MINERALOGICAL CHEMISTRY.ALTHOUGH the year has not been marked by any outstanding dis-covery serving to open up an entirely new field of research, theprogress to be noted along the familiar channels may be consideredto be more than satisfactory. A short appreciation of the generaltendencies may fitly precede a discussion of the work in detail.The X-ray methods of investigating crystal structure have natur-ally received much attention, both on the part of continentalworkers, who appear to be chiefly attracted by the more theoreticalaspects, and also by W. H. and W. L. Bragg, who, in a series ofbrilliant experimental researches, have succeeded in unravellingthe structure of several elementary and compound substances.Certainly the most striking feature of the work is the remarkabledegree of intimacy to which these authors have carried their struc-tural dissections.They have not been merely content t ofamiliarise us with the general nature of crystal structure, buthave carried their explorations to the utmost limit of chemicalinterest, namely, the atoms. The further development of thisabsorbingly interesting field of work must surely throw light onmany questions now in dispute; in particular, its application tothe simpler cases of polymorphism may well be expected toaccelerate a decision between the present conflicting theoreticalviews.There has been a very marked activity in the province ofchemical crystallography proper. The foundation by Fedorov ofthe method of crystallochemikal analysis very materially enhancesthe value of a crystallographic description, proving, as i t does, thatany substance which has once been measured can be readilyidentified on any subsequent occasion, and without involving theslightest loss of any valuable material.The brief accounts of twocases in which a crystallographic identification of several productswas quite indispensable to the successful issue of a biochemical re-search will, perhaps, convinm the most hardened Philistine thatcrystal-gazers have occasionally their lucid intervals in which theymay be of some use to the more wideawake of their chemical28MINERALOGICAL CHEMISTRY. 239brethren. Among the more conventional pieces of work, the in-vestigations of Wahl are, perhaps, preeminent as being valuablecontributions to our knowledge of the crystalline form of sub-stances which are liquids, or even gases, a t the ordinary tempera-ture.On the theoretical side, there has been a dignified inter-change of opinion between Richards and Barlow and Pope on thechemical aspects of the vaiency volume theory.The systematic physico-chemical investigation of minaralogicalproblems continues to make very obvious advances, especially inthe hands of the American geophysical school. Bowen’s masterlyinvestigation of the ternary system, diopsids-f orsterite-silica,shows both a clear appreciation of the difficulties to be overcomeand a nice discrimination of the methods which are the most likelyt o make for success. The, research may be upheld as a mo’del whichother workers would do well to copy.The exact petrological in-formation that is already beginning to accrue from such investiga-tions is in itself a sufficient repayment for all the trouble involved.Another important work from the same school is that of Fenner onthe various polymorphous modifications of silica ; a significantfeature of this work is the astounding sluggishness of certain poly-morphous transformations, which may be used as an argument infavour of polymerism.The last general topic to be discussed is the problem of “liquidcrystals.” Few subjects have ever awakened the general interestaroused by the discovery of these much-discussed substances byOtto Lehmann; there was a time when it was impossible t o attenda conversazione without being asked t o give an account of theirproperties, or, in default, to swell the audience of a colleague’sdemonstration.It is therefore not surprising that the generalinterest soon began to flag. A new stage was, however, inaugu-rated by the chemical researches of Vorlander, who prepared avast number of new examples in a high state of purity, and alsoadded considerably to the general stock of knowledge, showing thatthe production of the turbid liquid phase is a constitutional featureof a peculiar type of stereochemical configuration. It was, how-ever, reserved for a pure physicist, Bose, to lay the foundation ofa general theory of “anisotropic liquids,” and to substantiate hisCheory by means of a carefully planned series of experiments.Itis indeed fortunate that these experiments, so untimely interruptedby BOSB’S early death, have been continued by others. Mauguin’selectromagnetic researches, in particular, afford a very strikingconfirmation of Bose’s views. Again, the definite proof that thecubic modification of silver iodide does not in reality form liquidcrystals is very suggestive as indicating that a molecular configura240 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.tion of some chemical complexity is indispensable to a regularorientation of liquid molecules. It is perhaps only to be expectedthat Lehmann should find a difficulty in adapting himself t o thenew views, however much the new researchee add lustre to his dis-coveries. The history of liquid crystals affords a good example ofthe rapidity with which a scientific enigma may be compelled togive up its secret when once it has attracted the attention of anumber of differently equipped workers.The writer believes thatthew Reports will have conspicuously failed in one of their mainobjects if they do not serve t o stimulate a more catholic interestand in a small measure help the specialist' to take up a less detachedattitude towards science as a whole.X-Ray Methods of Exploring Crystal Structures.The progress made in this branch of investigation has alreadyreached a stage a t which the relative merits of the reflection anddiffraction methods may be appraised. The rapid advances madeby the users of the X-ray spectrometer1 are to a large extent duet o the simplicity of the method involved.The various glancingangles, el, 8,, O,, etc., a t which a plane of atoms serves as anefficient reflector, are first determined, and the results are theninterpreted by means of the equation nh= 2d sin 8. I n this equa-tion the factor n has the value l when the reflection is due to amutual reinforcement of pulses provided by successive chemicallyand cryst allograp hically identical planes (first-order reflection atthe angle 0,). Similarly, the co-operation of alternate planes leadsto the second-order reflection, O2 (?t=2), the mutual assistance ofevery third plane to the third-order reflection, O3 (n=3), and soon. The intensity of the reflection diminishes regularly in passingup the orders, except in cases where successive atomic planes differeither as to composition or to their distances apart; even ordersmay become stronger than odd orders, and certain orders inayvanish altogether. An admirable illustration of these effects isgiven by rock salt, the relative reflection intensities from the cubeand octahedral planes being :1st order.2nd order. 3rd order. 4th order.Face (100) ...... 100 18.7 6.25 too weak to measureThe regular decrease of intensity as the scale of orders isascended indicates both a chemical identity and an equality ofspacing of successive cube planes, whilst the periodical rise andfall of intensity in the octahedral plane premises some structuralpeculiarity or other. A glance a t the figure given in last year's,, (111) ......16'5 24'4 3 *1 4 -2W. H. Bragg, Phil. May., 1914, [vi], 27, 881MINERALOGICAL CHEMISTRY, 241Report will reveal the fact that successive cube planes are identicalin every respect, whereas octahedral planes are alternately occupiedby chlorine and sodium atoms. An example in which an “orderanomaly” has necessarily to be referred to’ a difference of spacingis afforded by the diamond (p. 243); in this case the entireabsence of the second-order reflection from the octahedral planeimmediately led t o the correct elucidation of the peculiar andhiglily interesting structure involved.The diffraction method of exploring crystal structure is, ofcourse, based on the production of a Laue radiogram, by placinga crystal in the path of a primary pencil of &--rays and receivingon a photographic plate both the undeviated pencil and thesecondary, deviated beams which are due t o the diffraction or re-flection effect of the internal planes of the crystal.An analysisof the radiogram is then made both with respect t o the patternand the relative intensities of the spots.A cursory examination of the published material is enough t oshow that the employment of the Bragg X-ray spectrometer leadsto a rapid determination of the general features of a particularstructure, and in the simpler cases may even obviate any appealt o the evidence afforded by the Laue