XXVII. On the Radioactive Matter in the Earth and the Atmosphere. By A. S. Eve, M.A., McGill University, Montreal [Communicated by Professor E. Rutherford, F.R.S.]. Numerous observers have found that part of the ionization occurring within closed vessels is due to penetrating radiation. These rays are attributed to the presence of radium in the earth and of radium emanation in the atmosphere. Experiments are described in this paper which lead to an estimate of some of the magnitudes of the quantities involved. Let K denote the number of ions, due to the [gamma] rays only, generated per cubic centimetre per second at a distance of one centimetre from one gram of pure radium bromide, supposed concentrated at a point, and so placed that all the rays are absorbed in air. Then Q grams at a point will generate at a distance r, by the [gamma] rays alone, a number of ions per c.c. per second denoted by N and given by, [formula redacted] where [lambda] is the coefficient of absorption of the [gamma] rays by air. The value of K has been approximately determined by finding the number of ions produced per c.c. per second in closed vessels made of various metals. These vessels were placed at a distance of 303 cms. from 14*3 mg. of radium bromide sealed in test-tubes, and placed behind lead 7 mm. thick. The radium bromide used was that which Professor Rutherford found gave per gram a heating effect of 110 gram-calories per hour. The experiments were made in the Chemistry Building, where radium emanation has not been artificially introduced. Vessels were made of clean commercial sheets of lead, copper, zinc, iron, aluminium, and tinned iron. These vessels were 51 cms. high. and 23 cms. in diameter (rig. 1). Two tubes were soldered into the sides, so that dry filtered air could be drawn into the vessels when required. Above each vessel was a small electroscope, 8 cms. high and 6*5 cms. in diameter, made of clean tinned iron. A rod within the electroscope carried a Dutch-metal leaf observed by a microscope with a graduated eyepiece. This rod was supported by sulphur, originally poured when in a molten state into a larger cylinder of ebonite. This arrangement gives excellent insulation and firm support. The rod terminated within the lower vessel in a hook to which a rod 190 [header] or wire cage was suspended. A rod, one centimetre in diameter , was first used, but difficulties arose in obtaining a saturation current. A cylindrical cage of clean copper wire about 40 cms. long and 7 cms. in diameter, was therefore substituted for the rod. Each vessel had then a capacity of about 11 electrostatic units. The electroscopes could be interchanged when required, and insulated from their respective vessels by mica sheets. This most convenient method was recommended to me by Professor Rutherford. Bragg has shown the difficulty of securing complete saturation in the case of weak currents due to feeble radiation. But, by the arrangement described, the lower vessels could be earthed, the case of the electroscope maintained at a high potential by a constant battery, and the w ire cage and goid-leaf charged to a yet higher potential. In this way saturation may be ensured between the large vessel and wire cage, whilst the gold-leaf is deflected to a convenient amount due to a moderate difference of potential between the gold-leaf and the case of the electroscope. After deducting the corrections for natural leaks the results were as follows : — Substances. Thickness in mm. Ions per c.c. per sec. [table redacted] There is clear evidence of much secondary radiation from the heavier metals. In the lead vessel 45 per cent, of the ions must be attributed to this cause, for there is no other reason why the ions should be more numerous in the lead than in the aluminium. It will be noted that tinned iron takes the place due to tin rather than iron. The order of ionization need not necessarily be that of atomic weight, [header] 191 because the secondary radiation due to transmitted rays is a function of the thickness traversed [citation redacted]. If the penetrating radiation, discovered by McLennan, and by Rutherford and Cooke, consists of [gamma] rays, it follows that the weak ionizations resulting from it within closed vessels made of different metals should have the same characteristics as those determined in these experiments. It is now possible to calculate the value of K in formula (1) . The value of N in the thin aluminium vessel was 297, and this is probably not more than 10 to 20 per cent, in excess of the value in free air [formula redacted] Hence [formula redacted] or adding a small correction for the absorption by 3 metres of air [citation redacted], [formula redacted] But the [gamma] rays have been absorbed to a considerable extent during their passage through 7 mms. of lead, and if we take the coefficient of absorption of lead to be initially '01 J, the intensity has been diminished by the factor [formula redacted], so that the value of K is about [formula redacted] This