137 Li.. ...... No ....... I(.. ...... Rb ....... Ga ....... General and Physical Chemistry. 28586 109625 24476 110122 21991 114450 20939 121193 19743 I 122869 I Spectra of the Alkali Metals. By H. KAYSER and C . RUNGE (Ann. Phys. C'hem. [a], 42, 302--320).-The alkali metals or their salts were volatilised in the electric arc, and their spectra, obtained by means of a Rowlnnd grating, were examined. It was found that lines belonging to any one series could be expressed in terms of X-', the reci- procal of their wave-lengths, by the formula X-' = A + Bn-2 + Crt-l, which is a modification of Balmer's formula, and in which n is a whole number, which may vary from 3 to 16. All the alkali metals have a number of reversible lines, which occur in pairs (except in the case of lithium), and are divided over the whole length of the spectram ; these form the chief series.The formulae for the first lines in each series of pairs are as follows:- Lithium,. .. At. wt. 7.01 A - ' = 4358493 - 1331369n-~ - 1100084n-'. Sodium.. ... ,, 22.995 A - ' = 4153631 - 1299851t-: - 803301n-'. Potassium .. ,, 39.09 h" = 35086'53 - 126983n-2 - 625318n-4. Rubidium . . ,, 85.2 A-' = 33762.11 - 125521n-2 - 5622551t-4. Cesium .... ,, 132.7 h-' = 33501.56 - J25077n-' - 489885n-4. The wave-lengths are given in &stroni units. It will be seen that all three constants in the formula decrease as the atomic weight rises. The wave-lengths, therefore, increase with rising atomic weight, as Boisbaudran has already pointed out. The difference between the values of X-' for lines in each pair is inversely pro- portional to the fourth power of ?L, the lowest value of n being always 3.In addition to the above, there are also two series of lines for lithium, rubidium, and cEsium, which (again with the exception of lithium) form one series of pairs ; and there are four series of lines for sodium and potassium, forming two series of pairs. The first lines in each pair me given by the above formula, the value of the constants being Second Series. I First Series. I 1 A. I B. 3267 111241 133207 311224 28666 122391 231700 24549 120726 197913 22021 1 119393 1 63243 - 1 - 1 - I - 1 - l - The difference between the values of X-' for lines in the same pair It is equal to that of the first VOL. LX. 1 is alw%ys the same for each element.138 ABSTRAOlS OF OHEMIOAL PAPERS.series when n = 3, and appears, therefore, to be characteristic for each element. The mean values of this difference are Na, 172, K, 568, Rb, 234.2, Cs, 5450, in wave-numbers. These numbers are very nearly in the proportion of the squares of the atomic weights, for if their equnre roots be multiplied by 1.706 w e get Na, 23.0 K, 40.6 Rb, 82.6 Cs, 126.0, instead of 22.995 39.99 85.2 132.7. For lithium, assuming the above law to be general, the calculated difference between the pairs is one which would quite fall within the liniits of ohservcttion, but as the lithium lines do not foym pairs, this metal seems to be an exception to the rule. A comparison has been made between the lines here measured and the Fraiinhofer lines in Rowland’s solar atlas. Only two pairs in the Ghief series of sodium lines could be detected, the lines of all the other elements being apparently absent.H. C. Dispersive Power of Organic Compounds. By R. NASIXI (Guzzetta, 20, 356-361) .-A claim for priority as against Barbier and Roux (Abstr., 1890, 1:353). Relation between the Refractive and Rotatory Powers of Chemical Compounds. By I. I. KANOKNTROFF ( J . Russ. Chem. Soc., 22, 85--96).-1n two previous papers (Abstr., 1888, 326, 453), the author has shown that on expressing the relation of the refractive and rotatory powers of a substance by the equation a = A$ - B, the relnt,ion A/B = C, a constant peculiar for the solvent, and independent Qf the optically active substance. The author has investigated solu- tions of camphor and turpentine in over 70 different organic solvents, and gives his results in tabnlar form. It is found that in homologous compounds, such as aliphatic alcohols, ethercd salts of fatty acids and their halogen derivatives, the free acids, the aldehydes, chlorides, bromides, &c., the constant C (as a mean = about 26) increases with every increase of CH2, the differences diminishing from the lower to the higher members.and rarying between 1.35 and 0.46, or, as a mean = 0.85. The difference for an increase of H, = 1.4 in gonetically connected compounds, and -2.5 for compounds of dis- similar constitution. Similar values are found for other changes in composition and constitution (double linkage, isomerism, polymerism. substitution), but it would occupy too much spane t o give the results in detail.For aroniatic compounds, the equation a = A$-- B of the fatty series is converted irito a = A$ + B. But where in aliphatic compounds increasing complexity in composition is regularly nccom- panied by an increase of the value of C, in the case of aromatic com- 1)ounds a decrease is ohserved. The author proposes to investigate the influence of inorganic solvents on the value C. B. B. New Photographic Method. By A. G. GREEN, C. F. CROSS, and E. J. BEVAN (Ber., 23, 3131-3133).-The diazo-compounds of de- hpdrot hiotoluidine and its condensed derivatins, which form theGENERAL AND PHYSICAL CEEMISTRY. 139 dyes of the primuline group, can be employed for photographic pnr- poses ; as the sensitiveness of these compounds is increased by com- bination with the complex colloids which constitute animal or vegekable textile fabrics.The sensitive surface is prepared by colonr- ing a cotton or silk fabric with primuline ( 1 to 2 per cent.), and then diazotising. Such a surface will give a complete positive picture after 40 to 180 seconds exposure, that is t o say, in the bright lights the diazo-compound is completely, in the half lights only partially, decomposed, so that a perfect reproduction of the original is obtained in the form of diazo-primuline. The picture can be developed with any of the various amines or phenols which form a dye with the diazo- compound. The authors' experiments have already brought to light the follow- i n g facts :-(1,) The action of light consists in the decomposition of the diazo-group, with evolution of nitrogen, probably with formation of the corresponding primuline phenol. (2.) The rapidity of the action of light varies, cceferis paribus, with !he nature of the substance with which the diazo-compound is combined.(3.) Photographic reproductions of the spectrum show that, as regards intensity of action, the various rays of light are not in the same order as that in wliich they stand with reference to halogen salts of silver. F. S. K. Action of Borax in Developers for Photographic Plates. By P. MERCIER (Compt. rend., 111, 644--645).-Borax, a.lthough an alkaline salt, acts as a retarder of development when mixed with pyrogallol or catechol. The author points out that this is doubtless due to the formation of the conjugated acids described by Lambert, (Abstr., 1889, 864).Quinol, resorcinol, sodium amidonaphthol-P- sulphonate (eikonogen) , and hydroxylamine hydrochloride do not form similar conjugated acids, and with these compounds borax acts as a very good accelerator of development. By E. WIEDEMANN (Ann. Phys. C'hern. [2], 41, 299-301). The violet colour of iodine dissolved in carbon bisulphide changes to brown when the solution is cooled by means of ethcr and solid carbonic anhydride (Abstr., 1888, 543). The author now finds, in accordance with a statement by Liebreich, that. when the brown solutions of iodine in ethereal salts of the fatty acids are heated to about SO" they become violet, provided the solutions be not too concentrated. Solutions of eosin or Magdala-red in alcohol, heated in capillary tubes, are found to exhibit a very marked fluorescence at temperatures above the critical.Experiments with saffranine failed owing to its Contact Difference of Potential of Metals. By F. PASCREN (An.ia. Phys. Chem. [2j, 41, 186-209).