General a n d P h y s i c a l Chemistry, Refractometric Observations. By JAX F. EIJKMAN (Eec. TTav, Chim., 1895, 14, 185-202; compare Abstr., 1894, ii, 173j.-The author has determined the refractive indices for the hydrogen lines a and p and A of a large number of organic substances, each being examined a t two temperatures differing by SOo to 125". A t high temperatures, Gladstone and Dale's molecular refraction formula gives too low values, whilst Lorenz's expression gives too high values; it should therefore be possible to obtain a moleciilar refraction formula which should give concordant values a t all tempera- tures. By calculating from the experimental numbers now obtained, the author finds that the results are best represented by the expres- sion (n2 - l)M/(n + 0*4), in which n is the refractive index, and M the molecular volume, and he therefore proposes to take this expres- sion as the molecular refraction.If the new formula is employed, the agreement between the observed and calculated inolecular refrac- tions becomes very close, and tlie dispersion equivalent for the increment oE an homologous series, CHP, acquires a constant value. W. J. I?. Molecular Origin of the Absorption B_ztnds of Salts of Cobalt and Chromium. By ALEXANDIIF, L. ETARD (C'ompt. rend., 1895, 120,1057-1060) .-Violet solutions of chromium sdphate and nitrate, and of chrome alum, exbibit a fine absorption band in the red, X = 678-6iO. Addition of a nitrite changes the colour of these solutions to lilac, and an arsenate turns them green, but the above characteristic band remains in the absorption spectrum, and is only shifted a little towards the red, X = 687-680.Chromic acid in concentrated solution also shows this band. On the other hand, anhydrous chromjl chloride, potassium chromate and dichromate,. and roseochromic sulpliate give no distinct band. The blue chromium potassium oxalate gives absorption bands in the red a t X = 700-693 and X = 732-729. The red solution of cobalt sulphate giws a band, h = 654-650, and the red solution of cobalt chloride a band, X = 667-642. If these solutions are turned blue by heating and adding a little con- centrated hydrochloric acid, two additional bands appear in the red, but the bands of the red solutions still remain visible. The author concludes that the absorption bands of the chromium and cobalt compounds are not due to the atoms of the metals, but to the internal arrangement of the molecules.As these absorption spectra resemble in character those of the mre earth and of uranium compounds, the hypothesis that each band in the spectrum of a rare earth corresponds with some element, is not iieceesarily true. Anomalous Rotatory Dispersion of M'alic acid. By RAFFAELB NASIKL and G. GENNARI (Gazzetta, 1895, 25, i, 41 7-438) .-Employing a Landolt-Lippich polarirneter, fitted with the ray-filters recently H. C. VOL. LXX. ii. 11134 ABSTRACTS OF CHEMICAL PAPERS. described by Landolt (Abstr., 1895, ii, 1): the authors haye examined the rotatory dispersion of malic acid dissolved in various solvents, under different conditions of concentration and temperature ; the mean wave-lengths, ,up, in millionths of a millimetre of the various rags employed, are 665.9, 591.9, 553.0, 448.5, and 448.2.I n a 4.6 per cent. aqueous solution a t 20": the specific rotation for all these rays is a h v o one; for red light of p,u =665*9, the value [a] = - 1-87', whilst for p,u = 448.2 [p] = -2.51'; as the concentration increases, the temperature remaining the same, the solutions become more and more dextrorotatory, until i n a 72.8 per cent. aqueous solution, the values of [a] for the above wave-lengths become + 1.80" and + 6.39' respectively. At intermediate concentrations, the solutions become inactive for one or other wave-length of light, although the particular solution for which [a] = 0 for light of one wave-length is strongly active towards a different coloured light.As the temperature of the solution rises, the value of [a] becomes more negative; thus a t 7" a 33.24 per cent. aqueous solution has a specitic rotation of [a] = 3-0.44" for the ray ,up = 665.9, and- of +2*63O for the ray pp = 448.2, whilst a t 41.5' these rotations become -5.96' and -5.84' respectively. The addition of boric acid to the solution acts in the same way as n rise in temperature. Malic acid has nearly the same specific rota- tion in both methylic and ethylic alcohols, and the variation in rotation with varying concentration is of much the same kind as when water is the solvent; the solutions are, however, much more laevorotatory than aqueous solutions. The ,satno seems to hold for propylic alcohol and acetone, although the laevorotation is not so high as when the solvent is methylic or ethylic alcohol.A number of solutions of sodium malate were examined, showing t h a t the specific rotations f o r light of various wave-lengths change in much the same way as with the acid itself. The dispersion coefficients of aqueous solutions of malic acid change very irregularly with the concentration, wbilst those for sodium malate vary much less with the concentration. After a full discussion of these anomalies, the authors are unable to furnish any explanation of them ; they can hardly be due to changes in the degree of ionisation, malic acid being so slightly dissociated in aqueous solution. There is also no evidence indicating the existence of hydrates or polymerides in solution, as cryoscopic determinations show that, in concentrated aqueous solutions malic acid has the normal molecular weight.New Examples of the Superposing of the Optical Effects of Two Asymmetric Carbon Atoms. By PHILIPPE A. GUYE and C. UOUDET (Compt. reic.d., 1895, 121, 827-829).-The amylic amyl- acetate of the formula CHMeEt*CH,-CH,*COO*CH,*CHMeEt contains two asymmetric carbon atoms, each of which should behave towards polarised light as though the rest of the molecule were inactive (Abstr., 1895, ii, 149). The authors have prepared compounds of the above formula from ( a ) a mixture of dextrorotatory amylacetic acid and racemic amylic alcohol, ( b ) a mixture of racemic amylacetic W. J. P. * i kGENERAL AND PHYSICAL CHEMISTRY.135 acid and lsvorotatory amylic alcohol, and (c) a mixture of the dextrorotatory acid and the lt-evorotatory alcohol. The first com- pound has the rotation [ a ] ~ = +4.36, and the second [a]D = +2*54. Theoretically, therefore, that of the third should be +4*36 +1*54 = +5*90, and the number actually found was [a]= = +5*64, in close agreement with the theory. The amylic amylmalonate, CHMeE t*CH,*CH(COO*CH2*CHSleEt),, contains three asymmetric carbon atoms. Compounds of this formula were prepared from ( a ) dextrorotatory amj-lmalonic acid and racemic amylic alcohol, ( b ) racemic amylmalonic acid and hvorotatot-y amylic alcohol, and ( c ) dextrorotatory amylmalonic acid and ltzvorotatory amylic alcohol. The first compound has the rotation [a]D = +6-10, and the second [a]= = t3.48.The rotation of the third should therefore be +9*58, and the number actually found was [ a ] , = + 9-68,. the hheory being thus again confirmed. H. C. By PAUL WALDEN (Zeit. physikal. Chem., 1895, 17, 245-266).-The product of asymmetry, P, of a substance containing four groups of molecnlay weights gl, g,, gs, and g4 attached to one asymmetric carbon atom, is obtained from the equation ., ., Optically Active Derivatives of Succinic acid. omitting the constant term ( 1 sin a)G, the product of asymmetry is supposed by Guye and his supporters (Abstr., 1893, ii, 561) to be a measure of the rotatory powers of optically active organic substances. The author has prepared a large number of derivatives of malic and succinic acids, and shows that in the several series of cornpounds thus obtained, the product of asymmetry does not, in most cases, even indicate correctly whether the specific rotation of a substance will increase or decrease when the mass g of one of the four groups is altered by substitution.Thus the specific rotations [aJD, and the molecular rotations [MI,, of many of the substances mentioned in the accompanying table should be negative in sign, judging by the products of asymmetry, whilst others which should be either Izevo- rotatory or inactive are highly dextrorotatory. An inspection of the table shows that the ethereal salts of malic acid are dextrorotatory, the specific rotation rising as the mass of the nlkyl group increases, until the maximum is attained a t about the propylic salt ; the amides are more lzevorotatory than the salts, and their specific rotation increases with the mass of the substituted amidogen.The alkylic salts of the substituted malic acids are 13evo- rotatory, and have about double the specific rotations of the parent alkylic malates ; the derivatives of chlorosuccinic and bromosuccinic acids prepared from ltwo-derivatives of mdic acid are all highly tlextrorotatory, and the specific rotations of the bromo-compounds are higher than those of the corresponding chloro-derivatives. Iso- merides, or two compounds which contain groups of approximately the same mass, although of different kind, hare not the same rota- tory powers. Two substances of the same molecular weight., differ- 11-2136 ABSTRAOTS OF CHEMICAL PAPERS. Substance.Dimethylic malate.. . . . , Dietliylic ,) . . . . . a Dipropylic ,, . . . . . . Diisopropylic malate.. . , Diisobutyric ,, . . .. Dicaprylic ,, . . . . Dimethylic acetylnialate ,, propionjlma lot0 ,, butjrojlma- late ,, isobutyrojl- malate ,, isoraleropl- malate ,, cliloracetjl- malate ,, broinacetyl- late Diam ylic ,, . . . a Diethylic acetjlmalato . . ,) propionyliiia- late , , hu t j roy lm alatc ,) isobutyrojlms. late ,, isowJc~ojliiia- late ,, bromacetylma- late ,, bromopropio- ii y 1 ma late ,, broruobutjroyl malate . ,, bromisobutj- royliiialate ,, ethoxgsuccin- ate ,, ch1or:tcetj 1- Dipropylic acetylmalate . nisla t e late late malate ,, butj1.o~lllla- ,, iaocalerojlma ,) broniacetgl- Diisobutjric sect) linalatt ,, butyrojlma ,, isoraleroyl- ,, bromacetgl- late maln t e mlilnt e - 6-85' - 10 '18 -.11 '62 - 10 '41 -1L.14 - 9.92 - 6-92 - 22 -92 - 22 *94 - 22 -44 -22 '36 - 22 *39 - 23 '80 - 22 -40 - 22 *52 - 22 *20 - 22 -22 - 21 -99 - 22 *07 - 22 -48 - 22 '48 - 24 -76 - 22 '57 - 1.44 - 22 *85 - 23 -52 - 22 -40 - 21 -68 - 22 '24 - 2 1 -88 - 21 -68 -19 -91 - 20 *38 1.2317 - 19.35 - 25 '32 -22.69 - 27 -39 - 27 '19 - 24.77 -46'76 - - 52 -07 - 51 '86 - 55 -07 - 55 -56 - 63 -38 - 52 '25 - 54 -6% - 57 9s - 57 -17 - 60 '46 - 69.92 - 73 -05 - 83 *93 - 76 '50 - 3-13 - 59 '40 - 69 -26 - 64 '50 - 65 -47 - 76 -41 - 63 '01 - 68 '52 - 65 -70 - 74 'SO 1 -1294 1 *0745 1 -076 1 -0418 1.079 0.9761 1.1975 1'1317 1 *12% 1 -1034 1 *3862 1 -5072 1 *1168 1 '0938 1 * 0736 1 *(I688 1 *0605 1.3936 - 1 '3325 1 -3059 1 '2850 1 .I015 1 *0724 1.1566 1 -0417 1 a0263 1.3150 1 * 0362 1 '0146 1 -0045 1 -2022 - n.