210 



NATURE 



[October 13, 192 1 



The first column contains measured wave-lengths 

 of the Pickering series in the star, the second the 

 computed values, the third contains the measured 

 wave-lengths of the Balmer series in the star, and 

 the fourth the residuals, star — laboratory, using 

 Curtis 's wave-lengths. In view of the closeness 

 of the agreement between the observed and com- 

 puted values, there can be no doubt as to the 

 existence of the helium components of the Balmer 

 series, and hence no doubt that the Pickering series 

 is due to enhanced helium. 



Using these preliminary wave-lengths, corrected to 

 vacuum, the value of the constant Nhc in Bohr's 



formula, « = NhcI -^ — „ ), for the Pickering series 



has been computed for all lines except those with 

 footnotes. The weighted mean, using in addition 

 the measured wave-length 468576, comes out 

 NHe = 438 890-3, with a total range of i6-8. From the 

 well-known formula : 



m^_ Nhc-Nh 



M NH-iNHfe' 

 arising from the correction to Bohr's simple theory on 

 account of the motion of the nucleus, the mass of 

 the electron «»„ in terms of the mass M of the 



The revised values 



hydrogen atom comes out —^ ;. 



*^ 1851-6 



of the wave-lengths which will result from additional 

 plates and re-measures should give a value of the 

 electron mass of a high degree of accuracy. 



In a recent paper (Proc. Roy. Soc., A, vol. 99, p. 135, 

 192 1) Saha has predicted that at about the stage Ob 

 in stellar spectra the Balmer series disappears and 

 is replaced by the Pickering lines 6560, 4859, 4339, 

 4100. Reference to his tables shows that the 

 enhanced Mg line 4481 should disappear at a stilt 

 higher temperature than the Balmer series. In the 

 star 9 Sagittae the line 4481 has almost disappeared. 

 In fact it has only been glimpsed on one of several 

 plates. However, at H7 there are lines at 43404 

 (Hy) and 4338-8. Further, Hy is about six times as 

 intense as the Pickering component. This non-veri- 

 fication of Saha's prediction is scarcely surprising 

 when it is recalled that the nebulae, which are prob- 

 ably more advanced than O type stars, show the 

 Baimer series (Lick Observatory Publications, vol. 13). 



H. H. Plaskett. 



Dominion Astrophysical Observatory-, 

 Victoria, B.C., September 17. 



Radiation and Chemical Action. 



In the Philosophical Magazine for November, 1920, 

 Prof. Lindemann has shown that if Prof. W. C. McC. 

 Lewis's hypothesis is true, i.e. that the velocity of a 

 chemical reaction is proportional to the intensitv in 

 the system of the radiation of a wave-length which 

 is characteristic of the reaction, then in the case of 

 the inversion of cane-sugar by dilute acid, exposure 

 to the radiation of the sun should increase the reaction 

 velocity 5 x 10" times. 



Recently Prof. Lewis has suggested that any such 

 effect would not be true for the conditions ' under 

 which such a reaction is usually carried out, because 

 the activating rays lie in the region of i/m, and at this 

 wave-length water, the solvent, would absorb the 

 radiation almost totally in the first thin layer, so that 

 the bulk of the liquid would remain unaffected. 



In order to test the radiation hypothesis, paying 

 heed to Prof. Lewis's suggestion, the following simple 

 experiment has been carried out. A solution con- 

 NO. 271 1, VOL. 108] 



taining 100 gr. of cane-sugar and 365 gr. of hydrogen 

 chloride per litre was made up and divided into two 

 portions. One was left to stand indoors at room- 

 temperature (14-7° C), while the other was forced 

 upwards through four fine jets made from 3 mm. 

 diameter glass tubing drawn out to the narrowest of 

 capillaries ; this treatment took place on the roof in 

 full sunshine. The liquid from the jets formed fine 

 columns about 8 cm. high, which then broke up into 

 clouds of small drops, the drops rising a further 

 40 cm. They were allowed to fall, and collected in a 

 large dish ; the temperature of the resulting liquid 

 had risen to 19-1° C. 



Now for the cane-sugar at the concentrations used 

 the period required for half to be inverted is- at 20° C. 

 from Lewis's measurement of the velocity constant, 

 I -6 X 10" seconds, or approximately 47 hours. On 

 Lewis's hypothesis, illumination b\' the sun should 

 reduce this to 3 x 10-" seconds ; hence it was reasonable 

 to expect some indication of a more rapid change in 

 the exposed portion if the hypothesis were true. Two 

 polarimeter tubes were filled, one with the exposed 

 portion, the other with the unexposed. The rotations 

 measured were : — Unexposed solution, 13-06° ; exposed 

 solution (measured 2 min. later), 12-82°. In other 

 words, illumination by the sun had had no appreciable 

 effect on the reaction velocity. 



The size of the drops in which the one portion had 

 been exposed was measured later by catching drops of 

 the same solution formed under ihe same conditions 

 on black velvet and measuring them under a micro- 

 scope. They were of pretty constant size, the 

 average diameter being 0-015 cm. It is thus difficult 

 to imagine that the absence of any effect was due to 

 absorption by the solvent. Further, from Stokes's 

 formula the size gives an average time of fall 

 0-68 second, so that the time of exposure is 

 ridiculously ample. For even if the radiation density 

 at ifjL had been reduced inside each drop, by absorp- 

 tion, to 10- ' of its value, which is unlikely, then a 

 rough calculation, assuming the truth of the hypo- 

 thesis, shows, that an exposure for this time should 

 result in a difference in polarimeter reading for the 

 two solutions of about 13°, instead of the observed 

 0-24°. 



It is, of course, realised that this simple experiment 

 has no bearing upon M. Perrin's version of the radia- 

 tion hypothesis, in which the reactants are supposed 

 to absorb several quanta of considerably longer wave- 

 length than I-2/Z. But this also can be tested directly 

 by experiments, which it is proposed to carrv out here 

 as soon as possible. T. W. J. T.wlor. 



Brasenose College, Oxford, October 2. 



Qualities of Valency. 



In his article in Science of July 22 (see Nature, 

 September 15, p. loi), and in his address to Sections 

 A and B of the British .Association at Edinburgh, 

 Dr. Irving Langmuir asks us to believe that the 

 sodium and chlorine atoms in sodium chloride are 

 never united by a chemical bond, i.e. that the salt 

 is ionised in its synthesis and remains ionised under 

 all conditions. This appears to be a denial of the 

 existence of NaCl molecules, yet such molecules exist 

 in the state of vapour at cir. 2000° C. Are we, then, 

 to suppose that sodium and chlorine ions are held 

 together at this temperature by electrostatic attraction 

 only? Difficult to reconcile with this idea, if it is 

 applied to salts in general, are the phenomena of 

 aqueous solutions of some salts — mercuric salts, for 

 example — which are attributed to imperfect ionisation 

 and the increase of specific conductivity of imperfect 



