November ii, 1897] 



NATURE 



The double refraction is' also progressively affected by the 

 replacement of potassium by rubidium and the latter by coesium, 

 the amount varying inversely as the atomic weight of the metal. 

 Some very remarkable phenomena follow from this rule, owing 

 to the fact that the initial amount of the double refraction in the 

 potassium salt of any series or group is already very small. For 

 a comparatively slight diminution is sufficient to bring it to zero, 

 and even to go further and reverse its sign. As the three 

 refractive indices of any salt are not equidistant from each 

 other in value, and are affected to different extents by the same 

 chemical change, they never in either the rubidium or caesium 

 salts become reduced to absolute identity at the same time, but 

 become equal in pairs for specific wave-lengths of light and 

 degrees of temperature. When this occurs the crystal becomes 

 for that wave-length and temperature temporarily uniaxial in its 

 optical comportment. The particular salt in which this occurs 

 varies in the different series. In the sulphates it is the rubidium 

 salt, in the selenates the caesium salt, both at temperatures 

 slightly elevated above the ordinary ; in the double sulphates it 

 occurs in caesium magnesium sulphate and at the ordinary 

 temperature for blue light, while in the other groups of this 

 series the progression never quite reaches the state of equality 

 of two indices. It will be evident that the distribution of the 

 directions corresponding to the a, y8 and 7 indices, and also the 

 optic axial phenomena in convergent polarised light, vary in the 

 three salts of each series or group to a quite extraordinary 

 extent ; so much so that without the discovery of this rule, 

 which affords the key to them, the vagaries must have been in- 

 comprehensible. For whenever two indices become identical the 

 third is also very nearly so, and consequently the slightest change 

 of wave-length or temperature brings about most drastic changes 

 in the optical phenomena dependent upon the mutual relations 

 of the three. The most striking case is that of caesium selenate. 

 At the ordinary temperature the sign of double refraction and 

 disposition of the optic axes are already reversed from what they 

 were in potassium and rubidium selenates, and the three indices 

 are so nearly identical that a section plate a centimetre thick 

 is necessary to produce an interference figure in convergent 

 polarised light. On heating gradually to 250°, the sign of double 

 refraction changes twice over, and the optic axes move so 

 rapidly that their acute bisectrix occupies the direction of each 

 of the three morphological axes in turn. That these remarkable 

 phenomena are precisely what are demanded by the rule of 

 progression above enunciated, given the initial optical conditions 

 of the potassium salt, is perhaps the most striking proof of the 

 validity of the rule. 



It may be here mentioned that double sulphates of potassium 

 and manganese, and potassium and cadmium, containing six 

 molecules of water, have not yet been obtained, although the 

 corresponding rubidium and caesium salts are readily formed ; 

 yet from the above rule, and others which have been indicated 

 for the other projDerties, the author has been able to predict the 

 morphological and optical constants which these salts will 

 probably exhibit if ever they are obtained. 



The determination of the densities of the crystallised salts has 

 enabled the molecular optical constants to be calculated, with 

 the aid of the formulae of Lorenz and of Gladstone and Dale. 

 It has been found that they exhibit a similar progressive increase 

 following the increase in the atomic weight of the alkali metal, 

 and that the increase is always greater when caesium replaces 

 rubidium than when the latter replaces potassium. The mole- 

 cular refraction for the crystallised state of the sulphates and 

 selenates was also compared with that for the state of solution in 

 water, specific determinations of the density and refraction of 

 highly concentrated solutions of known strength being made for 

 the purpose. It was found that while the values for the two 

 states are approximately the same, there are slight differences, 

 due to the change of state, which observe a distinct progression ; 

 for they vary directly as the specific refractive energy and in- 

 versely as the atomic weight of the alkali metal. In both series 

 the refraction equivalent of the potassium salt rises by 2J per 

 cent, when the salt is dissolved in water, and that of the 

 rubidium salt by i per cent., while that of the caesium salt 

 decreases by \ per cent. This interesting order of the differences 

 clearly demonstrates the quantitative influence of the nature of 

 the alkali metal upon even the smallest details of the physical 

 properties ; and the fact that the sign of the differences passes 

 during the progress from positive to negative, indicates the sub- 

 stantial accuracy of the principle enunciated by Dr. Gladstone 

 in 1868 that " the refraction equivalent of a solution is the sum 



NO. 1463, VOL. 57] 



of the refraction equivalents of the solvent and the substance dis- 

 solved," by the application of which principle the calculations 

 for the state of solution were made. 



