XVII. The [gamma]-Rays of Thorium and Actinium. By Alexander S. Russell, M.A., B.Sc, and Frederick Soddy, ALA., F.R.S. [Communicated by the Authors.] SIMILAR investigations to those described in two previous papers [citation redacted] on the [gamma]-rays of uranium X and radium C have been carried out with the [gamma]-rays of actinium C (or possibly it may prove to be actinium D) and with the two types of powerful [gamma]-radiation in the thorium series, namely that given [header] 131 by the short-lived [beta]-ray product between mesothorium 1 and radiothorium, which we shall refer to as mesothorium 2 [citation redacted], and that given by the last known product of the disintegration series, thorium D [citation redacted]. The paper conveniently divides itself into three sections. The first deals with the relative intensity of the [gamma]- and [beta]-rays of these substances. In the second section a number of so far unexplained effects in the measurement of the absorption coefficients of the [gamma]-rays are described in detail. In the last section the penetrating power of the actinium and thorium types of [gamma]-rays are compared with that of radium C. A brief indication of the general character of the results may conveniently precede their detailed consideration. The two thorium products resemble radium C remarkably closely both in their [gamma]/[beta] ratio, and in the penetrating power of their [gamma]-rays, and, although interesting differences exist, these are comparatively small. The most penetrating [gamma]-ray known is that given by thorium D, that of mesothorium 2, speaking in a general sense, being about as much less penetrating than that of radium C as that of thorium D is more penetrating [Apparently nothing has been previously published with reference to the [gamma]-rays of mesothorium 2 ; but it should be mentioned that Eve [citation redacted], in comparing- the [gamma]-rays of a preparation of radiothorium (thorium D) with those of radium, found that they were almost identical in penetrating- power, although the measurements indicated that the radiothorium [gamma]-rays were a little the more penetrating - . The difference, however, Eve considered to be within the error of the experiments.]. These three bodies are sharply distinguished from all the other [beta]- and [gamma]-ray products by their high and similar [gamma]/[beta] ratio. At the other extreme are radium B and radium E, which we have not examined, the [gamma]-rays from which are, either not at all, or only barely detectable [citation redacted]. It is perhaps still natural to leave radium B out of the comparison, as its [beta]-rays are excessively feebly penetrating; but the established very low value of the [gamma]-rays of radium E, taken in conjunction with the character of its [beta]-rays, which, although feebly penetrating, are of the same order as those of actinium C and mesothorium 2, furnishes another example of the lack of connexion between the two types of rays. We have to consider actinium C and uranium X, each of which differs in this respect from all the other types. For uranium X, as we have previously found, both [beta]- and [footer] 132 [header] [gamma]-rays are similar in penetrating power to those of thorium and radium C, but the [gamma]/[beta] ratio is of an entirely different and lower order of magnitude. For actinium C, both [beta]- and [gamma]-rays are less penetrating, and the [gamma]/[beta] ratio, although of the same order, is distinctly greater than for uranium X. Indeed there is no rule about the matter. From the fact that hard [gamma]-rays usually accompany hard [beta]-rays, it might be supposed that thorium D, which gives the most penetrating [gamma]-ray known, would have also the highest [gamma]/[beta] ratio. Whereas it is distinctly lower, both than that of radium C and that of mesothorium 2. Owing to the importance of the [gamma]-ray method as a standard means of comparison of radioactive substances, we have thought it advisable in the second section to collect a number of observations showing how the most curious changes of penetrating power of the [gamma]-rays are brought about by the slightest change in the experimental disposition, although we have no general explanation of these effects to offer. Section I. — The Ratio of the [gamma]- to the [beta]-rays of Mesothorium 2, Thorium D, Actinium 0, and Radium C. In a previous paper (I. p. 629) the ratio of the [gamma]- to the [beta]-rays of uranium X has been accurately compared under the same conditions with the same ratio for radium C. It was found, assuming the rays to be homogeneous, by measuring the [gamma]-rays through 1 cm. of lead and calculating the initial intensities from the known absorption coefficients, that the [gamma]/[beta] ratio for uranium X was about 50 times smaller than that for radium 0. When the [gamma]-rays were measured through CH) cm. of aluminium, the ratio, uncorrected for absorption, was 18 times less. Similar measurements have now been extended to actinium and thorium. It may be pointed out that we are measuring a highly complicated effect in this [gamma]/[beta] ratio, and that before and T absolute comparison is possible, it would be necessary to have a good deal more data than are at present available. We have been concerned only with the relative order of magnitude of the ratio sought, and have attempted to get at least a rough idea of this order by carrying out the [gamma]- and [beta]-ray