result has been calculated from the ionization within a thin aluminium vessel, and it is, therefore, owing to secondary radiation, probably about 10 to 20 per cent, in excess of the value of K in the open air. The value of K within a zinc or copper vessel is 3*6 x 10°. On the Total Number of Ions generated per Second by the [gamma] rays from a Gram of Pure Radium Bromide. It is now possible to determine the total number of ions which would be generated every second by the [gamma] rays from Q grams of pure radium bromide, surrounded entirely to a great distance by air, so that the rays are all absorbed by it. If [lambda] is the coefficient of absorption of the [gamma] rays in air, then at a distance r the number of ions produced per second per c.c. by the [gamma] rays from Q grams is [formula redacted] , and in a 192 [header] spherical shell of thickness dr the total number of ions per second would be [formula redacted] times as great. Hence N, the total number of ions produced per second in all the surrounding air, is given by [formula redacted] But McClelland has shown that the coefficients of absorption of the very penetrating rays vary approximately as the densities of the absorbing substances. The value of A for water is *034, hence A, for air is about '000044. Thus the [gamma] rays are reduced to half value after penetrating 157 metres of air, and to one per cent, after passing 1000 metres of air. From (3) we find the total number of ions per second, due to the [gamma] rays from 1 gram of pure radium bromide, when the rays are wholly absorbed in air, equal to [formula redacted] and this is probably an overestimate of the true value. A comparison can therefore be made between the total number of ions, due to the a and [gamma] rays respectively, produced per second by a gram of radium bromide, when all the rays are absorbed in air. Professor Rutherford has found that the [alpha] rays from one gram of radium bromide in radioactive equilibrium produce about 1*24 x 10 16 ions per second if wholly absorbed in air. By an independent method the writer found about 1*67 x 10 16 ions per second due to the same cause. Hence the [gamma] rays appear to produce less than one-sixteenth part of the total number produced by the [alpha] rays. Without insisting on the exact value of this ratio, it is clearly seen that all the ions due to the [alpha] rays of a given mass of radium are more numerous to a considerable extent than all the ions due to the [gamma] rays. And this might be expected, inasmuch as the heating effect of the [gamma] rays is only a small percentage of that of the [alpha] rays. On the Ions due to the [gamma] rays from the Emanation in the Atmosphere. In a recent paper [citation redacted] in this Magazine it has been shown that the equivalent amount of radium bromide which would be required to provide the emanation actually present in one cubic kilometre of the atmosphere near the earth's surface is [header] 193 between * 1 4 and *5 gram. This quantity, which may be denoted by R, has been determined by a method which is quite independent of any theory of the radioactivity of the atmosphere. It was found by measuring the radiation from the active deposit on a negatively charged wire placed in a large vessel. The lower value, '14, was that obtained by observations in the open air, and is probably the more correct. But since it is desired at the present moment to find a higher limit, let the larger value of R, namely *5 gram, be taken. Then from the result (4) it is easy to calculate the total number of ions produced per second in a cubic kilometre of the atmosphere due to the [gamma] rays from the active matter in it. The calculation will be made on the supposition that an infinite volume of air contains emanation, to the amount per c. k. above stated, uniformly distributed throughout it. A little consideration will show that a volume distribution gives results identical with distribution at points. The number of ions produced per c. k. per second is [formula redacted] or .44 ions per c.c. per second. It is clear that at the earth's surface, which is a boundary of the atmosphere, one-half of the above value should be taken, or '22 ions per c.c. per second. A more probable value is 14/50 of this, or '06 ions per c.c. per second. In any case we must conclude that the ionization due to the [gamma] rays of the active matter in the atmosphere is almost a negligible quantity and quite inadequate to account for the relatively large effects due to penetrating radiation observed by Cooke. Campbell, and others. For example, in a well-cleaned brass vessel, H. L. Cooke found q = 13*6 *, and by a screen of about 5 cms. of lead he was able to reduce this value to 9*1. No further reduction could be effected by thicker screens, so that about 4*5 ions per c.c. are produced in a brass vessel every second by the penetrating radiation near the surface of the earth. Campbell appears to find even larger results from this cause. Since it has been shown that