-The author shows that an amalgam prepared by the electrolytic deposition of zinc on mercury changes in the electromotive properties which it a t first exhibits, on merely being allowed to stand for some time. I n order to restore its original properties, it must be submitted to a new and longer electro- C. H. B. Optical Notes. decomposition. H. c.140 ABSTRAOTS OF CHENIOAL PAPERS. Iysis. Thus 705.8 grams of mercury were placed in a solution of zinc sulphate of sp. gr. 1.288, and zinc deposited by a current from two Daniel1 cells for 30 seconds.The amalgam thus produced would contain 0.0000656 gram of zinc to 100 grams of mercury. The E.M.F. of freshly-prepared amalgam I ZnSOa I , amalgamated zinc, was then measured and found to be 0.14 volt. After remaining for three hours, the E.JY1.F had risen to 1.1291 volts, and a further electrolysis for 26 seconds was necessary to restore it to its original value. Tho above change is, however, only exhibited by an amalgam which con- tains very small quantities of zinc, and by increasing the quantity of zinc to a sufficient degree the property of the amalgam becomes practically constant. The suggestion is made that dropping electrodes, similar to those described by the author (Ann. Phys. Chem. [2], 41, 62), might, be used in determining the contact difference of potential of metals, if filled with the molten metals, and these be then allowed to flow into some suitable liquid electrolyte.If, as in the case of mercury, there is no potential difference a t the place where the metal enters the electrolyte, the potential difference between two such electrodes will be that of the metals which they contain. Great practical difficulties lie in the way of such experiments as those here suggested, but it may in some cases be possible to use in place of the metals themselves the amalgams which they form with mercury. The author describes a number of experiments made in this manner with zinc amalgam, and shows that the E.M.F. amalgam I mercury, varies with the amount of zinc contained in the amalgam, tbe variation in these experiments being from 0.021 to 0.156 volt.Electrical Conductivity of Precipitated Membranes. By G. TAMMANN (Zeit. physikal. Chem., 6 , 237--240).-A solution of cuprlc sulphate superposed on a solution of potassium ferrocyanide precipitates at the dividing surface an exceedingly fine membrane of cupric ferro- cyanide, which permits the transfusion of water, but not of any of the salts present in the solutions. Notwithstanding this, the author finds that, the presence of such membrane in an electric circuit does not increase the resistance. His mode of experiment was as follows. He prepared solutions of the above-mentioned salts, having equal electrical conductivity, and snperposed them in an electrolytic cell, so that one horizontal electrode was in one solution, the other electrode in the other.The current had thus to traverse the precipitated semi- permeable membrane, and it WRS found that the resistance remained exactly as before. A membrane of zinc ferrocyanide behaves similarly at first, but after some time it increases in thickness, becomes opaque and permeable for the salt^, its resistance meanwhile growing greater, and attaining a maximum i n about 1 5 minutes. Precipitated membranes of zinc and cupric hydroxides thicken rapidly and diminish the con- ductivity by from 5 to 8 per cent. Membranes of insulating material, mch as pyroxylin, increase the resistance enormously, the only con- duction being probably t'hat through the pores. (Compare Ostwald, Abstr., 1890, 1354.) J.W. H. C.GENERAL AND PHYSICAL OHEMISTRY. 141 Influence of Water of Crystallisation on the Electrical Conductivity of Salt Solutions. By J. TROTSCH ( A m . Phys. Chem. [ZJ, 41,259-287).--The conductivities of solutions of a large number of different salts have been determined for temperatures ranging from 10" to 80". The Kohlrausch telephone method was employed, and in order to obviate the difficulty arising from evaporation of the solutions at the higher temperatures the top of each solution was covered with a layer of molten paraffin. The conductivities were measured at every lo", and the difference d between consecutive readings is taken as a mean temperature coefficient for the 10" rise of temperature. Solutions of salts which are ordinarily anhydrous in the solid state, such as KCI, NaCI, KN03, have temperature coefficients which rise continually with the temperature, or which attain a maximum and then remain constant.On the other hand, the temperature coefficient of solutions of hydrated salts at 6rst increases, reaches a maximum, and t,hen decreases, the author attributing this last behaviour to the loss by the salt of its water of crystallisation. Calcium chloride forms ;t:i exception to this rule, tbe three solutions examined, which con- tained 4.5, 19.2, and 32 per cent. CaC12, behaving throughonit as solutions of anhydrous salts. The temperature coefficient in the case oC the second solution, however, only' undergoes a slight increase with rising temperature, and i3 far smaller than that of the first or third solutions.Five solutions of cupric chloride were examined, having the percentages 1.35, 9, 18.2, '28.7, 35.2. The t w o concentrated solu- tions are green, the two dilute solutions blue, and the colour of the third solution is intermediate between the two. All these solutions behave as solutions of hydrated saits, the temperature coeflicient in each case reaching a maximum at between 40" and 50". At the same temperatures, a change in colour is also observed, the green solutions becoming yellow and the blue solutions becoming green. I n each cas9 these changes seem, therefore, to be conditioned by a dehydration of the salt taking place as the temperature rises. Cobalt chloride shows a somewhat similar behaviour, the colour of the solutions changing from red to blue on heating, and at the same time the tem- perature coefficient reaching a maximum.The temperature of the change is, however, higher in this case, and therefore it is not so readily observed. The author concludes that salts are contained in solution partly as hydrates and partly i n the anhydrous state. At high temperatures, the hydrates part with their water, this taking $lace the more readily in the more concentrated solutions. The water of hydration exercises a specific influence 011 the electrical conductivity of solutions. H. C. Electrical Conductivity of Saline Solutions. By P. CHROU- SHTCHOFF and W. PASHKOFF (J. Russ. Chem. Soc., 22, 110-115), and by CHROVSHTOHOVF (ibid., 115--116).-The two papers contain tt large number of experimental data as regards the conductivity of aqueous solutions of salts and mixtures of salts and acids, but the conclu- sions arrived at are the same as those contained in Chroushtchoff's previous papers (Abstr,, 1P89, 808-809).B. B.14 2 ABSTRACTS OF CHEMICAL PAPERS. Solubility of Mixtures of Electrolytically Dissociated Substances. By A. A. NOYES (Zeit. physikal. Chem., 6,84i -267) .- It has been shown by Nernst (this vol., p. 3 ) that the principles regu- lating the influence of two salts on each other's solubility are those deduced from the general law of mass action as interpreted in the light of the electrolytic dissociation theoiy. The author, in the present paper, contributes an account of his experimental work on the subject. He investigated 11 pairs of salts, and fintls the results of his experiments in very good agreement with the theoretical values.Most of the work was done with binary electrolytes, for example, AgBrO, : AgNO?, TlSCN : T1N03, but a few ternary electro- lytes were shown to give results equallp in harmony with the theory. Experiments were made not only with pairs of salts con- taining one ion in common, but also with pairs whose ions were all diff wen t . Reckoning back from the solubilities, i t is possible to calculate the dissociation constants of strong electroljtes. This is a fact of COD- siderable importance, for the ordinary method of calculation from the electrical conductivity fails in such cases to give a constaat number at all. J. W. Method of Determining Thermal Expansion for Equal Quantities of Heat.By E. J. DRAGOCMIS (Zeit. physikal. Chem., 6, 281--284).