measurements respectively for both substances under exactly the same conditions. Considerable differences will be found in the ratio according to the method of measurement employed, as is only to be expected; but the general order of the effects measured is sufficiently correctly indicated by the measurements here given. [header] 133 For measurement of [beta]-rays a cylindrical brass electroscope of the ordinary type, 13 cm. high and 10'8 cm. diameter, was used. The thickness of the walls was 0"32 cm. The base consisted of a layer of aluminium foil, 0095 mm. thick. In the first series of measurements the preparation was placed at a distance of 60 cm. below the base, but later various shorter distances were also used. The apparatus was set np so as to reduce secondary [beta]-radiation to a minimum. The electroscope was flush with the table, which had a large hole cut in it, and was supported by light steel brackets from the wall. The preparation was supported centrally below the electroscope on a light framework of brass rods, made in sections screwed together, and attached to the under side of the table, so that any distance between the preparation and base of the electroscope could be employed. For measurements of [gamma]-rays three different dispositions were used, as follows ; — ([gamma]1) The active preparation was placed at a distance of 8*6 cm. below an electroscope of lead of the usual type having wall and base thickness 0*3 cm. of lead ; ([gamma]2) at a distance of 3'3 cm. below a lead electroscope whose wall was 0*65 cm. and whose base 0*U75 cm. thick. ([gamma]3) At a distance of 8*7 cm. below the same electroscope as in [gamma]2 . The radium C was prepared by exposing one side of a negatively charged brass disk to a quantity of radium emanation for 20 hours. After the exposure the disk was placed in a brass cell and a piece of 0'095 mm. thick aluminium foil was cemented over the cell to prevent any possible escape of adhering emanation. One hour after the preparation of the film of radium C, alternative measurements of the [beta]- and [gamma]-rays, in the four different dispositions detailed above, were taken over a period of three hours, and from the four decay curves obtained from these measurements the [beta]- and the three different [gamma]-activities could be readily deduced. The four decay curves obtained were found to be very nearly exponential with values of [lambda] very approximately the same, [formula redacted]. The relative values therefore of the intensity of the preparations measured under the four different dispositions could be accurately obtained. For thorium D the procedure of preparation of the disk was exactly the same, the disk being exposed negatively charged to the thorium emanation from a preparation of radiothorium spread out on a shallow dish for 20 hours under conditions, such that only the one side of the disk was coated. The disk was then mounted and covered with the same 131 [header] thickness of aluminium as the radium C disk. Fifteen minutes alter the disk had been removed from the emanation the first measurements were taken. Experiments in each of the four different dispositions were made with thorium D, just as they had been made with radium C, and the relative intensities of the rays at each disposition accurately determined. The mesothorium 2 used in the experiments was separated chemically from a preparation of mesothorium 1 previously prepared from several kilograms of thorianite. The precipitate, which weighed only a few milligrams, was evaporated down on a small watch-glass of diameter exactly the same as that of the disks used for the other two active bodies. The quantity of matter was so small that self-absorption can be neglected. The watch-glass was covered with a piece of aluminium foil, 0*095 mm. thick, and several measurements were made for each disposition over a period of 3 1/2 hours. The four decay curves were exponential [formula redacted]. The residual activity remaining two days later was negligible. In the following table the actual results found for each disposition for each source are given : — [table redacted] Or, taking the radium values as the standard :- [table redacted] For these dispositions therefore mesothorium 2 gave 13 per cent, more [gamma]-rays per [beta]-ray than radium C, and thorium D about 25 per cent. less. Also, the variation of the ratios thorium D to radium C, or mesothorium 2 to radium C, with the thickness of the lead base is slight, if any. This is to be expected because, as will be shown in [header] 135 another part of the paper, the penetrating powers of all three types of [gamma]-radiation are very similar. A second series of experiments, which embraced in addition the actinium active deposit, was started in order to investigate various influences which are likely to affect the values of the [gamma]/[beta] ratio. These are the difference in penetrability of the hard [beta] rays of each element and the effect of scattering of such rays by the air between the preparation and the electroscope. Some metal must necessarily intervene between the preparation and the electroscope, but the amount used was the minimum, so that the absorption of [beta] rays by it was small. When the ratio of [gamma]- and [beta] rays for uranium X was measured in a former experiment it had been necessary to place the preparation at a great distance from the electroscope in order not to give too large a [beta] ray effect. This was not essential in the case of active deposits having a quick rate of decay, so that the values of the [beta] activity of the preparations, placed at five different distances from the electroscope, has been measured in order to see how the distance affects the values of the [gamma]/[beta] ratio of thorium D, mesothorium 2 and actinium active deposit relatively to that of radium C. The distances used were 51*1 cm., 35*9 cm., 20*6 cm , 13"0 cm., and 5'37 cm. These positions are denoted as [beta]1, [beta]2, [beta]3, [beta]4, and [beta]5 respectively. The disposition [gamma]1, used before, was used again. The mesothorium 2, separated chemically as before, was this time mounted on a brass disk, exactly the same as those on which the active deposits of the other bodies were collected, placed in a brass cell and covered with the usual aluminium foil. The substitution of the brass for the glass had only a small effect on the value of the [beta] rays. The active deposits were prepared and mounted as before, the actinium active deposit being obtained in the same manner as that of thorium. The very fine actinium preparation, lent by Messrs. Buchler & Co., and described in Section III., was employed. Measurements were started in the [gamma]1 and [beta] positions, and, when the preparation had become too weak to be easily measurable in the latter, it was placed in the ft position, and so on till the ft position. The residual activity was also measured in the [beta] position ; but since the [beta] rays of radium C and mesothorium 2 reach a minimum and then again increase, owing to the formation of later [beta] ray products, the correction for the residual activity cannot be directly applied, but must be extrapolated. From the decay curves of all the bodies experimented with in the five different dispositions, corrected, where necessary, for residual activity, the [gamma], [beta] ratios in the 136 [header] different positions are readily obtainable. The final results are given in the following table, in terms of the ratio for radium C which is taken in each position as unity. [table redacted] For actinium the [gamma]-rays were m ensured also through 0*95 cm. zinc (74) as well as through 0*3 cm. lead (7,). This was done because, as was shown previously by Godlewski and confirmed later in this paper, the [gamma]-rays of actinium are abnormally highly absorbed by lead. Radium C, the ratio for which is again taken throughout as unity, was also measured under the same conditions for comparison. The final results are: — [table redacted] It will be seen that the general effect of decreasing the distance of the preparation in the [beta]-ray measurements is to decrease the [gamma]/[beta] ratio, to about one half, over the range of distance examined, for mesothorinm 2 and thorium D ; but for actinium C the difference is less marked. ]So doubt the causes of this are very complex. In the first place it is to be expected that the scattering of the [beta]-rays by the air between the preparation and the electroscope will be the greater the less penetrating the [beta]-rays, which, in descending order of penetrating power, are-— radium C, thorium D, mesothorium 2, actinium C. Then, with diminishing distance and greater angle of the cone of rays entering the electroscope, the effect of u reflexion " of the rays from the inner sides [header] 137 of the electroscope will come in. According to the experiments of Kovarik and Wilson [citation redacted], this reflexion increases with the penetrating power of the rays tip to a maximum, at X (cm.) -1 aluminium = 20, and then again diminishes. This effect for most of the rays therefore would oppose that of scattering. An attempt to evaluate the absorption in the thin aluminium foil covering the preparation led to the unexpected result that the ionization was slightly greater with the foil than without. This was at the time ascribed to a possible generation in the foil of secondary [beta]-rays, an effect which had frequently been looked for previously but never actually obtained. The same effect is recorded by Kovarik [ciation redacted], who made a closer examination of it than we have done, and ascribed it to an effect of scattering. None of the effects considered affect the order of the [gamma]/[beta] ratio to a serious extent, and they are relatively unimportant. Probably from the theoretical point of view the original ratios obtained with the preparation at a great distance from the electroscope are to be preferred. For practical purposes the [beta] 4 position may be selected, as here the distance, 13 cm., is a very usual one. The ratios, relatively to that of radium C, are for these positions : — [table redacted] The [gamma]/[beta] ratios are thus of the same order for radium C. mesothorium 2, and thorium D; while for actinium C it is only about one-eighth to one-sixteenth of radium C and therefore more nearly approaches the order previously found for uranium X. Experiments with Thorium Minerals. — Before any mesothorium or radiothorium preparations were ready, Mrs. Soddy had carried out a comparison of the [gamma]/[beta] ratio for Ceylon thorite (containing about 66 per. cent, of thoria) and Joachimsthnl pitchblende (containing about 40 per cent, of uranium oxide). The former mineral contains practically no uranium, and the latter no thorium, so that they represent respectively the thorium and uranium disintegration series in equilibrium. The