beyond our knowledge at present. This consideration will explain many disagreements in the results obtained by observers using different instruments. It is simplest to suppose that the effect of transmission through a metallic screen upon any one component of a heterogeneous beam is independent of the presence or absence of the other components. In other words, the effect of a given screen upon the transmission of a particular sort of ray is measured by an absorption coefficient which may change with the conditions of the experiment only in so far as those conditions affect the ray in question. An attempt has been made to test the constancy of these coefficients under changing intensity of the rays, by the experiments where the dependence of the ratio of transmission upon the distance between the tube and the screen was investigated. These experiments are given in detail on pages 682, 683. It seems very probable that in experiments so conducted the variation in the intensity of the rays incident on the screen, involved in moving the tube, very nearly fulfils the condition that the changes in the intensities of all the components of the beam shall be in one ratio. Except for the absorption by the air, this ratio is doubtless fixed for all the different sorts of rays by some function of the distances involved; and from numerous experiments with various instruments 20 it appears that the absorption by air of Röntgen rays in general is negligible for such short distances as these. Of course a real disturbing factor is the changing behavior of the tube, for which correction is made, as well as possible, by alternating and checking readings. Granting that these sources of error have had no appreciable effect, we learn from the observations that the ratios of transmission of the metallic screens examined are independent of the total intensity of the incident radiation, when that intensity changes in such a way that the intensities of all the components change in the same ratio. If we examine this result in the light of the conclusions already reached in this paper, its practical importance will appear. It is plain that if we knew in advance that all of the absorption coefficients for the metallic screen and the beam in question are constant so far as the intensity is concerned, the constancy of the ratio of transmission would necessarily follow, provided we conclude from the data of Table IV that all of the absorption coefficients of platinum are independent of the 20 J. Trowbridge and J. E. Burbank, Amer. Jour. Sci., 157, 396 (1899); A. M. Mayer, Amer. Jour. Sci., 151, 467 (1896); C. G. Barkla, Phil. Mag., 7, 555 (1904). intensity. But to draw conclusions as to the constancy of any or all of the separate coefficients from the constancy of the ratio of transmission is not in strictness possible. It might be that two or more of the coefficients changed together in such a way as to keep the ratio of transmission constant. But the repetition of this coincidence in experiments with different metals and with rays from tubes in very different conditions is exceedingly unlikely, and the constancy of the ratio of transmission with varying intensity is at least very good presumptive evidence that the coefficients characteristic of the absorption of Röntgen rays in metallic sheets are all constant with varying intensity of the rays. It may be of interest to examine the consequences of supposing that the constancy of the ratio of transmission is not a result of the constancy of each of the coefficients. To simplify the argument, let us suppose that the metallic screen is very near the instrument. Let s and s represent respectively the fraction of a certain sort of ray which the metallic screen transmits at the large intensity and at the small intensity. Let p be the fraction of the same sort of ray which the platinum of the instrument transmits at all intensities. Let I, and I represent the two intensities of this sort of ray at the instrument. Then the ratio of transmission for this ray alone at the large intensity would be I1 81 (1 − p). or s1; and at the small intensity, I1⁄2 81⁄2 (1 - p) or s. If the ratios of transmission for the whole beam are found to be the same at the two intensities, the explanation must be either that each 8 is equal to its corresponding s2, or that some of the si's are larger than their corresponding sa's while others are smaller. The latter explanation is equivalent to saying that the absorption coefficients for some sorts of rays are increasing functions of the intensity, while others are decreasing functions; and this seems highly improbable. The argument of this section may be summarized in another form, as follows: The possible effects of transmission through a metallic screen upon a beam of Röntgen rays are three : (1) An effect produced upon the beam by transmission across the surfaces of the screen. (2) An effect of transformation suffered by the several components of the beam in passing through the substance of the screen. (3) An effect of absorption suffered by the several components of the beam in passing through the substance of the screen. Experimental results rule out, more or less certainly, the first two effects. Granting these results, a theory of relatively selective absorption is necessary to explain the reduction of penetrating power for one metal by transmission through another. By somewhat indirect experimental evidence, this absorption is shown to be governed by coefficients which are probably constant with varying intensity of the rays. VI. CONCLUSIONS. The following conclusions are reached in this paper: (1) The approximate measurement of the energy of Röntgen rays is of the same order of magnitude as the earlier measurements of that energy, and to that extent confirms them. (2) In the transmission of Röntgen rays through metallic sheets, the effect of the surfaces of the metal is small. (3) In the transmission of Röntgen rays through metallic sheets, the probability that one sort of ray is transformed into another sort of ray, to an appreciable extent, is small. (4) The absorption of any particular sort of ray by a metallic sheet is measured by a coefficient which is probably independent of the intensity of that ray. (5) In one special case, at least, the general effect on absorption by one metal of passing the rays through another metal is not, so far as we can judge, a decrease of absorption, but an increase; and this fact tends to support the theory that the radiation from an ordinary tube is heterogeneous in character, and that the absorption in metallic sheets is more or less selective. It is hoped to continue this research in the direction suggested by the conclusions of this paper: to test the hypothesis of relatively selective absorption by direct experiment; 21 to obtain, if possible, beams of Röntgen rays which will yield transmission curves of a simpler form than those heretofore found; to study the properties of those beams; and to study the indications of the instrument used in this research in comparison with those of other instruments heretofore used, especially the ionization-electroscope and the photographic plate. 21 For a preliminary note on this experiment, see Amer. Jour. Sci., 173, 91 (1907). |