LONDON. Royal Society, November 2.—Sir Charles Sherrington, president, in the chair.—Lord Rayleigh: Polarisation of the light scattered by mercury vapour near the resonance periodicity. White light scattered at right angles by dense mercury vapour is to a first approximation completely polarised. Ultra-violet radiation of the mercury spectrum line λ2536, when examined immediately it enters mercury vapour in an exhausted vessel at room temperature, gives a scattered radiation which is slightly though definitely polarised. This polarisation has been observed to increase as the beam is filtered by penetration of a considerable depth of vapour. After penetration of 27.5 cm. of vapour the weaker polarised image had 6o per cent, only of the intensity of the stronger one, instead of 90 per cent. as at first. The radiation removed by the filtration appears to lie within a spectral range of about 1/100 Ångstrom.—G. P. Thomson: The scattering of hydrogen positive rays and the existence of a powerful field of force in the hydrogen molecule. At a pressure of less than 1/100 mm., hydrogen positive rays of 10,000 volts mean energy suffer considerable small-angle scattering in a distance of cm. This scattering is 10.20 times greater than would be expected on theoretical grounds. There must, therefore, be a field of force in the hydrogen molecule at distances of the order of io from a nucleus which is much stronger than would be expected from the inverse square law. A subsidiary experiment throws great doubt on Glimme and Koenigsberger's “Stossstrahlen.”-H. D. Smyth: A new method for studying ionizing potentials. Positive ray analysis is used to study the ions produced in a gas or vapour by the impact of slow- speed electrons of known energy. This requires that the density of gas be considerable where the energy of the impacting electrons is known, and as small as possible where the energy conditions are not known. In the case of mercury such a localisation of vapour density was obtained by using a unidirectional molecular stream similar to that employed in a mercury diffusion pump. Ions were produced by electrons from a hot filament, and after acceleration by a large electric field were analysed by a magnetic field. In this way the values of me were determined approximately. The experiments on mercury indicate the formation of doubly charged ions at 19+2 volts.