radiogram. On the otherhand, although an unaided analysis of the radiogram is apt to leadto inconclusive and even erroneous results, it cannot be deniedthat it not only serves as a check on the spectrometric method, butalso brings with i t an additional quantitative precisioii in the finerdetails.This point will be emphasised under pyrites and hauerite(p. 246). It is obvious, then, that the two methods admirablysupplement one another.An important contribution is that of Friedel,2 who has pointedout t h a t the radiogram does not admit of a decision as t o whethera crystal is endowed with a centre of symmetry. An immediateresult of this is t o rule out any hope of obtaining direct’ evidencesof enantiomorphism, and the announcement in last year’s Report,for example, of a difference between 6- and I-crystals of quartz,must accordingly be withdrawn. A useful table of the theoreticalsymmetry types of radiogram is appended by Friedel, from whichi t appears t h a t the thirty-two classes of crystal symmetry can onlyyield a total of eleven types.I n the cubic system, for instance,the holohedral, holoaxial, and tetrahedral classes are indistinguish-able radiographically, all yielding a holohedral pattern, but thetetartohedral and pyritoliedral classes must both afford the pyrito-hedral type. This deduction, as far as pyrit.es is concerned, hasbeen fully confirmed. The keen disappointment felt by crystallo-a G. Friedel, Compt. rmd., 1913, 157, 1533.REP.--VOL. XI. 242 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.graphers when it was first announced that the undoubtedly tetra-heldral zinc blende yields a holohedral radiogram must now betempered by the knowledge that' no other result could in the natureof things be expected.The year has been characterised by a considerable amount ofwork, and apparently of quite a fundamental character as far asphysics is concerned, on the Debye effects-the influence oftemperature on the atomic vibrations in the crystal edifice.It isinteresting to note that Debye's theory has received a qualitativeconfirmation.4 An obvious diminution in the intensity of thesecondary beams is exhibited by mica when it is kept a t a tempera-ture of 400°, and spots that are faintly visible at the ordinarytemperature are not a t all represented on the photograph. Again,there is no trace whatsoever of any spots when X-rays are passedthrough a crystal of rock salt heated a t 620O.The spectroscopicmethod also indicates a weakening in intensity of reflected rayswith rise of temperature; moreover, the increase of the value d,owing to the general expansion of the crystal (rock salt), is clearlyindicated by a corresponding diminution of the glancing angle, 8,a result which is to be expected from the form of the reflectionequation.X-Ray spectra have been mapped out by de Broglieb by thefollowing ingenious device. Rays are allowed to impinge on acrystal which is mounted on a spectrometer, rotated by clockworka t a rate equal to loo per hour. An intense reflection of rays isthereby produced, and is registered on a photographic plate, a t eachshort interval of time during which the glancing angle has a valuesatisfying the reflection equation nh = 2d sin 8.Since both plateand anticathode are kept fixed throughout ths experiment, thepoint of impingement of the reflected rays gradually moves acrossthe plate, and the latter, on development, reveals the spectrum asa series of bands and lines with great definition. The spectrumappears t o range over several octaves.Stereochemical Deductions from the X-Rag Methods ofInvestigation.We may now proceed to review in detail those crystalline sub-stances the structures of which have been elucidated with a reason-able degree of certainty. I n addition to those now discussed, thereare a certain number of others, amongst them being sulphur andquartz,6 which are admittedly imperfectly characterised ; and theVarious papers in PhysikaZ.Zeitsch. and Ann. Physik.M. Laue and J. St. van der Lingen, Phpika2. Zeitsch., 1914, 15, 75.M. de Broglie, J. Phys., 1914, [v], 4, 101.W. H. Bragg, Proc. Roy. Xoc., 1914, [ A ] , 89, 575MINERALOGICAL CHEMISTRY. 243same is perhaps true for the calcite group of minerals and sodiumnitrate, although the results already obtained vindicate the closeststructural analogies which have ever been claimed for these sub-stances. It seems advisable to postpone discussion of such casesuntil additional evidence is available.Copper.7-Although in point of time the latest to be investi-gated, it is convenient to make a commencement with this element,since the structure proves t o be the simplest of all crystals so farexamined.I n spite of this structural simplicity, however, theactual exploration was accompanied by considerable experimentaldifficulties, owing to the softness of the material, which renders i tvery susceptible to profound structural derangements whenattempts are made to grind artificial planes with appropriatecrystallographic orientations. This difficulty was finally sur-mounted by etching away the external layers so as t o lay bare theFIG. 1.unimpaired structure. The results unhesitatingly point to anarrangement of atoms on the pattern of the centred-face cubeshown in Fig. 1. This arrangement corresponds with the closestpacked cubic arrangement of equal spheres, and with a rhombicdodecahedra1 domain of influence for each copper atom.Diamond.*-The main structural peculiarity of the diamond isthe tetrahedral environment of each carbon atom by its fourimmediate neighbours, a, feature which is obviously of prime stereo-chemical significance.I n the absence of a model, the real natureof the structure is perhaps best realised by the following construc-tion. Take eight sinall cubic cells, four being empty, whilst theremaining four are of the kind shown on a large scale in Fig. 2,that is, provided with a whole carbon atom, 2, a t the centre, anda quarter of an atom a t each of four corners, numbered 1, 6, 7,8, selected tetrahedrally. Now interpose the two kinds of cells7 W. L. Bragg, Phil. illng., 1914, [vi], 28, 355; A., ii, 775.8 W.H. arid W. L. Bragg, Proc. Boy. Soc., 1914, [ A ] , 89, 277.R 244 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.in the way shown in Fig. 3, so that the occupied cells are identicallyorientated, but separated from each other by the empty cells. Itis seen that four quarter-atoms now meet a t the centre of thefigure numbered 1, and thus make up a whole atom. (It needscarcely be added that an infinite repetition of the eight cellsthroughout space would lead t o a similar fourfold completion ofall the quarter atoms, so that the assemblage consists in its entiretyof whole atoms.) An imaginative chemist will doubtless see clearevidences of a six-carbon ring-shown, as t o five atoms, by, say,7, 2, 1, 3, 9-and may accordingly be disposed to trace an embryonicbenzene configuration in the structure. Without laying too muchstress on this point, it will be generally agreed, in view of theperfect tetrahedral environment of all the atoms, that the diamondstructure can be regarded as a perfect symbol of the chemistry ofcarbon.I n addition t o its four immediate neighbours a t the distanceFIG.2. FIG. 3.12 71&.a/2, where n denotes the edge-length of the cubelet, eachcarbon atom is symmetrically surrounded by twelve other atoms atthe somewhat greater distance J 5. C I . These twelve atoms have tobe taken into account in any determination of the sphere of influenceof each atom. It has been pointed out by Poppls that the sphereof influence is accordingly a regular tetrahedron, each of the fourcorners of which has been modified by the dodecaliedron to suchan extent that the original triangular faces of the former are cutaway to regular hexagons.Inasmuch as this modified tetrahedrondeviates considerably from the form of a sphere, it is evident thatthe diamond structure corresponds with an extremely loose-packedsystem of spheres. This peculiarity has led Barlow10 t o offer theopinion t h a t X-ray methods may be incapable of correctly de-lo W. Rarlow, Proc. Roy. Xoc., 1914, [ A ] , 91, 1 ; A, ii, 843.L. Foppl, Physikal. Zeitsch., 1914, 15, 191MINERALOGICAL CHEMISTRY. 245ciphering the intimate structure of crystals, and to propose certainmodifications which might tend to bridge the apparent gulfbetween the requirements of the Barlow-Pope theory and the ex-periments of W.H. and W. L. Bragg.Although in the past the many crystallographical and physicalaspects of the diamond have been the subject of repeated inquiries,it cannot be said that the, results obtained are altogether satis-factory. Thus, the frequent twinning on the cube plane wouldseem t o negative the structural subsistence of a correspondingplaiie of symmetry, the effect of which would be to relegate thecrystal t o the tetrahedral class. This tetrahedral symmetry would,however, in turn imply electric polarity along the trigonal axes ;the recent exhaustive work of van der Veen,ll however, failed totrace any such polarity. The stereochemical elucidation of thediamond appears to have furnished a satisfactory answer to theseFIG.4.contradictions the cube planes of the structure are not planesof symmetry, nor are the trigonal axes polar, for althoughsuccessive octahedral planes of atoms are alternately spaced a tdistances in the ratio 1: 3, the structure is nevertheless identicalwhen viewed along the trigonal axes in opposite directions.Ziizc Blei~cbe,l3 ZnS.