37 MAXWELL'S LAW 



79 



sound the pressure alone is transmitted, whereas in effusion 

 the transmission of pressure can only be effected by the gas 

 itself flowing from the compression vessel into space of 

 lower pressure, and thereby raising the density in this outer 

 space. Energy and mass therefore flow out together with 

 equal speed, and the gas would flow out with the speed of 

 sound if there were no obstacle in the way. 



This assumption, however, is untenable, for on the one 

 hand the friction which the gas experiences from the rim 

 of the orifice retards it, and on the other the pressure in the 

 space into which it flows acts as a resistance. 



The first disturbance, that by friction, is avoided as much 

 as possible by making the orifice small and using a thin 

 wall. For if the wall is not very thin the orifice acts as a 

 tube in which the flow would be considerably diminished by 

 the internal friction of the gas. And the orifice must be 

 small, as otherwise it would offer room for eddies and 

 vortices to form which would hinder the regularity of the 

 efflux ; for the turbulent motions that must occur in wider 

 orifices would cause heating effects which would consume 

 a large part of the energy present. 



The other disturbance, that by the back pressure of the 

 space into which the efflux occurs, is removed when the gas 

 flows into vacuous space. In this case the issuing mass of 

 gas appears to be almost exactly proportional to the pressure 

 which drives it out, as was a priori to be expected. Since 

 the density of a gas is proportional to its pressure, we must 

 from this experimental result draw the conclusion that the 

 speed with which a gas flows into vacuous space is inde- 

 pendent of the pressure which causes the flow. But in this 

 conclusion lies a confirmation of the inference already drawn 

 from the theory, viz. that the speed of efflux of a gas is 

 determined only by the speed with which sound travels in 

 this gas, or, what amounts to the same thing, on the mean 

 speed of molecular motion in the gas. 



The manner of dependence of the speed of efflux on that 

 of the molecular motion could be assigned if we knew how 

 many molecules issue per unit of time, and what is the 

 pressure in the orifice itself. The former question may be 



