294 



NA TURE 



[July 30, 1908 



deal with the simple and then proceed to the complex, 

 and he would consider only the secondary X-rays. 1 grant 

 this principle, of course, but I object entirely to the applica- 

 tion which he makes of it. It is the 7 rays which give 

 the simpler effects, and the hardest 7 rays which give 

 the simplest, for the obvious reason that such rays ignore 

 atomic structure altogether even in the case of the heaviest 

 atoms. The X-rays are soft, and therefore atomic struc- 

 ture influences and complicates the effects to a remarkable 

 degree, as Dr. Barkla's own work shows. A true applica- 

 tion of the principle would lead us to work out the laws of 

 the hard 7 rays first, and then to consider the X-rays in 

 the light of the knowledge we have obtained. This is 

 what 1 have tried to do. The 7 rays suggest a corpuscular 

 hypothesis, and on turning to the X-rays it is at once clear 

 that a large proportion of the effects which they show may 

 also be simply explained on the same hypothesis. 



Yet I am willing to meet Dr. Barkla even on the narrow 

 ground on which alone he has chosen to risk encounter. 

 He states certain pieces of evidence, numbered i to 9 in 

 his letter, which should show convincingly that a theory 

 of ether pulses is to be preferred to one of neutral pairs. 

 Let us consider these. 



No. 2 refers to the equal penetrating powers of primary 

 and secondary rays in certain cases. This is a natural 

 consequence of almost any theory, certainly of a cor- 

 puscular one, and the argument may be set aside at once. 



No. 3 refers to the equality in the proportion of rays 

 of different penetrating power which are scattered. The 

 theory was given by Prof. Thomson in his " Conduction 

 of Electricity through Gases," and experiments have been 

 made by Dr. Barkla (Phil. Mag., May, 1904) and Mr. 

 Crowther (Phil. Mag., November, 1907). The latter found 

 notable exceptions to the rule, of which no theoretical 

 explanation has yet been offered. The experiments are not 

 easy, and there is enough chance of error to cover a con- 

 siderable departure from the law, especially considering 

 that no great variation of quality is possible with X-rays 

 alone. In any case, there seems to be no reason for sup- 

 posing the effect, if a true one, to be a special consequence 

 of the pulse theory. It might well hold for a corpuscular 

 theory, at least over the same limited range. 



I have discussed No. 5 in a previous letter. The actual 

 distribution of the secondary scattered rays agrees with 

 the rule deduced by Dr. Barkla in special cases only. In 

 others it does not, and the pulse theory does not say why. 

 One might reasonably expect the rule to be of very partial 

 application, for the secondary kathode radiation has such 

 an extremely asymmetrical distribution that it is hard to 

 believe in a complete symmetry of the remaining secondary 

 radiation. Again, there seems to be nothing irreconcilable 

 with a corpuscular theory. 



As regards No. 6, it was asserted by Dr. Barkla as the 

 result of his experiments that the ratio of the quantity 

 of the secondary radiation to the quantity of the primary 

 depended only on the density of the gas producing it. Prof. 

 Thomson (Phil. Mag., June, 1906) used the quantitative 

 result as the base of one of three proofs that the number 

 of electrons in an atom was nearly equal to the atomic 

 weight. So far as I can judge, the proofs are not con- 

 clusive, and are not generally held to be so. In repeating 

 Dr. Barkla's experiments, Mr. Crowther found the rule 

 to be true only over a very limited range. C, N, and O 

 did, indeed, give the same ratio, but all other atoms gave 

 different ratios. In the case of H it was 70 per cent, 

 larger, of He 10 per cent., of S 40 per cent., of Br about 

 8000 per cent., and so on. Any theory would show a 

 constant effect over so narrow a range. Dr. Barkla con- 

 siders his result to be evidence of value because it fits in 

 with the result which Prof. Thomson derived from the 

 ratio determined for air (N and O) ; but the strength of 

 the other two proofs of Prof. Thomson's theory is hardly 

 enough to permit this one to be removed and used to sup- 

 port Dr. Barkla's. 



Nos. I and 4 refer to the well-known polarisation effects 

 found by Dr. Barkla. The latter has been confirmed by 

 Dr. Haga, who, however, rejects the former. I have 

 already shown that it is not impossible to explain such 

 effects on a neutral pair theory. 