-The atomic arrangement in this mineral isprecisely analogous to that of the foregoing, and it is thereforeonly necessary t o suppose the quarter atoms on the cube cornersto be composed of zinc, whilst the central atoms are sulphur. Thestructure indicated in Fig. 4 is thus obtained. Although thearrangement of the material is obviously the same, the difference incomposition brings about a fundamental structural difference ascompared with the diamond, for successive octahedral planes areA.L. W. E. van cler Veen, Zeitsch. Kryst. Min., 1913, 51, 545.l2 P. P. Ewald, Ann. PSysik, 1914, Liv], 44, 257.J3 W. L. Bragg, Proc. Bop, Soc., 1914, [ A ] , 89, 468; A,, ii, l S l 246 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.now strewn with zinc and with sulphur atoms respectively; thetrigonal axes are now polar, and the symmetry of the assemblagecorresponds exactly with that of the mineral. It will be seen thateach zinc atom is tetrahedrally surrounded by four sulphur atoms,and more distant’ly by twelve other zinc atoms. It must be addedthat a careful analysis of the original Laue radiograms nowcorroborates Bragg’s structure in every detail.12FZuorspur,l3 CaF,.-The structure of this substance has also becninvestigated by W.L. Bragg, and apparently with satisfactoryresults. The atomic arrangement advocated may again bevisualised by the construction adopt’ed for the diamond and zincblende. I n the latter assemblage, substitute F for S, and Ca forZn, so that each of the four occupied cubic cells corresponds witha group -CaF. I n order to make up an assemblage correspondingin composition with CaF,, it is obviously necessary t o allocate anFIG. 5. FIG. 6.additikal number of fluorine atoms; if each of the four emptycube cells of the diamond and zinc blende structures be now sup-posed to contain a centrally situated fluorine atom, the assemblageof Fig. 5 is finally obtained.The symmetry of the assemblagecorresponds with the holohedral cubic class, and is in agreementwith all available crystallographic knowledge.Pym’tes,l3 FeS,, and Hauem’te, MnS,.-The geometricalform and physical properties of these two isornorphous mineralsare so replete with items of interest as to warrant an expectationthat the underlying structure should also offer unique peculiarities.This expectation has been conspicuously realised, for Bragg’sanalysis, effected by the help of the X-ray spectrometer, proves theintimate atomic structure to be in accordance with one of theFedorov-Schonflies point systems, namely, the asymmorphoussystem ( 2 5 ) ~ ’ of Fedorov’s ~lassification.1~ Moreover, a detailedl4 E. 8. Fedorov, Zeitsch.Kryst. Min., 1914, 54, 163.IroMINERALOGICAL CHEMISTRY. 241analysis and comparison of the radiograms obtained by Ewaldserve to fix the relative spacing of the iron (or manganese) andsulphur atoms with great precision.It betrays a generalsimilarity to the structure adopted for fluorspar, but the follow-ing important difference must be noted : each sulphur atom is notsituated a t the centre of its cubic cell, but has been moved alongan appropriate cell diagonal away from an iron atom towards anunoccupied corney, so that its distance from the iron atom is aboutfour times its distance from the unoccupied corner. It will beobserved that the displacements of the sulphur atoms are notalong a series of parallel diagonals, but rather according to adefinite scheme in which the various diagonals are favoured inturn.The symmetry of the structure is in complete harmony withthat of the crystal. Preliminary experiments carried out withcrystals of hauerite also led to the same general conclusions.The publication of these results furnished Ewald 16 with an indis-pensable foundation for the exhaustive analysis of Laue’s radio-grams. This analysis proves the essential correctness of Bragg’sdeductions, the sole modification being that in hauerite the posi-tion of the sulphur atom is obtained by dividing the cube diagonalin the exact ratio 4: 1, whilst in pyrites the sulphur is a littlenearer to the metallic atom than Bragg supposed, the correspond-ing ratio being 7 : 2.It may be mentioned, in concluding this section.that a valuableappreciation of the recent advances in stereochemical crystallo-graphy has been contributed by Groth,lS who concludes that regu-larities, like the symmetrical environment of each sulphur atom inzinc blende by four zinc atoms, must be held to indicate that nosulphur atom is specifically united to any single zinc atom in thechemical sense, and it must therefore be inferred that chemicalmolecules do not exist in the crystalline condition. The writerbelieves this conclusion to bs not a little premature. Although thefrequent formation of crystalline double salts, like the alums,clearly proves that crystallisation is accompanied by a certainamount of pooling of chemical affinities, i t must nevertheless beapparent that t o deny the existence of the chemical molecule wouldopen up most extensive possibilities of isomeric change, optical in-versions, and so forth, whenever the crystalline structure is brokendown by dissolving or fusing the substance.Such changes havenever been observed. The reader need hardly be reminded thatX-ray methods of investigation are still in their infancy, and that,The structure is represented in Fig. 6 .l5 P. P. Ewald, Physikal. Zeitsch., 1914, 15, 399l6 P. Groth, Zeitsch. Kryst. Min., 1914, 54, 65 ; A . , ii, 719..248 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.however remarkable the progress already made, it must be stronglyemphasised that all theoretical interpretations must be regarded asprovisional.Happily, the originality of the workers who aredevoting themselves t o the study of this new branch of investiga-tion is in itself a sufficient guarantee, not only t h a t steady pro-gress will be made, but also t’hat entirely new developnients maybe expected.Crystullo-chemical Analysis.The trustworthiness of Fedorov’s method of identifying a sub-stance by virtue of its characteristic crystalline form has been putt o a further searching test, with results that demonstrate itsextremely practical value to the chemist. The following two ex-amples taken from a brochure written by Fedorov17 will be enought o show that the method can be successfully applied even when theavailable amount of material will not admit of a chemical analysis.I n a research on the chemical processes that accompany theripening of seeds, N.R. Nedokuschaev was seriously impededowing t’o the small quantities of the products isolated. The latterhe eventually handed over t o M. B. Orelkin and G. Pigulevski,pupils of Professor Fedorov, in order to see whether crystalmeasurements would afford any clue to the identity of the sub-stances, and the crystal analysts succeeded in identifying asparagine, C4H,03N, ; allantoin, C4H&,N4 ; histidine, C,H,0,N3 ;arginine, C,H,,0,N4 ; and betaine, C,H,,O,N. Of these com-pounds, histidine was identified in the form of its hydrochlorideand betaine as platinichloride.The same chemical crystallographers rendered inestimable assist-ance t o A. A. Pomaski in a research on the teleutospores of therust fungus (Pucciuia graminis Persoon), and identified d-mannitol,C,H,(OH),, and ‘‘ vitaline.” The latter substance belongs to theclass of the crystalline albumens, and according t o Weyl has thecomposition C = 52.43, H= 7.12, N = 18.1, S = 0.55, 0 = 21.8 percent.Its molecular weight is unknown. The amount of vitalineplaced a t the crystallographers’ disposal was 0.05 gram, and wasfound t o be quite sufficient.It is evident that the method is destined to be of a special valueto the bio-chemist, and it is therefore desirable that every well-crystallised substance of complex chemical structure should begoniometrically measured, its geometrical form properly analysed,and its characteristic constants tabulated. It may be stated thatthe immense work connected with the tabulation of the thousandsof substances which have been measured and described by successivegenerations of crystallographers is drawing to a satisfactory corl-l7 “ Crystallochemiccd Analysis ” (Russian), 1914MINERALOGICAL CHEMISTRY.249clusion, the tables being already in the proof stage. A discussionof some of the points arising out of the study of these tables hasbeen recently given by Fedorov,ls and the