-- 1 '4&25 1 *4362 1 *4380 1 -4392 1 '4438 1 -4500 1.4318 - - 1 -4342 1 *4310 1 *4350 1.4530 1 '4680 1 *4295 1 -4305 1 *4315 1 -4285 1 -4338 1 '4610 1-43.61 1 *A568 1'4520 1.4320 1.4315 1 -4465 1 '4348 1 -4352 1,4608 1'4330 1.4352 1 -4353 1 -4520 -__ Observ -- 34 '78 44 *oo 53-26 62 -14 71-48 98 -57 44 -17 53 '41 53 -35 58 - 18 49 -35 52 -20 53.61 58 9 9 62 5'5 62.65 67 *26 61 '24 66 -31 '70 -68 - - 71 -17 54 '10 62 -82 67 *9R 72 '11 76 -81 $0 -71 72 '32 81 -30 85 -78 82 *36 R. Calc. -- 34 -98 44 -18 53 -40 62 -60 71 -80 99 -4.2 44 -52 - - 53 -73 53'73 58 *34 49-48 52 -41 53 '73 58 -33 62 -94 62 *94 67 *55 61 *61 66 '22 $0 -82 $0 -82 53 *55 62 '94 67 -83 72 -1.5 76 -75 70 -82 72 -15 81 '35 85 9 6 80 -03GENERAL AND PHYSICAL CHEMISTRY.137 Substance. Chlorosuccinic chloride. . Dimethylic chlorosuccin- Diethy1 ic chlorosuccinate Dipropylic ,, Diisobutyric chlorosuc- Diamylic chlorosuccinate Dimethylic brornosuccin- Diethylic bromosuccinate Dipropylic bromosuccin- Diisobutyric broinosuc- ate cinate ate ate cinate C a b + 20 -53 + 41 '42 + 2'7 '50 +25-63 + 21 '57 + 21 '56 + 51 '18 + 4.0 *96 +38-05 + 23'66 W I D . --- + 55 -93 + 74 -76 + 57 -33 + 60 -61 + 5'7 *05 + 63 '07 t 114 -37 t 103 *63 t 106 *9 + 72 -80 d. 1 '5003 1 '2555 1.1493 1 '0925 1 '03.24 1 *0319 1 .i5050 1.3550 1 -3010 1,2394 n. -- 1 *4840 1 -4436 1 -4372 1 a 4 4 1 2 1 '44.03 1 a4436 1 -4618 I *4550 1 -4592 1 * 4580 R. 3bserv. 36.12 33 -16 47 -55 57 -19 66 *20 75 -24 41 *32 50 -66 59-07 68 -03 CdC.-- 35 -73 38.40 4'7 '61 56 '82 66.03 73 -23 41 -33 50.24 59 *74 68 -94 - ently distributed in the two cases between the four molecular groups, have sometimes nearly the same, and sometimes very different rotatory powers; in some cases, a considerable a,lteration in the mass of a gvoup causes only a very small change in the rotatory power. In addition to the specific and molecular rotations of the various substances determined at 20°, the table gives the densities d, at 20' referGed to water at O', and the refractive indiccs n, for the D line at 20°, together with the observed and calculated molecular rotations R, obtained from Gladstone's formula. JV, J. P. Optically Active Derivatives of Rhenylacetic acid : Optical Superposition. By PAUL WALDEN (Zeit.y h y s i k d Chein., 1895, 17, 705-7'24 ; compare preceding Abstr.) .-Starting with mandelic acid, in which the four groups attached to the asymmetric carbon atom have the masses 77, 45, 17, and 1, the author has prepared and ex- amined the rotatory powers of a number of derivatives in mliich the masses of the above four groups vary; from the data thus obtained, the author shows that, as in the case of the malic acid derivatives, the product of asymmetry affords no criterion of the rotatory power, and deduces conclusions similar to those stated in the preceding ab- stract. The carboxyl group in mandelic acid is of practically the same mass as the group CONHZ in mandelamide, so that these two substances should have almost the same rotatory power. Table I, however, shows [a]a to be very different in the two cases.Similarly, tartaric dianiicie has a specific rotation in water or saturated boric acid solution of [ a ] ~ = 4-108' to 109.4", whilst for tartaric acid The very different specific rotations possessed by ethylic mandelate in acetone and carbon bisulphide solations seem not t o be due to a difference in molecular weight, as this snbstance depresses the boiling point of the two solvents normally. [ a ] D = +14*93.185 ABSTRACTS OF CHEMICAL PAPERS. Substance. ---- Mandelic acid . . . . , . . Mandelamide . . . . . . . . Methylic mandelate . . Ethylic ,, .. Isobutyric ,, .. Methylic acetylmande, Amylic Y Y ' * late mandela te mandela te ,, propionyl- Ethylic propionyl- ,, raleroylman- delute Acetylmandelic acid..Phenylchlometic anid ,, chlor. ide Methylic phenylchlor- Ethylic phenylchlor- PropyIic phenylchlor- Amylic phenylchlor- Phenylbromacet,ic acid Methylic phenylbrom- Ethylic phenylbrom- Isobutyric phenyl- Phenylbromacetic bro- acetate acetate acetate acetate acetate acetate broniacetat e mide TABLE I. Calu. - 153 *lo in H,O. . . - 66.7 ,, COMe, -110.2 ,, COMe, - 123 * 12 liquid . . . -100'73 ,, . .. - 96'46 ,, . .. -146.37 ,, . .. -135'5 ,, ... -113.7 ,, ... - 97.06 ,, . .. - 156 *4 in COMez. + 158 * 3 ,, CS, . . . . + 107 -55 liquid . . . + 25-19 ,, . .. + 23-94 ,, ... + 23.41 ,, . ,. + 45 -4 in CGH,. . . + 28 -82 liquid . . . + 16-56 ,, . .. +131*8 ,, CsHs a . + 9.77 ,, * I . -t 44.53 ,( . * . - 148 *Oo in COMe, -214.1 in CS,. - 88.8 ,, COMe, - -145 '0 ,, CS,.- - 110 in CHCI, . . -117 ), csp + 131 -3 in CS,.. , . I + 26 '39 in CSn.. . Ca!D. -180.0" in CS?. -129 in CS?. t 107.9 in CHCI,. In order to test the principle of " optical superposition," which states that the seve~al optically act.ive groups in a given molecule act additively and in such a way that the specific rotation of a sabstance becomes the algebraic sum of the specific rotations of two of its stereoisomerides, the authoy has examined a number of suitable salts and obtained results which are summarised in Table 11. The density d, the molecular rotation [MI,, and the specific rotation [a]D, of each series of three salts was determined ; the last column contains the specific rotation of the third salt of each series calculated as the sum of the specific rotations of the first and second.The specific rota- tions of lavo-amylic laevo-lactate is thus the sum of the specific rotations of inactive amylic leevo-lactate and lavo-amylic inactive lactate ; the agreemeut, as will be seen, is very close.GENERAL AND PHYSICAL OHEMIISTRY. TABLE 11. i-amylic Z-lactate.. ................ 1- ,, i- ................... I- .. 1- .................... i-amylic 1-mandelate .............. 1- 3 1 i- ,7 1- $ 9 1- ,, .............. .............. f i-amylic d-phen~lchlorscetate ...... ,, 7 5 ...... I:: 1; 2 ...... 1 1 - .. i- .................. i-diamylic Z-malate ................ Lr- .. 1- .................. i-diamylic d-chlorosucciiiate ........ 1- ), i- ........ I. ), d - ........ 7 9 , i-diamylic d-tartrate ..............I - ,, racemate I. .. d-tartrate .............. ............... -( Cl. 0 -9719 0 *9672 0 -9667 1 -0531 1 -0520 1 -0530 1 -0828 1 -0832 1 -0826 1 .om0 1 -0180 1 -0176 1 '0319 1 *0314 1 -0303 1 '0637 1 -0640 1 -0636 [ Af] D. - 10 -21C + 4-22 - 6.29 - 214 '14 + 6-12 - 208 '72 + 56'03 + 7.79 + 64.42 - 27-19 - 18-85 + 67.03 + 10.98 + 73.53 -- - - - 139 Observ. -- - 6 ' 3 8 O + 2-64 - 3.93 - 26 '46 + 2-76 - 94 -02 + 23 -31 + 3'23 + 26 -79 - 9'92 + 3'50 - 6.88 + 2 1 -56 + 3.75 + 25 *15 + 14 -10 i- 3.37 + 17,73 W. J. P. Optically Active Halogen Compounds. By PAUL WALDEN (Ber., 1895, 28, 2766--2773).-See this vol., i, 139. The Birotation of Glucose. By HEINRICH TREY (Zeit. yhysikal. Chenz., 1895, 18, 193-218) .-The birotation phenomena of glucose were investigated in aqueous and other solutions, both the anhydride and the hydrate being employed. I n solutions in methylic and ethylic alcohol, birotation occurs as in water, but more slowly, whilst also the final valh3 is hig!ier than in aqueous solution, both initial and final values being higher in ethylic than in methylic alcohol.By the addition of water to the alcoholic solution, the retrogression was accelerated and the final value also reduced to an extent correspond- ing with the quantity of water added. Chemically indifferent com- pounds cause a retardation in the methylic alcohol solution of the anhydride, and a slight increase of the end value. I n aqueous solu- t'ions of both anhydride and hydrate, acids cause an acceleration, the effects in t h k respect being in the same order as the aflinity con- stants.By the addition of hydrogen chloride, even in small quanti- ties, to the methplic alcohol solution, the rotation was reduced to zero, this being probably due to decomposition. By the solution in water of the amorphoiis residue left on evaporating an alcoholic solntion, the end value of the rotation was at once obtained, and the author considers that his experiments indicate that the explanation of the birotation is t.0 be sought for, not in the hydration of the compound,140 ABSTRACT3 OF CHEMICAL PAPERS. but in a change from a crystalline to an amorphous variety, or in some such alteration of the molecular configuration. (Compare also Levy, Abstr., 1895, ii, 586.) L. M. J. Theory of the Decomposition of Racemic Compounds.By CHR. WIKTHEIC (Bey., 1895,28,300O-3U2~).-T he aut’lior enunciates a general theory respecting the decomposition of racemic compouiids into their optically active constituents either by means of active bases or by crystallisation. The theory only holds good for com- pounds in the solid state or in saturated or supersatumted solutions, The atoms or groups attached to the asymmetric carbon atom of one molecule are supposed to have certain affinities for the corresponding atoms or groups in a second molecule, These affinities the author terms “ secondary,” and supposes they are of two kinds. For example, in the case of a compound containing an asymmetric carbon atom to which hydrogen and hydroxyl me attached, the affinity between the H and H, or between OH and OH, is termed “ racemic ” affinity, and that between H and OH “ contrary ” affinity.If, under given cir- cumstances, the racemic affinities of a compound are greater than the contrary affinities, it will be found impossible to split up the compound into its active constituents by the above means ; in order to bring about such a decomposition, energy, either thermal or chemical, must be sup- plied to the system. Under a certain set of conditions there will always be an equilibrium between the two secondary affinities. It is shown &hat the theory agrees with the facts hitherto known regarding the decomposition of racemic compounds, and the paper concludes with an index to the literature of the subject. J. J. S. Flames and Illuminating Gases. By C. BOHN (Zeit.physiktil. Chem., 1895, 18, 219--239).