-Let V be the volume of a substance; g its weight ; a its coefficient of cubical expansion ; c its specific heat ; g its specific gravity ; At the rise of temperature, and A the expansion caused by the communication of 1 cal.; then A = Vaht. But At = l/cg and V/g = 11s; therefore A = ales. In the case of gases, A is evidently inversely proportional to the molecular heat, for a is constant for all gases, and s varies inversely as the molecular weight. The author determines A in the following manner : A dilatometer packed in cotton wool contains the substance whose expansion is to be measured, and also a platinum spiral, the ends of which are fused through the walls. By means of this spiral the substance is heated, a current of about 0.2 ampere being passed. From the current, the heat communicated is easily calculat'ed, and the expansion for this amount is obserred.Comparative experiments with various liquids were executed, and the results found to be satisfactory. J. W. Estimation of the Specific Gravity of Frothy Syrups. By A. GENIESER (Zeit. ang. C'hern., 1890, 44-45).-A tared pyknometer is about two-thirds filled with the syrup, in which air-bubbles are entangled, and the weight is noted. It is then carefully heated in a salt bath, and maintained in ebullition for a few moments. The whole of the air rises to the surface, where it forms an extremely thin layer of froth. After cooling, water is added, so as to float on the surface without mixing. The froth readily dissolves, and the air escapes.The pyknometer is then filled to the mark with water, and weighed. On deducting the excess of weight, above that of the syrup taken, from the total amount of water which the pyknometer willGENERAL AND PHYSICAL CHEMISTRY. 143 contain, the remainder gives the weight of water equal in volume to the syrup taken, and thence the specific gravitty. By A. W. v. H o F b r m N (Ber., 23, 3303-3319).-Dissociation of Cadonic Anhydride.--In 1860, the author published a paper in conjunction with H. Buff, in which i t wits stated that carbonic anhydride is gradually decomposed by the passage of a series of electric sparks through the gas, and that after a time the carbonic oxide and oxygen recombine with explosive violence. A repetition of these experiments has shown that the explosion only occuis under certain special conditions.6-10 C.C. of dry carbonic anhydride under a pressure of 650-700 mm. are brought into a stout glass tube standing over mercury; a short piece of platinum wire is fused into the shorter limb of a, thin U-shaped tube, the tube is filled with mei-cury, and a second piece of wire wound spirally round the outside of the shorter limb, which is then passed up into the vessel containing the gas; in this way the length of the spark may be readily regulated; in general it should be 2.5-3 mm. Connection is made b7 two wires dipping into the mercury contained in the U -tube and trough respectively. The electric current is obtained from two Bunsen’s elements of medium size, which are connected with a Rnhmkorff’s coil and a small Leyden jar, the coil being 30 cm.long and 10 cm. in diameter. The first explosion usually occurs after 15-20 minutes, aud the subsequent ones at shorter intervals, since the regeneration of the carbonic anhydride is not complete. The dissociation of carbonic anhydride may be shown by passing the gas tbrough a glass tube, in the middle of which two plat<inum terminals are fused ; a series of sparks is allowed to paw, and the issuing gas collected over potash ; part of the gas remains undissolved, and is found to be explosive. Carb- onic anhydride does not appear to be at all affected by a glowing spiral of wire ; it was not found pomible to prepare the gas free from air. Dissociation of Steam-The accompanying illustration (next page) shows a form of apparatus which niay be employed for the purpose of showing the dissociation of steam at varying pressures.The wide glass tube is 2.5 cm. in diameter and 20 cm. in length, the lower tube is 1 cm. in diameter and 40 cm. in length ; the apparatus is filled with moist mercury and heated with steam; instead of fixed terminals, the U-tnbe and wires described above may be employed ; in one experi- ment 2.9 C.C. of gas were obtained after ten minutes ; no regeneration of water occurred, as in the case of carbotiic anhydride. The experi- ment may be varied by allowing the apparatus to cool whilst the electric current is continued ; the dissociated gases gradually combine, and the whole tube becomes refilled with mercury.The current employed is obtained from three Bunsen cells, with the coil and Leyden jar as before. Steam may also be dissociated by means of a glowing white hot spiral of platinum wire; the two ends are con- nected with accumulators, stearn is passed over the coil, and the mixed gases are collected over cold water, which serves to condense the excess of steam. Dissociation of Gases and Vapours by the Silent Discharge.-Experi- M. J. S. Dissociation Phenomena.144 ABSTRAOTS OF OHEMIOAL PAPERS. ments in this direction show that ozone is produced by the decompo- sition of carbonic anhydride, the results of Andrews and Tait, Brodic, and others being thus confirmed. Steam may also be decomposed by passing it through 8 Siemens ozone tube, or by the use of the modi- fied apparatus devised by Berthelot ; various experiments were made to prove that the explosive gas obtained was really derived from the steam, and was not due to electrolysis.Berthelot’s results on the decomposition of ammonia by means of the silent discharge are con- firmed. The vapours of methyl alcohol, ethyl alcohol, and ethyl ether may also be dissociated by mems of the silent discharge. Influence of Mineral Acids on the Velocity of the Reaction between Brornic and Hydriodic Acids. By G. MAGNANINI (Gazzeffa, 20,377-393).-The reaction between bromic and hydriodic acids was shown by Ostmald (Abstr., 1885, 1024) to form an excep- tion to the ordinary rule of mass action, and neither Meyerhoffer’s (Abstr., 1889, 9) nor Burchard’s (Abstr., 1889, 208) expressions are found to satisfy the experimental data respecting the vmiation of the speed of the reaction.These discrepancies are evidently occasioned by secondary reactions, which alter the velocity of thc changes at every instant. Ostwald found that mineral acids increase the speed of the reaction, and that the increments are sensibly proportional to the affinity coefficients of the respective acids, or to the quantity of hydrogen electrolytically dissociated from them. The author has continued Ostwald’s investigations on the accelerating or retarding J. B. T.GENERAL AND PHYSICAL CHEMISTRY. 145 effecto. of mineral acids on this reaction, esperimenting with hydro- chloric, nitric, sulphuric, and bromic acids, with mixtures of some of these acids, and with potassiuni bromate ; from the tabulated results, he draws the following conclusions.During tohe course of the different changes, the reaction is influenced in the same way by the secondary actions. The reciprocal values of the times required for the separation of a determinate quantity of iodine vary as the velocity of the respective reactions. The velocity of the reaction is accelerated by hydrochloric acid, but the acceleration is not propor- tional to the quantity of acid present. The quantity of iodine deposited after equal times in presence of equivalent quantities of hydrochloric and nitric acids is the same. Mixtures of hydrochloric and nitric acids, in any proportions, are equivalent to either of the acids, so that, independently of the nature of the electro-negative radicle, it may be said that the velocity of the reaction depends entirely on the quantity of hydrogen electrolytically dissociated, without, however, being proportional t o it.The action of sulphuric acid is more complicated, on account o€ t.he incomplete dissociation of that acid. The :tccelerating effect of bromic acid is almost six times that of hydrochloric or nitric acid. S. B. A. A. Velocity of the Halogenisation of Fatty Hydrocarbons. By If. WILDERMAN" (BAT., 23, 317&3175).