The molecular volumes of the members of each series likewise 

 exhibit the order of the atomic weights of the alkali metals, but 

 the progression again proceeds more rapidly than in simple 

 arithmetical proportion to the latter. A point of paiticular 

 interest was elicited with respect to the volumes of the double 

 sulphates, which throws a strong light upon the nature of those 

 salts. It was found that the vplunie of the alkali sulphate is the 

 same in the double salt as it is for the simple salt itself, that is 

 to say, the simple alkali sulphate enters into the structure of the 

 double salt without suffering any contraction in volume. This 

 fact, contrasted with the very large contraction which ac- 

 companies the chemical union of the elementary constituents of 

 the simple sulphates, negatives the possibility of chemical com- 

 bination of the molecular constituents of the double salts. This, 

 together with other facts which the investigation has brought to 

 light, has led to the important conclusion that the composition 

 of these double salts is simply a result of the aggregation of the 

 molecular constituents in a particular type of homogeneous 

 structure, in which they find stablest equilibrium, the very nature 

 of this structure ensuring that the component simple molecules 

 are always present in the same, the observed, proportion. The 

 elucidation and definition of all the possible types of homo- 

 geneous structures have recently been elaborately worked out by 

 Federow on the continent, and Barlow in this country, and the 

 precise correspondence between the possible types of homo- 

 geneous structureskand the observed varieties of crystal symmetry 

 is now established beyond all doubt. The author has been able 

 to indicate, moreover, the particular member of Barlow's 

 classification corresponding to both the simple and double 

 salt series. 



The conclusion just referred to, regarding the nature of the 

 double sulphates, taken in connection with other facts which the 

 author has established, leads to a further one of a still more far- 

 reaching character, namely, that the units of a homogeneous 

 crystal structure are the simple chemical molecules themselves, 

 and that the current assumption that the crystal unit is a more 

 or less complex aggregation of chemical molecules is quite un- 

 necessary and in general erroneous. This conclusion is further 

 strongly supported by some recent work of Fock, on the solu- 

 bilities of mixed crystals, based upon the theory of solid 

 solutions. 



Before leaving this interesting subject it should be mentioned 

 that a method has been found, by combining the molecular 

 volume with the morphological axial ratios, of determining the 

 relative distances apart of the centres of contiguous chemical 

 molecules of the sulphates and selenates, and of contiguous 

 groups of the eight component chemical molecules of the double 

 sulphates, each such group corresponding to the generic formula 

 R2SO4.MSO4.6H2O. A comparison of these "distance ratios" 

 shows that the replacement of potassium by rubidium, and of 

 the latter by caesium, is accompanied by a progressive increase 

 in the separation of the structural units or groups of units in 

 every direction, corresponding to the progress of the atomic 

 weight of the alkali metal, and that the latter replacement 

 always gives rise to a greater extension of the structure than the 

 former. 



In conclusion, the net result of the investigation has been to 

 show that the whole of the morphological and physical pro- 

 perties of the crystals of each of these isomorphous series ex- 

 hibit progressive variations, which follow the order of progres- 

 sion of the atomic weights of the alkali metals which the salts 

 contain. Hence it may be said that these variations are func- 

 tions of the atomic weight of the alkali metal, and it has been 

 shown that the function is usually one which involves higher 

 powers than the first. Of course atomic weight is only one of 

 the numerous properties of an element, but it is doubtless the 

 most convenient reference constant that could be chosen to ex- 

 press fundamentally the difference in the essential nature of the 

 atoms of different elements. It is this difference in essential 

 nature which gives rise to the rules which have been brought to 

 light by the investigation ; and the author desires to make it 

 quite clear that atomic weight is merely employed as the basis 

 of reference because it is the aptest expression of such difference, 

 and not because of any virtue in atomic weight per se. The fact 

 that the rules are equally applicable to series so widely different 

 as the rhombic sulphates and selenates and the monoclinic double 

 salts, appears to indicate their application to isomorphous series 