Lastly, there are the three statements Nos. 7, 8, and 9. 



NO. 2022, VOL. 78] 



They refer to certain remarkable effects observed by several 

 workers, particularly M. Sagnac, Dr. Walter, Mr. Adams, 

 and Dr. Barkla himself. Dr. Barkla says that they can 

 be explained on the pulse theory. If any such explana- 

 tion has been given I have not seen it, and I have not 

 been able to devise any such explanation myself. The 

 secondary X-rays from a substance like copper are homo- 

 geneous, and specially able to penetrate screens of the 

 same substance. The homogeneity cannot result from sift- 

 ing either primary or secondary rays, since it is complete 

 when the radiation leaves the radiator, to say nothing of 

 the difficulty of understanding how outer layers of a sub- 

 stance could sift rays emitted by inner layers of the same 

 substance in the same condition. The effect cannot be 

 due to anything like selective reflection, for then the 

 secondary would be strongly turned back by screens of the 

 same substance. For a similar reason it cannot be a true 

 secondary. It must therefore be a transformed primary, 

 transformed not by the conversion of primary energy into 

 energy of secondary vibrations, but by a true change in 

 its own properties. What can be suggested on the pulse 

 theory as to the nature of this process? .Again, in the 

 case of the primary rays, a screen of any one substance 

 has in nearly all cases the power of rendering the rays 

 more penetrating to all other substances, but especially to 

 that substance. It is true that this can be explained by 

 sifting alone, e.g. a substance .\ might absorb soft rays, a 

 substance B medium rays, and neither hard rays. But it 

 can also be explained by true transformation of the primary 

 as M. Sagnac and Dr. Walter have suggested, yet the 

 transformation must not be accompanied by much scatter- 

 ing of the new radiation. I am aware that Mr. .Adams 

 (Amcr. Journ. Sci., xxiii., p. 376), unlike M. Sagnac, did 

 not find any effect due to reversing two screens, but I am 

 inclined to think that there is really some transformation 

 of this sort. If that is so the effect will be very hard, if 

 not impossible, to explain on the pulse theory. It is con- 

 ceivable on a neutral pair theory, since the pair has proper- 

 ties which can be altered without disturbing the velocity 

 and line of flight, so that the primary can be transformed 

 without much scattering. Whether these surmises are 

 correct or not, it seems to me that these particular pheno- 

 mena give no support to the pulse theory. 



Finally, it may be pointed out that the pulse theory will 

 need radical alteration if it is to explain the asymmetrica' 

 effects which Dr. Madsen and I have lately investigated, 

 and it is not clear that the revised theory will fit Dr. 

 Barkla's experimental facts even as well as it does now. 



W. H. Bragg. 



The University of .Adelaide, South .Australia, June 25. 



The Discovery of the Weight of the Air. 



The discovery, in the first half of the seventeenth 

 century, that the air has weight is associated with things 

 of immense importance, for instance, the invention of the 

 barometer and the refutation of the dogm.a — dear to the 

 false science and the false philosophy of the day — that 

 " Nature abhors a vacuum." In a new edition of the 

 " Essais de Jean Rey," reviewed in Nature of July 9, an 

 attempt is made to assign this discovery to Rey, and, so 

 far, to regard Torricelli, Galileo, Pascal, and Descartes as 

 his disciples. Without claiming to be an authority upon 

 Rey or upon Galileo, I would direct attention to the state- 

 ment, made in " Galileo — his Life and Work," by J. J. 

 Fahie, that Galileo's way of determining the specific gravity 

 of the air was first described in his letter to Baliani dated 

 March 12, 1613. Rey's " Essais " was published in the 

 year 1630. 



Erroneous suppositions regarding Rey being frequently 

 made, I inay be allowed to quote Humphry Davy's de- 

 scription of the " Essais " as " a mere logical exercise in 

 physical science." The fact that metals on calcination in- 

 crease in weight was known to Cardan, .Scaliger, Fachsius, 

 Cresalpinus. Hamerus Poppius, and Libavius, who are 

 mentioned by Rey. For confirmation of the fact he relied 

 upon the statements of his friend " le sieur Brun," and 

 altogether the evidence th.nt Rey made experiments of 

 any value in support of his doctrines is slight. 



.A. N. Meldrum. 



Tannachie, Whitburn, Linlithgowshire, July 11. 