--The following two laws are deduced from a study of the action of bromine or chlorine in sunlight on amyl bromide, amylene bromide, liquid and solid tri- bromopentane, tetrabromopentane, and amylene chloride :-( 1) Sub- stitution proceeds more slowly as the quantity of positrive hydrocarbon becomes smaller.(2) The larger the quantity of hydrocarbon present, the quicker the substitution. Cryoscopic Investigation of Colloidal Substances. By A. SABAN~EFF ( J . Russ. Chew. SOC., 22, 102--107),-The author has shown that Raoult's method may be conveniently employed for the determination of tbe molecular weight of colloidal substances (Abstr., 1890, 1215). Similar results were obtained by Morris and Brown, by Ekstraud and Mauzelius. 0 1 1 the other hand, Paternb, in his research on gallic and tannic acide, has arrived a t the conclusion that their molecular. weights cannot be determined by Rao ult's method. The values obtained by Paternb give 10 mols. of the first, and 109 ruols. of the second, as the molecular weights in solution. The author shows that Paternb's paper includes an error in calculation, and that the values greatly depend on the purity of the material.First the molecular weight of gallic acid was determined. It mas dried at 120°, iosing 9.65 per cent. water, corresponding with the formula C,H6OS + H20. In aqueous solution, containing 0.5238 per cent., the depres- sion was 0*06", and the rnolecular weight 166 ; i n acetic acid, con- taining 0.4107 part, the depression was 0*095", molecular weight = 168, whereas the value for C7H606 is 170; gallic acid exists, there- fore, as a. airigle molecule in solutions. Commercial tannin was found to contain some gallic acid, the quantity of which, as calculated from the depression of the impure J. B. T.146 ABSTRACTS OF CHEMICAL PAPERS.preparation, compared with that of the pure preparation, was found to be 2-39 per cent. It was dried for 20 hours at 120', and the loss was 10.66 per cent. Aqueous solutions with a concentration of 0-822-3.773 shorn a depression of 0*015--0~060, giving a molecular weight of 1044-1195 (mean, 1104). More concentrated solutions, contain- ing 5.5 to 9.5 per cent., gave higher molecular weights, 1497 to 2436, but the values are useless, as solutions of more than 4 per cent. of tannin in water become turbid a t O", tannin separating out. I n Paternb's solutions, containing 11.5-23 per cent. of tannin, $--3 Of the tannin must have separated tit 0" in the insoluble state, and h i s (corrected) molecular weights, = 2643-3700, are of no value. The author finds in acetic solution, molecular weights = 1105-1114 (mean 1113).Pure tannin was pwpared by Lowe's method, but it was impossible to work with aqueous solutions, as not more than 0-5 per cent. is dissolved in water at O", and even such weak solutions became turbid. I n acetic acid solution, as a mean, M = 1322 was found, whereas, M calculated for (C14Hlo09)4 = 1288, so that tannin exists in solutions as a quadruple molecule. The author thinks that Paternb's tannin was not quite dry, and shows by experiment that such a preparation causes a much larger depression, owing to the contamination of the glacial acetic acid by the water of crystallisation of the compound. The author says that the empirical formula of tannin requires confirmation by further research. Apparatus for Distillation under Reduced Pressure.By H. W ISLICENUS (Ber., 23,3292-3295).--The author describes two forms of apparatus for use with the the Bunsen pump, to prevent backward diffusion. The first consists of a tube, with one round and one pear- shaped bulb; in the depression between these a rubber ring is placed, one end of the tube is sealed, 5 small opening is made in the side and is covered with a piece of rubber tube, or an ordinary Bunsen valve may be attached; the other end of the tube is connected with the vessel to be exhausted, the pear-shaped bulb fits into a wide tube, the rubber ring serving to make the connection air-tight ; the second tube is joined to the pumpin the usual manner. The second form of valve consists of two tubes, one closely resembling a thistle funnel, the narrow end of which is attached to the pump; a bulb with a small aperture is blown a t the end of the second tube, and i t is covered with a rubber cap, through which an opening is pierced st a litt.le distance from the one in the glass; the rubber cap serves to make an air-tight connection between the bulb and the wide end of the first tube.An apparatus for fractional distillation under reduced pressure is also descrihed ; it consists of a combination of several of the first of these joints, and provision is made for changing the receiver without interrupting the distillation. Isomorphism. Part 111. By J. W. RETGERS (Zeit. physikaE. Chem., 6,193-236).-1n this communication (for previocs papers see Abstr., 1890,328,1208), the author first discusses the relations of morphotropy.He would limit the term morphotropic to such substances as show a B. B. J. B. T.GENERAL AND PHYSICAL CHEMISTRY. 147 total analogy of form, and not, for instance, merely analogy of angles in one zone. Isomorphons substances have not only this total form- analogy, but also analogy of chemical constitution. Morphotropic substances are not necessarily cheniically analogous, but must be chemically connected with each other. Substances which show a total form-analogy, but have no chemical rcsemblance, are termed isogonic. In the regular system, for instance, potassium chloride and rubidium chloride are isomorphous ; potassium chloride and sodium chloride, morphotropic ; and sodium chloride and sodium chlorate, isogonic. lsomorphous mixtures are proved by the continuous linear change of physical properties with the composition of thc mixture.No intimate crystalliue mixture can be obtained with merely morphotropic Substances, and in this case identity of system, and of degree of hemi- hedry, are not essential. The author rejects Marignttc and Klein’s conception of mas3 isomorphism, according to which an element, o r group, largely preponderating in a compound, determines the crystal - line form, no matter what the other components may be. It is the volume and not the weight of the group that is decisive. An investigation follows of the supposed isomorphism between potassium and sodium sulphates. The author proves that a definite double salt, 3K2S04,Na$04, crystallisev from a solution of the mixed snlphates.It usually crystallises in hexagonal prisms or pyramids, but when the mother liquor contains sodium chloride, i t separates in tables-the “ plate-sulphate ” obtained in the manufacture of iodine from kelp. Potassium chloride does not effect this change of form. The three simple forms are then- K2S04. Rhornbic, pseudo-hexagonal. Forms pyramids and prisms of hexagonal section. Optically biaxial. Weak birefringence. Sp. gr. = 2.666. 3KZSO4,Na,SO4. Hexagonal. Forms pyramids and prisms. Optic- ally uniaxial. Marked birefringence. Sp. gr. = 2.695. Easily fukible. NhSO,. Rhombic, but not pseudo-hexagonal. Forms only pyramids with rhombic section. Optically biaxial. Strong birefrin- gence. Sp. gr. 2.673. Fusible with great difficulty. The double salt is not an isomorphons mixture, as is shown by a consideration of its properties in relation to those of the simple salts.Each simple salt can take up a minute quantity of the other, which points to a very limited isodimorphism. Not easily fusible. The general results are- (1) K&Oa and Na2S04 are not isomorphous. (2) They only form one double salt, 3KzSOa,Na2S04. (3) From mixed solution the pure double salt separates out along- (4) K,SO, and the double salt are morphotropic. (5) NazS04 is not, morphotropic either with K2S04 or with thO double salt, but is crystallographically completely independent. The dolomite series is next discussed. In this series we have calcspnrs containing a little magnesium carbonate, magnesites con- taining a little calcium carbonate, and dolomites which have the two salts in nearly equal molecular proportions.Calculated from the side one or other of the simple salts.148 ABSTRACTS OF OHFJIICAL PAPERS. specific gravities of the component salts, that of dolomite shonld be 2.843 ; it is actually 2.872. The author considers the series not iso- morphoua, but merely morphotropic. J. W.137Li.. ......No .......I(.. ......Rb .......Ga .......General and Physical Chemistry.28586 10962524476 11012221991 11445020939 12119319743 I 122869 ISpectra of the Alkali Metals. By H. KAYSER and C . RUNGE(Ann. Phys. C'hem. [a], 42, 302--320).