ii. 65 General and Physical Chemistry. The Double Refraction of Vanadium Pentoxide Solution. H. DIESSELHORST and H. FREUNDLICH (Physikal. Zeitsch. 1915 16 419425).-The anisotropy exhibited under certain conditions by a colloidal solution of vanadium pentoxide has been further investigated. I f the sol is allowed to flow through a tube the axis of which is parallel to the line connecting crossed Nicols placed at opposite ends of tLhe tube the field of view is illuminated in convergent light and shows the cross and concentric rings which are observed when the light passes through a plate of a uniaxial crystal cut perpendicularly to the axis. By the use of a quarter- wave mica plate it is found that the streaming sol behaved like a positive uniaxial crystal. If a coiicentrated colloidal solution is made to flow through a tube of triangular cross-section which is used as a prism it is found that the red line of hydrogen is resolved into two oppositely polarised lines.The more strongly refracted ray has its vibrations parallel to the direction of flow and therefore parallel t o the major axes of the colloidal particles and in accordance with Babinet’s rule this extraordinary ray is more strongly absorbed than the other. Some of these effects have also been observed although less clearly with an old ferric hydroxide solution. The properties of such sols are obviously similar to those shown by crystalline liquids and in this connexion the general question of the anisotropy of the amorphous state is discussed. H. M. D. The Arc- and Spark-spectrum of Silver in International Normals.JOSEPH FRINGS (Zeitsch. wiss. Photochern. 1915 15 165-182).-Accurate measurements of the wave-lengths of lines i n the arc and speck spectra of silver have been made by the use of a large concave grating. The data are compared with the results of previous observers and the series relations exhibited by certain pairs of lines are discussed. H. M. D. Changeable Lines in the Arc Spectrum of Iron. JOSEPH LANG (Zeitsch. wiss. Photochem. 1915 15 223-252).-According to the measurements of Goos (A. 1912 ii 404 1016) the wave- lengths of certain lines in the arc spectrum of iron vary to an appreciable extent with the length of the arc which is employed as the source of the. radiation. I n view of the importance of this result in connexion with the determination of the wave-lengths of secondary and tertiary normal lines in the iron arc spectrum further experiments have been made in order to ascertain more exactly the relation between the wave-length and the conditions of the arc discharge.Observations were made with short (2-4 mm.) and long (9-16 VOL. CX. ii. 3ii. 66 ABSTRACTS OF CHEMICAL PAPERS. mm.) arc; the strength of the current was varied from 3 t o 15 amperes and the applied potential from GO t o 440 volts. I n the case of the long arc series of ineasurenients were also made with the light emitted in the neiglibourhood of the anode and of the cathode. The data obtained from observations in the greenish-yellow region ( ~ 5 3 6 0 to h5660) show clearly t h a t the wave-length of a large number of lines varies with the length of the arc the strength of the current and according to whether the light is emitted by the anodic or cathodic portion of the arc.These changeable lines are all more or less diffuse and this is more particularly the case for the short arc spectrum when a strong current is passed between the poles. This circumstance suggests t h a t the observed displace- ments are not real and are attribut'able to unsymmetrical broaden- ing of the lines but tlie author maintains that many of the observed changes in wave-length are greater than can be accounted for on this assumption. Although no explanation of the effects can be given it is quite improbable t h a t they are of the same nature as the displacements which are observed when the pressure is varied.The view t h a t they are pressure effects would neces- sitate the assumption of variations of pressure' within the1 arc of such magnitude as to preclude the possibility of this affording a reasonable basis of interpretation. Similar measurements were also made in the ultra-violet region ( ~ 2 4 0 0 to h3330) and variations in wave-length with the con- ditions of the arc discharge of the same kind were found. The results obtained in this region are however n o t nearly so' definite as those f o r the greenish-yellow and the author considers t h a t it is possible t h a t the apparent displacements may be due to un- symmetrical broadening of the lines. I n favour of this view attention is drawn t o the fact t h a t many diffuse lines in the ultra- violet are not displaced a t all whereas all the diffuse lines in the greenish-yellow region are found t o undergo a shift of measurable amount.The fact that all the diffuse iron lines in the greenish-yellow region are variable in wave-length leads to the conclusion that the spectrum of some other element should be employed in the determination of tlie secondary and tertiary normals in this region. I n regard to the region (X2400-/\3330) only certain lines are suitable and a table is given in which unsuitable lines are indi- cated. 13. M. D. Absorption Spectra of Aqueous Solutions of Poly- and Simple Hydrated Salts by Means of a Radiometer. E J. SCHAEFFER M. G. PAULUS and HARRY C. JONES ( Z e i t s c h . physl'kn7. Chenz. 1915 90 561-593).-The influence of a number of salts on the' int'ensity of the infra-red absorption bands of water has been yarefully studied and measurements of the intensity of the bands have been made by means of a radiomicrometer used in connexion with a Hilger spectrometer.The saltc; used were1 such as possessed no absorption of their own and included potassiumGENERAL AND PHYSICAL CHEMISTRY. ii. 67 chloride sodium chloride iiiagiiesiuin chloride magnesiuln bromide magnesium sulphate zinc sulphate calcium chloride zinc nitrate magnesium nitrate ainmoiiiuni bromide a i d sodium nitrate. The results are expressed by means of a large number of tables arid curves. It is showri that lion-hydrated salts which have no absorp- tion of their owii cause aii increase in the intensity or' the middle of the water bands.It is suggested that this may be due to a slight shift of the baiids toward the red end of the spectrunl. Hydrated salts cause a decrease in tlie absorption of tlie middle of the water baiids. 'To this observation nitrates form an excep- tion. I n general it can be stated that all aqueous solutions of noii-hydra ted and noii-absorbing salts are inore transparent than water f o r all the wave-lengths made use of in the present investigation ( A = 706-1426). This is particularly the case for the band 1 . 2 5 ~ . Magnesiunl nitrate presents a number of irregularities which could be due conceivably to an absorption by tlie salt. To settle this point the absorption of magnesium nitrate in both acetone and .ethyl alcohol was measured with tlie result t h a t this salt is shown t o have no absorptioii a t any point.Tlie general result of the work is to show t h a t combined water has a smaller absorptive power than free water and this is regarded as an additional piece of evidence in favour of the solvate theory. Absorption Spectra of a Series of Aniline Dyes and the Separation of Certain Portions of the Spectrum by Means of Gelatin Filters. ADOLF HNATEK (Zeitsclz. zuiss. Photochenz. 1915 15 133-148).-The absorption spectra of some sixty aniline dyes have been examined and of these twenty-four have been found suitable for use in the preparatjoii of gelatinised filters. The filters are arranged into three groups for which the range of transmitted light is ahout I000 A 5 0 O 4 0 0 A and 200-500 A respectively.In most cases the filters contain a single dye but in others tlie absorptioii is produced by suitable admixture of two cr inore dyes. Further reduction in the widtli of the transmitted bands can be obtained by contbination of two or more filters. 3. F. S. H. M. D. Colorations Produced by Some Organic Nitro-compounds with Special Reference to Tetranitomethane. 11. ALEXANDER KILLEN MACBETH (T. 1915 107 1824-1827. Compare T. 1915 107 87).-Some points of interest which arose in tlie earlier study on tlie coloratioiis produced by treating tetranitromethane with uiisaturatetl substances are discussed. 