when its energy had almost run out. But the curve shows that it is just the other way. The lagging ray, about to resign its ionising properties, becomes a much more efficient ioniser than it was at first. The maximum efficiency is, however, in the case of a bundle of parallel rays, not quite at the end of the range, but about half a centimetre from it. The increase to the maximum is rapid, the fall from the maximum to nothing is much more rapid. It can be shown that the ionisation effected anywhere along the path of the ray is inversely proportional to the velocity of the ray at that point. But this evidently does not apply to the last 5 or 10 mms. of the range where the rate of ionisation and of the speed of the ray change most rapidly. To what are the changing properties of the rays near the end of their path to be ascribed? It is only recently that this matter has been elucidated. When the alpha ray has sufficiently slowed down, its power of passing right through atoms, without appreciably experiencing any effects from them, diminishes. The opposing atoms begin to exert an influence on the path of the ray, deflecting it a little. The heavier atoms will deflect it most. This effect has been very successfully investigated by Geiger. It is known as "scattering." The angle of scattering increases rapidly with the decrease of velocity. Now the effect of the scattering will be to cause some of the rays to complete their ranges 217 or, more accurately, to leave their direct line of advance a little sooner than others. In the beautiful experiments of C. T. R. Wilson we are enabled to obtain ocular demonstration of the scattering. The photograph (Fig. 14.), which I owe to the kindness of Mr. Wilson, shows the deflection of the ray towards the end of its path. In {Fig. 14} this case the path of the ray has been rendered visible by the condensation of water particles under the influence of the ionisation; the atmosphere in which the ray travels being in a state of supersaturation with water vapour at the instant of the passage of the ray. It is evident that if we were observing the ionisation along a sheaf of parallel rays, all starting with equal velocity, 218 the effect of the bending of some of the rays near the end of their range must be to cause a decrease in the aggregate ionisation near the very end of the ultimate range. For, in fact, some of the rays complete their work of ionising at points in the gas before the end is reached. This is the cause, or at least an important contributory cause, of the decline in the ionisation near the end of the range, when the effects of a bundle of rays are being observed. The explanation does not suggest that the ionising power of any one ray is actually diminished before it finally ceases to be an alpha ray.