LONDON. Physical Society.—Special meeting held by invitation of Prof. Callendar in the Physical Laboratory of University College, March 2.—Prof. G. Carey Foster, F.R.S., Vice-President, in the chair.—Dr. F. G. Donnan read a paper on the relative rates of effusion of argon, helium, and other gases. The introduction to this paper contains a short account of the work which has been done on the effusion of gases. 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. This gives rise to formulæ which are different from the square root of the density law of Graham. 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. The constant derived from Parenty's work is 1.084. The theory therefore indicates that argon should effuse faster than would be calculated from Graham's law. 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. When the holes are large compared with molecular dimensions the phenomenon is one of efflux on a small scale. In the actual experiments this was the case, although the holes were sufficiently small to cause appreciable viscosity effects. 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. The observations showed that argon effused 3½ per cent, faster than as calculated from the densities alone. This agrees qualitatively with theory, and affords a confirmation of the high specific heat ratio of argon. 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. Carbon-dioxide when compared with oxygen appears to effuse 1 per cent, faster than as calculated from the densities. This result is not in accordance with the adiabatic theory of the efflux of ideal gases. The results for helium are not uniform, but show that its behaviour is unlike that of argon, a result not foreseen by the theory. If account be taken of the deviation of ordinary gases from the ideal laws, it is possible to obtain an expression for the efflux which contains a correction term involving the constant K of the Joule-Thomson effect. The sign of this correction term shows that a real gas will effuse more rapidly or more slowly than an ideal gas of equal density and specific heat ratio, according as K is positive or negative. The suggestion is made that possibly the anomalous results obtained with carbon dioxide and helium may be thus explained. The deviations of the observed results from the results calculated for an ideal gas are, in the case of carbon dioxide, in qualitative agreement with the theory proposed. In the case of helium, they would be so if that gas possesses a negative K.—Lord Rayleigh congratulated the author, and pointed out that in the case of very small apertures the gas laws might not be obeyed. The ratio of the dimensions of the aperture to the length of the mean free path determined this, and not the ratio of aperture to molecular dimensions.—Prof. Ramsay and Prof. Everett expressed their interest in the work.—Dr. Donnan thanked Lord Rayleigh for his correction, and stated that the apertures used were about 1/50 mm. in diameter.—Mr. E. C. C. Baly read a paper on the distillation of liquid air and the composition of the gaseous and liquid phases. From the experiments described in this paper, the author has drawn curves showing the relation between the composition of the gas evolved by boiling liquid air and the temperature, and between the composition of the liquid and the temperature, both at constant (atmospheric) pressure. These curves enable the temperature of boiling liquid air to be at once accurately determined by means of an analysis either of the liquid or of the gas evolved. The measurements of temperature were made with a Callendar compensated constant pressure hydrogen thermometer. The correction for the contraction of the glass bulb was determined by measuring directly the linear expansion of glass between - 190° C. and 20° C. This was found to be.0000073, Regnault's measurement between 0° and 10° being.0000085. The values for the boiling points of oxygen and nitrogen agree fairly well with those given by Olszewski and Estreicher. Boiling nitrogen has a great tendency to superheat. This can be obviated by passing a rapid current through the boiling liquid, or by dropping in pieces of copper. There does not appear to be any connection between the ratio of the vapour-pressures and the composition ot the gaseous phase in the distillation of oxygen and nitrogen at constant pressure. It is proposed to investigate the distillation at constant temperature. Prof. Ramsay drew attention to the uses of liquid air for carrying on researches at low temperatures. It is non-explosive, easy to work with, and is easily kept either by means of a vacuum jacket or by surrounding it with cotton-wool. Prof. Callendar referred to the question of superheating, and stated that the constant pressure thermometer was more accurate than the constant volume one for measuring low temperatures.—A paper on the reversibility of galvanic cells, by Mr. T. S. Moore, was read by Dr. Lehfeldt. In these experiments the reversibility of cells, such as the Daniell and the Clark, which are assumed to be reversible, was tested by allowing the cell to send a current, and by sending a current through the cell. The E.M.F.'s of the cells were determined by means of a Crompton potentiometer, and from the E.M.F.'s on open and closed circuits the internal resistances of the cells were calculated. Prof. S. P. Thompson asked if experiments had been made upon Leclanché cells where the products of the action escape. Dr. Lehfeldt said that experiments were not made upon these cells because they were known not to be reversible.—A paper on the damping of galvanometer needles, by Mr. M. Solomon, was postponed until the next meeting.