July 29, 1920] 



NATURE 



677 



follow out, in order to acquire methods of accurate 

 observation and technique. This work should be 

 written up and modelled in the form of a scientific 

 paper, and illustrated so far as may be desirable. By 

 means of such an essay the student will become fami- 

 liar with the elementary procedure in research work, 

 he will acquire some power of independent observa- 

 tion, and learn how to deal with entomological litera- 

 ture, thus gaining some idea of the sources where he 

 will find first-hand information. 



Furthermore, I would also insist upon the student 

 forming a small but thoroughly representative col- 

 lection of insects, so proving that he has had some 

 field practice in collecting, and is able to refer them 

 to their families and genera. By means of such a 

 course as I have outlined, it should be possible to train 

 good, all-round entomologists, capable of tackling a 

 problem unaided when out in the wilds of Africa or 

 the plains of India. 



If the student can spare a fifth year, it would be all 

 to his advantage, and the time would be most profitably 

 spent in prosecuting some line of independent entomo- 

 logical research. A. D. Imms. 



Institute of Plant Pathology, Rothamsted 

 Experimental Station, Harpenden. 



The Separation of the Isotopes of Chlorine. 



In order to prevent confusion of issues, instead of 

 CI and Cr let us write A and B. Then when we say 

 that A and B are identical, we mean that all the 

 properties of A and B are the same except that of 

 position occupied. Thus we are enabled to divide 

 the atoms into two groups, the A group and the B 

 group, in spite of their identity of properties. Then 

 it is quite certain that if the atoms exist as mole- 

 cules -A,, B2, and AB, in equilibrium by the reversible 

 reaction A, + B2::rL2AB, the equilibrium is given by 

 [ABJV[A,][B,]=K = 4. .^^ ^ ^. , 



The following considerations will, I think, meet 

 any difficulties that have been raised in reconciling 

 this reaction with Nernst's heat theorem. In the case 

 of complete identity, if we convert the solids Aj and 

 Bj into the solid AB by evaporation to the gases 

 A2, B2, transformation into the gas AB, and con- 

 densation to the solid AB, we obtain an increase of 

 entropy of Rlog4. But this solid is really a solid 

 solution or mixture, since, as we assumed that the 

 vapour pressure over it is equal to the pressure over 

 the solids A, or Bj, we must assume that the mole- 

 cules condense on its surface with " longitudinal in- 

 difference." The solid, then, is a solution of the 

 molecules AB in BA. 



Now the entropy of a body consists of two parts, 

 one depending on the distribution of velocities, the 

 other on the distribution of the co-ordinates of posi- 

 tion. The first term cannot give rise to any change 

 of entropy when the solids are transformed, irrespec- 

 tive of Nernst's theorem, but the second term is a 

 constant, and accounts for the change of Rlog4. 

 It mav, in fact, be calculated directly by statistical 

 methods. 



If we assume that the gas AB condenses to the 

 solid AB (or B.^) instead of into the solid solution, 

 then we must take the pressure over this solid as 

 double that over .A, or B„ and not equal to them; 

 because, consistentlv with the assumption of the 

 formation of the pure solid AB, we must assume that 

 the solid rejects half the molecules which strike its 

 surface; that is to say, the molecules AB condense, 

 but not the molecules BA. 



This double vapour pressure will make the entropy 

 of the two gram-molecules of k^ (or BA) equal to 

 the entropy of one gram-molecule of A, ^\us one 

 gram-molecule of B,. 



NO. 2648, VOL. 105] 



No essential difference in the argument is made 

 when Aj differs sjightly from Bj. 



Prof. Soddy throws out a suggestion for the 

 removal of the term Rlog4 which surely must be 

 erroneous. He seems to agree to the distribution of 

 molecules given by [AB]VlA,][B,] =4 (which must 

 result whatever kinetic process be assumed), but he 

 considers it wrong to write 4 as the equilibrium con- 

 stant of the reaction A^-t-B^z^zAB, as this gives for 

 the coefficients of the reaction velocities fe, = 4fej. He 

 therefore would write the reaction Aj-|-Bj = AB-hAB, 

 and then, taking half the concentration of AB, write 

 [i'\BBAB]/[.'\,][B,]=K=i. Therefore fe. = fe,. 



To write this reaction in this form is unjustifiable. 

 In the first place, that fei = 4fe2 in no way contradicts 

 the assumption of the iderLtity of A and B. For the 

 velocity coefficients do not depend only on the pro- 

 perties of the atoms or molecules involved, but con- 

 tain a factor depending on the statistics of the re- 

 actions. In this respect the direct and .reverse re- 

 actions may be different. This is better seen by com- 

 paring the two reversible reactions 2A = .'\, and 

 B-t-C = BC, where A, B, and C are identical atoms. 

 The two reverse coefficients are equal, fej^fe/, but the 

 two direct coefficients are not equal, for fej = ^fe '. 

 This is because n atoms B, together with n atoms C, 

 give twice as many B-C collisions as n atoms of 

 .A. give^ A-A collisions. If we write the reaction 

 A-t-A — .^2, and take half the concentration of A, we 

 still do not find fe, = fe/, but fei = 2fe/. 



In the second place, to write the reaction 

 A2-l-B2Z:^AB + AB suggests that we can divide the 

 molecules AB into two equal sets, and that a signi- 

 ficant collision only occurs when an AB molecule from 

 the first set collides with an AB molecule from the 

 second set. Finally, the semi-permeable membrane 

 that may be used in calculating the change of entropy 

 due to the gaseous reaction must be assumed perme- 

 able to all or none of the molecules AB, thus giving 

 an entropy change Rlog4. So that by no con- 

 siderations whatever are we justified in taking half 

 the concentration of the AB molecules when cal- 

 culating the change of entropy. Angus F. Core. 

 The University, Manchester, July 24. 



Anticyclones. 



Prof. Hobbs rn Nature for July 22 gives some 

 experimental reasons for contending that over large 

 ice-covered areas, such as exist in Greenland and the 

 Antarctic continent, the cooled lower layer of air 

 moves outwards in all directions from the centre of 

 the ice-covered area. Under the influence of the 

 earth's rotation the air thus set in motion is regarded . 

 as circulating as in normal anticyclones, and Prof. 

 Hobbs on that account speaks of such areas as being 

 anticyclonic. He remarks: "The centrifugal nature 

 of this motion tends to produce a vacuum above the 

 central area of the ice mass, and the air must be 

 drawn down from the upper layers of the atmosphere 

 in order to supply the void. It is here that is located 

 the 'eye ' of the anticyclone." He thus postulates an 

 anticyclone with a low-pressure centre. 



With the physics of Prof. Hobbs 's theory there 

 need be little criticism. The point really seems to be : 

 •Are the conditions described by him as existing over 

 an ice-cap anticyclonic? .An anticyclone has a high- 

 pressure centre, and a cyclone a low-pressure centre, 

 the surface air moving outwards in the former and 

 inwards in the latter, whereas the conditions 

 described by Prof. Hobbs are an outward flow and 

 a low-pressure centre. Would it not be well to 

 designate such conditions by some other Avord? 



R. M. Deeley. 



Tintagel, Kew Gardens Road, Surrey, July 23. 



