July 28, 192 1] 



NATURE 



683 



£3 



tfH = — 3-7- («H=radius of the electron in the normal 

 47rVw 

 state). 



= 0*532 X io~* cm. 



and 

 2 

 — =^(13-54 volts). Therefore for an element x with 



an ionisation potential of V,; we have 



(0-532 X 13-56)^^0-8 cm. 



Thus the atomic radius varies inversely as the ionisa- 

 tion potential. 



The atomic radii calculated according to this for- 

 mula are, in general, smaller than the atomic radii 

 calculated either from crystal data (Bragg) or from 

 the kinetic theory qf gases (Rankine and others). 

 They are shown in the appended table for the sake 

 of comparison : — 



Atomic Radius. 



_FromI.P. Crystal measures Viscosity 

 aXicfi IrXiifi data 



Eement I. P. 



H 13-54 0-530 — 



He 2540 0-28 — 



Ne 2280 0-33 065 



(Horton, Phil. Mag., May, 192 1) 



Li 5-40 1-34 1-50 



Na 511 1-41 177 



K 432 167 207 



Rb 416 1-73 2-25 



Cs 3-88 1-86 2-37 



108 



lOI 



TI 730 099 2-25 — 



Mn 738 098 1-47 — 



The values of V^. for copper, silver, and gold have 

 been calculated from Hicks's value of the (iS) term 

 for these elements. That for manganese has been 

 similarly calculated from Mr. Catalan's value of (iS) 

 for manganese (not yet published). For these data I 

 wish here to record my indebtedness to Prof, Fowler 

 and Mr. Catalan. The sources for the other values 

 are quite well known. Megh Nad Saha. 



21 Cromwell Road, London, July 13. 



American and British Superannuation Systems. 



I READ with great interest the article in Nature of 

 June 30 on the American and British superannuation 

 systems. The selection of a satisfactory scheme of 

 superannuation is a matter of great importance in 

 the organisation of a public service. On the one 

 hand, an age limit can be effectively enforced only 

 when suitable provision is made for those who are 

 forced to retire, and on the other the provision of a 

 pension conditional on the completion of a full term 

 of service is objectionable, because a public servant 

 who retires before that period is completed is penalised 

 by the loss of a portion of the consideration for which 

 he has given his labour. The result is that although 

 a man may feel that he would do better work in 

 another sphere, and has an opportunity of doing so, 

 he cannot bring himself to fore;-o the pension towards 

 NO. 2700, VOL. 107] 



which he has already contributed some years of 

 service. 



The recent Committee of the British Science Guild 

 on the Utilisation of Science in Public Departments 

 considered this question, and came to the conclusion 

 (Journ. Br. Sci. Gd., June, 192 1, p. 37) that the best 

 solution appeared to be to award at the end of every 

 year's service a pension (or alternatively an endowment 

 insurance) accruing at the age fixed for superannua- 

 tion (or in the case of the insurance at that age or 

 previous death), independently of whether the officer 

 had remained in the service or not. The advantages 

 accruing in respect of a single year's service would, of 

 course, "be comparatively small, but those for succes- 

 sive years would, when added together, furnish an 

 adequate provision for the old age of officials who had 

 served the full term, while they would be a welcome 

 addition to the resources of those whose later careers 

 had followed other directions. 



It is essential that these benefits should be secured 

 by public funds, and based on actuarial calculations at 

 current rates of interest. The amounts now quoted 

 by insurance companies are apparently calculated on 

 pre-war rates, and are far too low. 



This scheme could be adopted whether the basis 

 of the superannuation were contributory or not. 



July 19. John W. Evans. 



May I point out, in connection with the note ap- 

 pended to my letter printed in Nature of July 21, 

 p. 651, that if only one mutual life assurance company 

 were available the argument quoted in the leading 

 article of June 30 would be answered, for that argu- 

 ment implied that dividends necessarily go to share- 

 holders? The remark about expenses in the note 

 leaves the point of paragraph (3) of my letter un- 

 touched, and the final sentence of the note makes me 

 wonder whether the two-year-old American "Teachers' 

 Insurance and Annuity Association " will grow up and 

 prove itself to be more "philanthropic" than the 

 selected assurance companies in England. 



I am afraid that, quite unintentionally, my former 

 letter must have seemed offensive to have justified 

 your note to it. The hot, dry weather has been, and 

 still is, trying to us all, et tout comprendre c'est 

 tout pardonner! W. Palin Elderton. 



July 22. 



A Novel Magneto-Optical Effect. 



In connection with the very interesting observations 

 communicated by Dr. R, Whytlaw-Gray and Mr. 

 J. B. Speakman (Nature, July 14, p. 619), I should 

 like to point out the close similarity of the phenomena 

 which they have observed with those observed in the 

 case of soap solutions (Proc. Roy Soc, A, 192 1, vol. 

 xcviii., p. 395; and Journ. Chem. Soc, 1920, vol. 

 cxvii., p. 1506). 



Gray and Speakman describe the formation of 

 flexible strings or fibres in clouds of various metallic 

 oxides, these fibres being made up of particles of col- 

 loidal dimensions which stil! retain their individuality. 

 Miss Laing, in her study of gelatinisation, was led 

 to the conclusion that such conjunction or orientation 

 of colloidal particles forms the mechanism of gelatini- 

 sation. For instance, in a soap solution the individual 

 colloidal particles are otherwise the same in the 

 liquid sol as they are in the elastic jelly. In the letter 

 referred to it is pointed out that the particles in a 

 cloud of cadmium oxide have an exceptional tendency 

 to form such strings, and this agrees with the striking 

 behaviour of Svedberg's sols of the same substance 

 in alcohol, which on standing gelatinise, but on slight 

 shaking revert to the fluid condition, an alternation 

 which can be indefinitely repeated. 