--The form of Bunsen burner devised by T e c h (Abstr., 1892, 768) was used for the experiments. The ap- pearance of the flame is first described, five parts being recognised- (1) the inner cone surrounded by (2) the mantle, around which lie (3) the outer cone with (4) its border, above which is (5) the cap. The variations in the several parts according to the air supply is recorded, and then the spectroscopic examination. The cap gave n feeble continuous spectrum with 110 red and but little blue; the border also gave a continuous spectrum with the red feebly cle- veloped ; that of the outer cone was also continuous, dark and bright lines being absent, and the red being well deyeloped. The mantle, however, gave a band spectrum with well-marked green and indigo or violet bands, and under some circumstances a blue stripe also, and a dark band close to the D lines.The inner cone appeared to give a feeble band spectrum, most probably, however, due to the mantle. The author considers the mantle t o be the place of explosive combus- tion and of the greatest development of energy, although not neces- sarily the hottest part of the flame. Sulphur, hydrogen, carbon bisulphide (by a wick), and carbonic oxide mere also burnt, and for all these flames, the spectra were continuous. The measurements of the band spectra .are recorded, the results being compared with those obtained by Swan. The spectroscopic measurements are alsoGENERAL AND PHYSICAL CaEBlISTR T. 141 recorded in the case of Geissler tubes containing various carbon com- pounds, and the author concludes that the discontinuous spectra of carboniferous gases are not identical, the cliffereuces being greater than those occasioned by alterations of temperature and density.It is hence not possible to define a carbon band spectrum, and even sharp- ness of the less refrangible cdge of the bands, and gradual fading of the more refrangible edge, does nct exist with all carbon compounds. Cause of Luminosity in the Flames of Hydrocarbon Gases. By VIVIAN B. LEWES (Proc. Roy. Xoc., 1895, 57, 4,50--468).-Accord- ing to the " solid part,icle '' theory of luminosity, i t is to be expected that the luminosity of different flames of the same size and burning from the same kind of jet, would be governed (1) by the tempera- ture of the flame; (2) by the number of carbon particles in a given area.I n order to determine the temperatures of different flames, the author has made use of a very small and thin Le-Chatelier thermo- couple. Preliminary experiments showed that the diameter of the wire seriously affected the temperature recorded, and the author concludes $hat the temperatures indicated by the finest wires which can be used without fusing are probably 100--200' too low. The following results were obtained. L. M. J. Acetylene. Ethylene. Coal gas. Non-luminous zone.. . . . . 459' 952" 1623' Commencement of lumin- osity . . .. .. .. .. . . . . .. 1411 1340 1658 Near top of lnniinous zone 1517 1865 2116 As regards luminosity, however, the three gases stand in exactly the reverse order; and as there appears to be no apparent relation between the temperature of the flame, or the probable number or carbon particles contained in it, and its illuminating value, it is sug- gested that the luminosity must be in great part governed by some thermochemical changes taking phce in the flame and yet not, appre- ciably affecting the average temperature.It is thought that as acetylene is formed when hydrocarbons are burnt, and as i t is an endothermic substance, the heat liberated during its decomposition endows the carbon particles with a high incandescence. In sup- port of this view, the author shows that acetylene, when decom- posed by a detonator or merely by strongly heating it in a glass tube, develops light. It is also shown that acetylene, when largely diluted with hydrogen, carbonic oxide, carbonic anhydride or nitrogen, burns with a non-luminous flame (compare P.E'ranklaiid, Trans., 1884, 30 and 227). It has been found possible to make such mixtures burn with luminous flames by externally increasing the tem- perature. The luminosity of a flame, therefore, depends not so much on the percentage of acetylene in the gas, but rather as to whether there are many points at which the temperature is su6cientlg high to bring about decomposition of the acetylene. The flame of alcohol contains as much acetylene as a good coal gas flame, and yet is non-, or only142 ABSTRACTS OF CHEMICAL PAPERS. slightly, luminous, because the temperature is too low to decompose the acetylene. When burnt in oxygen, the flame becomes brightly luminous, owing to the increase in temperature.Cyanogen, which is even more endothermic than acetylene, burns with a non-luminous flame. This is due to the fact that cyanogen requires a much higher temperature before it is decomposed, and i t is shown that when suffi- ciently strongly heated, it can be made to burn with a luminous flame. J. J. S. Red and Yellow Mercuric Oxides. By WILHELM OSWALD and THOR MARK ( Z e d . physikal. Chem., 1895,18, 159-160) .--The experi- ments of Varet proved the identity of the heats of formation of the two mercuric oxides, hence the total energies are equal for these com- pounds (Abstr., 1895, ii, 305). A galvanic cell of mercury-red oxide -caustic potash-yellow oxide-mercury, was found by the author to give no E.M.F., whilst, also, no change of the E.M.F.of various cells occurred when red and yellow oxides replaced one another. Hence, the author points out, the free energyof the two forms are alsoequal, and the compounds are not isomeric but identical. L. M. J. Peroxide Electrodes. By O L ~ N FREEMAA’ TOWER (Zeit. physikal. Chem., 1895,18,17-50) .-Examination of varions peroxide electrodes showed those of lead, silver, and thallium t o be unavailable, as with the first a long time is necessary before a constant E.M.F. is obtained, whilst the two peroxides last mentioned are unstable and readily acted on by dilute acid solutions. Manganese peroxide, being free from the above disadvantages, was therefore selected for the experiments. The E.M.F. with this electrode is given by the formula n ~ ~ 7 r = Rt log (C,C$/Cm~CoQ), where C, + Co are the concentrations of the manganese and hydroxyl ions respectively, the other letters having the usual significance.k/C, may be substituted for C,,, and the formula reduces to In acid solutions, C, x CH being constant. 7 = -0.0286 log (CmCH4/CmCH4). Experiments were made with solutions of varying nitric acid and manganoua nitrate conient ; the observed and calculated results show satisfactory agreement, the differences being probably due to incom- plete dissociation. As t,heoretically indicated, the effect of the salt concentration was found to be four times t h a t of the acid. Experi- ments were also made with phosphoric, malonic, tartaric, formic, acetic, salicylic, orthamidobenzoic, mono-, di-, and tri-chloracetic acids, the concordance between the observed and calculated numbers being again, for the most part, satisfactory. The influence of temperature was next determined.This, in the case of nitric and sulphuric acids, is very slight., rather more for phosphoric and the chlorrtcetic acids, and very marked with acetic, formic, and tartaric acids. No reason for these differences is, however, indicated. The determination of the E.M.F. is also employed for the determination of the degree of dissociation of the sodium salts of 16 organic acids at varying concentrations, and the results compared with those obtained by Trevor (Abstr., 1893,GENERAL AND PHYSICAL CHEMISTRY. 143 ii, 62). The only cases in which satisfactorp agreement is not obtained are those of the phthalates and isophthalates, where the author finds the former salt to be the more highly dissociated.Barium salts were also employed in a few cases, and the dissociation is determined for acid snlphates of potassium, magnesium, sodium, aluminium, and copper, hydrogen sodium selenate, and dihydrogen sodium phosphate. I n a1 kaline solutions, the electrolyte is a solution of Mn(OH),, hence C,,, >: C,Z = k, and C,C$ = 3i/C,,L, hence Z- = -0°.0286 log (Co2jC2). The availability of the formula is shown by experiments with potas- sium, sodium, barium, strontium, calcium, and ammonium hydr- oxides. By the use of a perfectly neutral solution, the dissociation o€ pure water is obtained as 4.4 x lo+, a number which, although of the same degree of magnitude, is smaller than that which has been obtained by other methods.IJ. M. J. Potential Differences between Metals and Electrolytes. B y GEORG MEYER (Ann. Phys. Chem., 1895, [2], 56, 680-699).-Roth- mund (Abstr., 1823, ii, 35) has calculated the E.M.F. of various cells from his electro-capillary measurements of the potential differences between the metals and electrolytes composing them, and in a number of cases the values thus obtained do not agree with those directly measured. The author has repeated, and, in general, confirmed Rothmund's measurements. He therefoi-e concludes that the E.M.F. of a cell M, I F, I F2 I M2, composed of the metals M, and M, and the electrolytes F, and Fa, is not given by the difference between the forces necessary to produce the maximum surface tension of polarisa- tion of M, in F1 aud of M, in Fz.The potential difference between the two metals and two electrolytes must be taken into account', and to the difference of E.M.F. above spoken of must be added the potential difference between two dropping electrodes, which, contain- ing MI, and Mz, are respectively placed in the electrolytes F, and F,. It is only when this last term approaches zero i n value that it can be neglected . H. c. Electro-capillary Properties of Dilute Sulphuric acid. By A. GOUY (Compt. Tend., 1895, 121, 765-768)-'l'he author has made a number of determinations of the relationship between the height h of the column of mercury balanced by electro-capillary forces, and the difference of potential V between the mercury and the electrolyte, using solutions of sulphuric acid of various concentrations as the electrolyte.The ralues of h decrease vith the coucentration of the solution, but, in the more concentrated solutions, exact measurements become impossible, owing to the occurrence of electrolpsis. The second derived function d'hldV? is always negative, so that the curve of h has no point of inflexion, and does not tend to any limiting value. 'l'he actual value of the function is not constant, but is subject to complex variation. H. C. The Passage of Electricity through Gases. By OTHO LEHMANN (Zeit. yhysikal. Chew ., 1895,18, 97-11 7).-Experiments are described on the electric discharge through various gases, vacuum tubes,144 ABSTRACTS OF Cl3JZMICAL PAPXRS.mixed gases with different forms of elcctrodes, and the discharge iii a strong magnetic, field. The paper is illustrated by a number of figures of discharge phenomena, aiid the author considers the views of Goldstein aiid Hertz, that the discharge takes place iirto the ether and not into the gas, to be erroneous. Relation between the Di-electrical Constants of Gases and their Chemical Valency. By ROBERT LANG (Ann. Phys. Chem., 1895, [2], 56, 534-545).