3IO 



Supplement to " Nature ^ August 25, 1923 



the surfu<c by the well-known method of dipping the 

 wire in u hot solution of radium (.'. In this case the 

 (lifTiculty due to recoil is absent, but the number of 

 singly charged particles was the same as Ijcforc. 



It is very significant that the relative number of singly 

 and doubly charged particles is about the equilibrium 

 ratio to be expected when the wire, after being activated, 

 is coated with an appreciable thickness of copper or other 

 material. We can scarcely suppose that singly as well 

 as doubly charged particles are actually liberated from 

 the radioactive nucleus itself, for even if it l)e supposed 

 that an u-particle with an attendant electron is ex- 

 pelled, the electron must be removed in escaping 

 through the very powerful electric field close to the 

 nucleus. It is much more probable that the doubly 

 charged a-particle in passing through the dense distri- 

 bution of electrons surrounding the radioactive nucleus 

 occasionally captures an electron, and that the process 

 of capture and loss goes on to some extent in escaping 

 from the radioactive atom. This seems at first sight 

 rather unlikely when we consider the relatively large 

 number of atoms an a-particle ordinarily passes through 

 before equilibrium between capture and loss is estab- 

 lished, but it is well known that the chance of effective 

 electronic collisions appears in general to be greater for 

 a charged particle expelled from the central nucleus 

 than for a similar particle passing from outside through 

 the electronic distribution of an atom. It may be 

 that those electrons, the orbital motion of which round 

 the nucleus is comparable with the speed of the^ a- 

 particle, are particularly effective in causing capture or 

 loss. 



So far we have dealt mainly with the distribution in 

 a ipagnetic field of the particles in a vacuum after their 

 escape from a mica surface. Some very interesting 

 points arise when the distribution is examined in the 

 presence of sufficient gas to cause a rapid interchange 

 of capture and loss along the path of the a-particle in 

 the gas. This is best illustrated by a diagram, Fig. 4, 

 in which the results are given for a-particles escaping 

 through mica with a maximum emergent range of 

 about 4 or 5 millimetres in air. Curves A and B give 

 approximately to scale the distribution of He+ and 

 He++ particles in a vacuum, while C gives the 

 relative number of neutral particles under the experi- 

 mental conditions. Suppose now sufficient air is intro- 

 duced into the vessel to cause many captures along the 

 gas but yet not enough to reduce seriously the velocity 

 of the tt-particles. The first salient fact to notice is 

 that the distributions A, B, C vanish and there remains 

 a distribution of particles (curve D) about midway 

 between A and B. This band is narrower than either 

 A or C, and its height at the maximum much greater 

 than either. It is evident that the particles have been 

 compressed into a band of much narrower width than 

 the normal distribution in curve B. 



This is exactly what we should expect to happen. 

 The swifter particles present suffer less capture than the 

 slow ; consequently the average charge of the swifter 

 a-particles along the gas is less than 2e, and their de- 

 flexion is less than the swiftest particles shown in 

 curve B. On the other hand, the slower a-particles 

 have an average charge nearer \e than 2e and are 

 relatively still less deflected than the swifter particles. 

 It is thus clear that the resulting distribution of par- 



ticles with air inside the ve .\(T 



a much narrower width tb. *t • 



particles. From calculation iKiscd on the laws of 

 capture and loss, the width of the liand under the ex- 

 perimental conditions can l>e deduced and is found to 

 be in good accord with experiment. It will be seen to 

 be significant that similar results have been oljserved 

 for iiydroL'cn under cf^r- ■ • • '• ' •■ -- . 



General Discussion of Kesults. 



Attention may now be devoted to a consideration of 

 the results so far obtained and the possibility of their 

 explanation on present views. In tlie first place, it is 

 important to emphasise the large number of capture 

 and losses that occur during the flight of an a-jjarticle 

 from radium C. While the mean free path of the a- 

 particle from radium C of 7 cm. range is about 3 mm. 

 in air, its value rapidly decreases with lowering of the 

 velocity of the a-particle and is probably about 0-0015 

 mm. for a velocity of 0-3 V^. It is not difficult to 

 calculate that not far short of a thousand interchanges 

 of charge occur during the path in air of a single particle 

 between velocities V,, and 0-3 Vq. While the data so far 

 obtained do not allow us to calculate the num1)er of 

 interchanges of charge that occur between velocities 

 0-3 Vq and o, it seems probable that the number is con- 

 siderably greater than a thousand. We have already 

 pointed out that for low velocities the interchange 

 He+^HCo predominates. When we consider the 

 rapidity of interchange of charges of the a-particle at 

 average velocities, it seems clear that we cannot exf)ect 

 to observe any appreciable difference in power of pene- 

 tration between a beam of rays of the same velocity, 

 whether consisting initially of singly or doubly charged 

 particles. It is clear that a singly charged particle 

 after penetrating a short distance is converted into a 

 doubly charged particle and vice versa, and that the 

 effects due to the two beams should be indistinguish- 

 able. Henderson tried such absorption experiments, 

 using the photographic method, but witli indc finite 

 results. 



When an a-particle captures an electro... i..^ ...ncr 

 presumably falls into the same orbit round the helium 

 nucleus as that which characterises an ionised helium 

 atom, i.e. an atom which has lost one electron. When 

 the a-particle with its attendant electron passes swiftly 

 through the atoms of the gas in its path, it will not only 

 ionise the gas but will also occasionally be itself ionised, 

 i.e. will lose its attendant electron. When we take 

 into account the strong binding of the first electron to 

 the helium nucleus — ionisation potential about 54 volts 

 — the mean free path for loss of the captured electrons in 

 air is of the right order of magnitude to be expected 

 from considerations based on the ionisation by the a- 

 particle per unit path in air. ^^^hile we can thus offer 

 a quantitative explanation of the mean free path for 

 loss observed experimentally, the inverse problem of 

 the capture of an electron by the flying a-particle pre- 

 sents very great diflliculties. 



In the actual case, the a-particle is shot at high speed 

 through gas molecules which for all practical purposes 

 may be supposed to be at rest. For convenience of dis- 

 cussion, however, it is preferable to make an equivalent 

 assumption, namely, that the a-particle is at rest and 



