This still is a modification of Hulett's pattern in which mercury is distilled in a current of air at greatly reduced pressure. The current of air has two uses; first it prevents bumping and so is led in beneath the surface of the boiling mercury, and secondly, it oxidizes most of the zinc, cadmium and lead which go over. The oxides float on the distillate and are easily removed by straining it. A still made according to Hulett's directions and having a boiler of half a litre capacity will distil satisfactorily about 750 grams per hour. If the rate of distillation is increased much, some of the other metals go over, and if more air is admitted to prevent this, the boiling mercury is thrown about very violently in the boiler, which breaks sooner or later.

The source of light used as a background is an important factor in determining the convenience and accuracy of ultra-violet absorption work, etc. If a source of light of constant intensity is available, a direct comparison method can be used and it is only necessary to calibrate the plates. If the source of light is not constant in intensity, it is necessary to divide the light into two beams and use one to check the variations of intensity while the other goes through the absorbing substance or (during calibration) the reducing sector or wedge. This latter method requires much more complicated apparatus and if the variations in the source are at all large it becomes inaccurate. In addition to being constant in intensity a good background for ultra-violet absorption spectra should possess the following qualities:

(1) Most of the energy should be emitted in the form of a continuous spectrum.

(2) It is desirable to be able to use one photograph of the whole region to be investigated. For this purpose it is necessary that the variations of intensity in different parts of the spectrum should be small enough for it to be possible to arrange the exposure so that all parts of the spectrum are within the correct exposure range, i.e. it must not be necessary to over-expose any part in order to get a strong enough intensity at another wave-length.

The hydrogen continuous spectrum possesses both these qualities and is an excellent background for the region on the short wave-length side of 3200 A.U. It may be used for longer wavelengths, but the hydrogen secondary lines are apt to prove trouble-some unless a fairly large dispersion is used.

The assumption that the compressional waves of earthquakes follow the ordinary laws of refraction, the energy within any pencil of rays remaining permanently within that pencil, has been found to lead to too small amplitudes for the indirect waves from near earthquakes. In this paper a system consisting of two superposed compressible, but non-rigid, media is considered; the lower is supposed to transmit compressional waves with the greater velocity. It is found that ah explosion within the upper medium produces a disturbance at the upper surface involving the direct wave and all the reflected waves that might be expected; but in addition a wave is found that appears to have travelled along the interface with the velocity of sound in the lower medium. This indirect wave would have zero amplitude on the simple laws of refraction with plane boundaries. The variation of its amplitude with distance from the focus is in reasonable accordance with seismological observation, and its time of arrival agrees with that inferred from the laws of refraction for boundaries with slight curvature. But if the direct wave begins with a finite velocity, the indirect one will begin with a finite acceleration; in seismological language an iPg will be associated with an ePn. The indirect wave will also take longer than the direct one to give its maximum displacement.

The early experiments of Lenard(1), Leithauser(2), and Becker(3) on cathode rays led to the exponential law of absorption; those of Rutherford(4), Schmidt(5) and Hahn and Meitner(6) on β-rays indicated that the rays from some radioactive bodies were absorbed exponentially and that those from others were not. It was considered by Hahn and Meitner that the exponential law of absorption was an indication of the homogeneity of the rays. It was found also that the exponential law was not disturbed by the thickness of radiating material and this led to the conclusion that the rays are not appreciably reduced in velocity on their passage through matter. The theories of McClelland(7) and Schmidt(8) attributed the absorption to the stoppage of particles in mid-career.

In the course of an investigation of the β-ray spectrum of Ra (B + C) by the ionisation method, Chadwick(1) concluded that the line spectrum is superimposed on a strong continuous background. The ionisation method, however, is not one which admits of high resolving power, and doubts have been entertained whether the continuous spectrum has in fact any real existence. In this connection the case of Ra E is one of special importance, since no line spectrum from this body has been detected by the more sensitive photographic method. Experiments have therefore been undertaken with the object of determining the distribution with velocity of the numbers of particles in the β-ray spectrum of Ra E.