111 the first place a rough examination suggested that the deptli of colour oktaiiied might he ail iiidicatioii of tlie relztivef reactivity or of the amount of residual affinity of a snbstance.A quantita- tive spectrographic study of the colours produced by tetrahydro- I 4-thiopyran 1 4-dithian and 1 4-thioxan has shown t h a t the reactivities of these compouiids are of the same order a;> Clarke determined in the case of their combination with bromoaceto- phenone (T. 1912 101 1788). 3-2ii. 68 ABSTRACTS OF CHEMICAT PAPERS. The existence of inhibitory factors is also revealed. The pre- sense of electronegative atoms (such as Cl) in ethylenic com- pounds for example is one factor and no colours are obtained also in the case of unsaturated compounds containing conjugated double linkings (ethylene with carbonyl or nitrogen with carbonyl) except in the aromatic series which exhibits many anomalies. The effect of the nature of the solvent medium has alyo been examined.Light petroleum and chloroform appear t o hinder the development of colour but water accelerates it. It seems that certain solvents have the power of preventing the ‘‘nitrite isomer- isation ” of the tetranitromethane. J. C. W. Phosphorescence. K. SCHERINGA (Chens. TVeekbZad 19 15 12 1106-1109).-An account of the close relationship between phos- phorescencel and fluorescence. A. J. W. The Photobromination of Toluene. K. ASDRICII and M. LE BLANC (Zeitsch. wiss. Photochem. 1916 15 148-164 183-223).-The bromination of toluene has been examined with special reference to the influence of light of varying wave-length on the yield of benzyl bromide. I n a preliminary series of experi- ments the influence of various solvents on the absorption spec- trum of bromine was examined.The curve of absorption for a solution of bromine in carbon tetrachloride shows a maximum in the blue (400-450pp) and a minimum in the neighbourhood of A =313 pp. Similar selective absorption is also exhibited by solutions of bromine in hexane. On the other hand the curves for ethyl acetate benzene and toluene solutions show gradually increasing absorption with diminishing wave-length. The absorption spectra of solutions in hexane and carbon tetrachloride resemble that of bromine vapour and the authors draw the conclusion that the bromine in these solutions is largely. present in the free condition whereas in ethyl acetate benzene and toluene solutions it is present in the form of complex molecules resulting from the combination of the bromine with the solvent.Comparative bromination experiments in presence of hexane and ethyl acetate indicate that the free bromine molecules are photo- sensitive whereas thei solvated molecules are not. This relation affords an explanation of the fact that the photo-sensitiveness of the reaction between bromine and toluene diminishes rapidly as the wavelength of the incident radiation decreases. The yield of benzyl chloride is found to depend on the presence of oxygen in the reaction mixture the yield increasing with the amount of oxygen. This has been traced to the oxidation of the hydrogen bromide whereby bromine is regenerated. Phenolic substances are also formed in these circumstances and there is some evidence of the formation of intermediate peroxides.Both toluene and hexane which have been brominated in presence of pure oxygen exhibit oxidising properties. The yield of benzyl bromide is nearly independent of the wave-length of the incidentGENERAL AND PHYSICAL CHEMISTRY. ii. 69 light between h 579 and A 325 pp; the reaction velocity decreases rapidly below 40Opp and is not affected by light of wave-length ~300-1202. The addition of water t o the reaction mixture reduces the velocity considerably in presence of pure oxygen but the effect is small in presence of air. Some convenient light filters and an arrangement for producing spark discharges of high intensity are described. Reference is also made to the action of anhydrous phosphoric acid on potassium bromide as affording a simple method for the production of hydrogen bromide.H. M. D. Attem ptcJ made t o Influence the Velocity of Radioactive Transformations by Means of a-Rays. JEAN DANYSZ and LOUIS WERTENSTEIN (Compt. rend. 1915 161 784-787).-The authors have endeavoured to increase the velocity of transformation of uranium-Z t'g uranium-X and of mesothorium-l t o mesotliorium-11 by means of a-rays but without success. I n the first case a tube containing a t the commencement 18 millicuries of emanation was allowed t o act f o r six days on 3 layer of ura.nium oxide TT308 just sufficiently thick t o absorb the a-rays. I n the second case an ampoule. containing 30 millicuries of emanation was allowed t o act for thirteen hours on 3 mg. of mesothorium-radium con- tained in a capsule covered with mica 0.005 mm.thick. I n neither case could the slightest i r crewe in the velocity of transformation be detected and the authors consider from a calculation of the1 possible collisions that even the most intimate contacts of the a-rays with the atoms are unable t o provoke an artificial radio- activity o r a premature transformation of the atom struck. W. G. The Charge of Radioactive Recoil. LOUIS WERTENSTEIN (Compt. rend. 1915 161 696-699).-The author has studied the charge of recoil in the case of radium-D projected during the transforma- tion of radium-C. Measurements were made of the charge received by a cylinder receptor the source being a deposit of radium-C (1) when the rays had no obstacle t o traverse; (2) when they were filtered through a screen of aluminium foil 0.7 p thick; (3) when they had to traverse a sheet of aluminium foil 0.05 mm.thick which arrested all the a-rays. The whole apparatus was contained in a vessel having a high vacuum maintained by a Gaede molecular pump giving 0.6 dyne/cm.2 as measured on a Knudsen's absolute manometer. All the three measurements were made in one experi- ment the screens being moved by an electromagnetic control out- side. The difference between the values from (1) and (2) gave the charge of the a-rays and that between (2) and (3) the charge of the a-rays. In the high vacuum 0.6 dyne/cm.2 the atoms of radium-D when projected do not carry an electric charge but as the pressure rises they acquire a gradually increasing positive charge which may equal or even exceed that of the a-rays. This positive charge is a consequence of the collisions of the atoms of radium-D with the molecules of the surrounding gas one or more electrons beingil. 70 ABSTRACTS OF CHEMICAL PAPERS.sq.arated from the radioactive atom. The pressure under which the recoil attains a charge equal to half of that of the a-rays corresponds with a free path of the atoms projected equal t o tile distance of the source from the receptor. The experimental results give the value of the iiieaii free path a t a pressure of 1 dyne/cm.2 as 100 nim. whilst tlie value calculated from the ionisation pro- d u c d is 90 nim. The free path of a particle of radium-C calcu- lated in the same way f o r the initial portion of its path is 440 mm.and the n-rays ioiiise a t their commencement five times better than the a-particles. A t the atmospheric pressure the atoms of recoil behave as univalent ions and the dissociatioiis of the radioactive atoms resulting from the collisions with the molecules inust therefore be followed by recombiiiations as the velocity of the atoms diminish this behaviour showing the close analogy existing between the n-rays and the canal rays. W. G. Influence of Foreign Substances on the Adsorption of Uranium-X by Charcoal. H. FRBXJNDLICH and H. KAEMPFKE (Zeitsclh. p h y s i l d . CJietu. 1915 90 681-’716).