-The alkali metals or theirsalts were volatilised in the electric arc, and their spectra, obtained bymeans of a Rowlnnd grating, were examined.It was found that linesbelonging to any one series could be expressed in terms of X-', the reci-procal of their wave-lengths, by the formula X-' = A + Bn-2 + Crt-l,which is a modification of Balmer's formula, and in which n is awhole number, which may vary from 3 to 16.All the alkali metals have a number of reversible lines, whichoccur in pairs (except in the case of lithium), and are divided overthe whole length of the spectram ; these form the chief series. Theformulae for the first lines in each series of pairs are as follows:-Lithium,. .. At. wt. 7.01 A - ' = 4358493 - 1331369n-~ - 1100084n-'.Sodium.. ... ,, 22.995 A - ' = 4153631 - 1299851t-: - 803301n-'.Potassium .. ,, 39.09 h" = 35086'53 - 126983n-2 - 625318n-4.Rubidium .. ,, 85.2 A-' = 33762.11 - 125521n-2 - 5622551t-4.Cesium .... ,, 132.7 h-' = 33501.56 - J25077n-' - 489885n-4.The wave-lengths are given in &stroni units. It will be seenthat all three constants in the formula decrease as the atomic weightrises. The wave-lengths, therefore, increase with rising atomicweight, as Boisbaudran has already pointed out. The differencebetween the values of X-' for lines in each pair is inversely pro-portional to the fourth power of ?L, the lowest value of n beingalways 3.In addition to the above, there are also two series of lines forlithium, rubidium, and cEsium, which (again with the exception oflithium) form one series of pairs ; and there are four series of linesfor sodium and potassium, forming two series of pairs.The firstlines in each pair me given by the above formula, the value of theconstants beingSecond Series. I First Series. I1 A. I B.326711124113320731122428666 122391 23170024549 120726 19791322021 1 119393 1 63243- 1 - 1 -I - 1 - l -The difference between the values of X-' for lines in the same pairIt is equal to that of the firstVOL. LX. 1is alw%ys the same for each element138 ABSTRAOlS OF OHEMIOAL PAPERS.series when n = 3, and appears, therefore, to be characteristic for eachelement. The mean values of this difference are Na, 172, K, 568,Rb, 234.2, Cs, 5450, in wave-numbers. These numbers are verynearly in the proportion of the squares of the atomic weights, for iftheir equnre roots be multiplied by 1.706 w e getNa, 23.0 K, 40.6 Rb, 82.6 Cs, 126.0,instead of 22.995 39.99 85.2 132.7.For lithium, assuming the above law to be general, the calculateddifference between the pairs is one which would quite fall within theliniits of ohservcttion, but as the lithium lines do not foym pairs, thismetal seems to be an exception to the rule.A comparison has been made between the lines here measured andthe Fraiinhofer lines in Rowland’s solar atlas.Only two pairs in theGhief series of sodium lines could be detected, the lines of all theother elements being apparently absent. H. C.Dispersive Power of Organic Compounds. By R. NASIXI(Guzzetta, 20, 356-361) .-A claim for priority as against Barbierand Roux (Abstr., 1890, 1:353).Relation between the Refractive and Rotatory Powers ofChemical Compounds.By I. I. KANOKNTROFF ( J . Russ. Chem.Soc., 22, 85--96).-1n two previous papers (Abstr., 1888, 326, 453),the author has shown that on expressing the relation of the refractiveand rotatory powers of a substance by the equation a = A$ - B, therelnt,ion A/B = C, a constant peculiar for the solvent, and independentQf the optically active substance. The author has investigated solu-tions of camphor and turpentine in over 70 different organic solvents,and gives his results in tabnlar form. It is found that in homologouscompounds, such as aliphatic alcohols, ethercd salts of fatty acidsand their halogen derivatives, the free acids, the aldehydes, chlorides,bromides, &c., the constant C (as a mean = about 26) increaseswith every increase of CH2, the differences diminishing from thelower to the higher members.and rarying between 1.35 and 0.46, or,as a mean = 0.85. The difference for an increase of H, = 1.4 ingonetically connected compounds, and -2.5 for compounds of dis-similar constitution. Similar values are found for other changes incomposition and constitution (double linkage, isomerism, polymerism.substitution), but it would occupy too much spane t o give the resultsin detail. For aroniatic compounds, the equation a = A$-- B of thefatty series is converted irito a = A$ + B. But where in aliphaticcompounds increasing complexity in composition is regularly nccom-panied by an increase of the value of C, in the case of aromatic com-1)ounds a decrease is ohserved.The author proposes to investigatethe influence of inorganic solvents on the value C. B. B.New Photographic Method. By A. G. GREEN, C. F. CROSS, andE. J. BEVAN (Ber., 23, 3131-3133).-The diazo-compounds of de-hpdrot hiotoluidine and its condensed derivatins, which form thGENERAL AND PHYSICAL CEEMISTRY. 139dyes of the primuline group, can be employed for photographic pnr-poses ; as the sensitiveness of these compounds is increased by com-bination with the complex colloids which constitute animal orvegekable textile fabrics. The sensitive surface is prepared by colonr-ing a cotton or silk fabric with primuline ( 1 to 2 per cent.), andthen diazotising. Such a surface will give a complete positive pictureafter 40 to 180 seconds exposure, that is t o say, in the bright lightsthe diazo-compound is completely, in the half lights only partially,decomposed, so that a perfect reproduction of the original is obtainedin the form of diazo-primuline.The picture can be developed withany of the various amines or phenols which form a dye with the diazo-compound.The authors' experiments have already brought to light the follow-i n g facts :-(1,) The action of light consists in the decomposition ofthe diazo-group, with evolution of nitrogen, probably with formationof the corresponding primuline phenol. (2.) The rapidity of theaction of light varies, cceferis paribus, with !he nature of the substancewith which the diazo-compound is combined. (3.) Photographicreproductions of the spectrum show that, as regards intensity ofaction, the various rays of light are not in the same order as that inwliich they stand with reference to halogen salts of silver.F.S. K.Action of Borax in Developers for Photographic Plates.By P. MERCIER (Compt. rend., 111, 644--645).-Borax, a.lthough analkaline salt, acts as a retarder of development when mixed withpyrogallol or catechol. The author points out that this is doubtlessdue to the formation of the conjugated acids described by Lambert,(Abstr., 1889, 864). Quinol, resorcinol, sodium amidonaphthol-P-sulphonate (eikonogen) , and hydroxylamine hydrochloride do not formsimilar conjugated acids, and with these compounds borax acts as avery good accelerator of development.By E.WIEDEMANN (Ann. Phys. C'hern. [2], 41,299-301). The violet colour of iodine dissolved in carbon bisulphidechanges to brown when the solution is cooled by means of ethcr andsolid carbonic anhydride (Abstr., 1888, 543). The author now finds,in accordance with a statement by Liebreich, that. when the brownsolutions of iodine in ethereal salts of the fatty acids are heated toabout SO" they become violet, provided the solutions be not tooconcentrated.Solutions of eosin or Magdala-red in alcohol, heated in capillarytubes, are found to exhibit a very marked fluorescence at temperaturesabove the critical. Experiments with saffranine failed owing to itsContact Difference of Potential of Metals. By F. PASCREN(An.ia. Phys. Chem. [2j, 41, 186-209).