiently to bring it within my 
present powers of solution. 