-A large number of experiments have been carried out 011 tlie influence of thorium salts and many other substances on the reversal of the adsorption of uranium-X1 by charcoal.These experiments are in part a repeti- tion of those of Ritzel (A. 1909 ii 851) and partly an extension of Ritzel’s work. It is shown t h a t the presence of small concen- trations of thorium salts causes a reversal of tlie adsorption of uranium-X1 by clmrcoai. Thus the presence of 0.0004 millimol. of thorium nitrate increases tlie amount of uranium-X which remains in solution after shaking with charcoal by about 20%. The action however is not specific f o r thorium or the radioelements for a large number of substances such as zirconium salts benzoic acid strychnine nitrate and basic dyes have a similar action. A number of points of difference between the action of thorium iiitza te and the other substances was noticed. The adsorption of benzoic acid strychnine nitrate and zirconium oxychloride was measured by means of th% radioactivity of uranium-X,.The method adopted consisted in determining the end concentration iii solutions of these substances of the uraiiium-LX which had been shaken together with these substances and charcoal and from the amount of reversal of the adsorption of the uranium-X the adsorption of the other substance was deduced. In the case of benzoic acid and zirconium oxychloride the usual adsorption isotherms were obtained by this method but in the case of strych- nine nitrate there was a tendency to reach a saturation value. In all three cases an elid condition was reached which was independent of time. When an attempt was made t o determine the ailsorption of thorium nitrate by the same niethod a difference was noticed in the beliaviour from that observed with the other substances in the sense that the end concentration appeared t o decrease with time.This decrease could not be confirmed by iiephelometric measurements on the oxalate and was explained by the fact thatGENERAL AND PHYSICAL CHEMISTRY ii. 71 a solution of thorium nitrate wliicli has been shaken with charcoal afterwards produces a niuch smaller reversal of the adsorption of uranium-S,. This corresponds with the observation of Ritzel which was confirmed that freshly produced uranium-X in uranium nitrate solution which has been shaken with charcoal is more strongly adsorbed by charcoal than uranium-Xl of long standing. The authors explain the above phenomena by the assumption that in thorium and uraiiyl nitrate solutions small quantities of still iindiscovered radio-elements exist) and t h a t these exert a sirong reversing action on the adsorption of uranium-S but t h a t they are completely adsorbed by charcoal when shaken with it.Thorium nitrate differs from the other substances which reverse the adsorption of uranium-S in the fact t h a t it not only diminishes the adsorption of uranium-S when it is adsorbed a t the same time as tlie uraniurii-S but it also causes a reversal of the adsorption if i t is added to the charcoal after the uranium-S has been adsorbed. Tlie latter effect is not observed with the other sub- stances. The difference is probably to be explained in the follow- ing way uraniurn-S is not only concentrated on the surface of tlic cliarcoal but also penetrates into tlie interior; the same is ])robably also true for thorium nitrate mid the unknowii sub- Ftances mentioned above.These substances can therefore exercise their reversing action on tlie nraiiium-,Y when they are added after it has been adsorbed whereas the other substances such as btliizoic acid by adsorpticn are only colicelitrated on the surface and cannot therefore exercise this action. The foregoing observa- tions lead to the conclusion that- the statement of Soddy (T. 1911 99 72) to the effect that the reversal of the adsorption of uranium-3 by thorium is due t o the isotropy of the two sub- stances is not in accordance with experimental facts. On the other hand the fact t h a t other substances have the same action as thorium does not disprove in any way the isotropy of these two substances.The coefficient of the absorption factor of the adsorp- tion isothermal for both thorium nitrate and uranium-3 was found t o be very nearly unity. The reversal of the adsorption of uraiiium-91 by other foreign substances does not depend wholly on the adsorption of these substances and is not purely a decreas- ing of the power of the charcoal t o adsorb uranium-X but rather a hindrance by the foreign substance to the penetration of uranium-X into the interior of the charcoal much in the same way t h a t adsorbed substances hinder the crystallisation of super- saturated solutions. Ritzel (Zoc. c i t . ) found that thorium emana- tion also caused a reversal of the adsorption of uranium-X,; this the authors find does not occur if the emanation is filtered through wool before it is led into the solution.The reversal observed by Ritzel is therefore t o be attributed to dust of thorium nitrate carried over by the emanation. J. F. S. Transmutation of Chemical Elements. IT. W. P. JORISSEN and J. A. VOLLGRAFF (Zeitsch. physikc11. CI~enz. 1915 90 557-560. Compare A. 1915 ii 134).-The authors have sub-ii 72 ABSTRACTS OF CHEMICAL PAPERS. jected bismuth to a stream of cathode rays with the object of ascertaining whether by this action bismuth beconies radioactive o r is converted into thallium. The experiments show that bis- muth which has been treated in this way has no action on the leaves of a charged electrc4scope and has also no action on a photographic plate.On testing the metal for thallium it is shown that in the untreated metal there is a small quantity of thallium which although not sufficient t o give the usual micro- chemical reactions can be detected spectroscopically. The amount of thallium is not increased by the treatment with cathode rays. J. F. S. Dielectric Constants of Some Compounds of Vanadium. ALBERT G. LOOMIS and HERMAN SCHLUNDT ( J . Physical Chem. 1915 19 734-738).-The dielectric constants of vanadium oxy- chloride vanadium oxybromide and vanadium tetrachloride have been determined by the Drude-Schmidt method a t a series of temperatures. The compounds examined were prepared by the Roscoe methods and full details of the preparation are given in the paper. The following values were obtained for the dielectric constant vanadium oxychloride 3.42 a t 21° 3.38 a t -7OO; vanadium oxybromide 3.62 a t 25O 3.93 a t -2O and 4.38 a t - 70°; vanadium tetrachloride 3.05 a t 25O.J. F. S. Electrical Conductivity and Luminosity of Flames Contain- ing Salt Vapours. H. A. WILSON (Phil. Truns. 1915 [ A ] 216 63-90).-The author has studied the influence of an electric field on the velocity of positive ions of salt vapours in flames together with the relationship of electric conductivity of flames contain'ing salt vapours to their luminosity. Further experiments have been made to show the variation of the conductivity with the con- centration of the salt vapour and to determine the relative con- ductivity due to salts of different metals. From the experimental results it is shown that the luminous vapours of salts in a Bunsen flame are not deflected appreciably by an electric field and that the positive ions present in the luminous vapour can be made to move out of it by an electric field.The positive ions are not luminous but can form luminous vapour after recombination. The velocity of the positive salt ions in flames is about 1 cm. per second for 1 volt per cm. and is the same for all salts. The con- ductivity and luminosity of sodium chloride both vary nearly as the square root of the concentration of the salt vapour in the flame. The conductivities and luminosities of chemically equiva- lent amounts of sodium chloride and sodium carbonate are equal. A considerable excess of hydrogen chloride does not change the luminosity due to sodium chloride but it slightly increases the conductivity of the flame.A large excess of hydrogen chloride obtained by the introduction of chloroform vapour diminishes the luminosity due to sodium chloride by about SO% and diminishes tihe conductivity about six times. The percentage changes in the liiminasity and Conductivity are independent of the concentrationGENERAL AND PHYSICAL CHEMISTRY. ii. 73 of the sodium chloride vapour. The chloroform increases the current due to large potential differences but diminishes the con- ductivity as measured by the ratio of the current t o the uniform potential gradient in the flame. When the drop of potential a t the negative electrode is got rid of by placing potassium carbonate on it the chloroform diminishes the current to about the same extent as the conductivity.The presence of a large amount of potassium carbonate does not change the luminosity due to sodium carbonate although the potassium carbonate greatly increases the total conductivity. The variation of the conductivity (c) with the concentration (k) f o r alkali salts can be represented by the equa- tion klO4= (c2- l ) / ~ . ( 6 + ac) using proper values for the constants n and b for each salt. This expression can be deduced from the ionic theory and the fraction of the salt molecules which are ionised is equal to b / ( b + ac). Alkali chlorides and carbonates impart equal conductivities to the flame for chemically equivalent concentrations. The percentage of the salt which is ionised when the concentration is small has the following values c s i u m salts 91%; rubidium salts 69%; potassium salts 58%; and sodium salts 1.6%.The constant b is proportional tp the molecular weight of the salt and according to the theory this shows that all salts give negative ions having the same velocity due to an electric field. The conductivity due to solutions containing a mixture of salts agrees with that to be expected on the ionic theory. J. F. S. The Influence of Some Hydroxy-acids on the Electrical Conductivity of Boric Acid. J. BOESEKEN [with R. DE BRAUW s. DE ~ ' A A R D and c. VAN LOON] (Rec. truu. chtim. 