412 



NATURE 



[December 25, 1919 



LETTERS TO THE EDITOR. 



[The Editor does not hold himself responsible for 

 opinions expressed by his correspondents. Neither 

 can he undertake to return, or to correspond with 

 the write,rs of, rejected manuscripts intended for 

 this or any other part of Nature. No notice is 

 taken of anonymous communications.] 



Polarisation of Light Scattered by Helium Atoms. 



About a year and a half ago I published an experi- 

 mental investigation of the degree of polarisation in 

 the light scattered at right angles by various dust-free 

 gases (Proc. Roy. Soc., A, vol. xcv., p. 155). 1 believe 

 the results then obtained to 1^ in the main quite 

 correct, but there is an important point on which I 

 have completely to withdraw what 1 then said. This 

 refers to the results on helium, which was then found 

 to behave differently from the other gases, giving 

 much less complete polarisation than any of them. 

 The result given was that the weak image (vibrations 

 parallel to incident beam) had 42 per cent, of the 

 intensity of the strong image (perpendicular vibra- 

 tions). This was given on the results of two inde- 

 pendent series of photographs, which were, indeed, 

 obtained under conditions much more difficult than 

 those for the common gases, but w^ere considered at 

 the time to give adequate evidence. I do not even 

 now know what was wrong with them ; but on 

 repeating the work with a much improved apparatus, 

 which it has taken many months to design and con- 

 struct. I have obtained an entirely different, I might 

 say opposite, result. 



I now find no intensity large enough to be observed 

 in the weak image, and certainly not 3 per cent, of 

 the intensity in the strong image. It may be possible 

 to lower this limit still further; but, in anv case, if 

 helium is outstanding at all, it is in the direction of 

 polarising more, and not less, completelv than the 

 Srenerality of gases. The details will be published 

 later, but I write to make the correction as soon as 

 possible, so that no one who speculates theoreticallv 

 on the subject may be misled by reliance on mv former 

 result. ' Raylrigh. 



December 21. 



Gravitation and Light. 



It should perhaps l)e stated, in connection with Mr. 

 Cunningham's remarks (Nature, December iS, 

 p. 395), that my difficulty with regard to Dr. Ein- 

 stein's theory must extend to the deviation of light 

 by the sun as well as to its change of period. Ac 

 cording to the theory, the velocity of light diminishes 

 near the sun; on the other hand, the scale of time 

 is increased, so that the wave-length is not altered. 

 Now. the space being nearly flat, the path of a rav 

 is, with such heterogeneous time, determined funda- 

 mentally by minimum number of waves, and not bv 

 minimum time; therefore, it should not be altered. 



On the other hand, passing from kinematics to 

 dynamics, Dr. Einstein requires in another connection 

 that light should consist of discrete bundles or quanta 

 of energy. I^t it afso be granted that inertia and 

 gravitation are attributes of all energy. It seems to 

 follow that each of these bundles of energv will swing 

 round the sun in a hyperbolic orbit, and that its 

 velocity will be increased when near the sun. It is 

 well known that this would account for half the 

 observed deflection. But, again, physical optics could 

 not exist without the idea of transverse waves and 

 their phases, which must be grafted on somehow to 

 the bundles of energv. Now the supposed gravita- 

 tional derangement of the fourfold extension from the 



NO. 2617, VOL. 104] 



flat being very slight, it can be agreed that the change 

 in extent of each element of it is of the second order. 

 The expansion of scale of time near the sun requires- 

 thus a compensating shrinkage of radial lengths ; and 

 this second-order effect, the cause of the adjustment 

 for Mercury, will, on the phase-principle of Huygens,. 

 just double the previous result. This would amount in 

 all to the observed deflection of the rays. 



But amid these uncertainties and apparent contra- 

 dictions the view asserts itself that the very important 

 astronomical determination is to be regarded as a 

 guide towards future theory rather than as the verifica- 

 tion of the particular theory which suggested it. 



Joseph Larmor. 



Cambridge, December 20. 



Radio-activity and Gravitation. 



In connection with the interesting letter of Prof. 

 Donnan in N.wure of December 18, it may be of 

 interest to mention some experimental results which 

 have a bearing on this question. Some years ago 

 Dr. Schuster suggested to one of us that it would be 

 of interest to test whether the rate of transformation 

 of radio-active substances was influenced by the 

 intensitv of gravitation. .\n accurate method of 

 testing the rate of decay of radium emanation over a 

 period of about a hundred days was developed, and 

 it was intended to compare the rate of decay of 

 samples which had been transported to suitable por- 

 tions of the earth's surface. The outbreak of the war 

 interfered with this plan. 



Since, according to Einstein's theory, a gravitational 

 acceleration is in no sense different from a centrifugal, 

 acceleration, experiments have been performed in the 

 Cavendish Laboratory to test whether the rale of decay 

 of radio-active substances is affected by subjecting 

 them to the high centrifugal acceleration at the edge 

 of a spinning disc. For the purpose of measurement 

 the 7-ray activitv was determined by a sensitive-balance 

 method. .Although the radio-active material was sub- 

 jected to an acceleration of more than 20,000 times 

 gravity, the change observed, if any, was certainly l^s~ 

 than one part in a thousand. 



This result is not in disaccord with the relation- 

 deduced bv Prof. Donnan, for a simple calculation 

 shows that his relation predicts an effect very much 

 smaller than can be detected by measurements of this 

 character. 



E. RuTHERKORn. 

 A. H. COMPTOX. 



Civendish Laboratory, December iq. 



Mortality among Snails and the Appearance of Blue- 

 bottle Flies. 



The residential parts of Calcutta are remarkably 

 free, as a rule, from both house-flies (Musca. 

 spp.) and blue-bottles. This is doubtless due to the 

 excellence of the municipal sanitary arrangements, for 

 at Sibpur, a few miles away, blue-bottles (Pycnonoma 

 or Lucilia dux) are not only extremely troublesome in 

 the houses, but are also prob.-ibly connected with fre- 

 quent epidemics of enteric, unknown in the better parts 

 of Calcutta. For some years past I have noticed in the 

 compound of the Indian Museum that Pycnonoma 

 from time to time becomes relatively numerous, and 

 on several occasions I have been able to trace the flies 

 to their breeding-ground. This has always been the 

 dead bodies of the snail .Ichatina fiilica. the largest 

 land mollusc in Bengal. 



A. fulica, the shell of which may attain a length 

 of at least 4 in., is not an indigenous species, but 

 was introduced for purposes of dissection by a keen 