-Between the dielectrjcal constant, I<, of a gas, that of the ether being unity, and the sum of the valencies of the atoms in the molecule of the gas, s, the author finds that there is the following general relationship. I < - 1 ~ loG = 123. L. 31. J. S This relationship does not always hold if. the simple chemical molecule of the gas is taken, but, in such cases, the assumption is made that a number of theso molecules are combined to form a com- plex group, when agreement is obtained.The values of I< are taken tor 0" and 760 mrn. pressure. The term dielectricaE valency constcmt is proposed f a - the constant, the value given abore for this being provisional only. H. C. Determination of High Resistances. By MARGARET E. &fA4LT131' (Zeit. physikal. Chem., 1895,18, 133--158).-A method for the deter- mination of high electrolytic resistances is described, previous methods being considered unsatisfactory. A Wheatstone bridge is employed with four carefully calibrated adjustable electrolytic resistances, and when a balance is obtained, the unknown resistance is placed in one arm (1) in series (2) parallel, and that resistance again adjusted for a balance.F u l l details of bhe construction, calibration, and working of the instruments are given in the paper. Comparison of numbers obtained with those due to Kohlrausch indicates the availability of the method. Experiments are recorded with a solution of hydrogen chloride in ether a t its critical temperature, the resistance increasing from 641 divisions a t 20' to 24,180 a t .197", a t which temperature critical phenomena were observed. Solutions of trichloracetic acid in ether, and of potassium chloride in water, were also examined a t high temperatures. The former gave a continuous decrease until 75O, when the resistance commenced to increase ; with the latter salt, the resistance decreased regularly to 297".The method is stated t o be available for the determination of the conductivity of pure water, but no measurements are recorded. L. M. J. Determination of some Conductivities. By A. E. BACR (Zeit. physikal. Chem., 1895, 18, 183--184).-The Conductivity of tetrazole compounds was first examined with the following results : Tetrazole, CN4H2, , L L ~ ~ ~ ~ = 37.96 ; amidotetrazotic acid, CZHGNIO, plOgl = 11.61 ; sodium azotetrazole, C2N,,Na2,5H,O, = 103.6 ; sodium amidotetr- azotate, CH,N,Na,SH,O, plmi = 93.27. The conductivities of the chlor- ates of caesiuni, rubidium, and potassium were fouild t c be 137.5,134*9, and 129.9 respectively (ploz4), and hence, taking the ion velocity ofGENERAL ASD PHYSICAL CHEMISTRY.145 potassium as 70.6 and pm = + 3, the following ion velocities are obtained: Rb, 75.6 ; Cs, 78.2 ; Clod, 62.3. The ralues of the Yeloci- ties of the inetdlic ions were also determined by means of the chlorides aich the results Rb, 76.5; CS, 79.3. L. M. J. The Dilution Law of Salts. By JACOBUS H. VAN'T HOFE' (Zeit. p7~ysikaZ. Ohem., 1895,18, 300-304) .-Rndolphi found (Abstr., 1895, ii, 490) that the relation between the molecular conductivity and the instead of by Ostwvald's expression X: = The author v ( l - X7./A.-)* shows that in Rudolphi's experiments a slightly -be;& constancy of (X,i'h,)4 d v ( l - h4Am)' k is obtained by the. expression ---e--- -__- which leads imme- diately to t h e equation ci3/cs2 = constant where ci and c.* are the concentrations of ions and non-dissociated substance respectively, although, as the author points out, the physical interpretation is not very clear.L. 31. J. Specific Heat of Mercury between 0" and 30". By ADOLFO BARTOLI and ENRICO STRACCIATI (Gazzettn, 1895, 25, i, 380-388). -After a summary of the work previously done, the authors give an account of their determination of the specific heat of mercury between 0" and 30" made by cooling a mass of platinum, previously heated in a steam bath, in Titer, and in pure mercury, successively; using the values for the specific heat of water pre- yiously obtained b.y them (Abstr., 1895, ii, 5 ) , the specific heat of mercury between 0" and 30' is given by the equation in \vhich T is the temperature for which the specific heat C is ibequired. The numbers agree well with those obtained b,y Naccari and by Winkelmann. Specific Heats of Platinum, Silver, Tin, Lead and Copper.By Auor,~o RAR'l.OL1 and EXRICO STRACCL4TI (Gazzetta, 1895, 25, i, 359-393 ; compare preceding abstract) .-The authors have obtained. the following values for the specific heats of various metals between 15" and 100" as the means of a large number of measurements ; the values are referred t o water a t 15", and the impurities are given in percentages. C = 0.033583 + 0~00000117 T - 0-0000003T2, W. J. I?. PIatinuin (containing traces of Ir) ............ 0.63223s Silver ( ,, 0.047 Cu and 0.016 Au) . . 0.056250 Tin ( ,, 0.030 Fe and 0.008 Pb) . . 0.055550 Lead ( ,, 0.030 T1 and traces of other metals) .............. 0.030887 Copper ( ,, 0.12 Sn and 0.12 Au) ....0.093392 9 1 ( ,, 0.005 Sn and traces of other inetalsj ........ 0.093045 W. J. P.'146 ABSTRACTS OF CHEMICAL PAPERS. Latent Heats of Vaporisation of Ketones and other Carbon Compounds. By WLADIMIR F. LOCGDININE (Compt. rend., 1895, 121, 556-558).-1n the following table, column I contains the boiling point under a pressure of 760 mm., colnmn I1 the latent heat of vaporisation, arid column 111 the ralne of the constant in Trouton's expression MS/T = constant, where M is the molecular weight, S the latent heat of vaporisation, and T the boiling point on the scale of absolute temperatures. I. 11. 111. Dipropyl ketone ............ 163*90° 75.94 Cnl. 20.76 Methyl butyl ketone.. ....... 127.61 82.91 .. 20.70 Diethyl ketone.. ...........102.46 90.54 ,, 20- 74 Methyl isopropyl ketone.. . . 94-04 88.67 ,, 20.78 Nethyl ethyl ketone. ........ 79-54 103.44 .. 21.13 Decane .................... 159.45 60.83 .. 19-98 Octane (normal). ........... 124.90 70.92 ,, 29-32 Diethylic carbonate ......... 126.28 72.80 .. 21-53 Dimethylic carbonate. ....... 90.30 87.87 .. 21.76 These results, with those of previous observers, show that in each group of compounds the value of MS/T is almost constarit, whilst i t differs considerably i n different groups. Ostwald has already pointed out that this expression makes it possible to calculate the latent heat of vaporisat'ion with a probable error of not more than 15 per cent. Since, however, the variations in its value are very small in a group of homologous and isomeric compounds, the latent heats of vaporisation of all the memhers of a series can be calcu- lated with an error of not more that 1 to 1.5 per cent., when the Talue has been experimentally determined f o r one member of the series.C. H. B. Determination of Transition Points. By A. E. BAUR (&it. physiknl. Chem., 1895, 18, 180-182) .-The transition temperature was determined by the electrical method of Cohen and Bredig (Abstr., 1894, ii, 407) in the change Na,HP04,12Hz0 - Na2HPOJ,7Hz0, and the method mas used to investigate the cause of the colour change experienced by the solution of the double salt HgIz,2AgI at about 50'. The transition temperature in the first case was found to lie between 36.5 and 36.8, whilst in the second case the colour change from yellow to red is due to t,he formation of mercuric iodide owing to the decomposition of the double salt.L. M. J. Improved Calorimeter for the Application of the Method of Mixtures. By F. A. WATERMAN ( P M . Mag., lb95, [ 5 ] , 40, 413- 4 21).--For the purpose of avoiding the radiation correction in the determination of the specific heat of solids by the method of mixtures, Hesehus has suggested that the calorimeter cup be introduced into the bulb of an air thermometer and maintained a t a constant tem- perature by the introduction of a sufficient amount of cold water, ofGENERAL AND PHYSICAL CHEMISTRY. 147 known temperature, immediately after the introduction of the heated solid. By this means, both the radiation correction and the water eqnivalent of the calorimeter cup are aroided, the heat received by the cold water being equal to that giveu out by the heated sub- stance. The author has constructed a calorimeter on the above principle.The water cooler and dropper is supported upon a vertical rod in such a manner that i t may be quickly turtied about the rod as an axis, and may deliver water directly to the calorimeter cup. An electric heater is supported upon a second vertical rod and may be turned about t,he rod as an axis until it is directly over the calorimeter cup, allowing the heated body to be transferred directly to the cup. A summary of some trial determinations made with this apparatus is given, showing that it gives concordant results. H.. C , Thermal Unit. By ERNEST H. GRIPFJTHS (Phil. Mag., 1895, [ 5 ] , 40, 431-454).-The capacity for heat, of water, has been invariably adopted in defining the thermal unit,, and its capacity at O", 4', 15O, and its mean capacity from 0' to looo, have been variously selected as standards by different observers, This selection is unfortunate, as at present our knowledge of the comparative value of these standards is vague, and comparison of the results of the different investigators becomes impossible. 'J'he range 10' to 30" is of particular import- ance, as the majority of thermal determinations are expressed i n terms of the capacity of water at some point within this range, and as a consequence great attention has been devoted to i t ; but in spite of this, a comparison of the results of those observers whose work appears to be most worthy of attention, shows that a discrepancy which may be as great as 1 in 300 exists between their results.Even greater uncertainty attaches to the determination of the " mean calorie," as the determinations of the ratio of the mean calorie t o the thermal unit at 15O, give results varying from 1.013 to 0-99.57, H difference of Z part in 60. I t appears, therefore, that the endeavours to establish the heat capacity of water as the standard of calori- metric measnrements have, so far, not, met with success, It has been pointed out by Gray, that even if the specific heat oE water was accurately known it would not, for any reason, be arith- metically commensurable with any other definite physical quantity, but i t would be a purely arbitrary quantity. But it is evident that the ideal thermal iinit should be n natural, not an arbitrary, one, and have some real relation with other units of energy.Further, it should not be dependent on the observations or conclusions of any one investigator ; i t should be of a convenient magnitude, and should cause as little disturbance as possible in the numerical values result- ing from our present system of thermal measurements. The author proposes that the value of this ideal unit should be first defined, that value being some convenient multiple of an absolute unit. A first approximation could then be made to the physical unit thus selected, and this improved, if necessary, by subsequent measurements. It is proposed that the unit adopted should be a thermo-dynamic one, and that this ideal unit should be tcrmed a ''Howland." The name148 ABSTRACTS OF OHEMIOAL PAPERS.