Hydrogen, oxygen and carbon-mon-oxide effuse in the manner predicted by the theory for gases having the same, or nearly the same, specific heat ratios. This agrees qualitatively with theory, and affords a confirmation of the high specific heat ratio of argon. The observations showed that argon effused 3½ per cent, faster than as calculated from the densities alone. By employing two or more observations in conjunction with the relative viscosities of the gases used, an apparatus constant was determined which allowed! these effects to be eliminated. In the actual experiments this was the case, although the holes were sufficiently small to cause appreciable viscosity effects. When the holes are large compared with molecular dimensions the phenomenon is one of efflux on a small scale. The gases used were oxygen, hydrogen, nitrogen, carbon-mon-oxide, carbon-dioxide, cyanogen, argon and helium, and they effused through small holes pierced in platinum foil. The theory therefore indicates that argon should effuse faster than would be calculated from Graham's law. The constant derived from Parenty's work is 1.084. The formula derived from Hugoniot and Reynold's work gives the ratio of the times of effusion of two gases whose specific heat ratios are 1.408 and 1.67, equal to 1 06 times the square root of the ratio of the densities. This gives rise to formulæ which are different from the square root of the density law of Graham. This is followed by a theoretical investigation of the subject, upon the assumption that the ideal gas laws are obeyed, and that the back pressure never rises above a certain fraction of the internal pressure.

The introduction to this paper contains a short account of the work which has been done on the effusion of gases. Donnan read a paper on the relative rates of effusion of argon, helium, and other gases. Carey Foster, F.R.S., Vice-President, in the chair.-Dr. Callendar in the Physical Laboratory of University College, March 2.-Prof.

Physical Society.-Special meeting held by invitation of Prof.

Crompton potentiometer theory