in minerals containing uranium and thorium assumed a totally new interpretation, borne out by the spectroscopic proof of the production of helium from radium by Sir William Ramsay and myself, and later from actinium, polonium, and even from uranium and thorium, all at the rates to be expected from radioactive data. The identification of the a-particle with helium, after the weight of the a-particle had been shown by new physical methods to be four times that of the hydrogen atom, was accomplished by enclosing the radium emanation. in a glass tube thin-walled enough to allow the a-particle to go through, but perfectly impervious to the passage of gas. In these circumstances, helium in spectroscopically detectable quantity was proved by Rutherford to make its appearance outside the tube.

Such confirmations by the spectroscope, welcome and gratifying as they are, are nevertheless in a sense subsidiary to the main problem, namely, the task of unravelling the complicated series of changes into its individual steps, and the characterisation by their radioactivity of the several intermediate members of the series, such as by the determination of their periods and the physical constants of the radiation a-, B-, or y-, to which they give rise. The determination of their chemical character, although equally important, was only later fully accomplished.

THE RADIATIONS.

In the successive radioactive changes, a- or B-particles are expelled, one a-particle per atom disintegrating for each change, although for the B-particles our knowledge is less exact. In some cases, certainly, although these are exceptional, B-particles seem to be expelled along with a-particles.

NATURE OF THE RADIATIONS

121

The a-particle is an atom of helium charged with two atomic charges of positive electricity, or, as we should now say, is the helium nucleus, deprived of the two electrons which are combined with it in the helium atom. The B-particle is the negative electron, and when expelled with sufficiently high velocity is accompanied with y-rays. The latter are X-rays of exceedingly short wave-length, varying from 1.3 to 0.1 Ångström units.1 A connection exists between the speed of the change and the speed of the particles expelled, and the more rapid the change the faster in general and the more penetrating are the attendant a- or B-particles. In the case of the a-particle, an empirical logarithmic relation, known as the Geiger-Nuttall relation, enables us to calculate approximately the period of the changing element from the velocity or range of the a-particle, and vice versa, and by this means periods too long or too short to be directly measurable have been estimated. In the case of the B-rays, no definite quantitative law has yet been made out, but it is clear that a similar relationship must exist. One of the important corollaries is that changes much slower than the slowest known, namely, those of uranium and thorium, would probably not be detectable, as, even were a- or B-particles expelled, they would be of too low velocity probably to ionise gases or show fluorescent or photographic actions. Indeed, for mesothorium-I and actinium this appears to be the case. No detectable radiation is expelled, although the products conform to what would occur

1 The shortest wave-length so far resolved by the crystal reflection method is 0.072 Å. in the spectrum of the y-rays of radium-C. Ishino and Rutherford have recently concluded, however, that the main y-radiation of radium-C must have a wavelength lying between 0.02 and 0.007 Å. (Phil. Mag., 1917, [vi.] 33, 129; 34, 153).

in B-ray changes. The period of both substances is long, and it is probable that the B-particle is expelled, but is undetectable by ionisation methods. For the slowest B-ray change, that of radium-D, with a period of twenty-four years, the B-radiation is of such low velocity as to be only capable of detection by special care, and is far less penetrating than average a-rays. These facts serve to show that changes may be going on in the non-radioactive elements which at present are beyond experimental means of detection.

PERIOD OF AVERAGE LIFE.

The law of radioactive change, which is the same for all cases, is that of unimolecular reaction, the rate of change, or quantity changing in unit of time, being a fraction, designated by λ and known as the radioactive constant, of the amount present. The value of λ, although vastly different for different radio-elements, is an absolute constant, so far as is known, for any one element, independent of every consideration whatever. The period of average life is the reciprocal of this constant, but the actual life of any one atom may assume any value. This is an experimental fact very difficult to account for. For example, it is quite easy to compare the value of for a collection of atoms (1) only just produced and not in existence a short interval before, and (2) that have remained undisintegrated from an originally very much greater number, and each of which has been in existence many times the period of average life. In both cases the value of λ is the same. This fact excludes from consideration as a conceivable cause of disintegration any gradual progressive alteration in the atom during its period of existence, as, for example, was at one time suggested, a gradual

DUAL DISINTEGRATIONS

123

radiation of internal energy by the electrons in their orbits within the atom. So far, we must admit, the cause of atomic disintegration remains unknown, although Lindemann (Phil. Mag., 1915, [vi.], 30, 560) has attempted, with some success, to frame a theory to account for it.

BRANCH SERIES.

The development of the various radioactive sequences revealed that sometimes the series branches, and that in the change of one radioelement sometimes two products result, in general, in different amounts. Thus the uranium series at one point branches into the radium and actinium series, in proportion 92 to 8 out of 100 atoms disintegrating. Again, in the case of radium-C and thorium-C a similar branching occurs, and here in one branch an a-ray change is followed by a B-ray change, and in the other branch the sequence is reversed. These cases are sufficiently explained if it be supposed that two simple radioactive changes are in progress in the same substance simultaneously, and that each obeys the law of simple change as though the other did not occur. The distribution of the original substance into the two products is then proportional to the relative rates of the two changes. If 1 and λ are the radioactive constants of the two changes, the proportion between the two products is as λ to λ, and the constant of the double change as a whole, +2. For thorium-C, the ratio is as 65 to 35, but for radium-C 99.97 to 0-03. The first is relatively easy, but the second extremely difficult to follow experimentally. It is, for example, impossible to follow further what occurs to the minor branch owing to the minuteness of the quantity of material, and although this has to be represented as

not further changing, we have only negative evidence to go on. This branching is very important as showing how from one element two products or more in very different quantity may result, and may be the explanation of the excessive rarity of certain of the elements in nature.

HISTORY OF THE ANALYSIS OF MATTER.

The second, and in many respects even more revolutionary phase in the development of the study of radioactive change arose out of the chemical characterisation of the successive products, but some historical comment on the various influences which have gone to shape the current conception of the chemical element may be of interest before dealing with this development.

The analysis of matter into different chemical elements was at first concerned with known materials obtainable in abundance. The question, then, was not as to the existence or otherwise of certain elements, but whether certain thoroughly well-known substances were elements or compounds. Boyle's original celebrated definition was a purely practical one. That was to be regarded as elementary which could not by any means be separated into different substances. Almost at once, however, there crept into the interpretation of this conception two fallacies, or two aspects of the same fallacy, implicit in all the later characterisations of the elements, right up to the present time, namely, first, that chemical analysis was necessarily the most fundamental and searching kind of material analysis, known or to be discovered, and, secondly, that chemical compounds were necessarily more difficult to resolve than simple mixtures. Any means soon came to mean any chemical means, and the element, in consequence,