“ Therm ” is taken to indicate the quantity of heat required to raise 1 gram of water (measured in vacunm) through 1” of the nitrogen tbermometer a t a temperature to of that thermometer. The Rowland should be of such a magnitude that a therm a t some convenient temperature would be its heat equivalent. The parkicular therm which is the exact equivalent of a Rowland could be denoted by the phrase “ standard theym.” (This paper was read a t the 1895 meeting of tho British Associa- tion. The consideration of the whole matter has been referred to the Committee on Electrical Standards.) Heats of Combination of Substances in the Liquid and Solid Conditions. By P. SPEXCER U.PICKERIKG (PhiZ. Hug., “51, 39, 510) .-The author has investigated several so-called molecular com- pounds to ascertain whether, as in the case of the monhydrate o€ sulphuric acid, their heats of formation were the same in both solid and liquid conditions. Their heats of combinatioii as liquids were deter- mined directly, and the value for the solids deduced indirectly from the heats of fusion of the constituents and compounds. The heat ca- pacities of the substances in both conditions had also to be determined i n order to reduce the heats of fusion to the same temperature. The heat of combination thus calculated is, however, the true heat of combination only if the heat of fusion of the compound is equal to the sum of those of its constituents, and the results obtained by the author prove that this is not the case, the heat of fusion of the compound being generally the smaller quantiiy.The true heat, of combination in the solid condition can, therefore, not be obtained. The substances investigated were : Compounds of sulphuric acid, pinacone, stannic bromide, and sodium hydroxide with water ; benzene with azobenzene ; and dinitro- and meta-dinitrobenzene with naph- thalene. The existence of the two last-mentioned compounds was investigated and established by series of f reeaing point determina- tions ; so also was the existence of the octobydrate of stannic bromide, which had not before been isolated. s. u. P. combination of Mercuric Cyanide with Iodides. By h o u r , VARET (Compt. r e d . , 1895, 121, 499-501).-l’he first column gives the heat of dissolution of the salt in water; the second, the heat developed on mixing solutions of mercuric cyanide and the particular iodide ; and the third the heat of formation of the solid salt from its proximate constituents (solid salts and liquid water).H. (3. 1. 2. 3. Hg(CN),,2NnCN7Hg1,,4H,O. . . . -22.8 +5.3 +24*7 C d . Hg;(CN)?,BNH,I,HgT,,~H,O . . . . -23.5 +4*5 +15*0 ,, Hg(CN),.2LiCN,Hg12,7H,0 . . . . - 20.7 +5*5 +50*0 ,, Rg(CN),,Ba(CN),.Hg12,6H20 . . -22.0 +5*3 +31.6 ,, Hg(CN),,Xr(CN),,HgI,,~~*O . . -21.8 $5.5 +41.8 ,, Hg(CN),,Ca(CN),,HgI,,/H,O . . -22.4 +5.5 +49*5 ,, Hg(CN),,nlg(CN),Hg1,,8HzO . . -20.0 f 5 . 3 + 69.1 ,, Hg(CN),,Cd(CN),,HffI,,8H20 . . -22.5 +2*0 +17*3 ,, At 30”, the solutioiis of the iodocyanides, unlike those of the chloro-GENERAL AND PHYSICAL CHEMISTRY.149 cyanides and bromocjanides (this 1-01., i, 3 ; ii, SS), are strongly alkaline to litmus, and give the isopurpnrate reaction with picric acid, and hence it iollows that they contain salts of the type Hg(CN),, M”(CN),,HgI,. The conversion o€ the system SHg(CN), + M”12, into Hg(CN), + M’(CN), + HgL, absorbs about -9.3 Cal. in solution, whilst the heat of formation of the double cyanides, Hg(CN),,M”( CN),, is + 12.4 Cal., and their union with yellow mercuric iodide develops f 2.3 Gal. Lithium, Magnesium, and Copper Cyanides. By RAOUL VARET C. H. B. (Compt. relid., 1895,121, 598-599) .--Lithizhm Cyanide.-The heat of neutralisation of hydrocjanic acid solution by lithia is +5*85 Cal., and hence Li sol. + CN gas + Aq = I i C N diss ... develops +65*12 Cal.Magnesitm C‘yaszide.-The heat of neutralisation [Mg(OH),] is Mg sol. + 2CN gas + aq. = Mg(CN)a Cuprotss CIJ ankle. Hg(CN), diss. + CuJ, sol. = Hg12(red) Cu,O sol. + 2HCN diss. = Cu2(CN), sol. +3.0 Cal., and hence diss. ............................. develops +112*0 C d . + Cu2(CN), ...................... develops +12-8 Cal. + H20 liq. ........................ 9 , +28*8 9, Cu, sol. + 2CN gas = Cu2(CN), sol.. ... > 7 +29.8 9 , Hydrogen cyanide displaces hydrogen chloride from cuprous salts with development of +13*6 Cal., and hydrogen bromide with de- velopment of +10.8 cal., but is displaced by hydrogen iodide with development of +3*2 Gal. These phenomena are analogous to those observed by Berthelot with mercuric oxide, and by the author with mercurous oxide ; but whereas mercuric and mercurous oxides behave similarly, cupric oxide differs from cnprous oxide in that the heats of neutralisation of tha former by hydrogen chloride, homide, and iodide are practically identical.Depression of the Melting Point of Sodium Sulphate by the addition of Foreign Substances. By RICHARD L~WENHERZ (Zeit. physikal. Chem., 2895, 18, 70--90).--Tlie effect of non-electrolytes in lowering the melting point of sodium sulphate was first determined, urea, glycocine, cane sugar, formamide, and glycerol being employed, atid the mean value for the depression constant so obtained was 32.6. The effect of sodium Palts was next investigated, and, as might be expected, they behaved practically as non-electrolytes, sodium phos- phate giving the highest value (37*8), whilst the results indicate that the formula of sodiuni persulphate is NaJ3,0e.By the addition of sulphates (those of potassium, ammonium, and lithium), a double depression constant results. This may due either to dissociation into K and KSO, ions, since neither of these is present in the solvent, or t o an interaction ; K,SO, + Pu’a2S04 = 2KNaSOI. Potassium chloride C . H. B. VOL. LXX. ii. 12I50 ABSTRACTS OF CHEMICAL PAPERS. and nitrate gave double values, which may he explained either I)y dissociatiou 01- interaction with the sodium sulphate, whilst potassium chromate and potassium carbonate gave depression constants equal to three times the normal, and capable of a similar explanation. By the determination of the solubility of sodium sulphnte in (1) wa,fer, and in (2) a solution of urea, the temperature of transition to the anhydride was found to be (1) 32.46, (2) 29-26, the numbers obtained by the direct determinations being 33.39 and 29.26.The depression constant is calculated by use of the forrnula 16 = 0*02I’W/, and by Raoult’s tension law, the values 36 and 33.8 being so obtained. L. M. J. Pressures of Saturation of Oxygen. By THADDEUS ESTREICH ER (Phil. Mag., 1895, [5], 40, 454---463).--Tlic author has measured the temperatures of the saturul ed rapour of oxygen under pressures lower than one atmosphere by ineans of a hydrogen thermometer. Three series of determinations were made, and the results, which alee generally in close agreement with one another, are given in separate tables i n the paper. The value of f i n the Van der Wad’s formula was calculated and compared with that given by a number of other associating and non-associating liquids a t low pressures (compare Guye,.dbstr., 1895, ii, 153). The mettn value for oxygen is about 2-45. The author finds that the value of f always d.ecreases with increase of temperature, both for associating and non-associating substances. The association of the molecules of the liquid has an influence o n j , but is not, the only reason of its increasing. Of 10 ethereal salts, six have values of f much higher than 3-06, although, according to Ramsay and Shields, they are not associated. Perhaps there is some relationship with the molecular weight, as the highest of the alcohois examined, isobutylic alcohol, has the highest f.Thermal Properties of Vapours : Alcohol Vapour and its Relationship to the Laws of Boyle and Gay-Lussac. By ANGELO BATELLI (Ann. Cliirn. Phys., 1895 [7], 5, 2563--875).--The pressui e of saturated ethylic alcohol vspour shows a t high tempern- ture n behaviour similar to that which the author has observed in several other substances, the pressure increasing as candensation proceeds and the liquid accumulates. The maximum pressures between the temperatures -16’ and 240’ may be represented by thc formula of Biot, the values of the constants being H. C. logp = n + bzt + c p , n = 5.0751023 b = 0.0435271 The critical constants obtained from the isothermal curves are : tc = 241.4”, pc = 47,348 mm., and vC = 4.38 C.C.The coefficients of dilatation of alcohol vapour under constant pressure increase as the temperature diminishes, this taking place the more rapidly the nearer +#he vnpour is to the point, of liquefaction. The absolute log b = 2.6387597 log a = 0.00336681 = -4-0217800 10g c = 0.6044184 iogp = i . ~ g s w o i ~GENERAL AND PHYSICAL GELEMISTRY. 151 values of the coefficients and their variations between the same limits of temperature increase with the increasing pressure of the vapour. From the isochoric curves, the values of dp/pdt, the co- efficient of pressure change a t constantl volume, were calculated. The coefficieiit decreases with rising temperature, the variations becoming more mayked the smaller the volume. As the volumes increase, the absolute value of the coefficient diminishes.The formula of Clausius in the modified form, applies very closely to the results obtained with alcohol. Vapour Pressure of Concentrated Solutions of several Salts, especially Lithium and Calcium Nitrates. By JOHN WADDELL (Chem. News, 1895, 72, 201--203).