144 



NATURE 



[October 7, 1915 



case nor in the other can the equation (3) be made 

 to prove the quoted statement with which this article 

 begins. 



Although fhe above considerations are theoretically 

 of profound importance, they will obviously have no 

 influence whatever on astronomical calculations ; for 

 in all such calculations the product of mass Into New- 

 tonian constant is used. Suppose in the case of the 

 sun G,s>G.M and let S, S' be the true sun's mass and 

 its present accepted value, then 



Gs.S = Gm,.S'. 



In no calculation do we require S alone, but always 

 product Gs . S, so that the false values for the sun of Gs 

 and S do not lead to error. 



P. E. Shaw. 

 University College, Nottingham. 



The Spectrum of X-Rays. 



(i) KossEL (Verh. d. D. Phys. Ges., 1914), discuss- 

 ing the previous measurements of the frequencies of 

 the corresponding lines of the spectrum K and L 

 characteristic for different elements, as well as of the 

 limits of the respective bands of absorption, established 

 certain numerical relations among them which Wagner 

 {Ann. der Physik, 1915) confirmed in a more perfect 

 way and even generalised. In addition, the extremely 

 precise measures of Bragg and Pierce (Phil. Mag., 

 October, 1914) for the lines Ka, Kp, Ky in the Ag, 

 the Pd, and the Rh give additional confirmation. 



From among these relations, attention is specially 

 directed to the one obtained by generalising the second 

 law of Kossel — 



La = Kj3-Ka, viz., (l)Ma = Ly-La 

 In these^ equations each term expresses the frequency 

 of the lines. With its help we can foresee the fre- 

 quencies of the principal lines of the spectrum of the 

 M characteristic if Ly and La are known. 



If we admit Bohr's model for the atom as modified 

 by Kossel— that is to say, supposing that in the transit 

 of an electron from one ring to another a quantum 

 Is emitted, the frequency of which Is determined by 

 Bohr's formula, the radiation M will correspond to 

 the transit of an electron from the fourth to the third 

 ring, and the corresponding frequency, 



'' = ''o(3,-^)(N-ay=i.59Xio»(N-..)' 



will be emitted. 



To demonstrate the exactness of this prediction, I 

 have calculated the values of v contained in the follow- 

 ing table by means of the formula (i), starting from 

 the values of La and L^ given by Moseley, Ruther- 

 ford, and Andrade. The last row contains the values 

 of V obtained from the formula — 



(2) 1/= 1.39 X io^^(N- 2 1.8)2 



deduced by graphic Interpolation, where N is the 

 atomic number of the element. 



N ^cal. by(,) vcal. by(2) 



Pd 46 SixiqI^ 82x10^* 



Ce 58 189 182 



Eu 63 24s 236 



Ta 73 364 365 



Pt 78 438 440 



Au ... ... 79 452 455 



Kadmm B ... 82 501 505 



As can be seen, the formula (2), of similar type to 



those found by Moseley for Ka and La, show's well 



enough the change of Ma with N, and besides the 



constant 1-39 x 10^* coincides with the theoretical value 



NO. 2397, VOL. 96] 



1-59 X 10^*, between the limits to be expected on account 

 of the method of calculation. No data are at present 

 available to test this hypothesis. The spectrum M is 

 most likely to be observed in heavy elements and in the 

 radio-active bodies. 



(2) In the numerical relations obtained by Kossel and 

 Wagner, there does not appear the groups Laj, L^, 

 and Ls of the spectrum L, nor does there seem any 

 simple way of introducing them In another series of 

 formula Involving the frequency of the K radiations. 

 This suggests the possibility of the spectrum being 

 composed of two distinct parts, which might be Illus- 

 trated by supposing that two rings are situated In 

 distinct planes, from one of which (the one correspond- 

 ing to L2) no electron can pass to the ring K. The 

 correctness of this hypothesis could be tested by inves- 

 tigating experimentally the lines of the L spectrum 

 that accompany the spectrum K- — a study that does 

 not seem impossible. 



It is logical to suppose that this group L, should 

 be accompanied by others, M,, N,, ... In the ex- 

 terior spectrum, and even that In these should appear 

 some new ones — M3, N3 . . . N4 . . . In this way it 

 would be easy to understand how the spectrum gets 

 more and more complicated from the X-ray spectrum 

 to the visible, and even the existence of the well- 

 known series In the light spectrum. 



B. Cabrer.'V. 



Laboratorio de Investlgaclones Fi'sicas, 



Palacio de la Industria (Hip6dromo), Madrid. 



mignt easily oDtam, 

 ly two-thirds of his 

 chapter, the title of J 



Porlirft-ofi " ie rtlr\rt£t. -^ 



Studies of the Cotton Plant. 



I SHOULD be sorry If any reader of Nature were 

 to be prejudiced against perusal of my book on applied 

 plant physiology, called "The Development and Pro- 

 perties of Raw Cotton," by thinking that I had 

 attempted to write on the systematic botany of 

 Gossypium. Yet this Impression might easily obtain, 

 since your reviewer devotes exactly 

 review to a few pages of my first 



which, "The Development of Pedigree," Is alone 

 sufficient to indicate that it was not intended for 

 biologists. 



Still, technological treatment Is no excuse for in- 

 accuracy, even as enthusiasm for herbaria is no excuse : 

 for such a phrase as " peripatetic Mendellan cross- 

 breeding of undetermined stocks." Therefore I would 

 direct your reviewer's attention to the fact that my 

 description of Gossypiae as a sub-order (instead of a 

 tribe) is taken from an accepted authority on 

 systematy. 



He states that " the results and evidences of 

 systematic botany " with regard to cotton have been 

 before us for a long while ; the Hindi weed-cotton to 

 which he purposely refers is a sad example of the 

 result. I regret being obliged to mention It, but 

 he misspells It as " Hindu " weed, and then objects 

 strongly to my leaving it with such a loose designa- 

 tion. Now, in the first place, this name is on record 

 in Sir George Watt's standard monograph of the ^' 

 genus, and is therefore as definite as any other. Pos- 

 sibly more definite, for though the specimen used for 

 that determination came from a pure strain. Sir G. ^i 

 Watt called It " a ferine hybrid possessed of a strain 

 of G. vitifolium.'" Only by enlisting the aid of "peri- 

 patetic" Mendelism can one justify the apparent con- 

 tradiction, but nothing a priori can excuse the inclu- 

 sion of this Hindi weed in the " fuzzy-seeded cottons " 

 during the primary division of the genus In that mono- 

 graph. There is no cotton which has a more naked 

 seed than Hindi. 



It is clear that I took too much upon myself in 

 attempting to persuade the growers and spinners of 



