582 



NATURE 



[July 8, 1920 



the hyperbolic susceptibiHty-temperature relationship 

 for gaseous oxygen (Tak^ Son^, loc. cit.). 



(6) If the diamagnetism of hydrogen is attributed to 

 thermal oscillations or rotations, we might expect that 

 Xh for the gas at i6° C. would be greater than Xh for 

 liquid hydrogen at -253° C. Precisely the reverse 

 holds, according to the above data, and there is cer- 

 tainly no indication that at low temperature Xh is 

 tending to change sign. 



A variation ot Xh from — 19-8x10-^ to —27x10-' 

 for a temperature interval of 16° C. to -253° C, 

 together with the fact that xh in different types of 

 organic compounds is constant and equal to 

 "30-5x10-% points to the conclusion that thermal 

 oscillations and rotations have little to do with the 

 origin of diamagnetism in molecular hydrogen, and 

 that the Bohr hydrogen molecule will not account 

 for it. 



The present writer's view is that the free hydrogen 

 atom is probably paramagnetic, but the structure of 

 the hydrogen molecule must be such that by com- 

 pensation it is, as a whole, diamagnetic. A model of 

 the hydrogen molecule which satisfies these conditions 

 was suggested in Nature of May 13, 1920. In this 

 model the irtdividuality of the hydrogen atom is pre- 

 served, and this may. have some bearing on the origin 

 of the primary and secondary hydrogen spectra. 



A. E. OxLEy. 

 The British Cotton Industry Research 

 Association, 108 Deansgate, Man- 

 chester, June 28. 



University Stipends and Pensions, 



All university teachers will thank you for the 

 leading article in Nature of June 17 pointing out the 

 injustice done to them and to university education by 

 the exclusion of such teachers from the provisions of 

 the School Teachers (Superannuation) Act, 1918. On 

 one point, however, the article is misleading. It is 

 stated that " what complicates matters is the fact 

 that there exists a contributory pension scheme in the 

 universities — the federated superannuation scheme — 

 which is thought by some to be superior to the 

 Teachers Act in certain respects." It should be 

 made quite clear, however, that the governing bodies 

 of university colleges are at liberty to adopt the 

 federated scheme or not, and that the governing 

 bodies of some colleges have refused to adopt it, with 

 the result that the staffs of these colleges have no 

 prospect of any pensions whatever. The position in 

 the University of London is, therefore, even more 

 anomalous than was suggested, since some schools 

 come within the provisions of the Act, some have con- 

 tributory pension schemes, and some have none. The 

 Northampton Polytechnic Institute and the Imperial 

 College of Science and Technology have each an 

 engineering department the courses in which enable 

 their students to take the B.Sc. degree of London 

 University as internal students of the University. 

 The lecturers of each sit side by side on the Faculty 

 of Engineering and on the various Boards of Studies 

 of the University. The first-named institution is 

 included in the Teachers Act, but the last-named is 

 excluded. Lecturers at the former retire at the age of 

 sixty with a non-contributory Government pension, 

 whilst their confreres at the Imperial College may 

 w^ork as long as they are able with no prospect of 

 any pension whatever. G. W. O. H. 



There is nothing in " G. W. O. H.'s" letter to^ 

 support his statement that the article is misleading. 



NO. 2645, '^OL. 105] 



It is true that no university or university college is 

 compelled to join the federated scheme, but it is 

 equally true that such a scheme exists, and that most 

 universities and university colleges have adopted it. 

 The additional particulars which "G. W. O. H." gives 

 were known to us, but obviously in a short article 

 every variety of illustration could not be included.— 

 Ed. Nature'. 



The Separation of the Isotopes of Chlorine. 



In agreement with Prof. Soddy, I find myself un- 

 able to understand ^how it is possible to separate 

 isotopes by the method suggested by Mr. D. L. 

 Chapman in Nature of June 17. Nevertheless, a 

 certain paradox has been brought to light in con- 

 nection with Nernst's theorem the solution of which 

 is not without interest. 



The paradox to which I refer is this : Consider the 

 equilibrium in the gaseous reaction Cl2+Cr2:::r:2ClCl'. 

 If the gases behave as perfect gases, and if CI and 

 CI' are identical or differ only very slightly, then it 

 is easy to show by probability considerations that the 

 equilibrium must be given by 



[ClCl ? 

 [CyLCi'2]' 



:K=4. 



From this it follows that to convert a gram-molecule 

 of CL plus a gram-molecule of Cl'^ into two gram- 

 molecules of ClCr at the same temperature and pres- 

 sure, a will require an amount of available energy 

 equal to —RT log 4. 



Now if the chemical and physical properties of the 

 isotopes are truly identical, then from this and the 

 necessary equality of the vapour pressures it is readily 

 shown that to convert a gram-molecule of solid CI3 

 plus a gram-molecule of solid Cl'a into two gram- 

 molecules of solid ClCr also requires — RTlog4 of 

 available energy. Therefore the difference in the 

 entropies of the two sets of solids is Rlog4, which, 

 being' independent of the temperature, must exist at 

 the absolute zero. 



It is, however, unjustifiable to say that this con- 

 tradicts Nernst's theorem, and to deduce from this 

 theorem that K for the gaseous reaction must be i 

 in order to make the change in the entropy zero. For 

 if there is a true identity, then this implies that there 

 are no forces to guide the atoms into any particular 

 configuration, so that! even down to the zero of tem- 

 perature no true reaction is possible, and what occurs 

 is really of the nature of mixing. That a difference of 

 entropy occurs on mixing, even at the zero, is neces- 

 sary, and in no way. contradicts Nernst's theorem; in 

 fact, the case of mixtures is explicitly excluded by 

 Nernst. 



If, on the other hand, there is a real, but small, 

 difference in the two isotopes, then, as before, K will 

 very nearly equal 4. Now, in order to obtain the 

 difference in the entropies between the solids near the 

 zero of temperature, let us carry out the cycle 

 described by Mr. Chapman, but in the neighbourhood 

 of the zero. Then, in spite of the fact that the isotopes 

 differ very little, it is impossible to say that the vapour 

 pressures remain equal. Thus it is impossible, so 

 long as there is any difference at all between the 

 isotopes, to argue that because K = 4 for the gaseous 

 reaction there must be a finite change in entropy at 

 the absolute zero. 



Angus F. Core. 

 The University, Manchester, July 4.' 