quickest methods ofmeasuring a crystal and tracing its identity by means of the tableshas also been described by the creator of the method.19Crystal Symmetry and Molecular Symmetry.The exceedingly rapid development of stereochemical ideas sincethe enunciation of the tetrahedral hypothesis of carbon by van’tHoff and Le Be1 would lead one to suppose that the possibility ofthe existence of a close connexion between molecular and crystalsymmetry must have occurred to the minds of many workers.Asfar as published material goes, the credit of the first examinationof this subject appears to belong to the veteran mineralogistTschermak, who, in 1903, pointed out certain numerical regulari-ties, now known as “Tschermak’s rules,” that subsist between thegeneral composition of a substance and the form or symmetry ofthe crystal. Thus, the formulz Al,O,,NaNO,, Na10,,3H20, inwhich the number 3 constantly recurs, represent well-known casesof rhombohedra1 or trigonal crystals ; ZrSiO,,SnO,, tetragonal ;MgSnF6,6H,0, hexagonal; CBr,, cubic; and so on. Several con-tributions t o the same subject have been published during the lastten years, and have received due notice in these reports, so that i tonly remains t o give an account of some recent notable develop-ments.The general relationship between crystal symmetry and molecular symmetry has been discussed by Fedorov,20 who with allmodesty puts forward the conception of a crystal-molecule, consist-ing of 7% chemical molecules, where n = l in the exceedingly rarecase in which the symmetry of the chemical configuration corre-sponds exactly with the symmetry of the crystal.I f , on the otherhand, the molecule is completely asymmetric, and the crystal formis, say, rhombic sphenoidal, the number I I =4 ; more generally, ifp represents the “ symmetry magnitude ” of the crystal, m the sym-metry magnitude of the molecular configuration, then 11 =p/m.It will be observed that the possibility of the chemical symmetrybeing greater than the cryst.al symmetry is not considered ;Fedorov believes it t o be mathematically impossible, which is per-haps another way of stating that the assumption is contrary t o thedictates of common sense.It will be readily realised that in thosecases in which the symmetry of the chemical formula is incom-’8 E. S. Fedorov, Zeitsch. Kryst. Min., 1914, 53, 337.l9 Ibid., 1914, 54, 1 7 ; A . , ii. 718.2o Ibid., 1913, 52, 22; A,, 1913, ii, 305250 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.patible with the observed symmetry of the crystal, there must bean error somewhere-if not in the crystal determination, then inthe chemical formula.One or two cases are discussed, andchemists are advised to set to work on the congenial task ofdevising new f ormulz.Any discussion of a chemical formula obviously implies, or leadsup to, a discussion of the valency, and, stereochemically, the sym-metry of the atom. The main result of Fedorov’s analysis is thatcarbon amongst the quadrivalent ele’ments, and aluminiumamongst the tervalent element-s, exhibit the clearest, although notperfect, crystallochemical evidences of a symmetrical dispositionof the chemical valencies. Those who peruse Fedorov’s paper willno doubt come across one or two statements of doubtful cogency;in particular, the reporter does not see why the bromine atom inCBr(C,H,), and ths hydrogen atom in CHI, are “ certainly devoidof a threefold axis of symmetry.”The same problem, a t least as far as carbon is concerned, hasbeen attacked by Wahl,21 who, with rare experimental skill, hasnot only succeeded in determining the crystal system of some fortycarbon compounds which are liquids or gases a t the ordinarytemperature, but also the system of crystallised argon and nitrogen !ThO determinations were, of course, carried out optically.Argonand nitrogen are optically isotropic, and presumably belong to thecubic system. I n the case of carbon compounds, there is hmoderately general agreement between the symmetq of the crystaland the symmetry of the molecular configuration; for example,both methane and its tetrasubstituted derivatives, CBr4, CI,,C(N02),, and CMe,, are cubic ; the disubstituted products, CH2Cl,,CH,Br2, CH212, are rhombic, and the trisubstituted derivative,iodoform, and also ethane are hexagonal.The author accordinglydraws the conclusion or “t,heory” that “the symmetry of thechemical molecule of a substance determines its crystal symmetry.”The author is aware that some of the compounds examined will notfit in with the theory unless the usual spacial formula= are arbi-trarily deformed to correspond with the observed determinations.For example, methyl chloride, bromide, and iodide are monoclinic.These irregularities the author seeks to explain by means of certainideas on the magnitude of atomic spheres of influence, but theseexplanations are, of course, only paraphrases of the anomalies,and do not in any way restore the generality of the theory.Avaluable research which bears on the subject is that of Drugman,22W. Wahl, Proc. Roy. Xoc., 1912, [ A ] , 87, 371 ; 1913, [ A ] , 88, 3 4 ; 1913, [ A ] ,89, 327; 1914, [ A ] , 90, 1 ; A . , 1912, ii, 1044; 1913, i, 693; 1913, ii, 1031 ;1914, ii, 348.J, Drugman, Zeitsch, Kqjst. Min., 1914, 53, 240 ; A . , i, 140MINERALOGICAL CHEMISTRY. 251who has made an extensive study of certain symmetrical dibasic,aliphatic acids. Thus, malonic, glutaric, dimethylmalonic, diethyl-malonic, b&dimethylglutariu, ad-dihydroxy-ad-dimethylglutaric,and n-pimelic acids should all presumably crystallise in the rhombicsystem.I n reality, however, the respective systems of crystallisa-tion are triclinic, monoclinic, tetragonal (trapezohedral), triclinic,monoclinic, triclinic, and monoclinic.Attention may here be drawn to a study25 of a suite of fourisomeric dinitrobenzoic acids, and also to an exhaustive examinationof a series of monosubstituted benzoic acids.24 Certain morphotropicrelationships are discussed in the latter paper, but a careful perusalof the memoir shows that the allocation of indices is of a pro-nounced subjective character, and is directly responsible for someof the alleged morphotropic similarities.The Barlow-Pope Theory.An interesting and illuminating discussion on its merits andshortcomings has taken place between the authors of the theoryand Richsrds.25 The chief bone of contention is the question ast o the comprelssibility of atoms.It is well known that the Ameri-can chemist has from time t o time adduced a large amount ofexperimental evidence in f avour of atomic compressibility,26 whilston the othelr hand the propounders of the theory would seem tofind no difficulty in elucidating crystal structure without havingrecourse to any such assumption. Perhaps the most importantpoint made by Richards is the proof that Le Bas’ interpretationof a volume relationship, C =4H, is inconclusive, since alternativeassumptions that C=2H or 3H or 5H will fit in equally well withthe known molecular volumes of the paraffin hydrocarbons con-cerned.The theory has also been described by Pope in his PresidentialAddress to Section B a t the Australian meeting of the BritishAssociation ; although no new matter is introduced, this contribu-tion will be found t o be a useful outline of the present positionheld by the propounders of the theory.The practical development of the theory has been actively prose-cuted during the year in a series of papers constituting numbersV-VII of ‘‘ Morphological Studies of Benzene Derivatives,” inwhich an enormous number of new crystal measurements are23 R.Gossner, Zzitseh. Kryst. Min., 1914, 53, 489; A . , i, 406.25 T. W. Richards, J. Amer. Chem. Soc., 1913, 35, 381; 1914, 36, 1686;26 T. W. Richards, ibid., 1914, 36, 2417.H. Steinmetz, ibid., 463 ; A . , i, 405.W. Rarlow and W.J. Pope, ibid., 1675, 1694 ; A., 1913, ii, 483 ; 1914, ii, 719252 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.described and discussed in t’he light of the t’heory.27 I n additiont o crystal measurements, the seventh memoir contains an accountof observat,ions on the principal magnetic directions of a numberof monoclinic crystals, which were carried out in the hope that theymight provide critical criteria for the correct setting of crystals.If by the term “correct setting” the authors imply the same mean-ing as Fedorov, it’ would seem that they have omitted to take intoconsideration the fact that magnetic induction is an ellipsoidalproperty, and on that account is entirely useless for the end inview. The authors’ general conclusion is that their work providesa verification of the Barlow-Pope conception of valency volume.Thermal Studies of Mirberals.Considecrabla developments have taken place during the year inthe study of the equilibrium conditions of mineral mixtures, andas marked progress has to be noted both in the methods of workingand in the actual results obtained, it will be convenient