result may be briefly referred to here. The [gamma]/[beta] ratio of these two minerals was practically the same 138 [header] Bearing in mind the fact that the uranium mineral contains two products, uranium X and radium E, which contribute [beta]-rays with little or no [gamma]-radiation, this result is in agreement with what is to be expected from the determinations just described of the ratio for the three single products, radium C, mesothorium 2, and thorium D. The thickness of the lead through which the [gamma]-rays were measured also made no practical difference in the ratio. Eve has already shown that the [gamma]-rays from thorium nitrate and uraninite have practically the same penetrating power. The similarity of penetrating power of the [gamma]-rays from thorium minerals is to be expected from the result, detailed in Section III., that the radium C [gamma]-rays are intermediate in penetrating power between the two types of [gamma]-rays given by the thorium series. There remains one practically important question with reference to the [gamma]-rays of the thorium series. What is the relative intensity of the radiation contributed by the two products mesothorium 2 and thorium D? Mr. Alexander Fleck has done some experiments with Ceylon thorite with a view to obtaining information on this point, and although, so far, only preliminary results are available, these maybe given here. The thorite was dissolved in hydrochloric acid, precipitated with excess of ammonia, filtered, and the precipitated hydrates subjected to the same treatment five times, in all, without unnecessary lapse of time. The filtrates were collected, evaporated and ignited together, and sealed up in a box with a thin aluminium lid for [beta]-ray measurement. The filtrate from a sixth precipitation, carried out immediately after the fifth, treated separately in the same way, was inactive, showing that by this treatment the whole of the mesothorium and thorium X had been separated. The [beta]-ray decay curve of the preparation was taken for some weeks. The intensity of the rays rose to a maximum after 2 days (due to thorium 1)) and then fell, until a constant minimum, due to mesothorium only, was attained. The decay curve for the maximum onward was extrapolated back to an origin corresponding to the time of the last precipitation with ammonia, and from the initial and final value of the intensity of the [beta]-rays, the proportion due to each product could be obtained. The measurements indicate that the thorium D contributes distinctly more [beta]-radiation than the mesothorium 2. The difference is not great, and may be estimated provisionally as from 25 to 50 per cent. Since thorium D is richer in [beta]-rays, relatively to the [gamma]-rays, than mesothorium 2, it follows that the [gamma]-radiation from the two types must be very similar in intensity. This result, although only approximate, will prove [header] 139 useful in calculating the rise in intensity of the [gamma]-radiation of a mesothorium preparation with time, due to the generation of radiothorium and thorium D, provided that the fraction of the radiation due to radium is known. Section II. — Variations in the Values of the Absorption Coefficients of [gamma]-rays [This section may be regarded as a continuation of Section III. of the previous paper (II, p. 744).]. Effect of distance of the preparation from the electroscope. — In all this work a lead electroscope with detachable base, similar to that previously described (II. p. 752, fig. 17), has been used. It has been shown to exclude external secondary radiations effectually. Moreover it gives of all metals easily the maximum ionization for a given intensity of [beta]-radiation (I. p. 642), which Bragg has since explained on the view that the [beta]-ray, produced by the [gamma]-ray, possesses really a greater penetrating power in lead than in equal weights of other metals, though its trajectory is more entangled. As a result of experiments on the effect of the distance of the preparation from the electroscope, it was found that at a distance above 14 cm. the values of A, became constant, the beam being now practically parallel (II. p. 736). In most experiments the distance employed was therefore 14 cm. The following table gives the value of A at various distances for radium [gamma]-rays, a thick lead base being used with the electroscope, the absorbing lead being laid directly on the preparation. [table redacted] Similarly, for zinc, the value of A, at 13 cm. was 0278, and at 115 cm. 0*274, which agree within the experimental error. The u Hardening " of [gamma]-rays by passage through Lead. — Some further experiments were done on this effect. By passage through lead the value of A for lighter metals is reduced, though other bodies possess the same power as lead in lesser degree. In no case have we observed any " softening." Hardening is more pronounced when the rays pass first through the lead and then through the lighter metal, than vice versa. Thus, with a brass base to the electroscope. 