April 22, 1920] 



NATURE 



233 



ionisation, we can apply Nernst's theorem of the 

 "reaction-isobar" to calculate the amount of ionisa- 

 tion under any given thermal stimulus. The method 

 is based upon a remark of Nernst in his book, " Der 

 Neue Wiirmesatz ..." (p. 154), that the electron 

 may be regarded as a monatomic gas of molecular 

 weif^ht Ys'sff' ^nd that its chemical constant can be 

 (alculated according to the Tetrode-Sackur relation 



C = log^ 



It has recently been applied by Eggert {Ver. d. D. 

 Phys. Gesell., December 15, 1919) for the calculation 

 of the degree of ionic dissociation in the interior of a 

 star, as supposed by Eddington in his theory of stellar 

 structures. But Eggert calculates U in a rather 

 artificial manner for iron from assumed atomic 

 dimensions and structures of the iron atom. 



We can, however, calculate U directly from the 

 \alue of the ionisation potential as experimentally 

 determined by Franck and Hertz, MacLennan, and 

 others, or from the quantum relation 



V=- 



^(i-S,-^). 



Using the value of U determined in this way for 

 calcium, barium, strontium, hydrogen, and helium, 

 1 he following remarkable results appear: 



(i) About 30-40 per cent, of the Ca atoms are ionised 

 just over the photosphere ; in the chromosphere, when 

 the pressure falls to lo"" atms., almost all the Ca 

 atoms are ionised. The same conclusion holds to a 

 varying degree for Ba and Sr. 



(2) Hydrogen and helium are not ionised anywhere 

 in the solar atmosphere. (This is due to their high 

 ionisation potential. V is 13-6 for H and 205 for He, 

 while for Ca, Sr, and Ba the figures are 6- 12, 57, and 

 512.) Helium can become ionised only in stars of 

 which the temperature exceeds 16,000 K. 



(3) Pressure has a great influence on ionisation, a 

 duction in pressure causing great enhancement of 



.^'iiisation. 



It therefore appears that the ionisation in the upper 

 layers of the solar atmosphere, as revealed by the 

 enhanced lines of Ca, Sr, and Ba, and probably also of 

 Fe, Ti, and Sc, is due to reduced pressure and the 

 low ionisation potentials of these elements, and not 

 to an increased temperature. 



The full theory has been worked out in a paper 

 communicated to the Phil. Mag. M. N. Saha. 



University College of Science, Calcutta, 

 March 8. 



Gravitational Deflection of High-speed Particles. 



In a letter published in Nature of March 11 Prof. 

 Eddington has shown that the statement made by me 

 in an earlier letter to the effect that Einstein's law 

 of gravitation seems to lead to a zero deflection for a 

 material particle moving with the velocity of light is 

 not in accord with the exact equation of the orbit 

 contained in his report to the Physical Society, and 

 suggests that my approximations vvere not sufficiently 

 (lose to warrant mv conclusion. The line element 

 from which Prof. Eddington derives the equation of 

 the orbit is expressed in co-ordinates which make the 

 velocity of light different in different directions at anv 

 one point, whereas the one used by me requires that 

 the velocity of light should be a function of position 

 only, and not of direction. In terms of my co- 

 ordinates the equation of the orbit of a particle 

 moving with the velocity of light is 



NO. 2634, VOL. 105] 



which leads to the same deflection 4— for a material 



R 

 particle moving with the velocity of light as for a 

 light-ray. Hence it is clear that my previous con- 

 clusion was based on an insufficiently close approxima- 

 tion, and therefore erroneous. 



I am glad to see that Prof. Eddington has verified 

 the other principal conclusion of my letter. 



Leigh Page. 

 Sloane Laboratory, Yale University, New 

 Haven, Connecticut, March 29. 



Science and the New Army. 



It requires some courage to offer any opposition to 

 the chorus of approval which has greeted the sug- 

 gestion that a proportion of officers endov^-ed with the 

 scientific spirit should be included in the General Staff, 

 but I venture to think that it is by no means so easy 

 to give effect to this proposal as some correspondents 

 in Nature seem to suppose. No doubt it would be 

 delightful if we could have Staff officers who knew 

 all about everything, but in actual practice the man 

 who does useful work in the world is a specialist in 

 one particular subject or in one particular branch 

 of work. 



A good regimental officer lequires a particular kind 

 of training and possesses a certain set of qualifica- 

 tions. Similarly, a good Staff officer requires a 

 different training and possesses a different set of 

 qualifications. A man of science, again, is different 

 from either of the other two. 



The proportion of officers in the Army as a 

 whole who possess any scientific training is 

 comparatively small. There are a certain number 

 of specialists whose ordinary duties are of a technical 

 nature, and there are a few officers who take up 

 some branch of science as a hobby, but the work of 

 the average officer is not such as to bring him into 

 touch with scientific thought and scientific methods. 

 Men are to be found who are good Staff or regimental 

 officers and also scientific workers, but they are 

 exceptions, and it seems to me that a system which 

 demands a regular supply of exceptional men is not 

 one which is likely to work in practice. 



There is also a further difficulty. Granting, 

 for the sake of argument, that there are sufficient 

 officers in the Armv who possess both the scientific 

 spirit and the qualities necessary for potential Staff 

 officers, it is still necessary to devise a method of 

 selecting them from, their more ordinary fellows. 

 Two methods are in common use, namely, examina- 

 tion and nomination. 



An examination is a good method of testing that 

 form of knowledge which is acquired by study, but it 

 will be generally agreed that it is not a f^ood method 

 for detecting the scientific spirit. The difficulty in the 

 case of nomination is that the candidates must be 

 selected by ordinary regimental officers who can alone 

 be acauainted with the qualifications of the individual 

 candidates. The average regimental officer, however, 

 is not himself a man of science, and I cannot see that 

 he can ever become a judge of another officer's 

 scientific attainments. 



Without arguinj?, therefore, against the desirability 

 of a General Staff containing an appreciable propor- 

 tion of scientific officers. T suggest that the ideal is 

 unattainable except in so far as specialists are attached 

 to the Staff for their own particular work, and I 

 think the object in view must be attained by some 

 other means. It might be done by raising the general 

 standard of education in scientific matters throughout 

 the country, but this is a very large question, and 

 not a very easy one. 