to dividethe short r6sum6 into two sections.Methods.Whenever feasible, i t is bed to measure temperatures by meansof a platinum-platinum-rhodium thermo-element.The methods ofconstructing, calibrating, and using this invaluable form of thermo-meter have been recently described in a series of papers dealingwith calorimeters and thermo-electric instruments of precision byWhite.28 The experimental ’error varieB with the amount ofmaterial taken and the character of the contact between the elementand the mineral charge.With a careful worker the error may fallas low as +2O. The limit of applicability appears to lie a t about1600O (the melting point of platinum is 1755O). F o r higher tem-peratures recourse must be, had to far less accurate methods; it istherefore fortunate that the great majority of minerals melt below1600°, the orly exceptions being a few oxides and sulphides whichhave long been known to be highly refractory materials.Since the direct observation of the melting point, as, for example,by means of the Doelter heating-microscope, must always remain acumbrous and quite untrustworthy procedure owing t o the diffi-culty of seeing whether a viscous mass is solid o r liquid, it hasbeen necessary t o devise indirect or non-visual methods.The earlier27 H. E. Armstrong,’.R. T. Colgate, and E. H. Rodd, Proc. Roy. Soc., 1914, [ A ] ,90, 111 ; C. S. Mummery, ibid., 455 ; H. E. ArmArong and E. H. Rodd, ibid.,463 ; A., ii, 443 ; i, 1062 ; ii, 768.28 W. P. White, J. Ainer. Ckem, Soc., 1914, 36, 1856, 1868, 2292, 2313 ; A , ,ii, 767MINERALOGlCAL CHEMISTRY. 253method employed by the American workers was to raise the tem-perature of the crucible a t a regular rate by means of an electricsource of heat, and plot temperature-time curves; any absorption ofheat, whether due to liquefaction or t o a polymorphous change,brings about a decrease in the rate of the rise of temperature. Itmust be noted that the converse method, namely, that of coolingcurves, is untrustworthy owing to supercooling effects.A second andlater method, and one which is now extensively employed, is that of‘‘ quenching.” 29 The crucible is kept a t a constant temperaturefor an hour or so, and then suddenly quenched by dropping i t intoa dish of mercury. This sudden cooling transforms any moltenmaterial into glass, the presence of which is immediately recognisedby a subsequent examination of the product under the polarisingmicroscope. By trying successive temperatures a t 5 O or loointervals, the temperature a t which glass is first formed is thusdiscovered. By trying higher temperatures, the point may then bedetermined a t which the whole of the crucible contents, afterquenching, proves t o take the form of glass.The two points thusdiscovered are the solidus and liquidus temperatures for the givenmixture. It is obvious that the microscopic examination also leadsto the identification of the various crystalline phases which are inequilibrium with the fusion.Another point that has been subjected to a searching examina-tion is the all-important question of the condition of the materialunder investigation. It goes without saying that this materialshould be chemically pure ; t3he silicates employed have thereforeto be made artificially from pure silica and metallic oxides. Now,although i t is an easy matter to obtain a product of a definitecomposition, it is extremely hard to ensure that i t shall consist ofa single chemical species.Suppose, for example, i t is desired t oprepare a series of pure magnesia-lime pyroxenes ranging in com-position between diopside, CaXlg(SiO,),, and clino-enstatite,MgSiO,. It is not’ sumcient to fuse together the correct propor-tions of lime, magnesia, and silica and allow the mass to cool, forin systems which are characterised by the separation of mixedcrystals, the liquidus has not the same composition as the solidus,and crystallisation is accompanied by a gradual impoverishmentof the liquid portion in one or more of the components. It isimpossible t o keep down the rate of cooling to such an extent ast o allow a perfect attainment of true equilibrium conditions.Theoretica,lly, the first crop of mixed crystals should redissolve asthe temperature is lowered, if solid and liquid were in perfectequilibrium ; in practice, the final crystalline product must always?Y N.1,. Bowen, Amer. J. Xci., 1914, [iv], 38, 207 ; A . , ii, 772254 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.be extremely variable from point to point throughout the mass.The obtainment of homogeneous equilibrium material is of greatimportance, especially when it is t o be used for subsequentresearches. The only way to obtain it is to quench the fusedpyroxene to a homogeneous glass, and then maintain this glassfor some time a t a temperature which is just a little below the pointat which it commences to liquefy. The homogeneous glass thencrystallises to a conglomerate of homogeneous mixed crystals.Results.Binary System,30 MgO-SO,.-This system, which has been care-fully determined by Bowen and Andersen, is characterised by theappearance of two well-defined compounds-the orthosilicate for-sterite, Mg,SiO,, and the metasilicate clino-enstatite, MgSiO,.Thelatter substance was on a previous occasion shown t o occur in fourpolymorphous forms, t o which a fifth (“a-MgSiOS”) was addedsomewhat later; it is now pointed out that this fifth form isoptically identical with forsterite, and owes its origin t o the disso-ciation of clino-enstatite a t its reputed melting point into the ortho-silicate and free silica : 2MgSi0, Z Mg,SiO, + SiO,. The mainfeatures of the system may be epitomised as follows: Magnesia(periclase) has a very high melting point, perhaps 2800°, and formsan eutectic (1850O) with forsterite, which itself melts at 1890O.There is no eutectic between forsterite and clino-enstatite; theeutectic between the latter and cristobalite (silica) lies at 1543O.C’ristobalite melts at 1 6 2 5 O olr higher.Ternary System,29 Diopside-Forst erit e-Silica.-Although thescope of this system only extends over three-eighk of the systemCa0-Mg0-Si0, the most important portion of the latter is reallyincluded, f o r the part under consideration embraces the pyroxeneand olivine families of minerals.The quenching method wasemployed, the material used being a previously prepared series oflime-magnesia metasilicates to which appropriate amounts of pureforsterite or silica were added.The exact temperature needed forcomplete liquefaction and also the identity of the phase which wasthe last to disappear (or, alternatively, the first to reappear whenthe temperature was slightly lowered) were determined for eachmixture. Such resulh are best illustrated by a spa‘ce model con-structed on a triangular base so as to allow of an exact representa-tion of the relative concentration of the three components, thevertical dimension being reserved for temperature. In a case likethe present, in which the three components do not form a con-tinuous series of solid solutions, the top of the model (see Fig. 7)Do N. L. Bowen and 0. Andersen, Amer. J. Sci., 1914, [iv], 37, 487 ; A . , ii, 562MINERALOGICAL CHEMISTRY.255will be divided by boundary curves into a number of fields orareas, each representing the fusion surface of a solid phase. Inthe case in point there are four such fields: forsterite, pyroxene&,cristobalite, and tridymite, the latter area being very circum-scribed. Two important points may be noticed with regard to thepyroxenes : (1) the area extends uninterruptedly from diopsideacross the whole of the diagram, thus indicating the existence of acomplete series of mixed crystals ranging from CaMg(SiO,), toMgSiO, ; (2) the boundary line, forsterite-pyroxene, crosses thedotted line, MgSi0,-CaMg( SiO,),, a t a point corresponding withabout 70 per cent. of diopside, from which it follows that allmetasilicate fusions, excepting those very rich in lime, will a t thecommencement of crystallisation deposit forsterite (olivine).Pyro-xenes will only appear Then the silica-enrichment of the liquidresidue brings its composition to the f orsterite-pyroxene boundary.PIG. 7.CaMg( SiO&Mg,SiO, MgSiO, Sio,If the cooling is sufficiently slow, the originally deposited f orsteritewill then commence to dissolve in the silica originally liberated;if, however, the rate of cooling does not admit of the attainmentof equilibrium, the f orsterite will remain permanently, and theexcess of silica will eventually crystallise out. We have here anelegant explanation of two petrological enigmas: (1) the not infre-quent occurrence of olivine in the presence of free quartz; (2) theunmistakable signs in many rocks of a gradual resorption of olivinewith the production of pyroxene.Again, the zonal developmentof pyroxenes generally shows a preponderance of magnesia in thenucleus. This is what would be expected from Bowen’s work,since