140 [header] the values of [lambda] for zinc were (1) 0*23, (2) 0*21 when 1*24 cm. of lead were placed (1) between the zinc and the electroscope, (2) between the zinc and the radium. Similarly the values for iron were (1) 0'304, (2) 0'280, 1 cm, thickness of lead being used. These results indicate at once that the values found for X are to some extent influenced by the nature and thickness of the base employed. Thus the results given, for example, in Table II. (I. p. 644), in which the mean value of [lambda]/d for Class II. bodies ([gamma]-rays of radium) was 0*040, refer to a lead base (0*975 cm.). The mean values for other bases were as follows : — - [table redacted] An experiment was conducted with zinc to see to what degree the process of hardening could be pushed. The disposition used in this case was different to that of the last, the radium being placed 25 cm, below the usual lead electroscope and the absorbing zinc clamped up to form the base. Seven thicknesses of lead were placed in turn over the preparation and the absorption by the zinc for each thickness of lead measured. The results are given in the following table[* Incidentally it may be noticed that the value of [lambda] for the base preparation (0-325) is 7 per cent, greater than the value found in the standard series of measurements given later (Table B, p. 148). The cause of this difference is not yet clear, but may be due to the fact that in this experiment the preparation rested in a groove in a lead disk, which somewhat confined the beam of [gamma]-rays.]. Thickness of Lead . . . [table redacted] Thus the final value is 0'79 of the initial, and there does not seem to be a limit to the degree to which the hardening process may be carried. The remarkable point is that the curves are in every case, as nearly as can be seen, exponential with the new value of [lambda], after from 1 to 035 cm. equivalent thickness of zinc has been penetrated, according to the amount of lead covering the preparation. For lead the value of [lambda] is, as we have shown, independent of thickness up to 22 cm. (II. p. 752). Provided there is a base of lead 1 cm. thick [header] 141 and 0'5 cm. of lead over the preparation, the absorbing plates of lead may be placed anywhere between the preparation and the base, the value of [lambda] is constant at 0'50 (cm.) -1 . These two facts are extraordinarily difficult to reconcile. The first indicates that each successive thickness of lead traversed modifies the nature of the radiation remaining, continuously without limit, while the second indicates that the effect of each successive thickness is to absorb a definite fraction of the radiation without modification of the nature of the remainder. This difficulty suggested the possibility that a heterogeneous beam, consisting of two types of about equal energy but different penetrating power, might be transformed by a layer of the absorbent so as to act subsequently as a homogeneous beam. A heterogeneous beam of [gamma]-rays was made by placing two point sources, one of radium and the other of mesothorium, side by side. The activity of the mesothorium was 40 per cent, of the total, measured through 1 cm. of lead. Lead laid over the preparations was used as absorbent. X for lead for radium for this disposition is O50, and for mesothorium 0*62. The differences are thus of the same order as that produced in the value of [lambda] (zinc) by hardening the [gamma]-rays of radium with lead. The absorption curve for the heterogeneous beam, over the 5 cm. of lead investigated, was not exponential and by no allowance for experimental errors could it be made so, the curve being plainly convex to the origin [This experiment suggests a possible means of detecting- the adulteration of radium by mesothorium without opening the tube.]. The theoretical values of the ionization of the beam for the different thicknesses of lead were calculated from the data known, and they agreed perfectly with the observed values. The only way of escape from the difficulty seems to be to deny that an exponential absorption curve necessarily means the stopping by the metal of a definite proportion of the incident radiation without modification of the fraction emerging. A similar state of things exists with regard to the [beta]-rays. It has been fairly conclusively shown that an exponential absorption occurs with a heterogeneous beam, and that the part unabsorbed suffers reduction in velocity and penetrating power. But a true exponential curve for the [beta]-rays has never been obtained by anyone over any considerable range, whereas with the [gamma]-rays under proper conditions the absorption curve never departs measurably from the exponential form. Initial Part of the Absorption Curves. — The deviations 142 [header] from the exponential curve over the first part of the range (0 to 1 cm. thickness of lead or its equivalent) now call for remark. If we place a source of radium at a distance of 13 cm. below a lead electroscope and clamp up as the base varying thicknesses of absorbing material, we invariably get a similar result. Between a thickness sufficient to absorb all [beta]-rays and a thickness equivalent to 1 cm. of lead the absorption curve is convex to the origin, i. e. the value of [lambda] decreases continuously. From about 1 cm. onwards [lambda] is constant. Experiments under identical conditions were carried out with lead, tin, zinc, and aluminium as absorbers. The absorption curves for each consist of two parts, an upper steep curved part and a lower straight part. The greater the density of the body the steeper is this upper part and the longer it takes to join the straight part of the range. If now 1*3 cm. of lead be placed on the radium and the results repeated, the general character of the curves is the same as before. The slopes of the lower part are less, since the