the [gamma] rays from the active matter in the atmosphere are quite insufficient to account for such large effects, let us consider the penetrating rays due to the active matter in the crust of the earth. If, for instance, radium is mixed with solid matter such as sand, or found in an ore, the penetrating radiation from the radium in the body would be larger than [The amount originally stated was 7*5, but on the assumption that the ionic charge was 6-8 X 10- 10]. [footer] 194 [header] the radiation from the limited quantity of emanation which escapes from it. It is not unreasonable to expect a much larger effect from the [gamma] rays of the radium (and its successive products) in the earth than from the [gamma] rays due to the emanation (or its quick transformation products) found in the atmosphere. The question may be investigated quantitatively. Before doing so it is well to note the rate of absorption of the [gamma] rays under different conditions. Table II. Substance. Density. [lambda] Thickness in cms. to half value. [table redacted] In this table the value of [lambda] for water is taken from McClelland' s results [citation redacted], and the values of [lambda] for air and for the constituents of the upper crusts of the earth are deduced from the density law which he established. It is assumed that 2' 7 is a fair average value to take for the density of the constituents of the upper crust of the earth. It follows that the [gamma] rays are cut down to one per cent, after traversing half a metre of the earth's crust, or 1000 metres of the atmosphere. The Amount of Radium in the Earth estimated by the Penetrating Radiation. We have seen that the number of ions produced by the penetrating radiation near the earth's surface is found from H. L. Cooke's experiments with a brass vessel to be about 4*5 per c.c. per second. If a small hollow cavity were taken many metres below the earth's surface, we should expect the ionization due to the penetrating radiation to be twice as great, because rays would come from above and below, not from below only as on the earth's surface. If [formula redacted] be the number of ions which would be produced per c.c. per second in a brass vessel, supposed to be within such a cavity, by the [gamma] rays from the active matter in the earth surrounding- it ; if Q' be the equivalent number of grams of radium bromide which is a measure of the active matter per c.c. in the earth's crust ; then, by a calculation precisely similar to [header] 195 that made for the atmosphere (3), we have [formula redacted] 111 this case K' has been taken as for a zinc or copper vessel (Table I.), because these would involve some secondary radiation to an extent approximately the same as brass. It is interesting to compare this value of the equivalent amount of radium bromide per c.c. of the earth's crust near the surface, found by measurement of the penetrating radiation, with that deduced by Professor Rutherford from consideration of the earth's temperature gradients. His result is about seventy times as small, or 2'6 X 10 -13 , also expressed in terms of radium bromide. But Strutt has just published a paper [citation redacted], stating that he finds by direct observation of rock-specimens that the radium present in the upper 45 miles of the earth's crust is alone sufficient to account for the existing temperature gradients in the earth. This indicates that he finds actually present thirty times as much radium as Rutherford showed was necessary to account for the earth's temperature gradients on the assumption that radium was distributed uniformly throughout the whole earth. In fact Strutt finds about 8*5 X 10~ 12 grams of radium bromide, in equilibrium, as the average equivalent of the active matter per c.c. of the earth's crust. But this result needs a further correction, because Rutherford and Boltwood state [citation redacted] that 1 gram of uranium is associated with 3*8 x 10~ 7 grams of radium, not with 7*4 x 10 -7 as previously stated. Therefore, from Strutt's investigations we may conclude that 4*25 x 10 -12 grams of radium bromide is a fair average measure of the active matter per c.c. of the earth's crust, and this result is four times as small as the amount which I have calculated as sufficient to account for the penetrating radiation. The discrepancy is not large considering the uncertainty of the distribution of active matter in the earth's crust [Measurements of the penetrating radiation in various mines would lead to a valuation and survey of the active matter in the earth's crust.]