the magnesia pyroxenes have higher melting points than thelime pyroxenes. The paper includes amongst other importantmatters an extensive series of optical determinations of the wholeseries of pure lime-magnesia pyroxenes256 ANNUAL REPORTS ON THE PROGRESS OF’ CHEMISTRYPolymorphism.Since it is generally agreed that two or more materials of thesame percentage composition must always be regarded as ( ( poly-morphous modifications ” until definite experimental evidence neces-sitates the employment of different formulae, it is obvious that theterm polymorphism can only be defined in negative chemical terms.The history of chemist’ry abounds with numerous proofs that thepolymorphs of yesterday are the isomerides of to-day: as chemicalmeans of detecting differences of structure have become more andmore refined the more imposing has become the list of instancesin which polymorphous modifications have finally received the hall-mark of chemical respectability.Especially resisting to this civilis-ing influence, howe’ver, are the old stack cases of polymorphismin which the physical differences are satisfactorily interpreted asmerely due to a different space-lattice arrangement of the samechemical molecule. The interpretation is sufficient because thereis no evidence of any persistence of the characteristic physicaldifferences when once the c7ystal structure is broken down ; i t musthold the field until such time as chemical differences have beenproved by tests or reactions applied in the crystalline state.It is,however, very questionable whether the chemistry of the solidcondition has yet seen the light of day; if in truth it has, the infantmust indeed be preternaturally silent.One of the consequences of the rapid growth of physical chemistryhas been the increasing tendency t o apply physical evidence as ameans of attacking such chemical problems as present insuperableexperimental difficulties when approached in more orthodox ways.I n the abstract, this is all to the good, but in practice there mustalways remain a doubt as to how far any such physical considera-tions have any real chemical bearing.This remark appears to beparticularly relevant to Tammann’s (‘ atomic theory of poly-morphism.” 31 Tammann believes that certain cases of polymorph-ism are to be regarded as simply due to a space lattice rearrange-ment, whilst others result from chemical polymerism. The criterionto be applied in the discrimination between the two categories isto study the stability relationships of the polymorphs over anextended range of temperature and pressure, and in this waydiscover whether the two forms can coexist in stable equilibriumunder varying conditions. According t o Tarnmann, stable co-existence implies polymerism, whilst cases, in which one form isentirely metastable with regard to the other, are deemed to beinstances of a space lattice rearrangement.The results of an31 G . Tammann, Zeitsch. physikal. Chem., 1913, 82, 172 ; A., 1913, ii, 193MINERALOGICAL CHEMISTRY. 257examination of some thirty organic compounds lead Tninmaniito the conclusion that most cases of polymorphism belong t o thelatter category. Ice, however, is held to be a notable exception.Although the five forms of ice, which are now known as a resultof the careful work of Bridgman,32 would be regarded as examplesof polymerism, a sixth form (the so-called modification III’), whichTammann has recently added t o the list,33 is, according to thesame authority, t o be regarded as a space lattice rearrangement ofthe form 111.Bridgmaii34 is, however, of the opinion that 111’is not merely metastable, but has no existence whatever. Withregard to the validity of his physical criterion, Tammann adducescertain thermodynamical considerations in its support, but Smits 35is of the opinion that the process of reasoning amounts to puttinga preconceived idea into the thermodynamica1 machine and subse-quently extracting the same after a respectful interval.Of an entirely different nature is Smits’s 36 theory of dynamicpolymorphism (or ‘‘ allotropy ”). Polymorphism is held to be initself a definite indication of the existence of more t’han onechemical species, and the two (or more) isomerides are supposedto co-exist both in the fused substance and in the various poly-morphous modifications in a state of dynamic equilibrium.If thetwo forms are denoted by A and B, then fall of temperature isaccompanied by a steady shifting of the equilibrium in the sense,A -+B. The transition temperature is held t o be the point a twhich one crystal structure will not allow of any further change,A -+B; accordingly, a new space lattice structure of a moreaccomodating nature makes its appearance. By way of criticism,i t is only necessary to recall and contrast the perfect limpidity ofall polymorplious modifications with the turbidity that invariablyaccompanies any serious contarnination of any crystalline substancewith another (even if the contamination be isomorphous), t o seethat Smits’s theory can scarcely commend itself to the pure crys-tallographer; but, however this may be, it must be stated thatthe theory has served to interpret a considerable number of curiousfacts. The theory can be invoked,37 for example, in order t ointerpret the fact that the melting point of rhombic sulphur varieswith the temperature a t which the specimen has been previouslymaintained.It is true that the observed temperature difference52 P. B. I3ridgni;in, Proc. Asizer. Acnd., 1912, 47, 441, 558 ; Zeitsch. anor!].3J G. Tamm:inn, Zeitsch.physikcc1. Chem., 1913, 84, 257; A . , 1913, ii, 935.3-1 P. R. Bridgman, ibid., 1914, 86, 513 ; A., ii, 254.85 A. Smits, ibid., 1913, 82, 657 ; 84, 250; A , , 1913, ii, 393, 933,s6 Ibid., 1911, 76, 421, and various subsequent papers.::7 Idem., ibid., 1913, 83, 221 ; A., 1913, ii, 499.Chem., 1912, 77, 377 ; A , , 1913, ii, 39.REP.-VOL. XI.258 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.only amounts t o about 0*5O, but in the case of the polymorphousforms of silica, to be described presently, certain temperaturefluctuations have been observed, according to circumstances, in thevalues of one of the t'ransition points, and these variations are notso open to doubt, a t any rate on the score of minuteness.Fenner's Work o n Quartz, Tridymite, and Cristobalite.3~--Thetransition temperatures and general conditions of stability of thenumerous modifications of silica are indicated in Fig. 8, in whichthe experimentally determined temperatures are plotted againsttheoretical (imaginary) vapour pressures, all that.is implicd beingthat the curves express the general experimental work. It will beobserved that each mineral species exists in a t least two modifica-FIG. 8.Ptions (a and P ) . Now Fenner found the change u Z fl to differfrom a change of one mineral species into anot'her in the followingimportant particular : the former takes place rapidly immediatelya transition temperature is overstepped, whereas the latter changeis excessively slow. Quartz, tridymite, and cristobalite can co-existtogether without undergoing any serious transformation a t all,except a t very high temperatures; and in order t o determine thesetransition fxmperatures it was found necessary t o add a smallamount of sodium tungstate, which apparently exercises a cata-lytic function.The same stubborn persistence of metastable phasesis also characteristic of the devitrification of silica glass, and evenin the presencs of sodium tungstate the first crystallised product isC. N. Penner, Amer. J. Xci., 1913, [iv], 36, 331MINERALOGICAL CHEMISTRY. 259generally a metastable one. Fenner is therefore driven to theconclusion that the extreme slowness of the changes, quartztridymite t cristobalite, demands the assumption that each is adifferent chemical species. It is then possible t o refer thesluggishness of each transformation t o the time demanded by pro-cesses of polymerisation and depolymerisation. On the other hand,the two kinds of quartz (a and /3) are supposed to be composedof the same chemical molecules, the transformation being due toa slight modification of the crystal lattice; and the same is truefor the three tridymites (a, P1 and P2).The cristobalites, however,exhibit the iollowing curious behaviour : the transformation pointa Z fl is extremely variable and depends entirely on the previoushistory of the specimen. The higher the temperature a t whichi t was originally obtained from amorphous silica, the higher isthe observed transition temperature ; a-cristobalite changes intoP-cristobalite a t temperatures which range regularly between thelimits 220--274°, according as the original temperature of itsformation had a low or high value between the limits 1050-1600O.Fenner holds that this behaviour points to the co-existence incristobalite of two kinds of material in the sense of Smits’s theory,the idea being that if the equilibrium condition between thedynamic isomerides responds very slowly to a change of temperature,then the relative amounts in a given specimen will depend on itspast thermal history.The inimpretation of the quartz and cristobalite inter-relation-ships, based on the work of Endell and Riecke,39 has been indepen-dently undertaken by Smits and EndelL40 They conclude that thebehaviour of a- and P-quartz and cristobalite is referable t o theexistence of two kinds of chemical molecules, the relative amountsof which can be indicated by arranging the minerals in the order :a-cristobalite, P-cristobalite, P-quartz, a-quartz.Relationships like those subsisting between a- and /3-quartz, inwhich the two forms are extremely similar in optical propertiesand apparently only differ in symmetry, have long been knownamongst laboratory products. An excellent example of this‘‘ polysymmetry ” has been described during the present year.