1915 35 211-224. Compare A. 1915 ii 136 667).-Measurements were made of the electrical conductivity of boric acid solutions con- taining varying molecular proportions of 8-hydroxybutyric acid a-hydroxyisobutyric acid benzilic acid and a-hydroxydiphenylene- acetic acid.The three last named all give a marked increase in the conductivity the results showing that the amount of the increase is not directly connected with the dissociation constant of the acid. Apparently the group adjacent to the a-hydroxy-group is of much greater importance in this connexion. With P-hydroxy- butyric acid there was a marked diminution in the conductivity the difference being however less than in the case of the other three acids. W. G. Conductivity and Dissociation of Some rather Unusual Salts in Aqueous Solution. CHARLES WATKINS and HARRY C. JONES (J. Amer. Chem. Soc. 1915 37 2626-2636).-The elec- trical conductivities of solutions of sodium bromate thiocyanate thiosulphate dithionate pyrophosphate tungstate f ormate chromate dichromate dihydrogen phosphate and phosphate ammonium iodate thiocyanate chromate and dihydrogen phos- phate rubidium iodide lithium chromate and potassium f erri- cyanide have been measured a t various concentrations and a t Oo 15* 25* and 3 5 O .From these data temperature-coefficients are 3*ii. 74 ABSTRACTS O F CHEMICAL PAPERS. calculated and the influence of t,emperat,ure on the ionisation is discussed. H. M. I). Electrical Conductivity of Certain Salts in Pyridine. EDWARD X. ANDERSON ( J . Physicrrl Chem. 1915 19 753-789).-The electrical conductivity of pyridine solutions of silver nitrate lithium chloride bromide and iodide sodium iodide potassium thiocyanate ammonium thiocyanate mercuric chloride bromide and iodide copper chloride and nitrate cadmium nitrate cobalt chloride and lead chloride have been determined a t Oo 25O and 50° for a series of concentrations in each case.These substances constitute two series of electrolytes the one consisting of salts the equivalent conductivity of which increases with increasing dilu- tion and the other consisting of salts the conductivity of which drops to a minimum. The temperature-coefficients have been calculated and it is shown that the coefficieiits between Oo and 25O are larger than those between 25O and 50°. Solutions of lithium bromide sodium iodide cobalt chloride and cadmium nitrate have negative temperature-coefficients. Explanations are given for the minimum conductivity and f o r the differences in the temperature-coefficients.It has been shown that the anomalous behaviour in equivalent conductivity is due entirely to the presence and properties of tlie ionisable polymerised solute which pre- dominates in concentrated solutions. J. F. S. The Conductivity and Viscosity of Solutions in Formamide. P. B. DAVIS W. S. PUTNAM and HARRY C. JONES (J. Frankliir Znst. 1915 180 567-601).-1n a further investigation of the properties of non-aqueous solutions of electrolytes measurements have been made of tlie electric conductivity and viscosity a t 15O 25O and 35O of salts dissolved in formamide. By repeated distilla- tion of formamide dried over anhydrous sodium sulpliate under reduced pressure in a current of dry air the1 conductivity of the solvent has been reduced t o a very much smaller value than that previously recorded by Walden (compare A.1912 ii 2G). The lowest value observed was K=2.8 x 10-6 mho. a t 25O. The salts for which data are recorded are sodium bromide iodide and chromate potassium chloride iodide and thiocyanate ammonium bromide and iodide tetramethylammonium chloride and iodide rubidium chloride bromide iodide and nitrate caesium chloride and nitrate lithium nitrate barium chloride mercuric chloride and cobalt bromide. The results afford further support for the view that the ionising power of a solvent increases with the dielectric constant and the association factor. Salts which form hydrates in aqueous solution afford evidence of combination with formamide. The solutions of these solvated salts show a larger percentage change in conductivity with change in temperature as compared with solutions of non- solvated salts.The percentage temperature-coefficient is found to be approximately proportional to the viscosity of the solutions. Formamide is the only solvent for which tlie viscosity of theGENERAT AND PHYSICAL C'HEMISTRY. ii. 75 solutioiis exainitied is less tltan tliat of tlie pure solvent. This is supposed t o be connected witli the large value of the association factor. Mercuric chloride is ioiiisecl to a greater exteiit in formamide than in water. H. RiI. D. Hydrates of Alcohols and Fatty Acids. Ballistic Electric Investigations. [Mrss] JOHAXNE CHRISTIANSEN (Ze itsch. physikal. Ckem. 1915 90 628-640).-Christiansen has shown (;1 I ? ? ? .PJrysiE 1913 [iv] 40 107) t h a t when aqueous solutions of salts. to which small quantities of non-electrolytes such as alcohol have been added are sprayed agaiiist a plate tliey give up considerable quantities of electricity. This electricity is termed 7)ollo-plpcfi.icity. The present paper deals witli the potential of the electricity given up when solutions o f calcium chloride or sodium chloride in water containing varying quantities of alcohols or fatty acids are sprayed against a platinum plate coniiected to a sensitive quadrant electro- meter. Measurements were made with methyl alcohol ethyl alcohol propyl alcohol butyl alcohol and amyl alcohol formic acid acetic acid propionic acid and butyric acid. I n the case of all the substances examined it is shown t h a t the voltage at first increases t o a maximum and then falls again to zero with increase in the percentage of the noii-electrolyte added.The voltage then changes sign and slowly increases numerically. The point of zero voltage is termed the ballo-electric neutral point. I n the case of the first three alcohols the mixtures a t which this point is reached are 1 mol. of methyl alcohol to 1.95 mols. of water 1 mdl. of ethyl alcohol t o 3.98 mols. of water and 1 mol. of propyl alcohol t o 8.38 mols. of water that is tlie number of moleculds of water have approximately the relationship 2 2? 23. In the case of tlie acids the ratio of tlie number of molecules of water is 3 :3?:33 etc. f o r tlie ascending members of the series. The action is held to be due t o a mechanical separation of neutral molecules into ions.J. F. S. A Hydrogen Electrode Vessel. WM. ~IANSFIELD CLARK (J. Bid. C ~ L A ~ . 1915 23 475-486).-After a description of a number of vessels wliicli liave been used for the cletermiiiation of hydrogen ioii potentials the author describes a piece of apparatus which is based on the principle of tliat devised by Hasselbach (A. 1911 ii 182). The object of the modification is to enable measure- ments of hydrogen ion concentration t o be made in bacterial and other liquids which are undergoing fairly rapid change. The apparatus consists of a length of fairly wide glass tube to one end of which a small bulb is attached which is fitted witli a neck a t the top t o take the platinum electrode and with a three-way tap immediately below the neck f o r the purpose of connecting witli the calomel electrode.The other eiicl of the tube is fitted with a three-way tap which is connected t o a siiiall reservoir containing the liquid under examination a i d to a hydrogen generator. The 3"-2ii. 76 ABSTRACTS OF CHEMICAL PAPERS. apparatus is used as follows thO apparatus containing only the platinum electrode is filled with hydrogen and then the solution is added until it is about one-third full. Then the apparatus which is mounted on a rocker is rapidly rocked for five to ten minutes so tha€ the electrode is alternately exposed to the gas and immersed in the solution. The measurement is then made and is found to be at once constant to about 1 millivolt. I n many cases a much more constant result is obtained.A series of measure- ments with milk gelatin cultures meat infusion and peptone solu- tion are given to illustrate the type and constancy of the results. J. F. X. The Anodic Solution of Lead. Nomm MURRAY BELL (Trans. Faraday SOC. 1915 11 79-90).