July 14, 192 1] 



NATURE 



619 



Letters to the Editor. 



[The Editor does not hold himself responsible for opinions ex- 

 pressed by his corre'pondents. Neither can he undertake to 

 return, or to correspond with the writers of, rejected manu- 

 scripts intended for this or any other part of Nature. 

 No notice is taken of anonymous communications.^ 



The Separation of the Isotopes of Chlorine. 



The method outlined in our letter of September 30, 

 1920 (Nature, vol. cvi., p. 144), and used for the partial 

 separation of the isotopes of mercury, has enabled us 

 to accomplish a partial separation of the isotopes of 

 chlorine. When about half of a strong solution of 

 hydrochloric acid cooled down to about -50° C. was 

 evaporated in a high vacuum, the mixture of water 

 and hydrogen chloride being condensed on a surface 

 cooled with liquid air, the condensed part of the 

 hydrochloric acid was found richer, and the remaining 

 part poorer, as regards the lighter constituent of 

 chlorine than the ordinary HCl. 



Starting from about i litre of 8-6 mol. solution, we 

 obtained, by repeated separations, about 100 c.c. of the 

 lightest, as well as of the heaviest, fraction, the differ- 

 ence of which was examined by two different methods 

 after transforming the acid into sodium chloride. In 

 the first method the density of the two saturated NaCl 

 solutions was determined. The salts were precipitated 

 several times by alcohol from their aqueous solutions, 

 and density measurements carried out after each pre- 

 cipitation. We found uniformly a higher density of 

 the solution prepared from the residual acid, the mean 

 values at 20° C. being 



da= I'20722 

 fl?i.= I"20235 



from distilled and residual acid respectivelv. On the 

 assumption of equal atomic volume of the two isotopes 

 these figures correspond to a difference of 0024 unit 

 in the atomic weight of chlorine, or 65 per cent, in 

 the atomic ratio of the isotopes. 



In the second method equal quantities (57500 g.) 

 of the molten isotopic sodium chlorides were dissolved 

 in water and each precipitated with accurately the 

 same volume of 02 n. silver nitrate. The latter was 

 added in a slight excess. After precipitation and dilu- 

 tion to 2000 c.c. the approximate concentration of the 

 filtrate ^ was determined by titration with potassium 

 rhodanide, and the ratio of the silver concentrations of 

 the two solutions measured by combining them 

 to a concentration cell. From the concentration 

 c = 000040 n. and the electromotive force of the cell, 

 0001 1 volt at 18°, we calculated that the difference 

 in the atomic weight of the two samples was 0021 

 unit, in close agreenient with the result of the first- 

 mentioned method. 



The hydrochloric acid used in these experiments was 

 thoroughly purified with potassium permanganate in 

 order to remove bromine contingently present. More- 

 over, the repeated precipitation of the sodium chloride 

 by alcohol would have given decreasing values for the 

 estimated separation of the isotopes if any bromine 

 should have been present. We think ourselves 

 justified, therefore, in regarding the above-mentioned 

 results as conclusive. 



J. N. Bronsted. 

 G. Hevesy. 



Physico-Chemical Laboratorv of the Polv- 

 technic Institute of Copenhagen, June 29. 



A Novel Magneto-Optical Effect. 



Prof. Elihu Thomson's explanation of the in- 

 teresting magneto-optical effect which he describes 

 in Nature of June 23, p. 520, is supported by some 



NO. 2698, VOL. 107] 



experiments we have made recently on various oxides 

 dispersed in air. When the vapour of zinc ethyl 

 diluted with carbonic acid gas is mixed rapidly with 

 a large volume of air, a fine fume is produced the 

 particles of which when examined with the ultra- 

 microscope exhibit rapid Brownian motion. In a 

 short time the motion becomes slower and the par- 

 ticles brighter, but fewer in number. This continues 

 until the fume has aggregated into a number of 

 loose complexes formed of irregular chains or strings 

 of particles. These chains are flexible and whirl 

 and twist about under molecular bombardment in a 

 striking manner, but fall under gravity at a sur- 

 prisingly slow rate. In an electrostatic field the com- 

 plexes straighten out and arrange themselves parallel 

 to the lines of force, and on reversal of the field rotate 

 through 180°. 



When caught on a slide and examined with a high- 

 power objective the same structure is seen more 

 clearly. The individual particles are not in contact, 

 but appear to be held together by invisible threads, 

 consisting probably of strings of" molecules or fine 

 molecular aggregates. The zinc oxide fume given 

 off from a zinc arc in air behaves in a precisely 

 similar way. When a dense cloud is produced 

 initially the particles agglomerate to large and 

 irregular masses. By transmitted light the connect- 

 ing hairs are invisible, but by a strong beam of 

 reflected light of short wave-length obtained by suit- 

 able screens the particles appear to be surrounded by 

 a nebulous haze. That the particles in these large 

 complexes are really linked together can be demon- 

 strated in another way by allowing a drop of im- 

 mersion oil to flow slowly' across the slide on which 

 the deposit has been caught ; the particles as they are 

 lifted up by surface tension are seen to be attached to 

 constellations of others, and drag these with them 

 from a considerable distance in front of the advancing 

 oil. The individual particles are about loo/x/x in 

 diameter, and the complexes about 3o,a. Even after 

 several hours these clouds always contain a number 

 of single particles. 



The particles in clouds obtained by the arc dis- 

 charge between electrodes of other metals form com- 

 plexes of varying structure. The tendency to 

 aggregation seems weakest with the oxides of Pb, 

 Cu, Mn, and Cr. It is slightly greater with Fe, 

 whilst the oxides of Mg, Al, and Sb give similar 

 results to zinc oxide. The particles of CdO show a 

 great tendency to aggregate in strings of a remark- 

 able length, which under the microscope look like 

 beads strung on a thread. Clouds of this structure 

 might be expected to show in a strong electrostatic 

 field an optical effect analogous to that described by 

 Prof. Thomson, but so far we have not observed it. 

 The work is being continued. 



R. Whytlaw-Gray. 



J. B. Speakman. 

 Eton College, Windsor, July 4. 



In the former account of this novel effect (Nature, 

 June 23, p. 520) it was pointed out that a micro- 

 scopic examination of the iron arc smoke deposited 

 on a glass surface gave evidence of the existence of 

 fine particles of iron compound arranged in short 

 chain sections of bead-like relation. 



It is now thought that this peculiar formation may 

 have its origin in the outer envelope of the arc flame 

 where the particles are formed and where they are 

 lined up around the arc stream by the circular mag- 

 netism surrounding the current conducted by the hot 

 vapour stream of the arc. The particles, being mag- 

 netic, would tend to form chains or rings surrounding 

 the arc. These would not be stable, however, but 

 would float away as they became shattered by gas 