*lAbove a temperature of 1 8 O , butyranilide crystallises in the rhombicsystem, but the crystals are pseudotetragonal.On idlowing thespecimen to remain, a gradual equalisation of the angles takesplace, so that after a period of eighty-five days, the crystals havebecome truly tetragonal. On raising the temperature, the reverse39 I<. Endell and R. Riecke, Zcitsch. anorg. Clzem., 1912, 79, 239 ; A., 1913,ii, 134.4o A.Smits and K. Endell, ibid., 1913, 80, 176 ; A., 1913, ii, 318.41 C. T. R. Wilson, Proc. Roy. Xoc., 1914, [A], 90, 169.s 260 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.progress gradually sets in. The tetragonal form may be obtaineddirectly by keeping the crystallisation solution below loo.The work described in a previous Eeport 42 on the formation andstability relationship of pyrites and marcasite, FeS,, and zincblende and wurtzite, ZnS, has been considerably extcnded.43 Themethod employed in the estimation of the r e l a h e amounts ofpyrites and marcasite was that originally due to Stokes; thismethod has been carefully examined, and found t o be entirelytrustworthy when certain precautions are observed.44Anisotropic Liquids or '( Liquid Crystals."The important observation of Reinitzer that cholesteryl benzoatemelts at 140° to form a turbid, doubly refracting liquid, which,a t a temperature of 1 7 9 O , becomes clear and singly refracting,prompted Lehmann to undertake an extensive microscopic studyof the substance, Gattermann's syntheses of pazoxy-anisole and-phenetole, as also the preparation of several other compounds byother workers, soon led to a considerable augmentation of theavailable experimental material, and i t was not long beforeLehmann drew the conclusioa that such turbid liquid phases %reta be regarded as aggregates of "liquid crystals." It is well knownthat suggestions that the liquids are really emulsions were soonproved to be untenable.One cannot but infer from a close perusal both of the earlyand of the most recent books and memoirs, which have been SDlavishly written by Lehmann, that he has always been at least asmuch swayed by certain, peculiar theoretical considerations, asby his actual observations. If the term crystal connotes a homo-geneous mass of material, arranged on a space-lattice pattern, itis obviously no light matter to put a fluid mass to the proof oEexperiment; for many of the tests t o which a crystalline mediumcharacteristically responds are wholly or partly inapplicable, owingto this very fluidity : any att'empts to investigate elasticity,cohesion, or cleavage are quite unthinkable.The general opticalproperties, on the other hand, are certainly more hopeful, and i tis in the field of microscopic work that Lehmann has made mostvaluable observations.At the same time, however, it must beemphasised that' the possession of properties like double refractionand dichroism does not prove the liquid substances t o be crystalline;it merely indicates the existence of a definite regularity of struc-42 Aim. Rcpo~t; 1912, 266.43 E. T. Allen, J. L. Crenshaw, and H. E. Merwin, Amer. J. Xci., 1914, [iv], 88,44 E. T Allen a d ,J. L. Creiishaw, ibid., 371 ; A., ii, 850.393 ; A . , ii, 850MINERALOGICAL CHEMISTRY. 261ture of some kind or other, and exactly t o that extent, and nofurther, supports an analogy between them and crystals. It maybe noted that Lehmann, who is naturally never more clear-sightedthan when he is criticising the views of others, is himself fullyaware that optical evidence can never be regarded as complete:apropos, Vorlander’s assertion that his viscous substances must beregarded as uniaxial crystals because they exhibit a perfect uniaxialinterference figure, Lehmann states 45 : ( ( such a liquid mass canonly be regarded as a uniaxial liquid crystal provided it stroveduring its growth t o attain a square or six-sided outline.” Now,as a matter of fact, i t is just this evidence that one seeks for invain in Lehmann’s own published writings.Although the dia-grammatic drawings originally given for ammonium oleste areclearly hexagonal bipyramids,46 it is, nevertheless, stated in thetext that the outlines are really circular (‘owing to surfacetension,” and this admission is, of course, fully borne out by tlisphotographic reproductions.I n Lehmann’s latest book 45 thediagram no longer indicates any hexagonal character, whilst in aquite recent memoir47 there is even a mention of the form beingthat of a tetragonal bipyramid. Perhaps the clearest geometricalshapes of all are shown by Vorlander’s p-azoxybenzoic ester ; hiiteven in this case, as far as one can judge from a photograph, thcelongated shapes are due to a linear aggregation of more or lessglobular individuals, producing a segmentation effect.The cubic modification of silver iodide occupies an altogetlierexceptional position in the history of liquid crystals. It is onlyin recent years45 that Lehmann has given a new interpretatioiiof his early work on this substance and has claimed it’ t o be thefirst, known example of liquid crystals-because (‘ i t is impossibleby mechanical means t o prove the existence of an elasticity limit,”which means that silver iodide a t high temperatures becomes sosoft that it can no longer be regarded as a solid.Some remarlr-able researches on this point have been published during the year.48It has been shown that if every precaution as to chemical purityis empIoyed in the preparation and subsequent handling of tliesubstance, i t is much harder a t 550°, that is, 2 O below its meltingpoint, than is yellow phosphorus at the ordinary temperature.The addition of a small quantity of ordinary silver iodide immenselyincreases its plasticity.Since similar results were obtained by thesame authors with silver chloride and bromide, as well as with thehaloid salts of thsllium,49 it mnst be concluded that all these sub-45 ‘( Die neue Welt dei flnssigen Iiiistalle,” 1911. 41j ‘‘ Fiussige Kiistalle,” 1904.47 0. Lehmlarin, Zcitsch. Krgst. Min., 1913, 52, 597.48 C. Tubandt &d E. Loren;?, Zeitsch. physikal. Chem.) 1914, 87, 527 ; A , , ii, 516.49 €I. Stoltzenberg aLd M. E. Huth, ibid., 1910, 71, 641 ; A , , 1910, ii, 295262 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.stances must be definitely deleted from the list of liquid crystals.It must here be noted that t’he real point a t issue! is not so muchwhether any particular “ solid,” crystalline substance, say, silve’riodide, is so soft that i t would be a misnomer to term it a solid-a decision in either sense would not be particularly momentous-but rather whether the turbid, doubly refracting liquids arerestricted t o organic compounds of some molecular complexity, orwhether representatives are also to be found among simple binarycompounds.This is, indeed, a question of some Lheoreticalimportance.Enough has been said t o indicate that the use of the term liquidcrystal was quite premature from the outset. Some recent workon the geometrical form of ammonium oleate and pazoxybenzoicester has led the authors5O t o the conclusion that the shapes haveno similasity whatever with crystal forms, but are figures of revolu-tion of extraordinary complexity.Again, with regard t o theoptical properties of the more mobile substances,61 which, of course,cannot be said t o have any shape, it has been found that there isa continuous internal movement even in the smallest discreteglobule that can be isolated f o r purposes of observation. Themovement increases in intensity as the temperature is raised, anddoes not appear t o have any definitely regular character. Thecomposite nature of such globules was, of course, recognised byLehmann, who has frequently been compelled t o point out thata globule is not a liquid crystal, but an aggregate of liquidcrystals. It will be realised how impossible i t is to substantiatethe crystal character of such unpromising material.Homogeneous masses of material, suitable in every way forexperiments of some precision, first became available as a result ofVorlander’s 52 numerous synt’heses of complex aromatic derivatives,which, when melted on a glass slide, with or without a cover slip,yield a clear mass of doubly refracting liquid.When viewed inconvergent light there results in every case a normal uniaxial inter-ference figure with ail extraordinarily perfect definition ; moreover,if the subst-ance has an enantiomorphous molecular configuration,the interference figure shows rotatory polarisation. I n thicknessesgreater than 0.3 mm. the masses tend t o be slightly turbid.The values of the two indices of refraction E. and w have beendetermined with great accuracy.63 The fact that the existence of50 G.Friedel and 3’. Grandjean, Bdl. Sbc. f m i z c . Alin., 1910, 33, 192, 409:466 :A . , 1910, ii, 809, 1018.51 C. Manguin, Compt. ~ c n d . , 1912, 154, 1359; A., 1912, ii, 630.52 D. Vorlander, Ber., 1908, 41, 2033 ; A . , 1908, i, 641.5 j Aim. Beport, 1910, 11.