-Experiments have been made to test the supposition that univalent lead ions exist in aqueous solu- tions (Denham and Allmand T. 1908 93 424). The anodic sclution of lead is determined by measurements in solutions of lead acetate sodium acetate potassium hydrogen tartrate hydro- fluosilicic acid potassium bromide and sodium thiosulphate a t various current densities. Bars of cleaned “pure” lead are used. The loss of weight of the anode is generally greater than is indicated for bivalent ions.With the above solutions except the two sodium salts the loss is shown to be due to mechanical and chemical removal of lead but sodium acetate and thiosulphate give results which appear to indicate the formation of univalent lead ions t o the extent of 2.8% and 5.6% respectively. The chemical character of the black deposit formed on lead anodes in thiosulphate solutions has not been determined. C. H. D. Electrical Transference in Amalgams. GILBERr N. LEWIS ELLIOT Q. ADAMS and EDITH H. LAMMAN ( J . Amer. Chem. SOC. 1915 37 2656-2662).-When sodium is dissolved in liquid ammonia a conducting solution is obtained and according t o Kraus the carriers of positive electricity in this solution are the sodium ions whilst the negative carriers are free electrons.I n the expectation that a solution of a strongly electro-positive metal in mercury would show similar behaviour on the passage of a current experiments have been made t o determine the effect of the current on dilute sodium and potassium amalgams. Sodium and potassium amalgams prepared by electrolysis were subjected to the action of a current of from four t o eight amperes for several days under conditions which eliminated as far as possible the effect of convection and the alkali in the anode and cathode compartments was then estimated. The results of these experiments show that the passage of the current through the amalgam was accompanied by the transfer of alkali metal from the cathode to the anode. The number of equivalents transferred per faraday was found to be 2.9 x 10-6 in a sodium amalgam contain- ing 3.24 atoms per cent.of sodium and 0.29 x 10-6 in an amalgam containing 0.57 atom per cent. I n a potassium amalgam contain-GENERAL AND PHYSICAL CHEMISTRY. ii. 77 ing 2.16 atoms per cent. of the alkali metal the transfer per faraday was 3-6 x 10-6 equivalent. The diminution in the electrical conductivity resulting from the addition of the alkali metal is attributed to a diminution in the average mobility of the electrons. I f the atoms of sodium are supposed to form nuclei of large aggregates of mercury atoms which aggregates are less easily penetrated by the electrons than are free mercury atoms the negative transfer of the sodium can be explained. The fact that potassium has a greater transference number is on this view consistent with the fact that potassium produces a greater lowering of the conductivity of mercury as com- pared with that produced by an equivalent quantity of sodium.H. M. D. Application of the Theory of Allotropy t o Electromotive Equilibria. 11. Passivity of Iron. A. SMITS and A. H. W. ATEN (Zeitsch. physikal. Chem. 1915,90 723-743).-A theoretical paper in which anodic polarisation and the passivity of iron as brought about either by chemical means or by anodic polarisation are discussed. A number of theories are discussed and i t is shown that the primary manifestation of passivity which consists in the fact that a t moderately low current densities the potential differ- ence between iron and electrolyte increases so that the liberation potential of oxygen is rapidly reached is not in accord with the oxide theory and cannot be explained by this theory.The charg- ing of iron with oxygen and the resulting oxide formation are secondary manifestations of passivity. It is emphasised that passivity has its situation on the surface of the metal and i t is shown that the appearance of passivity due to all causes the change from the passive form t o the active form and abnormal cathodic polarisation can all be explained from a common point of view on the basis of the theory of allotropy. Polarisation occurs not because the heterogeneous equilibrium between the metal and the solution is set up too slowly but because the metal is com- pelled to dissolve more rapidly than the inner equilibrium is abls to set itself up whereby the metal surface becomes more noble.A strong experimental confirmation of the new method of considera- tion is found in the results of Finkelstein (A. 1902 ii 81) who showed that iron in a solution of ferric sulphate has a more positive potential than iron in a solution of ferrous sulphate of the same concentration. The formula A = - RT/nF log,K/C is not obeyed and the formula A = - R T / F log,(R . C / [ / / C / / ) may not be applied because iron in a solution which contains ferric ions cannot be in equilibrium and cannot act as non-soluble electrode. The iron is strongly attacked by such a solution and the above-mentioned phenomena are the result of a disturbance of the internal equil- ibrium of the metal surface. It is also shown that the activation curve points to a discontinuity in the mixture series of the pseudo- components.The abnormal cathodic polarisation and the con- nexion between the allotropy of iron and its passivity are dis- cussed in connexion with the theory of allotropy. J. F. 8.ii. 78 ABSTRACTS OF CHEMICAL PAPERS. The Thermomagnetic Properties of Various Compounds and the Weiss Theory of Magnetons. K~TARG HONDA and TORAJIR~ ISHIWARA (Sci. Rep. Tohokzc Imp. Univ. 1915 [ii] 4 215-260). -The experiments of Ishiwara a t lcw temperatures (A. 1915 ii 141) have n3w been extended t o higher temperatures. The susceptibility of - paramagnetic ccmpourids is not derivable from t h a t of their ccinpcrents by the additive law. Curie’s law is only applicable t o a very limited number of ccmpounds but a hyper- bolic law such as that given by K.Oiines and Perrier is generally applicable over tlie whole range of temperature within which the same phase is present. Either an abrupt departure from this law or a gradual one extending over a range of temperature indicates a phase change. The susceptibility of all the diamag- netic substances examined is constant for all temperatures within the range of existence of a single nhase. The effect of water of crystallisation is different in different compounds. Phase changes are readily detected by means of t h e m omagnetic obyervations. Weiss’s theory of magnetons is criticised the evidence on which it is based being considered insufficient and the magnetic moments of elements and compounds are calculated from the data in this paper and this vol.ii 105 leading t o the conclusion that the oxygen atom has a greater magnetic moment than that of iron whilst niolecules of tlie compounds of iron nickel and cobalt have magnetic momelits wliich are f a r greater thaii those of the ferro- magnetic elements themselves. An Electromagnetic Vacuum Balance. JOHN 8. RNDEHSC~ (Trans. Faraday SOC. 1915 11 69-75).-The balance is intended for repeated weigliings of silica gels and similar substances with- out removing from the vacuum. One af the scale-pans of a balance is replaced by a coil of wire with its axis vertical. This moves between two larger coils conc3ntric with it so connected that the movable coil may be attracted or repslled. The force is varied by varying the current.The balance is enclosed in a bell-jar which may be evacuated. The beam is released and fixed magnetically by means of an iron core and two solenoids. An electrical arrange- ment with slide resistances allows the current passing through the weighing coils t o be measured. This is calibrated by means of known weights. C. H. D. Relation between the Thermal Conductivity and the Viscosity of Gases with Reference to Molecular Complexity. J. A. POLLOCK ( J . Roy. SOC. N e w Soicth TT’cc7es 1915 49 249-252).