. It will be seen, therefore, that whilst the active matter in the earth is of the right order to account for the observed magnitude of the effects of penetrating radiation, that in the atmosphere is not sufficient to do so. This conclusion is capable [footer] 196 [header] of further verification by experiments, some of which I hope to undertake. For example, an electroscope well screened from the earth, vertically and obliquely, should show a greater decrease of ionization than one equally well screened from the atmosphere. Again, the penetrating radiation in a deep mine or well should be stronger than on the surface of the earth. Such experiments on penetrating radiation should be made out of doors, because the walls of a building made of brick or stone may contribute almost as much radiation as they absorb [Observations of the penetrating radiation at the top of the Eiffel Tower and in a deep mine would throw much light on this subject.]. As the matter stands at present the active matter in the earth seems to be the probable and sufficient cause of penetrating radiation. It should be added that experiments have been made by the writer with Ebert's apparatus at Montreal, and the average number of ions measured in the atmosphere indicate that the conditions at Montreal are normal and resemble those found over a large area in Europe. Activity of Metals. In the course of these investigations the " natural leak " of each vessel was determined before any radioactive matter was brought near them. The value of q was found from the relation [formula redacted] where q is the number of ions produced per c.c. per second in the volume S, due to whatever causes ; e is the ionic charge [formula redacted]: C is the capacity and V the fall of potential in time t. In most cases it was necessary to let the vessels remain for one or two days after filling them with fresh air, before a minimum value could be obtained. The cause of this initial decrease has not been determined, but it has been previously observed by McLennan and others. It cannot well be attributed to fine dust, for I have found that tobacco smoke introduced into an electroscope produces an opposite effect. The electroscopes were frequently interchanged between one vessel and another, so as to compare the natural leak due to one metal with that due to another, simply by measurement of the rate of fall of the gold-leaf of the electroscope, and apart from all calculations. The results were confirmed by repeated observations lasting for some months, but they are [header] 197 not altogether in good agreement with those found by several other observers. In the lead vessel there were 95 ions generated per c.c. per second. In the case of all the other metals examined, there were 21 ions per c.c. produced per second. No difference could be detected between clean zinc, copper, aluminium, iron, tinned iron, and aluminium in lead. It was natural to suspect that the lead was coated with radioactive matter, but after cleaning it with emery-paper and washing it with hydrochloric acid, there was no diminution in the value of g. On the other hand, when the lead was lined with aluminium *4 mm. thick, q was equal to 2-1. My results appear to agree with the work of other observers in the case of lead, for which all agree in finding a high value of y : and the effect is certainly not due to a surface-deposit merely . It was also found impossible to lower the value of q for zinc by a cleansing treatment of the surface. It appeared possible, from Campbell's work, that the results obtained in these experiments might be due to an accidental relationship between volume and surface. Cylindrical rolls both of aluminium and copper were therefore compared, 51 cms. long, having diameters 23, 17, O'S cms. successively, and the central wires or rods were charged to potentials varying from 150 to 120 volts. The results obtained for copper and aluminium, under any similar conditions, were identical in each case. If the ionization is due to intrinsic radiations from the metals, it is extraordinary that so many different metals should give equal values. If the ionization is due to radioactive impurities in the metals, it is no less strange that the impurities should be distributed in so uniform a manner. Corrections amounting to 7 or 8 per cent, were made for the current across the small electroscopes, after deducting which the results were those given in column I. Table III. Metal. Density. Thickness in mm. Natural Ionization Artificial Ionization. Q. [table redacted] 198 [header] In this table the values of Q have for convenience been repeated from an earlier part of this paper. Q denotes the number of ions produced per c.c. per second in the vessels when 14*3 mg. of radium bromide were placed at 303 cms. from the centres of the vessels. The six equal results in the third column are the mean of very many observations. Their equality is based on the fact that by comparative measurements no difference could be detected between the metals observed under identical conditions. Further experiments are in progress on the effect of screening the vessels and on the values of g for various diameters. In several cases a gradual increase of ionization occurred due to an emanation proceeding from the sides of the vessel. Sometimes the rise curves had the character of those of radium emanation. The presence of an emanation has been noted by McLennan, Burton, and others. Further