582 



NA TURE 



[May 6, 1922 



Perhaps the simplest way to test the postulate 

 directly would be to observe the Doppler effect with 

 a concave reflecting grating so set as to form the 

 image on the normal to the surface of the grating 

 (cf. Tolman, Phys. Rev., 35, p. 136, 1912) ; the 

 retardation then occurs entirely before reflection, and 

 it is the wave-length of incident light which is measured 

 by the deviation. Any uncertainty as to the relative 

 speed of the reflected rays can be removed by making 

 the line of motion of the source pass through the 

 centre of the grating, and then observing the effect 

 of the motion upon the position of the central image 

 when the grating is turned so as to bring this image 

 into the position formerly occupied by the diffracted 

 one. In these circumstances, for reasons of symmetry 

 the speed of the incident waves along two rays 

 equally inclined to the direction of motion must be 

 the same ; if it then turns out that the position of 

 the central image is unaffected by the motion, it will 

 follow that the speed must likewise be the same along 

 the two corresponding reflected rays. This conclusion 

 will hold also for the two diffracted rays which take 

 these paths in the main experiment. 



E. H. Kennard. 



Department of Physics, Cornell University. 



On the N-Series in X-Ray Spectra. 



With the new and very powerful X-ray-spectro- 

 scopic outfit constructed by Prof. M. Siegbahn 

 (described in Comptes rendus, 1921, p. 1350) I have 

 endeavoured to find a weaker group of lines in the 

 X-ray region than the lines previously known as 

 M-group. I have been able to find some lines which 

 most probably must be referred to the N-series of 

 the elements uranium and thorium. Hitherto, the 

 measured wave-lengths for these lines lie for uranium 

 between ^ 8-6-12 -o A.U. and for thorium between 

 9-4-I3-5 A.U. 



From the theoretical and experimental work done 

 by Coster and others, we are able to estimate the 

 wave-lengths of the lines in the N-series. For the 

 elements uranium and thorium we really find that 

 some of these lines must have wave-lengths of about 

 the measured value. For bismuth, however, and 

 the elements in its neighbourhood, all the N-lines 

 must have a wave-length of more than 13 A.U. so 

 that in the present state of spectroscopy it will be 

 very difficult to measure the wave-lengths for these 

 elements. 



I am continuing these researches. 



V. DolejSek. 



Physical Laboratory, The University, 

 Lund, March 31. 



A Proposed Laboratory Test of the Theory of 

 Relativity. 



With the present interest so strong in devising 

 experiments to test the theory of relativity, it 

 may not be amiss to suggest the possibility of yet 

 another method. According to recent hypotheses, it 

 seems that the stars are the factories producing com- 

 plex elements from simpler structures. Inside the 

 stars, hydrogen atoms may unite to form helium, and 

 with hydrogen and heUum as intermediates, the more 

 complicated atoms may be built. As pointed out by 

 Harkins, Eddington, Perrin, and others, the synthesis 

 of an atom of helium from four hydrogen atoms necessi- 

 tates the loss of 0-774 per cent, of the mass of the 

 hydrogen atoms. Since we cannot conceive of mass 

 being annihilated, the only obvious solution is to say 



that mass is electromagnetic in origin and that, in 

 the helium nucleus, the four protons are brought so 

 near to the two electrons that their fields overlap and 

 neutralise each other to some extent, accompanied by 

 a loss of mass. According to the theory of relativity, 

 I gram of matter is equivalent to 9x10^" ergs or 

 2-1 X 10" calories. Both Harkins and Perrin have 

 calculated the amount of heat that must be produced 

 by the transformation of four gram atoms of hydrogen 

 into one gram atom of helium. It has the enormous 

 value of 0-0078 X 2-1 X 10" or 1-6 x 10" calories. 



It may be possible for several helium nuclei to unite 

 to form heavier nuclei, such as oxygen for example, 

 without such a great evolution of heat. More accu- 

 rate determinations of the atomic weights of the so- 

 called " pure " elements would be necessary before 

 we could say much concerning the energy relations 

 in such sub-atomic reactions. 



When the nuclei become so large that they are 

 unstable, then the process of synthesis in the stars 

 would stop. But there might be an over-shooting of 

 the mark. With the enormous amount of energy free 

 in the interior of the stars, some of this energy might 

 be absorbed, according to the theorem of Le Chatelier, 

 in the formation of nuclei which would be unstable 

 in an environment not containing so much energy. 

 Energy would be considered as one of the terms in 

 a mass law equation, to use a well-known chemical 

 analogy. The result would be the radioactive ele- 

 ments — uranium, thorium, etc. 



Now let us calculate with the aid of the above 

 equation, derived from the theory of relativity, the 

 effect on the mass of a radioactive substance that 

 would be caused by this addition of energy. Ruther- 

 ford, in his book " Radioactive Substances and their 

 Radiations," p. 582, states that i gram of radium in 

 disintegrating to lead gives off 3-7 x lo* calories. If 

 I gram of mass = 9 x 10*" ergs = 2-1 x 10^^ calories, 

 then I gram of radium in disintegrating to lead would 

 give off 0-00017 gram and i gram atom of radium, 

 0-038 gram in the form of energy. If the atomic 

 weight of RaG (radium-lead) is taken as 206 exactly, 

 then the atomic weight of its parent, radium,- may be 

 calculated as follows : 



I gram atom of RaG 

 5 gram atoms of He 

 4 gram electrons . 

 3-7 X 10* calories . 



206-000 grams 

 20-000 „ 

 0-0005 t. 

 0-038 



226-038 



Therefore the atomic weight of radium should be 

 226-038. Calculations of this type for radioactive 

 substances have been made by Harkins, but he does 

 not state that they may be applied as a test of the 

 theory of relativity. 



This calculation involves six assumptions : (i) that 

 the weight of one gram atom of RaG is 206-000, 

 (2) that the atomic weight of He is 4-000, (3) that 

 the weight of 4 gram electrons does not exceed o - 0005 

 by any great extent when incorporated in the nucleus 

 of Ra, (4) that the amount of energy given off in the 

 disintegration of Ra is substantially that calculated 

 by Rutherford (a 20 per cent, decrease in the value 

 given by him would not change the value for energy 

 in grams in the second decimal place), (5) that the 

 relativity equation connecting mass and energy holds, 

 and (6) that the energy given off in radioactive dis- 

 integrations is derived from the atoms themselves and 

 not photochemically from Perrin's hypothetical radia- 

 tions of extremely short wave lengths. In trying to 

 verify the results of such an equation, there are two 

 more assumptions necessary : that the atomic weights 

 of RaG and of Ra are determined for the pure sub- 

 stances, that there are no contaminating isotopes. 



NO. 2740, VOL. 109] 