- The relation between the thermal ccnductivity I; the viscosity ?I and the specific heat cv of a gas is expressed by the equation k=fqcV in which the coefficient .f has the same value for gases which contain the same number af atcJms in the mclecule. By the use of Eucken’s data (A.1913 ii 474) it is shown t h a t the con- nexion between f and the specific heat ratio y=cC?,/cv may be writhen in the1 form f = 7 * 3 2 ( y - l ) / - y 1 . 3 . It is also shown that rn y k / 11 is approxiniately constant n/ denoting the molecular weight of the gas. C. H. D. H. M. D.GENERAL AND PHYSICAL CHEMISTRY. ii. 79 The Thermal Conductivity of Neon. ERICH BANNAWITZ ( A m . Physik 1915 [iv] 48 577-592).-From observations made according t o Schleiermacher's method the thermal conductivity of neon a t Oo has been found t o be 1~,=0*0001091 gr./cal. sec./cm. The temperature-coefficient of the thermal conductivity is 0.00259 which agrees closely with the values for air (0.00253) and argon (0*00260) but is smaller than the value f o r helium (0*00318).According t c Chapman (I'lid. Trans. 1911 [ A ] 211 433) the1 thermal conductivity I; the coefficient of viscosity 11 and the specific heat' a t constant volume cL. are in the case of monatomic gases connected by the equation K / 11 . c,.= 2.5. By substituting in this equation the' values 11 =3*036 x 10-4 (at 10*lo) c,,,=O*1474 /~=0~00011195 ( a t l0.lo) the value of l;,/qcL is found t o be 2.501. H. M. D. Determination of the Velocity of Sound and of the Ratio of the Specific Heats of Gases by the Method of Rundt's Dust Figuree. G u s ~ a v SCHWEIKERT ( A m . Physik 1915 [iv] 48 593-667).-Measurements have been made of the velocity of sound in various gases a t the ordinary temperature using Kundt's method in combination with tubes of varying diameter and notes of different pitch.Data are recorded for air nitrogen oxygen hydrogen carbon monoxide hydrogen clilor ide carbon dioxide nitrous oxide sulphur dioxide ammonia ethylene and acetylene. The data are applied to the derivation of 'the specific heat ratio k = c y / c and of the molecular heats. The following recorded values of the molecular heats at constant pressure c j a t 20° show that for diatomic and still more for triatomic gases the molecular heat increases with the density carbon monoxide 6.900 ; nitrogen 6.905 ; oxygen 6.924 ; hydrogen chlcride 7.046 ; carbon dioxide 8.904 ; nitrous oxide 9.214 ; sulphur dioxide 10.059 ; ammonia 8.933 ; ethylene 9.773 ; acetylene 9.783. These values are com- pared with the results obtained in direct measurements of the specific heats of the various gases.H. M. D. The Speciflc Heat of Platinum and of Diamond at High Temperatures. A. MAGNUS ( A m . Physik 1916 [iv] 48 983-1004).-By means of a modified form of the calorimeter described in a previous paper (A.? 1913 ii 103) measurements have been made of the specific heat of platinum and its variation with temperatures up t o about 900°. The lieat (Q) liberated by 1 grain of platinum in cooling from t o to f,' can be expressed by the equation Q =0*031590(t - to) + 0.0%9234(t2 - to2)) and from this the specific heat a t any temperature t is given by c1,=0*031590 + 0-0558468t. I n Weber's well-known measurements of the specific heat of diamond and graphite at high temperatures the temperature of the heated substance was determined indirectly by measuring the heat' liberated by platinum which had been heated t o the same temperature.I n view of the greater accuracy of the new measure- ments of the specific heat of platinum Weber's data have beenii. 80 ABSTRACTS OF CHEMICAL PAPERS. made use of in a recalculation of the specific heat of diamond and of graphite. F o r diamond c p = 0.31456 + 0*033824(t - 282) - 0*063168(t - 282)z and for graphite c,=0*28815 + 0*0346348(t - 225.3) - 0*063642(t - 225-3)z. The equation for the specific heat of diamond is in very good agreement with Weber’s data over the entire range of temperature but in the case of graphite there are discrepancies a t temperatures below 500O. A New Method for Determining the Specific Heat of Liquids. E. J.HARTUNG (Trans. Faraday Xoc. 1915 11,64-68).- The method requires the use of about 60 C.C. of the liquid and gives a high accuracy. It consists in introducing a known weight of ice enclosed in a bulb and noting the fall of temperature pro- duced. A small cylindrical glass bulb contains 1 t o 3 C.C. of water and as large a roll of silver gauze as possible. This bulb is sealed and will sink in the liquid. A glass tube 15 cm. long and 4 cm. diameter contains 100 C.C. of mercury and is provided with a rubber stopper holding the thermometer and having a hole large enough t o admit the ice carrier. Dry air may be drawn through the apparatus. The calorimeter is a thin copper cylinder silvered internally and supported in a Dewar vessel. The mouth is closed by a rubber stopper t.hrough which a thin glass stirrer passes.After placing 60 C.C. of the liquid in the calorimeter and the weight having become constant the ice carrier is secured under the mercury in the freezing tube and dry air is drawn through t o prevent condensation of water. The tube is cooled to -5O and kept a t about -lo after freezing has taken place. The tempera- ture being noted the bulb is introduced into the calorimeter which is stirred until the temperature reaches a minimum. The usual corrections are made. The values obtained for mixtures of sulphuric acid and water are in good agreement with those of Thomsen. C. H. D. A Simple Method of Obtaining Melting Points of Fats etc. ARTHUR W. KNAPP (J. Soc. Chem. Znd. 1915 34 1121-1122).- Very thin shavings of the fat wax or fatty acid are spread over about one half of the bulb of a thermometer which is fixed in a corked test-tube immersed in a transparent bath.It is easy t o determine when the sharp edges of the scrapings soften and when the substance is quite) transparent. For crystalline compounds the bulb of the thermometer may be rolled in the powdered substance when sufficient will adhere. The Freezing Point of Benzene as a Fixed Point in Thermometry. JULIUS MEPER (Zeitsch. physikal. Chem. 1915 90 721-722).-Richards and Shipley (A 1914 ii 765) recommend the use of the freezing point of benzene as a fixed point f o r Cali- brating thermometers. The author holds that benzene purified by the method indicated will have an indefinite freezing point on account of the varying amount of air and other gases dissolved in it.He recommends that it would be preferable t o use the melting point of pure benzene crystals as a fixed point. H. M. D. J. C. W. J. F. S.GENERAL AND PHYSICAL CHEMISTRY. ii. 81 The Solidification Curve of the System Ammonium Nitrate- Lead Nitrate. €3. BOGITCH (Compt. rend. 1915 161 '790-'791).- The author has determined the solidification points of mixtures of ammonium nitrate and lead nitrate containing from 0-55% of lead nitrate. The temperatures a t which solidification commenced and finished were determined in each case for the mixtures contain- ing from 33-45*5% of lead nitrate the differences between these two temperatures f o r each mixture increasing rapidly with increase in the lead nitrate content.The temperatures of commencement of solidification were plotted and by extrapolation the temperature at which lead nitrate should melt was found t o be 450-470O. The eutectic mixture contained 33% of lead nitrate. W. G. The Critical Constants of %-Butane isoButane and Propylene and their Vapour Pressures at Temperatures Between Oo and 120". F. M. SEIBERT and G. A. BURRELL (J. Amer. Ch,em. SOC. 1915 37 2683-2691).-The substances used were carefully. purified by fractionation a t low temperatures. The vapour pressures recorded were independent of the relative volumes of liquid and vapour and the critical phenomena were quite normal. The following values are recorded f o r the critical temperature ( O ) critical pressure (T) and the constants a and b calculated from the van der Waals's equation n-butane 8 = 153*2O r = 35-67 atm.a = 0.02883 b = 0.005470 ; isobutane 8 = 133'7O T = 36.54 atm. a= 0-02562 b = 0*005096 ; propylene 8 =92*6O r = 45-34 atm. a = 0.01668 b = 0.003692. H. M. D. The Vapour Pressures of Sulpbur Dioxide and Nitrous Oxide at Temperatures Below their Normal Boiling Points. G. A. BURRELL and I. W. ROBERTSON (J. Amer. Chem. Soc. 1915 37,2691-2694).