rays are hardened by the lead, and the upper part joins with the lower part at a smaller equivalent thickness (0*5 instead of 1*0), thus smoothing out somewhat, but not eliminating, the convexity of the curve. With lead as absorber the placing of 1 cm. of lead over the radium has little effect, the convexity over the first centimetre being only slightly diminished. This shows that the effect has nothing to do with a soft type of [gamma]-rays initially present with the hard type, as has previously been supposed. If we now carry out similar experiments with another disposition (permanent base of lead or brass, absorbing material laid directly on the radium) the results are very different. For lead the upper part of the curve is again convex, whether the base be brass or lead, the convexity being the greater for a brass base : with zinc and aluminium, the upper part is concave, the concavity being about the same with either base. Typical results are given below. The ranges are expressed in actual thicknesses of lead and zinc. 1. Lead. Brass base 1'2 cm. thick. [table redacted] [header] 143 2. Zinc. Brass base 1'73 cm. thick. Range (cm.). ... [table redacted] 3. Lead. Lead base 1 cm. thick. Range (cm.). ... [table redacted] 4. Zinc. Lead base 1 cm. thick. Range (cm.). ... [table redacted] But it is possible to find a disposition in which zinc absorbs the [gamma]-rays strictly according to an exponential law with the normal value of [lambda] from thicknesses sufficient to absorb all the [beta]-rays up to the greatest thickness it was possible to use with the disposition. This was briefly referred to before (II. p. 754) and may now be more fully described. A preparation of radium (7 mg.) was placed at the apex of a cone of length 11*5 cm. and base 3 cm. diameter cut out of a cylinder [figure redacted] of lead (length 12*7, diameter 10*5). 2 cm. from the mouth of the cone was placed a short cylindrical ionization chamber of lead, 9*2 cm. inside diameter and 4*? em. long, an electrode in 144 [header] the centre of which communicated with a leaf system contained in a lead electroscope. Fig. 1 (p. 143) shows the disposition. Absorbing screens could be clamped tightly against the ionization chamber in position A, A. BB was a thick lead plate. Experiments were carried out with lead, tin, zinc, and aluminium, first with the short cylindrical ionization chamber and secondly with one three times as long but of the same diameter. This was done to find out if the shape of the ionization chamber had an effect on the values of the coefficients of absorption. Results are given below. A. Single Ionization Chamber. B. Triple Ionization Chamber. [table redacted] Thus X/d increases with d for this disposition. All the curves are exponential, except that for lead which is convex. The length of the ionization chamber made no practical difference. These dispositions are analogous to those used in Table V. (II. p. 753). For these the rays were confined in the same cone, which however entered directly through the base of an electroscope. The distance between the base of the electroscope and the mouth of the cone was 3'5 cm. as compared with 2*0 for the present dispositions. The results, however, were quite different from those detailed above. Thus zinc had for [formula redacted], 3*0 to 4*2 over a range of 0*35 to 2' 6, the absorption curve being concave. These results show how meaningless a value of [lambda] for an absorbing substance is unless a full account of the experimental disposition under which the measurements are made- is given. [header] 145 Section III. — The relative penetrating powers of the [gamma]-rays of the Radio-elements. It is convenient to state at the outset that, as a result of the experiments to be described, the relative penetrating power of the [gamma]-rays is, in descending order, thorium D, radium C, mesothorium 2, uranium X, and actinium 0. Thorium D thus gives the most penetrating [gamma]-rays known, though the differences between the first four are not great. But although it is easy to arrange the various types of [gamma]-rays in order of their penetrating power it is a more difficult matter to assign accurate values for [lambda] to each as their values as we have shown (II. p. 754-755) depend greatly upon the conditions, and their ratio for different rays is by no means constant when the results for several dispositions are compared. Thus the values of X for the uranium [gamma]-rays in four combinations, of two different metals with two different dispositions, were respectively 46, 28, 25, and 18 per cent, higher than for those of radium, and similar considerations hold good equally when the thorium and radium [gamma]-rays are compared. The radioactive preparations used in the following measurements were : — 1. Radium : 7 mg. of radium bromide, and in some experiments 0'5 mg. of radium bromide. 2. Mesothorium : a single tiny grain of concentrated mesothorium obtained from Knofler and Co., equivalent in [gamma]-rays to about 0*31 mg. radium bromide measured through 3 mm. of lead. All these preparations are practically point sources in sealed glass tubes. 