-The author shows that anamalgam prepared by the electrolytic deposition of zinc on mercurychanges in the electromotive properties which it a t first exhibits, onmerely being allowed to stand for some time.I n order to restore itsoriginal properties, it must be submitted to a new and longer electro-C. H. B.Optical Notes.decomposition. H. c140 ABSTRAOTS OF CHENIOAL PAPERS.Iysis. Thus 705.8 grams of mercury were placed in a solution ofzinc sulphate of sp. gr. 1.288, and zinc deposited by a current fromtwo Daniel1 cells for 30 seconds. The amalgam thus produced wouldcontain 0.0000656 gram of zinc to 100 grams of mercury. TheE.M.F. of freshly-prepared amalgam I ZnSOa I , amalgamated zinc, wasthen measured and found to be 0.14 volt.After remaining for threehours, the E.JY1.F had risen to 1.1291 volts, and a further electrolysisfor 26 seconds was necessary to restore it to its original value. Thoabove change is, however, only exhibited by an amalgam which con-tains very small quantities of zinc, and by increasing the quantity ofzinc to a sufficient degree the property of the amalgam becomespractically constant.The suggestion is made that dropping electrodes, similar to thosedescribed by the author (Ann. Phys. Chem. [2], 41, 62), might, beused in determining the contact difference of potential of metals, iffilled with the molten metals, and these be then allowed to flow intosome suitable liquid electrolyte. If, as in the case of mercury, thereis no potential difference a t the place where the metal enters theelectrolyte, the potential difference between two such electrodes will bethat of the metals which they contain.Great practical difficulties liein the way of such experiments as those here suggested, but it mayin some cases be possible to use in place of the metals themselves theamalgams which they form with mercury. The author describes anumber of experiments made in this manner with zinc amalgam, andshows that the E.M.F. amalgam I mercury, varies with the amountof zinc contained in the amalgam, tbe variation in these experimentsbeing from 0.021 to 0.156 volt.Electrical Conductivity of Precipitated Membranes. By G.TAMMANN (Zeit. physikal. Chem., 6 , 237--240).-A solution of cuprlcsulphate superposed on a solution of potassium ferrocyanide precipitatesat the dividing surface an exceedingly fine membrane of cupric ferro-cyanide, which permits the transfusion of water, but not of any of thesalts present in the solutions.Notwithstanding this, the authorfinds that, the presence of such membrane in an electric circuit doesnot increase the resistance. His mode of experiment was as follows.He prepared solutions of the above-mentioned salts, having equalelectrical conductivity, and snperposed them in an electrolytic cell, sothat one horizontal electrode was in one solution, the other electrodein the other. The current had thus to traverse the precipitated semi-permeable membrane, and it WRS found that the resistance remainedexactly as before.A membrane of zinc ferrocyanide behaves similarly atfirst, but after some time it increases in thickness, becomes opaque andpermeable for the salt^, its resistance meanwhile growing greater, andattaining a maximum i n about 1 5 minutes. Precipitated membranesof zinc and cupric hydroxides thicken rapidly and diminish the con-ductivity by from 5 to 8 per cent. Membranes of insulating material,mch as pyroxylin, increase the resistance enormously, the only con-duction being probably t'hat through the pores. (Compare Ostwald,Abstr., 1890, 1354.) J. W.H. CGENERAL AND PHYSICAL OHEMISTRY. 141Influence of Water of Crystallisation on the ElectricalConductivity of Salt Solutions. By J. TROTSCH ( A m . Phys. Chem.[ZJ, 41,259-287).--The conductivities of solutions of a large numberof different salts have been determined for temperatures ranging from10" to 80".The Kohlrausch telephone method was employed, and inorder to obviate the difficulty arising from evaporation of the solutionsat the higher temperatures the top of each solution was covered with alayer of molten paraffin. The conductivities were measured at every lo", and the difference d between consecutive readings is taken as amean temperature coefficient for the 10" rise of temperature.Solutions of salts which are ordinarily anhydrous in the solid state,such as KCI, NaCI, KN03, have temperature coefficients which risecontinually with the temperature, or which attain a maximum andthen remain constant. On the other hand, the temperature coefficientof solutions of hydrated salts at 6rst increases, reaches a maximum,and t,hen decreases, the author attributing this last behaviour to theloss by the salt of its water of crystallisation.Calcium chloride forms;t:i exception to this rule, tbe three solutions examined, which con-tained 4.5, 19.2, and 32 per cent. CaC12, behaving throughonit assolutions of anhydrous salts. The temperature coefficient in the caseoC the second solution, however, only' undergoes a slight increase withrising temperature, and i3 far smaller than that of the first or thirdsolutions. Five solutions of cupric chloride were examined, havingthe percentages 1.35, 9, 18.2, '28.7, 35.2. The t w o concentrated solu-tions are green, the two dilute solutions blue, and the colour of thethird solution is intermediate between the two.All these solutionsbehave as solutions of hydrated saits, the temperature coeflicient ineach case reaching a maximum at between 40" and 50". At the sametemperatures, a change in colour is also observed, the green solutionsbecoming yellow and the blue solutions becoming green. I n eachcas9 these changes seem, therefore, to be conditioned by a dehydrationof the salt taking place as the temperature rises. Cobalt chlorideshows a somewhat similar behaviour, the colour of the solutionschanging from red to blue on heating, and at the same time the tem-perature coefficient reaching a maximum. The temperature of thechange is, however, higher in this case, and therefore it is not soreadily observed.The author concludes that salts are contained in solution partly ashydrates and partly i n the anhydrous state.At high temperatures,the hydrates part with their water, this taking $lace the more readilyin the more concentrated solutions. The water of hydration exercisesa specific influence 011 the electrical conductivity of solutions.H. C.Electrical Conductivity of Saline Solutions. By P. CHROU-SHTCHOFF and W. PASHKOFF (J. Russ. Chem. Soc., 22, 110-115), andby CHROVSHTOHOVF (ibid., 115--116).-The two papers contain tt largenumber of experimental data as regards the conductivity of aqueoussolutions of salts and mixtures of salts and acids, but the conclu-sions arrived at are the same as those contained in Chroushtchoff'sprevious papers (Abstr,, 1P89, 808-809).B. B14 2 ABSTRACTS OF CHEMICAL PAPERS.Solubility of Mixtures of Electrolytically DissociatedSubstances. By A. A. NOYES (Zeit. physikal. Chem., 6,84i -267) .-It has been shown by Nernst (this vol., p. 3 ) that the principles regu-lating the influence of two salts on each other's solubility are thosededuced from the general law of mass action as interpreted in thelight of the electrolytic dissociation theoiy. The author, in thepresent paper, contributes an account of his experimental work onthe subject. He investigated 11 pairs of salts, and fintls the resultsof his experiments in very good agreement with the theoreticalvalues. Most of the work was done with binary electrolytes, forexample, AgBrO, : AgNO?, TlSCN : T1N03, but a few ternary electro-lytes were shown to give results equallp in harmony with the theory.Experiments were made not only with pairs of salts con-taining one ion in common, but also with pairs whose ions were alldiff wen t .Reckoning back from the solubilities, i t is possible to calculate thedissociation constants of strong electroljtes.This is a fact of COD-siderable importance, for the ordinary method of calculation from theelectrical conductivity fails in such cases to give a constaat numberat all. J. W.Method of Determining Thermal Expansion for EqualQuantities of Heat. By E. J. DRAGOCMIS (Zeit. physikal. Chem., 6,281--284).