504 



NA TURE 



u. 



.v,...,R 6, 192 



of conditions. On the other hand, the symmetry 

 demanded by the X-ray work on sal-ammoniac is 

 higher than that of the crystaliographer. It is, 

 therefore, not the crystal symmetry (the complete 

 physico-chemical symmetry on which crystallography 

 and its offshoot stereochemistry are based) but a 

 pseudo-symmetry — a phenomenon with which the 

 crystaliographer is familiar in other connexions. It 

 may accortlingly well be termed X-ray symmetry, 

 in order to distinguish it from other pseudo-sym- 

 metries. 



Now, whatever may be the true cau.se of this X-ray 

 symmetry, the explanation given by X-ray workers 

 is singularly unconvincing. So far from harmonising 

 a previously organised body of fact and interpretation 

 of proven wortli with the new results, the explanation 

 relies wholly on the data obtained from the X-ray 

 tube, and discounts the value of surface evidence, 

 almost on principle — for however ambiguous surface 

 evidence may be m certain crystals, it cannot be fairly 

 held to apply to sal-ammoniac, in which plane faces 

 of growth, the run of their striations, and the orienta- 

 tion of etch figures all demand the same class of 

 symmetry. It is surely obvious that the real explana- 

 tion must take equal cognisance of all well-established 

 facts, including those collected by the aid of the 

 goniometer, microscope, polarimeter, and, last but not 

 least, the test-tube, all such facts being apparently 

 equally precious in this province of crystal symmetry. 

 The problem is to evaluate the results from all these 

 instruments without unduly elevating or depressing 

 this or that section. My own view is that the ultimate 

 solution awaits the discovery of a new method, which 

 shall tell us as much about the chemical aspect of 

 crystal structure as the X-ray method does about its 

 physical side. Thus warned, a reader will not expect 

 too much from the following paragraphs. 



As a preliminary, it seems necessary to clear up a 

 widespread misapprehension concerning the results 

 of X-ray investigation. It has not infrequently been 

 stated that the recent work on organic compounds has 

 proved the existence of the molecule in the crystalline 

 condition. I do not know how this misapprehension 

 arises, for a perusal of Sir William Bragg's original 

 paper, in this domain, shows that the molecule is 

 assumed as a working hypothesis. The real position 

 is that the X-ray method can scarcely ever be expected 

 to prove molecular structure. It is now generally 

 accepted that the origin of X-rays (as also the seat of 

 their diffraction) lies near the nucleus, and not in the 

 few peripheral valency electrons which provide an 

 occupation for the chemist. The X-ray method has 

 the defect of its qualities : in revealing the atomic 

 positions in a crystal, it ignores the molecular aspect 

 completely. In the case of naphthalene the method 

 rev'eals the presence in the crystal of pockets, each 

 containing a mass of material having the composition 

 C20H16 (if anything, then, in this case it reveals poly- 

 merised molecules). The interpretation naturallv 

 takes a molecular form, because any other would be 

 chemically absurd ; but so far as anything of the 

 nature of chemical " bonds " is concerned (or of a 

 union of certain atoms into a molecule) it is, relatively, 

 a waste of time to appeal to the X-ray bulb. The 

 proper appeal is to the test-tube. 



A similar remark applies to the typical inorganic 

 case of sodium nitrate. The crystal model of the 

 X-ray analyst allows an interpretation of a structure, 

 which has been electrically resolved into sodium and 

 nitrate ions. But it might also be interpreted as a 

 fine mixture of sodium, nitrogen, and oxgyen, or even 

 as an ionised sodium nitrite plus oxygen, in which the 

 latter strives after geometrical symmetry and succeeds. 

 There can of course be little doubt that we are dealing 



NO. 2814, VOL. 112] 



with the first alternative, because it ' - -> 



crystiillise a solution to dryness, and 



salt without any appreciable evolnti'-i 



formation of nitrite. It may l>e ;i 



ionisation of a crystal salt into • 



extremely probable, it has not >- 



the X-ray method (Debye and S 



lithium fluoride being generally held, in particular b\ 



W. L. Bragg, to be inconclusive). 



In this purely atomic reaction of molecular n 

 to X-ravs (proved to the hilt by the pioneering 

 of Barkla and later by Moseleyj there would setw. to 

 lie a pos.sibility of obtaining a higher symmetry- th.m 

 by the physico-chemical method of surface stu : 

 The rdle played by the physicist's atom in his >• • 

 ment of symmetr>' is that of a sphere. Thi 

 true .so far as X-rays are concerned, but s' 

 the crystal, for valency forces must be taiMu mtn 

 account. The question therefore arises whether the 

 replacement of a sphere by a humped surface "r. 

 alternatively stated, whether a consideration of • 

 movements of valency electrons) will serve to rl<- 

 the symmetry, not merely of the individii 

 (as it must) but also of the structure as a \'. 

 it might). If this is found to be the case there is an 

 obvious explanation of a pseudo-symmetry, obtain 

 able by the X-ray method. 



An examination of this problem shows that no 

 lowering of symmetry can result from single valencies 

 (I have, then, no explanation to ofter for sal- 

 ammoniac), but with the double bond (the doul'K- 

 sharing of electrons), which first becomes possitjlc 

 with a divalent atom, the symmetry may, indeed, 

 be degraded, always provided that the atorn occupies 

 a " specialised " position in the structure, i.e. a 

 f)Osition in which it may be the seat of centro-sym- 

 metry or intercepted by a plane or axis of symmetry. 

 In a crystal of sodium nitrate, for example, we have 

 probably to deal with Na* and NO,-. If the oxygen 

 be monovalent, or divalent with the double fx)nd 

 lying in the basal plane of the cr>'stal, the symmetry 

 is still that of the atomic crystal of the physicist ; but 

 if the double bonds be arranged in a manner suggested 

 many years ago, without any reference to X-rays, by 

 J. E. Marsh and myself (J. Chem. Soc., 1913, vol. 103, 

 p. 845), the NOj- group acquires the symmetry of 

 quartz, and so does the crystal as a whole (the new 

 " space group " or point system being : Fedorov, 



46 5 ; Schoenflies, D^). Such a crj'stal would be 

 indistinguishable by the X-ray method from the 

 atomic cr\'stal of the physicist, but would presumably 

 betray its lower symmetry, when allowed to grow or 

 dissolve in its solution. A similar theoretical p>ossi- 

 bility holds, of course, with the calcite group, the extra 

 electron given up by Ca+'' making up the corresponding 

 deficiency in an atom of carbon as compared with an 

 atom of nitrogen. 



The real state of affairs is evidently not as described 

 above, for the symmetry of the calcite group and of 

 sodium nitrate is not that of the quartz class. It 

 does not necessarily follow, however, that the crystal 

 is exactly as it has been left by X-ray workers. It 

 may be a " racemic " substance, consisting of alter- 

 nate basal strata of d- and /-carbonate or -nitrate 

 groups, interlaminated by charged calcium or sodium 

 atoms. Examination of the new model shows such 

 a crystal to have both the correct symmetry* and the 

 same space group as the purely atomic model (Fedorov 



47 h, Schoenflies Djd*), but with a cr\'stallographic 

 vertical translation equal to twice the old value. 

 From the X-ray point of view, however, the vertical 

 translation would be as before, for X-rays would 

 scarcely appreciate the fact that they are deeding with 

 enantiomorphous groups of valency electrons. The 