3. Radiothorium : this body mixed with moist thorium hydrate was contained in a sealed cylindrical tube about 6 cms. long and about 0*5 cm. diameter. It therefore differed from the others in not being a point source. It was equivalent in [gamma]-rays to about 0'56 mg. radium bromide measured through 0*3 cm. lead. It also was obtained from Knofler and Go. Dispositions. — Sketches of the three dispositions employed are shown in fig. 2. They hardly need further description. Disposition 1 corresponds to that used in obtaining Tables II. and III. (I. pp. 644 & 646). It is easiest to use in practice [footer] 146 [header] as the electroscope is not disturbed. Disposition 2 is similar to that used in obtaining Table I. (I, p. 633) except that the electroscope is of lead as in Disposition 1. It offers certain theoretical advantages, but is more difficult to work with. [figure redacted] The only reason for using Disposition 3 was that uranium X had once been examined with it. It was not possible to prepare uranium X for the present set of measurements, but many of the previous results can be utilized. Those of Tables II. and III. (I. pp. 644 & 646) only differ from the present Disposition 2 in that a wood stand instead of a lead stand was used for the preparation, but parallel experiments with mesothorium 2 showed that this had no effect on the values of [lambda]. The results with Disposition 1 are tabulated in Table A. After a thickness equivalent to 1 cm. of lead (total thick- ness, with base, 2 cm. of lead) all the curves are exponential. Lead is in a class by itself (Class I.) with high value for [lambda]/d throughout. The Class II. bodies have an approximately constant [lambda]/d, but the limit of density down to which this holds is greater for the more penetrating rays than for those less penetrating. The lightest bodies (Class III.) have again higher values for [lambda]/jd. [header] 147 Thorium D. Radium C. Mesothorium 2. Uranium X. [table redacted] The nature of the absorbent has a marked effect on the relative penetrating power of the four types of rays. Taking the values for radium G as unity, this is shown in the following table, mean values for Class II. being employed. [table redacted] In Disposition 2, with the absorbing plates clamped up to form the base of the electroscopes, two sets of experiments were performed, namely (2 A) with the preparation bare, and (2 B) with the preparation covered with 0'6I cm. lead, to see if the hardened rays produced would be exponentially absorbed by all bodies according to the density law. The result shows that this is the case except for lead. Table B (p. 148) shows the results obtained. All absorption curves for this disposition are also exponential after 1 cm. of equivalent thickness of lead has been traversed. The curves made by plotting equivalent thickness of metal against the logarithms of the ionizations are coincident. So they are also in the case of Disposition I. This holds both for Dispositions 2 A and 2 B. This is opposite to what was previously obtained with a brass electroscope for the absorption by bodies over small initial [footer] 148 [header] Radium C Thorium D Mesothorium 2 [table redacted] [header] 149 ranges of thickness of the rays of uranium X (II. p. 729). There the bodies arrange themselves one above the other in order of the density of the material, the lighter substances giving for equivalent thicknesses far more ionization than the denser. The mean values of [formula redacted] are tabulated below. [table redacted] It is seen that the values for lead are very near one another but the differences are quite real. For radium the value of [formula redacted] has been obtained at least six times for Disposition 2 A. Taking the values for radium as unity the results are summarized in the following table : — [table redacted] In Disposition 2 A, the values of [lambda] are all about 5 per cent, greater than in Disposition 1. In Disposition 2 B lead still absorbs normally, but both Class II. and Class III. bodies now obey the density law. Disposition 3 was only used for copper. The following values of [lambda] were obtained :— - [table redacted] 150 [header] The results are collected in the following table : — [table redacted] [Not exactly the same disposition.] Thus the value of X for thorium D is from 8 to 21 per cent, less, and for mesothorium 2 from 4 to 25 per cent, greater than for radium O. Disposition 1 might well be adopted as a standard disposition for [gamma]-ray comparison of the intensity of radioactive preparations. Preparations of widely differing activity might be compared by means of carefully prepared lead blocks of known thickness, adopting, for the value of X, 0'500 (cm.) -1 . The [gamma]-rays of Actinium. — Through the kindness of Messrs. Buchler and Co., of Braunschweig, who most generously lent for the purpose a very fine preparation of actinium, weighing 1*5 grams and equivalent in [gamma]-activity through 3 mm. of lead to 023 mg. of radium bromide, it has been possible to make a more extended examination of the [gamma]-rays of this substance than has hitherto been done. The prepa- ration was in a sealed tube for these measurements, and subsequently the tube was opened and the preparation used to obtain the active deposit for the determination of the [gamma]/[beta] ratio described in Section I. Godlewski [citation redacted] measured the absorption of the actinium [gamma]-rays over the range from to 3'5 mm. of lead and to 10 mm. of iron and of zinc, and obtained nearly exponential curves, the value of [formula redacted] for lead being 4*54, for iron 1*23, and for zinc 1*24. He thus showed clearly the markedly less penetrating power of these rays as compared with those o£ radium and the abnormally high absorption