-Let V be the volume of a substance; g its weight ; aits coefficient of cubical expansion ; c its specific heat ; g its specificgravity ; At the rise of temperature, and A the expansion caused bythe communication of 1 cal.; then A = Vaht.But At = l/cg andV/g = 11s; therefore A = ales. In the case of gases, A is evidentlyinversely proportional to the molecular heat, for a is constant for allgases, and s varies inversely as the molecular weight.The author determines A in the following manner : A dilatometerpacked in cotton wool contains the substance whose expansion is tobe measured, and also a platinum spiral, the ends of which are fusedthrough the walls. By means of this spiral the substance is heated,a current of about 0.2 ampere being passed. From the current, theheat communicated is easily calculat'ed, and the expansion for thisamount is obserred.Comparative experiments with various liquidswere executed, and the results found to be satisfactory. J. W.Estimation of the Specific Gravity of Frothy Syrups. ByA. GENIESER (Zeit. ang. C'hern., 1890, 44-45).-A tared pyknometeris about two-thirds filled with the syrup, in which air-bubbles areentangled, and the weight is noted. It is then carefully heated in asalt bath, and maintained in ebullition for a few moments. Thewhole of the air rises to the surface, where it forms an extremely thinlayer of froth. After cooling, water is added, so as to float on thesurface without mixing. The froth readily dissolves, and the airescapes. The pyknometer is then filled to the mark with water, andweighed. On deducting the excess of weight, above that of the syruptaken, from the total amount of water which the pyknometer wilGENERAL AND PHYSICAL CHEMISTRY.143contain, the remainder gives the weight of water equal in volume tothe syrup taken, and thence the specific gravitty.By A. W. v. H o F b r m N (Ber., 23,3303-3319).-Dissociation of Cadonic Anhydride.--In 1860, theauthor published a paper in conjunction with H. Buff, in which i twits stated that carbonic anhydride is gradually decomposed by thepassage of a series of electric sparks through the gas, and that aftera time the carbonic oxide and oxygen recombine with explosiveviolence. A repetition of these experiments has shown that theexplosion only occuis under certain special conditions. 6-10 C.C.ofdry carbonic anhydride under a pressure of 650-700 mm. arebrought into a stout glass tube standing over mercury; a shortpiece of platinum wire is fused into the shorter limb of a, thinU-shaped tube, the tube is filled with mei-cury, and a second pieceof wire wound spirally round the outside of the shorter limb, whichis then passed up into the vessel containing the gas; in this waythe length of the spark may be readily regulated; in general itshould be 2.5-3 mm. Connection is made b7 two wires dipping intothe mercury contained in the U -tube and trough respectively. Theelectric current is obtained from two Bunsen’s elements of mediumsize, which are connected with a Rnhmkorff’s coil and a smallLeyden jar, the coil being 30 cm. long and 10 cm.in diameter.The first explosion usually occurs after 15-20 minutes, aud thesubsequent ones at shorter intervals, since the regeneration of thecarbonic anhydride is not complete. The dissociation of carbonicanhydride may be shown by passing the gas tbrough a glass tube, inthe middle of which two plat<inum terminals are fused ; a series ofsparks is allowed to paw, and the issuing gas collected over potash ;part of the gas remains undissolved, and is found to be explosive. Carb-onic anhydride does not appear to be at all affected by a glowing spiralof wire ; it was not found pomible to prepare the gas free from air.Dissociation of Steam-The accompanying illustration (next page)shows a form of apparatus which niay be employed for the purpose ofshowing the dissociation of steam at varying pressures.The wide glasstube is 2.5 cm. in diameter and 20 cm. in length, the lower tube is 1 cm.in diameter and 40 cm. in length ; the apparatus is filled with moistmercury and heated with steam; instead of fixed terminals, theU-tnbe and wires described above may be employed ; in one experi-ment 2.9 C.C. of gas were obtained after ten minutes ; no regenerationof water occurred, as in the case of carbotiic anhydride. The experi-ment may be varied by allowing the apparatus to cool whilst theelectric current is continued ; the dissociated gases gradually combine,and the whole tube becomes refilled with mercury. The currentemployed is obtained from three Bunsen cells, with the coil andLeyden jar as before.Steam may also be dissociated by means of aglowing white hot spiral of platinum wire; the two ends are con-nected with accumulators, stearn is passed over the coil, and themixed gases are collected over cold water, which serves to condensethe excess of steam.Dissociation of Gases and Vapours by the Silent Discharge.-Experi-M. J. S.Dissociation Phenomena144 ABSTRAOTS OF OHEMIOAL PAPERS.ments in this direction show that ozone is produced by the decompo-sition of carbonic anhydride, the results of Andrews and Tait, Brodic,and others being thus confirmed. Steam may also be decomposed bypassing it through 8 Siemens ozone tube, or by the use of the modi-fied apparatus devised by Berthelot ; various experiments were madeto prove that the explosive gas obtained was really derived from thesteam, and was not due to electrolysis.Berthelot’s results on thedecomposition of ammonia by means of the silent discharge are con-firmed. The vapours of methyl alcohol, ethyl alcohol, and ethylether may also be dissociated by mems of the silent discharge.Influence of Mineral Acids on the Velocity of the Reactionbetween Brornic and Hydriodic Acids. By G. MAGNANINI(Gazzeffa, 20,377-393).-The reaction between bromic and hydriodicacids was shown by Ostmald (Abstr., 1885, 1024) to form an excep-tion to the ordinary rule of mass action, and neither Meyerhoffer’s(Abstr., 1889, 9) nor Burchard’s (Abstr., 1889, 208) expressions arefound to satisfy the experimental data respecting the vmiation of thespeed of the reaction.These discrepancies are evidently occasionedby secondary reactions, which alter the velocity of thc changes atevery instant. Ostwald found that mineral acids increase the speedof the reaction, and that the increments are sensibly proportional tothe affinity coefficients of the respective acids, or to the quantity ofhydrogen electrolytically dissociated from them. The author hascontinued Ostwald’s investigations on the accelerating or retardingJ. B. TGENERAL AND PHYSICAL CHEMISTRY. 145effecto. of mineral acids on this reaction, esperimenting with hydro-chloric, nitric, sulphuric, and bromic acids, with mixtures of some ofthese acids, and with potassiuni bromate ; from the tabulated results,he draws the following conclusions. During tohe course of thedifferent changes, the reaction is influenced in the same way by thesecondary actions.The reciprocal values of the times required forthe separation of a determinate quantity of iodine vary as thevelocity of the respective reactions. The velocity of the reaction isaccelerated by hydrochloric acid, but the acceleration is not propor-tional to the quantity of acid present. The quantity of iodinedeposited after equal times in presence of equivalent quantities ofhydrochloric and nitric acids is the same. Mixtures of hydrochloricand nitric acids, in any proportions, are equivalent to either of theacids, so that, independently of the nature of the electro-negativeradicle, it may be said that the velocity of the reaction dependsentirely on the quantity of hydrogen electrolytically dissociated,without, however, being proportional t o it.The action of sulphuricacid is more complicated, on account o€ t.he incomplete dissociation ofthat acid. The :tccelerating effect of bromic acid is almost six timesthat of hydrochloric or nitric acid. S. B. A. A.Velocity of the Halogenisation of Fatty Hydrocarbons. ByIf. WILDERMAN" (BAT., 23, 317&3175).