Previous work on the existence and period of radium C′ is discussed with reference to an experiment of Jacobsen which provides evidence that a γ ray transformation of period comparable with that of radium C′ precedes the expulsion of α particles. It is shown that, from Jacobsen's results, part of the γ radiation from a source of recoil atoms should originate in the space surrounding the source.

A rough calculation is made which shows that the γ rays above the source should be detectable by ordinary methods, and a description is given of an ionisation method capable of detecting the effect. The γ rays predicted by Jacobsen's experiment were tested for by using specially prepared sources of radium C. Phenomena associated with α recoil were also investigated for sources of radium (B + C) and thorium (B + C).

No evidence of a γ ray emission from the space above any of the sources was obtained. The negative result indicates that the interval between the departure of the disintegration particle and the emission of the γ ray quantum is considerably less than 10−5 second.

The interaction between gas atoms and solid surfaces has been studied experimentally for many years. The first detailed study of the nature of this interaction was perhaps due to Knudsen, who in his classical researches introduced the idea of the accommodation coefficients for energy or momentum exchanges between gas atoms and a solid surface, when the gas atoms and the solid are at different temperatures or possess different mass motions. Knudsen and other investigators have given numerical values for these accommodation coefficients for various gases and solid surfaces, which seem to indicate that the accommodation coefficients are never small and are often of the order unity. This means that the gas molecules before reflection accommodate themselves almost completely to equilibrium with the temperature or motion of the wall. Before the recent work of Roberts referred to below, which has inspired this paper, no special precautions had been taken in the preparation of the wall surface, and as we now know the walls used by all previous investigators must have been completely covered with at least a mono-molecular film of gas. Thus the old observations of the accommodation coefficient do not determine it under precise conditions, and find in fact a value many times larger than that found by Roberts for the energy exchanges between the gas and a clean tungsten surface. For helium and metal with a dirty surface the value of the accommodation coefficient is some 6 times as large as the true value for the clean surface. The older values of the coefficients were so large that there were apparently grave difficulties in the way of any simple theory, but this is so no longer.