NATURE 



[September 12, 191; 



formulated an empirical theory of variable valency. 

 Kekule thought that atoms could not vary in their 

 valency ; but the alternative formulae which he put 

 forward to explain cases of difficulty would appear to 

 be, rather, an attempted explanation of variable 

 valency. It might be more correct to say that Kekule 's 

 formulas constitute an anticipation of Werner's theory 

 of auxiliary valencies, the theory which seems to find 

 most favour at the present day. Fixed valency can 

 scarcely now be defended, in view of the existence of 

 such compounds, for example, as the two fluorides 

 and the two chlorides of phosphorus ; the two oxides of 

 carbon, ammonia and ammonium chloride; and, for 

 example, the two series of compounds respectively of 

 iron, mercury, and copper. Variable valency of atoms 

 is, empirically at least, an establish fact. 



By the latest conceptions of variable atomic valency 

 and its extension almost without limit — so that, for 

 example, oxygen may be regarded as quadrivalent 

 and even sexivalent — no doubt the e.xistence of 

 numerous compounds which previously presented 

 difficulties can be explained. There are, however, 

 others long known to chemists, such as double salts 

 and the combination of wat^r with salts, formerly 

 called "molecular compounds," definite and individual, 

 in which these views do not assist us. These com- 

 pounds do not exist as gases, and unless they admit 

 of experimental study by the methods of dilute solu- 

 tion, even their gaseous molecular weights cannot be 

 ascertained. 



It is noteworthy that in most of the instances 

 recently investigated where variable valency has 

 been assumed the compounds studied have been 

 easily decomposable solids or liquids, and for one 

 reason or another their gaseous molecular weights 

 could not be determined. Many of these compounds, 

 indeed, only exist at low temperatures. As instances 

 of work of this kind I may mention Collie and Tickle 

 on quadrivalent o.xygen in dimethylpyrone derivatives ; 

 Gomberg on triphenylmethyl ; Landolf on acetone di- 

 hydrofluoride ; Thiele and Peter on methyl-iodo- 

 dichloride ; and similar studies by Kehrmann, Will- 

 stiitter and Iglauer, Biilow and Sicherer, Baeyer and 

 Villiger, .Archibald and Mcintosh, Chattavvav, Pfeiffer 

 and Truskier, and others. 



.\nother most interesting class of solids which are 

 capable of existing in two isomeric forms distinguished 

 from each other by such physical properties asdensitv 

 or colour are the Schiff's basis or anils. Some of 

 these were studied by Hantzsch, who proposed to 

 explain their existence by the Hantzsch-Werner stereo 

 hypothesis : — 



HO.CeH^.CH 



II 

 NR' 



HC.QHjOH 



II 

 NR' 



But since only a few, and these not very satisfactory, 

 compounds show this isomerism, which do not con- 

 tain the hydroxyl group, other suggestions have been 

 put forward to account for the isomerism, by Ansel- 

 mino and by Manchot. 



In my own laboratory, associated with Mr. F. G. 

 Shepheard and also with' Miss Rosalind Clarke, I have 

 made a study of various Schiff's bases for the purpose 

 of investigating the remarkable property which some 

 of these bases exhibit of phototropy. By phototropv 

 is meant the capability of reversible change of colour 

 in solids depending upon the presence or absence of 

 light. Incidentally, too, I wished to study another 

 physical property which many Schiff's bases possess, 

 in common with other substances, of reversible change 

 of colour with raising or lowering of temperature. 

 This property we have called thcnnotropy, and manv 

 old instances will be remembered of substances of 

 simpler constitution which exhibit it : thus, when 



NO. 2237, VOL. go] 



subjected to the temperature of solid carbon dioxide, 

 ordinary sulphur becomes colourless, red oxide of mer- 

 cury becomes yellow, vermilion becomes scarlet, and 

 on return to the ordinary temperature the original 

 colours reappear. 



As has been pointed out in a recent communication 

 by Biilman, it is most important in these discussions 

 that we should be perfectly clear in the use of terms. 

 I take it for granted that isomerism is a general term 

 for compounds differing in some respect but having 

 the same composition. If the molecules (gaseous) have 

 the same weights they are metamerides ; if of different 

 weights they are polymerides. When solids crystallise 

 in more than one form they are polymorphous. Now 

 it does not seem reasonable to suppose that reversible 

 colour changes such as those exhibited by phototropes 

 or thermotropes involve such violent intra-molecular 

 changes as the breaking and reconnecting of atomic 

 linkages. For example, take the three bases, 

 salicylidene-»[-toluidine, which in the dark or imme- 

 diately it is exposed to light is yellow, but on continued 

 exposure to light quickly becomes orange, and 

 changes back again to its original colour in the 

 dark; salicylidene-n;-aminophenol, which at ordinary 

 temperatures is orange, but is much paler at the 

 temperature of solid carbon dioxide, on raising the 

 temperature to nearly the melting-point (]28'9°) 

 becomes orange red, and these changes take 

 place in the reverse order again on cooling ; 

 salicylidene-Zi-aminobenzoic acid, studied bv our- 

 selves and by Manchot and Furlong indepen- 

 dently, shows a wider range of thermotropic change 

 between bright yellow and blood-red, and is also 

 phototropic. To explain such changes as these and 

 the others of a similar nature previously referred to, I 

 think some less drastic hypothesis should be sought 

 than intra-molecular breaking, and consequent meta- 

 stasis or polymerisation. Though doubtless the hypo- 

 thesis of Hantzsch and Werner could be invoked, or 

 the modified hypotheses of Manchot or Anselmino, I 

 think there should be some simpler explanation. 

 Someone suggests polymorphism. Now polymorphism 

 means that a change of crystalline form takes place 

 which might doubtless connote change of colour. If 

 one watches phototropic crystals changing colour 

 under the influence of light from yellow to red, and 

 notices that after remaining in the dark the 

 same crystals have changed back to the original 

 colour, and, remember, that these changes can be 

 repeated with the same crystals apparently without 

 limit, it will not be considered likely that this pheno- 

 menon depends on a reversible change of crystalline 

 form. In a communication to the Chemical Society 

 some three years ago Mr. Shepheard and I put forward 

 the following suggestion : — " Evidence is accumulat- 

 ing of reversible isomeric reactions, like those described 

 in this paper, which are indicated by physical differ- 

 ences, such as changes of colour. It is possible that 

 these may be explained by hypotheses, similar to that 

 of Hantzsch and Werner, assuming intra-molecular 

 rearrangement ; but in the case of phototropy and 

 thermotropy it should not be forgotten that the sub- 

 stances exhibiting these phenomena are solids. No 

 one will doubt, however, that these differences of colour 

 depend on isomeric change of some kind, but in the 

 case of solids we know practically nothing of their 

 molecules, not even of their relative molecular weights. 

 The molecules of solids are probablv far more com- 

 ple.x than those of liquids or gases ; indeed, thcv may 

 be rather complex groups or aggregates of ordinary 

 gaseous molecules, which would give rise to far more 

 numerous possibilities of isomerism. It appears to 

 us that phototropic and thermotropic reactions are 

 more probablv due to isomeric changes affecting the 

 atrgregation of molecules in solids than to intra- 