-The observed vapour pressures of liquid sulphur dioxide between - 1l0 and - 64.5O can be represented by the equa- tion log p= - 1448*01/T+ 8.425 and the data for solid nitrous oxide between -90-6O and -144'1O by the equation The Clausius equation gives 6619 cal. for the mean heat of vaporisa- tion of liquid sulphur dioxide and 5632 cal. for the mean heat of sublimation of solid nitrous oxide.The observations give - l l . O o and - 88*7O for the normal boiling points of sulphur dioxide and nitrous oxide respectively. lOgp= - 1232'2/T + 9.579. H. M. D. The Vapour Pressures of Some Saturated Solutions. MALCOLM PERCIVAL APPLEBEY and WILLIAM HUGHES (T. 1915 107 1798-1814).-1f P is the osmotic pressure of a solution of concentration c then as a general rule P / c increases with the concentration of the dissolved substance. According t o the boiling- point determinations of Berkeley and Applebey (A. 1911 ii 1062) concentrated solutions of thallous nitrate appear to be exceptional in that the saturated solution gives a value for P / c which is less than that calculated from the gas equation on the assumption thatii. 82 ABSTRACTS OF CHEMICAL PAPERS.the salt is non-ionised. I n view of this anomaly measurements have been made of the vapour pressures of saturated solutions of sodium nitrate sodium sulphate and thallcus nitrate by a static method covering a range cf yeveral degrees in the neighbourhood oE the respective bciling points. The temperature-coefficients a t 760 nini. derived from the vapour- pressure curves are sodium nitrite 21.3 mni. per lo; sodium sulphate 29.2 mm.; and thallcus nitrate 22.3 mm. From the experimental data the osmotic pressures af the saturated solutions have been calculated and the results shcw t h a t for sodium nitrate and sulphate the value of P / c iiicreases continuously with the con- centration whilst for thallium nitrate Y/ e first increases reaches a maximum and then falls ccntinuously.It is supposed t h a t this anomalous behaviour is connected with the association of the dis- solved salt. Special experiments showed that' the low boiling point of the saturated thallous nitrate solutions is not due to volatility of the salt. A form of thermostat is described which gives constant tempera- tures between looo and 120° within 0.02O. H. M. D. Vapour Pressure of Liquid Substances at Low Tem- peratures. C. DRUCKER E. JIM~NO and W. KANGRO (Zeitsch. pTiysilaZ. CTiem. 1915 90 513-552).-The vapour pressure of hexane ethyl alcohol ethyl ether toluene chloroform acetone carbon tetrachloride and iodobenzene have been determined at temperatures from + 15O to - looo. The pressures were measured by either a McLeod manometer or a horizontal modification of the Huygens manometer.I n this way pressures from 40 nim. to 0-001 mm. were measured with an accuracy of 0.001 mm. The experimental results are given in tables and curves and are shown t o be in good agreement with the values previously obtained by Poung. Using the vapour-pressure formula of Kirchhoff-Dupre- Rankine with four terms it is shown t h a t the experimental results may be calculated with accuracy. This is attained when the four coefficients are derived from the curves and also when one of them is chosen arbitrarily. These constants are consequently t o be regarded as empirical since they may have different values without affecting the final results. This indicates that it is not possible from the vapour-pressure measurements alone to deduce the * ' Chemical Constants " which correspond with temperatures which are not far removed froin the absolute zero.To determine these values accurately it is necessary t o know the specific heat and its changes with temperature a t these temperatures. J. F. S. The Vapour Pressure of Concentrated Sugar Solutions. D. ORSON WOOD (Tra'ns. Faraday SOC. 1915 11 29-50).-New measurements have been made by a direct method the solutions being prepared in a glass vessel thoroughly freed from grease and provided with taps. Distilled water is freed from air by boiling in a similar vessel and introduced through one of the tails. Water is then removed by evaporation under reduced pressure and theGENERAL AND PHYSICAL CHEMISTRY. ii. 83 last traces of air removed by heating a t looo and pumping.The concentration is then determined by weighing. The bulb in which the measurements are made is sealed cn to this apparatus and is provided with a stirrer consisting of a sealed glass tube containing short soft iron rods packed in glass wool. This bulb and the first part of the manometer are enclosed in a water-bath. Experiments with pure water show t h a t the results are slightly too low. Solutions containing 48 61 and 69% of sucrose have been examined. The heat of dilution calculated by means of Kirchhoff’s equation is of the riglit sign arid order. When the results are tested by calculating tlie hydration Callendar’s factor is found t o be about 4 whilst the csinotic pressures are of tlie right order but diminish with rise of temperature.This may be due to increased association and tlie theoretical meaning of the change is discussed. C. H. D. The Vapour Pressures and Specific Volumes of Binary Mixtures of Volatile with Non-volatile Liquids. F. H. CAMPBELL ( l ’ m n s . Pccrcrrlny SOC. 1915 11 91-103).-The vapour pressures of the liquids are determined by measuring the rise of mercury in an open manometer caused by allcwing a mixture of known com- position to escape into a space previously filled with hydrogen. Mixtures of cleic acid with either ether or carbon disulphide give vapour-pressure curves which closely approach the diagonal whilst in the third pair of normal liquids aniline and ether the curve has a point of inflection and cuts the diagonal. It is possible t h a t in this case a measurably volatile solvate is formed.The oleic acid-acetone curve acetone being associated lies above the diagonal. Sulphuric acid and ether give a curve with a point’ of inflection probably due t o a compound (C2H,),0,H,S04 which may combine with a further mclecule of ether. Of pairs of associated liquids glycerol-methyl alcohol gives a curve like t h a t of oleic acid and acetone whilst the curve of glycerol and water lies below the diagonal. The specific volume curves are also given. Attempts to apply Dolezalek’s theory quantitatively fail even in the simplest cases. C. H. D. Some Difficulties in Van Laar’s Theory of the Vapour Pressures of Binary Mixtures. R. T. Lmrm (Traits. Faraday SOC. 1915 11 118-120).-Van Laar’s formula is shown t o be inconsistent with the effects of temperature on the vapour pressure of mixtures of partly miscible liquids and even with the experi- mental facts when the comparison is made a t constant tempera- ture.The data fcr methyl ethyl ketone and water mixtures are used for the purpose. C. H. D. Von Babo’s Law and Rirchoff’s Equation for the Latent Heat of Dilution. ALFRED W. PORTER (Tram. Farady SOC. 1915 11 19-28).-Two thermodynamical derivations of von Babo’s law are given in order t o bring out the assumptions on which it is based. An alternative method of finding the internal latent heakii. 84 ABSTRACTS OF CHEMICAL PAPERS. of dilution is also given. The hydration factor a used by Callendar is here employed as representing an effect of alteration of volume of the solvent molecules the volume of a hydrated salt being greater than that of the anhydrous salt.Callendar's interpreta- tion breaks down unless water be assumed not t o be associated. The value for the hydration is the same on either explanation. C. H. D. Heat of Neutralisation and the Quantum Theory. ADOLF HEYDWEILLER (Ann. Physik 1915 [iv] 48 681-692).-Accord- ing to the quantum theory the heat changes associated with chemical reactions ought to bear some definite relation to the changes which occur in the optical properties f o r these are deter- mined by the electron frequencies. On the assumption that the chemical changes are due to changes in the frequencies of the valency electrons the author has calculated the change in frequency characteristic of the formation of a molecule of water from the corresponding H' and OH' ions (1) from the heat of neutralisation and (2) from the change in the refractivity due to ionisation.The first method gives vm - vi = 0.2895 x 1015(sec. -1) and the second v,-vi=0*292 x 1015(sec.