-Into a wide-mouthed, closely stoppered bottle of about 200 to 300 C.C. capacity three small test- tubes were introduced, one of which contained water or alcohol, and each of the others one of the salts to be experimented with. After some experience had been gained, the liquid was frequently added directly to the salts, arid the third test-tube dispensed with. The salts were taken in molecular proportions, weighed in milligrams, but 1, 2, or 4 mols. of one salt were taken to 1 mol.of the other, the numbers obtained being all so reduced as to show the quantity of liquid taketi up per molecule of each salt. Experiments were first made with calcium and lithium nitrates. If a curve is plotted whose ordinates are the quantities of water ab- sorbed respectively by the lithium and the calcium nitrate, it does not differ very much from a straight line, although it is slightly con- cave towards the axis of the lithium nitrate. The ratio of the water absorbed by the lithium nitrate to that absorbed by the calcium nitrate ramged from about four-fifths to five-sixths. As, if all the molecnles of each salt were dissociated into their ions, there would be the same vapour pressure when the amouut of water absorbed per molecule by the lithium and calcium nitrates is in the ratio 2 : 3, it may be assumed that the lithium nitrate is dissociated to a greater extent than the calcium nitrate.When alcohol was employed as the liquid to be absorbed, there was less uniformity than in the case of water. Each molecule of lithium nitrate absorbs approximately four- fifths as much alcohol as each molecule of calcium nitrate. This condition would be fulfilled if all the lithium nitrate were dissociated, and one quarter only of the calcinm nitrate inolecules. A series of experiments was instituted for t'he purpose of comparing the nitrates of the calcium group of metals among themselves and with lithium nitrate. This has not been quite completed, but there seems to be little doubt from what has been done, that barium nitrate is the most absorbent, that the calcium salt comes nest, and that the stron- tium compound, instead of being intermediate between the others, is less absorbent than either.A series of experiments was also made in which the metal was the same, but the salt radicle varied, the halo'id H. C. salts of potassium being chosen for this pnrpose. The bromide and 12--2152 ABSTRACTS OF CHEMICAL PAPERS. iodide both absorb enough water to make a solution while the chlo- ride is still in the solid condition, but when the vaponr pressure from the bromide and iodide comes to be as great as that of the saturated solutiolz of the chloride, it remains constant until the chloride is all dissolved. Therefore, the three salts absorb nearly the same amounts of water, and it appears that these salts are very nearly equally dissociated, even in rather concentrated solutions, but if any- thing, the bromide is more dissociated than the others. Experiments on the resistance of the solutions showed that the amount of dissociation of calcium nitrate is about 45 per cent.as great in a concentrated as in a dilute solution, whilst the amount of dissociation in the case of the lithium salt is about 83 per cent. Determination of the Molecular Weights of some Inorganic Substances. By HEINRICH BILZ (Chem. Centr., 1895, i, 770-771 ; from Math. natw. Mitt. Berlin, 1895, 35--38) .-The author has determined the density of some of the elements at a very high temperature by means of an apparatus made of a highly refractory porcelain. The ex- periments with arsenic, thallium, cadmium, and zinc were made at 1732-1748'.The density of arsenic vapour was found to be 5.30-5-54, the theory for As2 being 5.2. The dissociation at this high temperature was not greater in hydrogen than in nitrogen. Thallium showed a density of 14-77, theory for TI, requiring 14.11. Cadmium had a density of 4.34-4-38, theory requiring 3.87 for Cd. Zinc had a density of 2.64, theory requiring 2.25 f o r Zn. Iridium and tin did not evaporate at these high temperatures. Arsenic trioqide has, be tween 500-770", a density corresponding with the formula AsdOs. The same diminishes with a higher tem- perature, and has, at 1732", the value 7.32, theory for As,03 requir- ing 6.84. It seems that, above 1770', only molecules of the com- pound AszOs can exist.The increase in the boiling point of aqueous solutions of both varieties of arsenious anhydride showed that decom- position into 2 mols. of arsenious acid, H3As03, had occurred. The solution of the crystallised variety of arsenious anhydride in nitro- benzene seems, on the other hand, to contain unaltered As406. Attempts were made to estimate the vapour density of the alkali metals and of magnesium, but constant values could not be obtained although any action of the metals on the porcelain vessel was By CHARLES T. BLANSHARD (Chem. News, 1895, '72, 230-231, 237-238). --Continuing his observations in reference to t.he genesis of the ele- ments (Abstr., 1895, ii, 340), the author now points out that the specific volumes in certain homologous series of organic compounds offer parallels to various conditions that obtain with the atomic volumes of series of related elements.D. A. L. H. C. prevented. L. DE & Specific Volume and the Genesis of the Elements. Molecular Volumes. By ISIDOR TRAUBE (Ber., 1895, 28, 2722- 2728 ; compare Abstr., 1895, ii, 209).-The author compares the ob- served and calculated values (Abstr., 1895, ii, 70) of thc molecularGENERAL AND PHYSICAL CHEMISTRY 153 solution volumes of a large number of organic compounds of very different types, and finds a very close agreement between them ; he gives a revised table of atomic solution volumes for the non-metallic elements, and states that the molecular contraction volume per gram- molecnle of substance dissolved in water shodd be 13.5 C.C.instead of 12.2 as previously given. The molecular solution volume is not a pnrely additive property, but a highly constitutive one, the atomic solution volumes for the various types of oxygen, carbon, nitrogen, sulphur, &c., in organic compounds having very different values ; the presence of a benzene or hexamethyleue ring in the molecule decreases the molecular solution volume of the compound by 8.1 C.C. The molecular solution volume is not simply the sum of the atomic solution volumes of the atoms forming the molecule, as an expansion always occurs during the formstion of a substance ; this “ molecular dilatation” is approximately the same for all substances, and is 12.4 C.C. per gram molecular weight for the gram molecular volume in aqueous solution at 15’; to obtain the n~olecular yolume this number 12.4 must naturally be increased by the addition of 13.5 C.C.; the molecular contraction volume for water thus becomes 25.9 C.C. W. J. P. Molecular Volumetric Determination of Molecular Weights. By ISIDOR TRAURE (BeT., 1895, 28, 2728-2730 ; compare preceding abstract) .--Prom a single specific gravity determination of an aqueous solution containing 0.5-3 per cent. of a substance, the molecular solu- tion volume of the latter can be calculated, and a correct molecular weight assigned to the substance, using the equation (Abstr., 1895, ii, 70)- v, = nt + ng/d - nqia = ZytC f 12.4. The influence of iouic dissocia.tion and of the presence of the various types of ring present i n the substance must, of course, be taken into consideration.In determining the molecular weight of an acid, it is usually con- venient to nentralise the solution with standard soda, using phenol- phthalejin as an indicntor, and to then determine the molecular solu- tion volume of the sodium salt, which in so dilute a solution is practically wholly dissociated. The molecular solution volume of sodium metamidobenzoate determined in a 3.030 per cent. aqueous solution was found to be 89.1 c.c., whilst the calculated valne is 89.2 C.C. ; very accurate results may therefore be obtained. Molecular-volumetric Method of Determining the Molecular Weight and Constitution. By ISIDOR TRAUBE (Ber., 1895, 28, 2924-2928) .-The author gives the following formula for calculat- ing the molecular volumes of the hydrocarbons at 15’.W. J. P. where ZnC is the sum of the products of the atomic volumes and the number of atoms present, and p , (1, T are respectively the number of rings, of double, and of triple linkings. The atomic volume of carbon is 9.9, of hydrogen 3.1, the decrement for each hexamethylene154 ABSTRACTS OF CHEMICAL PAPERS. ring 8.1, for each benzene ring 13.2, for each double iinking 1.7, and f o r each triple linking 2 x 1.7 = 3.4. The molecular volumes of 80 hydrocarbons calculated bF this formula are given in the paper arid compared wit,h those obtained directly from the specific gravities. The agreement between the two series is excellent, and in no case is a greater difference than 6.3 C.C. observed. Since doubling the molecular weight would cause a difference of 25.9 c.c., the applica- tion of the method t o the determination of molecular weights is evi- dent (compare Abstr., 1895, ii, 209).Initial Rates of Osmosis of certain Substances in Water and in Liquids containing Albumin. By W. S. LAZARUS-BARLOW (J. Physiol., 1895, 19, 140--166).-See this vol., ii, 196. H. C. Correct Formulae for Osmotic Pressure, Changes of Solu- bility, Freezing Point and Boiling Point ; and Heats of Solu- tion and Dilution in Dissolved Dissociated Substances. By J. J. VAN LAAR. IT. (Zeit. yhysikal. Chem., 1895, 18, 245-282).- The author investigates, thermodynamically, the above formulae, and obtains theoretically the following results. If a strongly dissociated compound is added to a dilute solution of a feebly dissociated sub- stance with one common ion, the dissociation degree of the former remains almost unaltered, that of the latter compound being dimi- nished.When also there is one ion common, the solubility of the compound is lowered, the least soluble undergoing the greatest relative change. Non-electrolytes do not affect the solubility, neither are they affected in this respect. When there is no conitnon ion, the effects are more complicated, and frequently undetermined. In the case of partition coefficients between water and other solvents: it is seen that, owing to the ions being absent in the other solvents, the apparent, partition coeEcient increases with dilution. The effect of association of the solvent molecules on the formuh deduced here and in the former paper (1895, ii, 107) is considered, and, where necessary, alterations for this given.It is also shown that the vapour pressure of water is not influenced by the presence of indifferent gases, and that the validity of Dalton's law is limited to the cases where the volume is great. The Absorption of Nitrous Oxide in Water and in Salt Solutions. By VICTOR GORDON ( Z e i f . physikal. Chem., 1895, 18, l-l6).-The author has determined the absorption coefficient of nitrous oxide in solntions of chlorides of potassium, sodium, lithium, calcium and strontium, and snlphates of potassium, sodium, lithium, and magnesium. The experiments were in each case performed for three or more concentrations, and at five different temperatures, ranging from 8.1' to 22*3O, and interpolation formule* are giveu for each solution examined.The lowering of the absorption coefficient appears to be proportional t o M3, where M is the number of gram f It is noticeable that these interpolation €ormule are in all cases of the form a = a - pt + yt2 indicating, if the formula holds for extrapolation, a minimum at the temperature /3/2y, which in almost all cases lies between 34" and 49'. L. M. J.GESEKAL A N ) PHYSICAL CHEMISTRI’. 