--The following two lawsare deduced from a study of the action of bromine or chlorine insunlight on amyl bromide, amylene bromide, liquid and solid tri-bromopentane, tetrabromopentane, and amylene chloride :-( 1) Sub-stitution proceeds more slowly as the quantity of positrive hydrocarbonbecomes smaller.(2) The larger the quantity of hydrocarbon present,the quicker the substitution.Cryoscopic Investigation of Colloidal Substances. By A.SABAN~EFF ( J . Russ. Chew. SOC., 22, 102--107),-The author hasshown that Raoult's method may be conveniently employed for thedetermination of tbe molecular weight of colloidal substances (Abstr.,1890, 1215). Similar results were obtained by Morris and Brown,by Ekstraud and Mauzelius. 0 1 1 the other hand, Paternb, in hisresearch on gallic and tannic acide, has arrived a t the conclusion thattheir molecular. weights cannot be determined by Rao ult's method.The values obtained by Paternb give 10 mols. of the first, and 109ruols. of the second, as the molecular weights in solution. The authorshows that Paternb's paper includes an error in calculation, and thatthe values greatly depend on the purity of the material.First themolecular weight of gallic acid was determined. It mas dried at 120°,iosing 9.65 per cent. water, corresponding with the formula C,H6OS +H20. In aqueous solution, containing 0.5238 per cent., the depres-sion was 0*06", and the rnolecular weight 166 ; i n acetic acid, con-taining 0.4107 part, the depression was 0*095", molecular weight =168, whereas the value for C7H606 is 170; gallic acid exists, there-fore, as a. airigle molecule in solutions.Commercial tannin was found to contain some gallic acid, thequantity of which, as calculated from the depression of the impureJ. B. T146 ABSTRACTS OF CHEMICAL PAPERS.preparation, compared with that of the pure preparation, was found tobe 2-39 per cent.It was dried for 20 hours at 120', and the loss was10.66 per cent. Aqueous solutions with a concentration of 0-822-3.773shorn a depression of 0*015--0~060, giving a molecular weightof 1044-1195 (mean, 1104). More concentrated solutions, contain-ing 5.5 to 9.5 per cent., gave higher molecular weights, 1497 to 2436,but the values are useless, as solutions of more than 4 per cent. oftannin in water become turbid a t O", tannin separating out. I nPaternb's solutions, containing 11.5-23 per cent. of tannin, $--3 Ofthe tannin must have separated tit 0" in the insoluble state, and h i s(corrected) molecular weights, = 2643-3700, are of no value. Theauthor finds in acetic solution, molecular weights = 1105-1114(mean 1113).Pure tannin was pwpared by Lowe's method, but itwas impossible to work with aqueous solutions, as not more than0-5 per cent. is dissolved in water at O", and even such weak solutionsbecame turbid. I n acetic acid solution, as a mean, M = 1322 wasfound, whereas, M calculated for (C14Hlo09)4 = 1288, so that tanninexists in solutions as a quadruple molecule. The author thinks thatPaternb's tannin was not quite dry, and shows by experiment thatsuch a preparation causes a much larger depression, owing to thecontamination of the glacial acetic acid by the water of crystallisationof the compound. The author says that the empirical formula oftannin requires confirmation by further research.Apparatus for Distillation under Reduced Pressure.By H.W ISLICENUS (Ber., 23,3292-3295).--The author describes two formsof apparatus for use with the the Bunsen pump, to prevent backwarddiffusion. The first consists of a tube, with one round and one pear-shaped bulb; in the depression between these a rubber ring is placed,one end of the tube is sealed, 5 small opening is made in the side andis covered with a piece of rubber tube, or an ordinary Bunsen valvemay be attached; the other end of the tube is connected with thevessel to be exhausted, the pear-shaped bulb fits into a wide tube, therubber ring serving to make the connection air-tight ; the second tubeis joined to the pumpin the usual manner. The second form of valveconsists of two tubes, one closely resembling a thistle funnel, thenarrow end of which is attached to the pump; a bulb with asmall aperture is blown a t the end of the second tube, and i t iscovered with a rubber cap, through which an opening is pierced st alitt.le distance from the one in the glass; the rubber cap serves tomake an air-tight connection between the bulb and the wide end ofthe first tube.An apparatus for fractional distillation under reduced pressure isalso descrihed ; it consists of a combination of several of the first ofthese joints, and provision is made for changing the receiver withoutinterrupting the distillation.Isomorphism.Part 111. By J. W. RETGERS (Zeit. physikaE. Chem.,6,193-236).-1n this communication (for previocs papers see Abstr.,1890,328,1208), the author first discusses the relations of morphotropy.He would limit the term morphotropic to such substances as show aB.B.J. B. TGENERAL AND PHYSICAL CHEMISTRY. 147total analogy of form, and not, for instance, merely analogy of anglesin one zone. Isomorphons substances have not only this total form-analogy, but also analogy of chemical constitution. Morphotropicsubstances are not necessarily cheniically analogous, but must bechemically connected with each other. Substances which show atotal form-analogy, but have no chemical rcsemblance, are termedisogonic. In the regular system, for instance, potassium chloride andrubidium chloride are isomorphous ; potassium chloride and sodiumchloride, morphotropic ; and sodium chloride and sodium chlorate,isogonic.lsomorphous mixtures are proved by the continuous linear changeof physical properties with the composition of thc mixture. Nointimate crystalliue mixture can be obtained with merely morphotropicSubstances, and in this case identity of system, and of degree of hemi-hedry, are not essential. The author rejects Marignttc and Klein’sconception of mas3 isomorphism, according to which an element, o rgroup, largely preponderating in a compound, determines the crystal -line form, no matter what the other components may be. It is thevolume and not the weight of the group that is decisive.An investigation follows of the supposed isomorphism betweenpotassium and sodium sulphates. The author proves that a definitedouble salt, 3K2S04,Na$04, crystallisev from a solution of the mixedsnlphates. It usually crystallises in hexagonal prisms or pyramids,but when the mother liquor contains sodium chloride, i t separates intables-the “ plate-sulphate ” obtained in the manufacture of iodinefrom kelp. Potassium chloride does not effect this change of form.The three simple forms are then-K2S04. Rhornbic, pseudo-hexagonal. Forms pyramids and prismsof hexagonal section. Optically biaxial. Weak birefringence. Sp.gr. = 2.666.3KZSO4,Na,SO4. Hexagonal. Forms pyramids and prisms. Optic-ally uniaxial. Marked birefringence. Sp. gr. = 2.695. Easilyfukible.NhSO,. Rhombic, but not pseudo-hexagonal. Forms onlypyramids with rhombic section. Optically biaxial. Strong birefrin-gence. Sp. gr. 2.673. Fusible with great difficulty.The double salt is not an isomorphons mixture, as is shown by aconsideration of its properties in relation to those of the simple salts.Each simple salt can take up a minute quantity of the other, whichpoints to a very limited isodimorphism.Not easily fusible.The general results are-(1) K&Oa and Na2S04 are not isomorphous.(2) They only form one double salt, 3KzSOa,Na2S04.(3) From mixed solution the pure double salt separates out along-(4) K,SO, and the double salt are morphotropic.(5) NazS04 is not, morphotropic either with K2S04 or with thOdouble salt, but is crystallographically completely independent.The dolomite series is next discussed. In this series we havecalcspnrs containing a little magnesium carbonate, magnesites con-taining a little calcium carbonate, and dolomites which have the twosalts in nearly equal molecular proportions. Calculated from theside one or other of the simple salts148 ABSTRACTS OF OHFJIICAL PAPERS.specific gravities of the component salts, that of dolomite shonld be2.843 ; it is actually 2.872. The author considers the series not iso-morphoua, but merely morphotropic. J. W