in lead. Eve [citation redacted] with a stronger preparation found for X 4*1 over a similar range of lead. At 3 mm. thickness a sudden change to X = 2'7 took place and continued [header] 151 up to 5" 7 mm., the limit of the measurements. We have been able to extend the range up to 2' 5 cm. of lead, or including the 3 mm. thickness of lead as base to a total thickness of 2*8 cm. Owing to the abnormally high absorption of the rays by lead, as noticed by Godlewski, Disposition 1 could not be employed, but Disposition 2 was possible. A new arrangement (Disposition 5) similar to Disposition 1 was used, in which the walls and base of the lead electroscope were 3 mm. thick and the preparation was mounted 8" 3 cm. below. The absorbing lead and zinc were placed directly over the preparation, and parallel experiments were done with radium for the same disposition. The results* for lead and zinc are plotted in the curves (fig. 3). The abscissae are centimetres [figure redacted] of thickness multiplied by the relative density of the absorber compared with that of lead. The aluminium curve would coincident with that of zinc if plotted. The figure shows how abnormal lead is. Although the measurements were made through a base of lead 3 mm. thick there is still the point of inflexion in the lead curve at 0*3 cm., noticed by Eve, and a still more decided point of inflexion at 0*85 cm. After that the curve is linear up to 2*5 cms. For zinc and aluminium the curves are exponential. In Disposition 2 the curves for copper, aluminium, and zinc are exponential after 152 [header] a small initial thickness (0*36 cm. for zinc), but for lead the curve is convex to the axis up to 0*8 cm., as in Disposition 5, and then is straight. The results are tabulated below, the results for lead referring to the straight part of the curve beyond 0*8 cm. Disposition 5. Values of [lambda] (cm.) Actinium Radium [table redacted] Disposition 2 [table redacted] The separate mean values of [lambda] for different thicknesses of lead are shown below. Disposition 5. Range (cm.). [table redacted] Disposition 2. Range (cm.). ... [table redacted] 153 [header] In Disposition 2 the curve is the more continuously convex to the origin. For both dispositions the ratio of the absorption coefficients is about 1 : 1*9 for Class II. bodies, and the effect of hardening the zinc is the same, the value of [lambda] being reduced to 80 per cent. But the actinium rays are, relatively to radium, hardened more in Disposition 5 than in 2. Godlewski concluded that the actinium [gamma]-rays have only one-tenth of the penetrating power of the harder [gamma]-rays of radium. As in the case of uranium X, however, work with more active preparations over greater ranges has reduced this difference, though actinium is still far removed in this respect from the other [gamma]-ray sources examined. Some Generalizations with regard to [gamma]-rays. The results for actinium help to accentuate several relationships which exist between the penetrating power of the [gamma]-rays of the radio-elements in solid bodies and other properties they possess. There is obviously an intimate connexion between penetrating power, abnormality of absorption by lead and hardening by lead. The rule is — the greater the penetrating power of the ray, the less abnormal is the absorption by lead as compared with that of Class II. bodies, and further the less hardening effect has lead on the rays subsequently absorbed by Class II. bodies. In Disposition 1, however, thorium D is an exception. Thorium Radium Mesothorium 2. Uranium X. Actinium C. [table redacted] Again, as regards penetrability, there is far more connexion between the [gamma]-ray and the [alpha]-rays preceding and following it in the disintegration series, than there is between 154 [header] the [gamma]-ray and the [beta]-ray which accompanies it. It is possible, from the results that have been given, to make a rule connecting the penetrability of the [gamma]-ray with the period of the change in which it occurs, analogous to Rutherford's well-known rule for the range of the [alpha]-particle. Actinium is an exception throughout, and has been omitted from the comparison. In ascending order the penetrabilities of the [gamma]-rays are : — Uranium X, Mesothorium 2, Radium C 1, Thorium D. This is also the descending order of their periods, and also of their probable atomic weights. If the conclusions of Hahn and Meitner [citation redacted] are correct, the [beta]- and [gamma]-rays of radium C come from radium C1 (half-period 19 minutes), while the [alpha]-rays come from a succeeding product, radium C2 (half-period, about 2 minutes). This accords with Rutherford's rule, as the still more penetrating [alpha]-ray of thorium C is, in all probability, derived from a change even more rapid than that of radium C2 . The above order of the penetrabilities of the [gamma]-rays is also that of the [alpha]-rays which precede and follow them, whereas the order of the penetrabilities of the [beta]-rays is quite different. The figures in brackets denote ranges of the [alpha]-rays in centimetres of air. [gamma]-rays [alpha]-rays preceding [alpha]-rays following [beta]-rays [table redacted] Neither in penetrability, relative intensity nor homogeneity are the [beta]-rays obviously connected with the [gamma]-rays, whereas there is a certain connexion between the [gamma]-rays and the [alpha]-rays. Physical Chemistry Laboratory, Glasgow University.