155 molecules of dissolved salt per litre, so that (a-aS)/W = const., a and a, heiiig the coefficient in water and tlie solution respectively. The value of the constant decreases as the temperature rises, and raries with different salts. For analogous salts, however, the con- stants are nearly equal, whilst the value for bivalent salts is double that for univalent salts.L. $1. J. The Partition Coefficients of Solutions in Liquid and Solid Substances. By JACOBUS h1. VAN BEMWELEN (Zeit. physikal. Chew., 1895, 18, 331-334).-The results and conclusions of G. C. Schmidt (Abstr., 1895, ii, 39) are contested. Neither in silicic acid nor in any other substance experimented with by the author, did he find the absorption took place in accordance with Henry’s gaseous law ( c * / c ~ = const., where c2 and c1 are the concentrations of dissolved substance in the liquid and solid respectively). The partition coe6cient was in all cases not constaut, but a complex function of the concentration, and dependent on the temperature and modification of tlie colloid. The coefficient is onIy approximately constant, when t,he concentration is small, so that the author considers Schmidt’s conclusions erroneons.L. M. J. Note by Abstractor.--The partition coefficient should, however, only be constant when the solid absorbent, that is, the colloid, remains of the same modification, and only for dilute solutions, as in strong solutions Henry’s law coulcl not be expected to hold. Self-recorded Breaks in the Properties of Solutions. By P. SPKNCER U. PICKERING (Phil. Mag., [ 5 ] , 40, 472--476).-By running n continuous stream of sulphuric acid into water in a calorimeter, and making a chart of the niotion of the thermometer, either by photo- graphy or by taking successive readings, a diagram is obtained which reproduces automatically the sudden changes of curvature shown by the author’s heat of dissolution determinations (Trans., 1890, 127).By adjusting the initial temperature suitably, the figures obtained are rectilineal, and the breaks become as clearly visible as those which are made on starting or stopping the stream of acid. 8. U. P. Cryoscopic Relations of Dilute Solutions of Cane Sugar and Ethylic Alcohol. By HARRY C. JONES (Phil. Mag., 1895, [5], 40, 383-393 ; and Zeit. ph ysikal. Chm., 1895, 18, 283-293).- Nernst and Abegg (Abstr., 1895, ii, 155) have attributed the high results obtained by the author for tlie molecular lowering of the freezing point in dilute cane sugar solutions to the use of a jacket a t a much lower temperature than the freezing point of the solution. The experiments have therefore been repeated, using a freezing mixture from 0.3’ to O.PO colder than the freezing point of the solution.A large volume of solution (1100 c.c.) was employed, thus diminishing ~ery.greatly the effect of disturbing influences from without. The stirring was carried out so gently that errors from this source could not have assumed any appreciable dimensions. The results obtained are as follows.156 ABSTRACTS OF CHEMICAL PAPERS, Grams in litre. Normal. Lowering found. Gram-mol. lowering. 3.8875 0*01136 04251 2.21 7.775 0.0227 0.C475 2.09 1.5-550 0.0455 0.0915 2-01 23325 0.0682 0,1333 1.95 31.100 0.0909 0.1734 1.91 The molecular lowerings are throughout somewhat lower than those obtained in former determinations, but they are still far above the theoretical value. Dilute solutions of ethylic alcohol gave similar results, and the author does not find sufficient justification for the conclusion that non-electrolytes in fairly dilute solutions give lowerings which conform to the equation t = 0.02 T2/W.Relations between the Cry oscopic Behaviour of the Phenols and their Constitution. By KARL AUWERS (Ber., 1895, 28, 2878-2882).-Although, as the result of former investigatioris (Abstr., 1894, ii, 133 ; 1895, ii, 41), the author was led to conclude that the cryoscopic behaviour of the phenols in benzene solution was normal, certain irregularities observed with paracresol and pzra- nitrophenol rendered this conclusion doubtful. The author has, therefore, submitted this point to the test of further experiment, but, as many of the phenols are only very slightly soluble in benzene at the freezing point, naphthalene has been substituted for benzene as R solrent.Forty-eight phenols were examined in all, namely, phenol and 6 of its homologues, 5 halogen derivatives, 9 nitrophenols, 8 hydroxyaldehydes, 14 yhenolcarboxylic acids, and 5 polybasic phenols. The acids were used in the form of their methylic or ethylic salts. The examination of these substances shows that the cryoscopic behaviour of the phenols is largely dependent on their constitution. Phenols, where substitution is in the ortho-position, exhibit the normal cryoscopic behaviour, but para-substituted phenols behave abcormally. The meta-compounds occupy a position between the ortho- and para-, but rather resemble the para-compounds i n their behnvionr. Among the substituting groups, the influence of the aldzliydic group *CHO is the most marked, and then in decreasing order of influence follow the carboxslkylic group *COOR, the nitro-group, the halogens, and, lastly, the alkylic groups.The influence of ortho-substitution is stronger than that in the metn- or para-position, so that if a phenol contains the same substituting group in both the ortho- and paya-positions, the cryoscopic behaviour of the substance will probably be nearly normal. No exceptions to the above regularities have as yet been met with, and it is suggested that the cryoscopic behaviour of a phenol may aid materially in the determiriation of its constitution. H. C. Cryoscopic Behaviour of Substances having Constitutions similar t o that of the Solvent.By EMANUELE PAT ERN^ (Garzetta, 1895, 25, i, 411--417).-Garelli and Montsnari (Abstr., 1895, ii, 205) showed that the phenols when dissolved in the corresponding hydro- carbons give abnormal depressions of the freezing points, and conclude that this is due to the close chemical relationship existing between H. C.GENERAL AND PHYSICAL CHEMISTRY. 157 the solvent and the dissolved substance ; the author shows, however, that this similarity of coiistitution is not the onIy cause of a small depression of the freezing point, but that the chemical nature of the dissolved substance has also to be considered. Thus, phenol behaves abnormally, whether dissolved in benzene or in paraxplene ; para- xylenol also gives abnormal freezing point depressions both in paraxylene and benzene soltltions ; the similarity of constitution of the solvent a,nd dissolved substance seems ir- these cases to be without effect,.Determinations of the depression in freezing point caused by phenol and benzylphenol in diphenylmethane solution show that phenol behaves fairly normally and benzylphenol quite abnormally, as would be expected from Garelli and Montanari's conclusion. W. J. P. Freezing of Solutions at Constant Temperature. By ALBERT COLSON (Compt. Tend., 1895, 120, 991-993).-An increase of pressure raises the freezing points of liquids which contract during solidification, whilst, on the other hand, the presence of dissolved foreign matter lowers the freezing point in inverse proportion t o the molecular weight of the dissolved substance.The author has endeavoured to ascertain experimentally whether any relation exists between the molecular weight of the dissolved substance and the pressure required to maintain the freezing point of the solvent constant. The solvent selected was benzene freezing at 5.i'. The results are given in the following table. Dissolved substance. M. p . t . P. Henzoic acid .. . . .. . . 122 2.5 0.53O 98 mm. Acetic acid.. .. . . . , .. 60 2.5 1.16 232 ,, Naphthalene ... . . . I I . . 128 2.5 1.06 219 ,, Pmadichlorobenzeue.. 137 2.5 092 180 ,, Paradichlorobenzme.. 137 5.0 1.85 410 ,, I\letadinit,robenzene .. 168 3.0 0.98 225 ,, Here 31 is the molecular weight of the dissolved substance, p the number of grams dissolved in 100 grams of benzene, t the depression of the freezing point, and P the pressure in mm.in Amagat's apparatus required t o iaaise the freezing point of the solution t o that of the solvent (5 mm. on this scale correspond with about 1 atmos.). It mill be seen that in the case of benzoic and acetic acids there is no direct relationship between the pressurea corresponding with a particular freezing point depression and the molecular weights of the acids. On the other hand, if we consider three of the soiutions in which the depression of the freezing point is about lo, we find that the pressure which may be regarded as equivalent t o a depression of 1' is in each case about the same, thus : acetic acid, 232/1*16 = 200 ; naphthalene, 219/1.06 = 206 ; chlorobeuzene, 180j0.92 = 194. H. C. Influence of Chemical Constitution of Organic Compounds 011 their capability of forming Solid Solutions.By FELICE GAHELL~ (Zeit. physiknl. C'henz., 189.3, 18, 51-60).-The paper contains further examples of abnormal depressions of the freezing point occurring when solvent and dissolved substancs are closely158 ABSTRACTS OF CHEJIICAL PAPERS. allied in constitution. Cumarone, indole, and indene in naphthalene give too high values for the molecular weight, as do diphenylene oxide and [j-naphthoquinoline in pheiianthrene, phenanthroline in the same soivent giving a normal result. Dithienyl gives an abnormal depression in diphenyl, but normal in benzene, whilst metauicotine gires a normal value in diphenyl, a result in accord with the views of Pinner (Abstr., 1894, i, 388; 1895, i, 116). The regularity observed in the cyclic compounds is markedly modified by the presence of side chains, thus methylpyrroline gives a normal depres- sion in benzene, as do pyrroline and thiophen in paraxylene, whilst aa-dimethylpyrroline and aa-dimethylthiophen give abnormal values in paraxylene.I n acetophenone, acetylpyrroline and acetothienone give, as expected, abnormal depressions. I n benzoic acid as solvent, a-pyrrolinecarboxylic acid, a-thiophencarboxylic acid, ortho- and rneta-hydrosybenzoic acids, and ortbamidobenzoic acid give abnormal depressions, whilst those caused by parahydroxybenzoic and f urfuran- carboxylic acids are normal, those due to meta- and para-amido- benzoic acids being nearly so. In phenol, the three dihydroxy- benzenes are slightly abnormal, whilst in resorcinol the ortho- and para-compounds give normal Talues.Some fatty compounds are also examined : maleic anhydride in snccinic anhydride giving an abnormal result, whilst those of olejic acid in stearic acid, butyric acid in crotonic acid, apiole and dihydro-npiole in isapiole are normal. I n a short note on Beckmann’s work (Abstr., 1895, ii, 383), the author does not consider that the abnormal results of iodine in benzene solution can be due to the forniatiofi of a solid solution. L. If, J. The Velocity Law of Polymolecular Reactions. By ARTHUI.: A. NOYES (and WALTER 0. SCOTT) (Zeit. physikal. Chein., 1895, 18, 118--132).