investigation of this emanation may throw light on the activity of metals. The effects were irregular, but they could always be eliminated by pulling dry clean air through the vessels for a few hours. Ionization in the Atmosphere. In a previous paper [citation redacted] the writer gave an account of the measurement of the number of ions produced per c.c. per second in a large iron tank, 2*5 cms. thick, [formula redacted] metres in dimensions. The value obtained was q = 9'6 ; and as the active matter collected in the tank was sufficient to account for such ionization, the greater part of it was attributed to the emanation in the air in the tank. Mr. N. R. Campbell dissents from this view, writing thus : — " Surely if the quantity of emanation in the small volume of air inside the vessel causes an effect so large, the far larger quantity outside the vessel must have an appreciable effect even though the walls are so thick as to cut out all [alpha] rays." This criticism arises from a misconception. The [alpha] rays are much more effective in producing ionization than the [gamma] rays. The experiments and calculations given in this paper show that the a radiation of the emanation in the tank, despite its limited range, produces a much greater effect in the tank than the y radiation from the emanation in all the air both inside and outside the tank to any distance, even supposing the walls of the tank were thin ; and they were 2*5 cms. thick. And since Cooke found that 5 cms, of [header] 199 lead cut off the penetrating rays entirely, 2*5 cms. of iron would reduce the penetrating radiation to about half value. It is, however, possible that I underestimated the radiation from the sides of the taDk and the penetrating radiation passing through them. On the whole, the tank experiments seem to prove that the emanation in the atmosphere is an important ionizing agent in virtue of its a radiation; and the excited activity actually collected in the tank bears out the results of direct ionization experiments. It is certain that emanation exists in the atmosphere and that it produces ionization, but more experimental work must be done before any exact value can be assigned to the number of ions thus produced. My work and calculations point to the conclusion that the [gamma] rays from the emanation in the atmosphere may be disregarded as a negligible quantity. The rays may probably be ignored in the same manner. On the other hand, the [alpha] rays from the emanation do produce several ions per c.c. per second, whilst the penetrating rays from the active matter in the earth probably contribute the balance of those observed near the earth's surface. Summary. 1. The ionization produced within closed vessels at a given distance from a given quantity of radium, and due to the [gamma] rays alone, is dependent on the substance and thickness of the metal employed, owing to the secondary radiation (see Table I.). 2. From the results obtained within a vessel of aluminium an approximate value of K was obtained equal to [formula redacted] , where K is the number of ions produced per c.-c. per second by the [gamma] rays from 1 gram of pure radium bromide at 1 cm. from the source, supposing the rays are all absorbed in air. 3. Hence the total number of ions produced by the [gamma] rays from 1 gram of pure radium bromide completely surrounded by air would be about [formula redacted] per c.c. per second. 4. Concluding from a former paper that the equivalent radium bromide required to supply the observed quantity of emanation per cubic kilometre of the atmosphere near the earth's surface is between *14 and "5 grain ; it follows that the number of ions per c.c. per second due to the [gamma] rays from the active matter in the air is between about *06 and *22 near the earth's surface. 5. The [alpha] rays from the emanation in the atmosphere produce much more ionization per c.c. than do the [gamma] rays. The 200 [header] ratio may be estimated at 16 to 1, so that the [alpha] rays of the emanation produce about 2 to 7 ions per c.c. per second. 6. The penetrating radiation, observed by H. L. Cooke in a closed brass vessel, produced about 4*5 ions per c.c. per second. This cannot be attributed to the active matter in the atmosphere, but the radium present in the earth's crust appears to be of the right order to account for it. 7. About 1*8 X 10 -11 grains of radium bromide is the estimated equivalent o£ the active matter per c.c. present in the earth's crust sufficient to account for the penetrating radiation. This appears to be about four times as large as the average amount found by Strutt by direct observation of rock specimens. 8. The ionization in the atmosphere is due partly to penetrating radiation from active matter in the earth, partly to a radiation from the emanation in the atmosphere. I am most grateful to Professor Rutherford for his assistance in difficulties and for valuable suggestions in experimental work and theoretical considerations. Montreal, June 1906.