“ KristallinischfliisvigeSubstanzen,” 1908MINERALOGICAL CHEMISTRY. 263biaxial liquid crystals has never been established whilst some fiftycompounds are definitely known to be uniaxial,5* is of superlativeimportance, inasmuch as it presents a statistical proof that thestructure is not crystalline (the proof is really very strong, sincenearly all chemical substances, having a complex chemical structure,crystallise in biaxial systems). The whole of ths firmly estab-lished properties, dichroism and the invariable straight extinctionwhen the “ c r y ~ t a l ’ ~ is resting on a “prism” face, the apparentabsence of double refraction in parallel light, and the perfectuniaxial figure in convergent light, when the (‘ crystal ” is restingon its base,” are in complete harmony with a structure analogousto that of an even-grained piece of wood.These considerations ledBose t o advance the “swarm theory,” which has dominated thework of the last few years.Bose’s Swarm Theory.65I n all theoretical discussions of the molecular theory of liquids i tis presupposed f o r purposes of simplicity that the molecule is spheri-cal. This supposition can, of course, only be true for liquefiedmonatomic elements; it can scarcely hold for a molecule like anis-aldazine, ~~eQ*C,H,*CH:N*N:CH*C,H,.QMe. If two or moreelongated molecules approach each other so closely that the meandistances of their centres of gravity are less than half ths lengthof the molecule, all free rotation must cease except about the direc-tion of elongation, and a stable tendency towards a parallelarrangement may set in.Now a large bundle o r swarm of parallelmolecules in a perfectly fluid condition and without any suspicionof a space-lattice arrangement will evidently possess the symmetryof a figure of rotation, and optically should behave as a uniaxialcrystal. Each swarm or homogeneous tract (“ liquid crystal ”) willbe quite clear and transparent, the turbidity of a large mass ofliquid being simply due to the reflection and diffusion of light a t themutual boundaries of the swarms.The average size of the swarmdecreases on heating; the point a t which the swarms become smallerthan the wavelength of light is the clearing point. Above thistemperature the liquid is to all intents and purposes singly refract-ing. A rough idea of the stiructures of isotropic and anisotropicliquids will be gained from Fig. 9, in which a is meant to illustratethe state of affairs in the ordinary isotropic liquid, which on coolingpasses into the swarms of parallel molecules shown in 6 ; the optical54 I). VorlZiuder, Physiknl. Zeitsch., 1914, 15, 141.55 E. Rose, ibicl., 1907, 8, 347, 513; A., 1907, ii, 443: ibid., 1908, 9, 708 ;A., 1908, ii, 1017264 ANNUAL REPORTS ON THE PROGRESS OF CHEMISTRY.behaviour of a single swarm, as exemplified by Vorlander’s viscoussubstance, is shown a t c.General Evidence in Support of the Theory.-The theory is ingood accord with all the general physico-chemical properties.Tobegin with, the change, crystal + anisotropic liquid, is not onlyquite sharp, but is also accompanied by a considerable absorptionof heat (about 20 calories). On the other hand, the change, aniso-tropic liquid + isotropic liquid, is not only chaxacterised by a lowheat absorption (about 1 calorie), but also appears to be a gradualprocess extending over one or two degrees of temperature.* Thisimportant fact has been brought out by Bose’s visconietric work.5GI n the case of anisaldazine the viscosity of the isotropic liquidincreases quite normally with the fall of temperature to the pointof turbidity, and then decreases very sharply to a minimum a t 2 Obelow the clearing or turbidity point.The viscosity then begins t osuffer the normal increase with fall of temperature, but it does notsucceed in attaining anything like the value enjoyed by the iso-FIG. 9.( a ) ( b ) (4tropic liquid. The lower value of the viscosity of the anisotropicliquid, as compared with the isotropic liquid, constitutes the so-called“ viscosity anomaly.” Its interpretation was given by Bose 57 ona kinetic basis. I f the form of the molecule be taken to be thatof an elongated ellipsoid of rotation, the viscosity of a swarm cantheoretically sink to two-thirds the value of the viscosity of the samesubstance in the isotropic condition; further, the supposition of .athree axial ellipsoid could account for even lower values.The ratioof the viscosities, anisotropic : isotropic is 0.65 for anisaldazine.It has also been found that the change, isotropic-+- anisotropic isalways accompanied by a moderate increase of density, as might beexpected from the closer packing that is obviously compatible withtl10 parallel arrangement of e810ngatred molecules.* The crystalline nature of anisotropic liquids was accepted by Roozeboom, whoaccordingly regarded the clearing point as the true melting point, the point a twhich the solid crystal melts to give the anisotropic liquid being regarded as apolymorphous transition point. This decision niust now be reversed ; a t any rate, itis more consequent to regard Roozeboom’s ‘( transition point ” as the true meltingpoint.E. Bose and F. Conrat, Physikal. Zcitsch,., 1908, 9, 169 ; A . , 1908, ii, 258.57 E. Bose, ibid., 1909, 10, 230 ; A , , 1909, ii, 383MINERALOGICAL CHEMISTRY. 265Magneto-optical Researches.-The investigation of the opticalbehaviour of anisotropic liquids under the influence of an electro-magnetic field has attracted the attention of several workers, andthe results are regarded by some, including Nernst,58 as affordinga decisive proof of the correctness of the swarm theory.It was early observed by Lehniann that a magnetic field has aclearing effect on droplets of p-azoxyanisole. Bose 59 investigatedthe effect on relatively large masses of liquid; with anisaldazine thOeffect begins with the application of 600 Gauss units of force, and,by means of a few thousand units, layers up to 4 mm. are imme-diately cleared when the direction of vision is along the lines offorce. On turning off the current, the mass immediately becomesturbid. Similar results were obtained with p-azoxy-anisole and-phenetole.The work was carried a step forward by Mauguin,60 who showedthat the very mobile p-azoxyphenetole may be obtained in clearmasses provided that both slide and cover slip are thoroughlycleansed with hotl sulphuric acid, distilled water, and ether; sucha preparation gives a normal uniaxial figure. The same substancewas then investigated under the action of a magnetic field, and thefollowing three fundamentally important observations were made :(1) layers up t o 2 mm. may be rendered quite clear and transparent,no matter whether viewed along or across the lines of force;(2) when exaniined between crossed nicols, the behaviour is in everyrespect analogous t o that of a uniaxial crystal-uniform darknesswhen viewed along the lines of force in parallel light and a perfect,uniaxial figure in convergent light'; normal extinction in fourpositions when viewed across the lines of force. The double refrac-tion is strong, and increases up t o the application of 5000 Gaussunits, and then remains constant up t o 7000; (3) when a homo-geneous film between slide and cover slip, as in Fig. 9c, is subjectedt o the action of a transverse magnetic field, the optic axis isgradually deflected away in the plane containing the lines of force.811 releasing the force, the optic axis iminediately returns t o thenormal position. Similar results, although not quite so complete,were independently obt'ained by Wartenberg 61 f o r both pazoxy-anisole and -phenetole; the double refraction of the latter wasfound t o be by far the stronger.T. V. BARKER.58 I!'. Nern>t, Zeitsch. Elektrochem., 1910, 16, 702.5g F,. Ijose, P?iysik-aZ. Zeeitsch., 1911, 12, 60 ; A . , 1911, ii, 184.C. Maugnin, Compt. rend., 1911, 152, 1680.H. v. Warteaberg, PhysiknZ. Zciitsch., 1911, 12 837, 1230 ; A . , 1911, ii, 952 ;See also ref. 51.1912, ii, 112