-For the determination of the order of a chemical reaction, the constancy of tlie velocity constant is not alone suffi- cient, but comparison should be made of the constants, a t a definite stage of the action, obtained in independent experiments with different initial concentrations.Examined thus, i t is seen that the action between hydrogen iodide and hydrogen peroxide is of the second order (see Magnaniui, Abstr., 1892, llO), whilst the same holds foy the action between hydrogen iodide and bromic acid. According to Schwicker’s experiments (Abstr., 1895, ii, 213), the decomposition of potassium hypoiodite is, if in the presence of free iodine, of the first order, b u t no definite conclusion can be drawn if alkali is in excess. The reaction between ferric and stannous chlorides is, however, of. the third order, as is the poly- merisation of cyanic acid. An explanation of tlie first three cases, may be that the reaction takes place in two or more stages, of which the first alone takes an appreciable time.For example, HI + H20, = HI0 + H20; HI0 + HI = H,O + I,. In the last two cases, van’t Hoes law, that the order is determined by the number of interacting molecules, is obeyed. A Reversible Reaction of the First Order. By FRITZ W. KESTER (Zeit. plysikaE. Chem., 1895, 18, 161-179) .-The reversible L. M. J.OEXEItAL AND PHYSICAL CHEJIISTRY. 159 N15 acid. Salicin hydrolysis. change of hexachlor-~-keto-/3-pentene 2 hexachlor-~-keto--,-pentene is considered. The estimation of the compounds is effected yeadily, owing to the slight solubility of the anilide of the p-compound. The /%compound was heated a t 210°, and the quantity of the y-compound estimated hourly, the final state being reached when the quantity was 0.386, whilst the value of l / t log l/(l + c'/c):c raried from 0.055 t o 0.035.The change from the y-compound to the p- was examined a t the same temperature, the final stage being reached when 0.623 of the p-compound was formed, a number agreeing exactly with the prcvious results. Experiments in different atmospheres showed that siuall quantities of aqueous vapour hare a high accelerative influence, the same obtaining to a sinaller extent f o r hydrogen chloride. An increase in temperature to 237.5" caused the relocity to increase tenfold, but the final ratio was only slightly altered, 0.63 to 0.65. At 300°, equilibrium was \-cry speedily reached, the final ratio being 0.85. L. M. J. Velocity of the Hydrolysis of Salicin by Acids.B y Arrmir A. NOPES and WILLIAM J. HALL (Zeit. physilial. Chem., 1895,18,240- 244) .--The investigatioiis were undertaken to determine whether the hydrolysis of a glucoside is in accordance with the reaction law. whicli obtains for the inversion of cane sugar, &c. Salicin was employed for the purpose, the formation of either saligenin or saliretin during the reaction being immaterial, as both are inactive. A 5 per cent. solution was first employed, the rotation of which was - 12-32', and when completely hjdrolysed + 6*00", the ratio being hence 0.487 = c. If the reaction takes place according to the equa- tion dT/dt = k(A - x), the ralue for the constant X; is given by l / t log { (ca, + a,)/(ca, + a ) ) , where a, is the initial rotation, cc1 that at the arbitraryzero of time, and a that after time t.The values thus obtained for k are in close accord, the greatest variation being about 8 per cent., and the reaction therefore, like the sugar inversion, is of the first order. The relative effects of the acids follow the same order as their effects in sugar inyersion, as is seen in the folloving table. Sugar inversion. Hydrochloric. .......... Sulphuric .............. Oxalic ................. Malonic. ............... 1000 499 223 45 1000 536 186 31 L. 11. J. Molecular Symmetry and Asymmetry. By PAUL GROTH (Ber., 1895, 28, 25lO--Z511) .-The generally accepted statement, con- tradicted by Ladenburg (Abstr., 1F95, ii, 489), that when the molecule of a substance contains no plane of symmetry, the substance exhibits enantiomorphism, is quite correct, for planes of symmetry are of two kinds, namely, simple and compound.A figure possesses a plane of simple or direct symmetry when it gives a superposable image on1 ti0 ABSTRACTS OF CHEMICAL PAPERS. reflection from that plane, whilst i t possesses a plane of compouiid or indirect symmetry when a superposable image is only obtained after reflection and rotation through 180' about the normal to the plane. Diketopiperazine, the example quoted by Ladenburg, possesses a plane of compound symmetry, and, therefore, cannot exhibit enantiomorphism ; the application of the term pseudosym- metry to such a case is undesirable, this term being already used in another way. The author points out that the whole question of the symmetry of geometrical figures is now worked out in crystallogra,phic text-books in such a way as to be immediately applicable to all problems con- cerning molecular symmetry, W.J. P. Size of Crystalline Molecules. By AKDREAS FOCK (Ber., 1895, 28, 2734--2742).--Nernst has shown (Abstr., 1892, 560) that when an aqueous solution, coiitaining c1 and cz molecules of two isomor- phous salts per unit volume, is in equilibrium with a solid solution composed of xl and x2 molecules per cent,. of the same two salts respectively, the ratios C J X , and cz/x2 ave constant for all concentra- tions if the niolecular weights of the two salts are the same in both the liquid and the solid solutions ; if, however, the molecular weight in the solid state is times that in the liquid state, TZ being greater or less than unity, then Cl?t/xl and c ~ ~ / z z are constant.If ionic dis- sociation occurs, the number of molecules in the liquid solution changes, and this alteration must be allowed for in using the above constants ; but inasmuch a s the extent of the electrolytic dissociation of two salts having a common ion, in aqueous solubion, is the same, the above ratio should remain practically constant when the solubility of the two salts in molecules per unit volume is nearly the same; if the dissociat,ioa is slight, i t can of course also be disregarded. Disregarding dissociation, therefore, the 'autlhor has calculated the values of the above ratios taking the pairs of isomorphous salts in- vestigated by Bluthmann and Kuntze (Abstr., 1895, ii, 7 ) and shows that potassium dihgdrogen phosphate and arsenate have the same molecular weight both in solution and in the crystalline state.Potas- sium pernianganate and perchiorate and rubidium permanganate have, however, twice the molecular weight when solid that they have when liquid. The author contends that Mu thmann and Knntze's numbers for the equilibrium between potassium and rubidium pernianganates and their aqueous solution show that the two salts crystallise together in all proportions, and not between very narrow limits as stated in the paper quoted above. TV. J. P. Running together and Healing of Crystals. By OTHO L E H w m v (Zeit. physikal. Chew, 1895, 18, 91-95) .-On warming potassium oleate, or even ordinary soap, on a microscope slide with insufiicient alcohol for its complete solution, and allowing the slide to cool, pointed tet,ragonal octahedra separate ; these crystals are 0.1 - 0.3 mm. in length and each usually consists of a string of several individual crystals; they arc best observed by using alcohol colourad red byGENERAL AND PHYSICAL CHEMISTRY. 161 some dye which is not taken up by the deposited crystals. When, by moving the cover-slip two of these crystals are so brought into contact that their long diameters become perpendicular, or in approximately the position of twinning, the sharp end of the one crjstal merely flattens itself against the mass of the other ; if, however, by manipulating the cover slip, the crystals be moved until their longer axes are at about 60°, the smaller crystal turns OF its own accord until in a parallel position to the larger one, and then the two crystals so join together as to form one homogeneous whole. The tendency which deformed soft crystals, such as these, have, towards again assuming a regular form, may be also observed by crushing one of the larger crystals of potassium oleate into small detached fragments by pressure on the cover slip ; the small frag- ments soon become of quite symmetrical form. The processes involved in these changes are obviously the same as those by which a broken crystal becomes whole, or heals itself when immersed in its crystallising solution ; in the latter case, the surface tension of the solid mass can only act through the agency of the solu- tion, whilst with soft crystals like the above, the surface tension of the solid is suEcientlp powerful to cause the arraugement of the fragments in parallel orientation. W. J. P. Convenient Forms of Laboratory Apparatus. By D. ALRER'L KREIPER (Amer. J. Sci., 1895, [3], 50, 132--134).-The author de- cribes a simple form of JzotJilter. The jacket consists of an inverted flask, the bottom of which has been removed, the top of the funnel fits into this opening, and the neck is closed by a stopper containing per- forations for the stem of the funnel, and for the steam and waste pipes. An improved form of the ordinary Banse.12 valve is also described. It consists of a stout glass tube sealed at one end and drawn out in the middle with au opening in the constriction, a piece of rubber tubing containing a smooth slit is placed over this. The collapse of the rnb- ber, which is so common in the Bunsen valve, is thus rendered impos- sible. The valve works much better when lubricated with glycerol. A convenientfowepump may be constructed by adjusting two of the valves just described to the opposite extremities of a T-tube, whilst the third limb is enlarged so :ts to permit the attachment of a large a d stout piece of rubber tubing closed a t one end; this tube being alternately compressed and released by the hand. A New Refractometer. By CARL PULFRICH (Zeit. physikal. Chem., 1895, 18, 294--299.-The author has devised some improvements on the older form of his well-known refractometer so that the in- strument is now a\-ailable for the determination of the refractive index and dispersion not only for sodium light, but also for the C, F, and G lines. A heating arrangement is a130 added for the inveatiga- tion of liquids at various temperatures, and of compounds which have high melting points, whilst the new instrument is also adapted for use as a differential refractometer-that is for the direct deter- mination of the refractive index or dispersion of one solid or liquid with respect to a second. J. J. S. L. M. J.