SECT. in. EFFUSION OF GASES. 1 1 3 



reside solely in that internal movement of molecules 

 which is now generally admitted as an essential con- 

 dition of matter in a gaseous state. The molecules and 

 atoms are assumed to be perfectly elastic and to move 

 in all directions with different velocities according to the 

 nature of the gas. Enclosed in a porous vessel the 

 moving atoms constantly strike against its walls and 

 against one another, but in consequence of their perfect 

 elasticity, no loss of movement results from the collision. 

 When the gases inside and outside of the tube are 

 of the same density and molecular movement, an ex- 

 change takes place without any perceptible change of 

 volume ; but when the two gases are of different den- 

 sities and molecular velocities, then the reciprocal pene- 

 tration ceases to be equal on the two sides. Reciprocal 

 diffusion of gases is accelerated by heat and retarded by 

 cold ; the tension of the gases is increased in the first 

 case, and diminished in the second. 



In Mr. Graham's experiments relating to effusion, 

 a gas under a constant pressure was on one side of a 

 minute opening in a very thin plate, and a vacuum on 

 the other. The rapidity with which air or gases enter 

 the vacuum depends upon their specific gravity. A gas 

 rushes into a vacuum with the speed acquired by a heavy 

 body in falling from the height of an atmosphere of the 

 gas in question supposed to be everywhere of the same 

 density. The height of this uniform atmosphere will be 

 in an inverse ratio to the density of the gas. An at- 

 mosphere of hydrogen, for example, will be 16 times 

 higher than one of oxygen. But the velocity acquired 

 by a heavy body not being in direct proportion to the 

 height, but to the square root of the height, it follows 

 that the rate of flow of different gases into a vacuum 

 will be in an inverse ratio to the square root of their 

 respective densities. The rate of flow of oxygen being 

 represented by 1, that of hydrogen will be represented 



VOL. i. I 