-l). The agreement is excellent and is considered to furnish evidence in favour of the correctness of the author's method. H. M. D. Contradictions between the Real and Calculated Solubility of Certain Sodium Salts. ALBERT COLSON (Compt. rend. 1915 161 787-790. Compare this vol. ii 15).-By careful experi- mental work the author finds that for sodium chloride the value of p a t Oo is 28 cal. this development of heat being quite opposed to the marked cooling effect found a t loo and above. I n conse- quence of this sodium chloride should have a maximum solubility a t lo which is contrary to fact.Similarly in the case of sodium carbonate whether anhydrous or hydrated its solubility increases steadily from Oo to 38O and in consequence p should be positive in both cases but it is not. A t the mean temperature 1 8 O the value of p for the anhydrous salt is 5500 cal. and for the decahydrate -13,776 cal. Thus then the ordinary law of solubility does not apply to either of these common salts. W. G. History of Specific Gravity Determination. HRRMANN SCHELENZ (Chem. Zeit. 1915 39 913-915).-A historical description of the methods and apparatus employed for specific gravity determinations from the earliest times to the seventeenth and eighteenth centuries. G. F. M.A Convenient Thermostat for Accurate Speciflc Gravity Determinations and a Gas Pressure Regulator. K. G. BROWNING and C. T. SYMONS (Trans. Faraday SOC. 1915 11 60-63).-A large copper tank is used as the thermostat provided with stirring paddles on a horizontal shaft. The shaft is driven by a chainGENERAL AND PHYSICAL CHEMISTRY. ii. 85 from a spindle driven through a spring coupling attached t o an electrically-driven worm reduction gear. The heating is eit'her electrical or by gas. For electrical heating a heavy platinum wire is fused into the bottom of a Lowry thermo-regulator where it is in contact with mercury. The other contact is a platinum wire mounted on a thin brass rod having a fine thread cut on it a $-volt accumulator being in series with the mercury thread the platinum point and a relay.When contact is made the relay turns on part of the heating current. By shunting the platinum- mercury cont'act with a pair of insulated aluminium plates each about 12 sq. in. immersed in water the sparking is so far decreased as not t o cause fouling of the mercury. A gas regulator is also described consisting of a brass cylinder with a. horizontal partition carrying the valve seating the valve stem being connected with a sheet of rubber insertion t o which it is attached by two concave-convex washers. The regulator is set by adjusting the weight on the valve. C. H. D. The Critical Point of Game Liquefied with Difficulty Nitrogen Carbon Monoxide Oxygen Methane. ETTORE CARDOSO (J. Chim. Phys. 1915 13 312-350).-A detailed account of work part of which has been previously published (compare A.1915 ii 411). By measurements of the densities of co-existing liquid and vapour a t various temperatures the critical density of carbon monoxide is found t o be 0.3110 and of methane 0.1623. The author considers that the precision with which the critical data are determinable is frequently over-estimated. The errors attaching t o the values obtained for the four gases by the author are probably less than +0.lo for the critical temperature and rt0.l atmosphere for the critical pressure. H. M. D. An Interpretation of Van der Waals's Equation from the Point of View of Volume Determined by Equilibrium of Pres- sures. W. V. METCALF (J. Physicai! Chem. 1915 19 705-719).-A theoretical paper in which reasoning from the point of view that the volume of a gas or liquid under given conditions is due t o an equilibrium of opposing pressures it is shown that this equilibrium is represented by the van der Waals equation thus p + a/u2= RT/ (v - b ) .The author names the various pressure quantities thus external pressure + cohesive pressure = elastic pressure. From kinetic considerations it is shown that in the case of a stable liquid the elastic pressure which tends to increase the volume increases with density more rapidly than does the cohesive pres- sure which tends to decrease the volume. The difference between the density coefficients of these two pressures decreases as the density of the liquid decreases. The above-mentioned propositions are used t o interpret the various parts of the van der Waals curves.Explanations are given for the metastable portions of both the PF' and TV curves. J. F. S.ii. 86 ABSTRACTS OF CHEMICAL PAPERS. [Compressibilities of the Elements and their Relations to Other Properties]. Correction. TH E O I ~ R E W. R,ICHARDS ( J . A mpr. Chfnt. Soc. 1915 37 2696-2697).-The substitntion of inore accurate values f o r tlie atomic volumes given in the previous paper (A. 1915 ii 518) does not in any way invalidate the coil- clusions which were drawn. In nearly every case the effect of the substitution is t o improve the ngreeinent between the calculated and observed values. H. &I. T). The Viscosity Coefficients of Mixtures of Helium and Hydrogen. ARTHUR GILLE ( A n n Physik 1915 [iv] 48 799-83’7).-The viscosities of mixtures of hydrogen and helium have been measured a t Oo 15O and looo by the transpiration method. Ex- periments with air which were made to test the apparatus gave q=1.819G x at 19.83O whence ~ / = 1 . 7 2 2 1 x a t 0 O . For helium tlie coefficieiits obtained were ’1 x l o ’ = 1.8925 at 0 ) 1-9Gll a t 15O and 2.3408 a t looo; and for liydrogeii 11 x 104=0.8772 a t 14.79O. The viscosity coefficients for iiiixtures are only approximately represented by Puluj’s formula; a much closer agreement is found when the results are applied t o Thiesen’s formula (T’~rh. J ) P u t . w h . Phys. Ges. 1902 4 357). The influence of temperature on the viscosity of the mixtures is satisfactorily expressed by Sutherland’s formula but there is some1 evidelnce that mixtures falling near the middle of the series would begin to deviate from the requirements of this formula a t temperatures below Oo.13. M. D. The Study of the Density and Viscosity of Aqueous Solu- tions with Special Reference to Nitric Acid. IT. Viscositiee. WILLIAM ROBERT BOUSFIELD (T. 1915 107 1781-1797. Compare ibid. 1407; A. 1915 ii 744).-An improved form of viscometer of the gravity flow type is described for which it is claimed that an accuracy of 1 in 1000 is readily obtainable. Water is the oiily liquid which is adrnissible for the calibration of viscometers which are t o be used in the investigation of aqueous solutions and con- trary to previous statements the. author has found it quite suit- able as a calibrating fluid although i t is admitted t h a t sinall variations in tlie time of flow do occur f o r whicli no explanation can be given. In coiinexioii with the use of water as a calibrating liquid in viscosity work it has been found that the expoiiential formuke giveii by Thorpe arid Rodger to represent the influelice of temperature on the viscosity do not represent their experimental results so closely as the formula 11 =O*OI0514 { 1 + 0.025l(Q - 18) + 0*000115(0 - 18)2) -l.J / 4 - s / 2 - and iT-solu- tions of nitric acid were measured a t 4O 1l0 18O and 25O. The influence of temperature on the variation of the viscosity with the temperature is shown by a comparison of tlie values of the equi- valent viscosity increment as measured by ( 1 1 ~ . - 1)/ t i t where ?lw is the viscosity relative to that” of pure water a t the same temper- The viscosities of Ar/ 32- Nil&,GENERAL AND PHYSICAL CHEMISTRY.ii. 87 ature and ni is tlie number of equivalents of acid per litre. A t 4 O the recorded values are all negative a t lS0 and 2 5 O all positive whilst at I lo the iiicreinent shows a change of sign. By means o€ curves i t is shown that the several values of the increments lie on smooth curves the interpretation of wliicli is discussed in terms of tlie theory that tlie viscosity of a solution is approximately a simple function of the relative sizes of tlie different groups of molecules in the solution. 111 tlie particular case under investiga- tion the three groups concerned are tlie nitric acid molecules t’he dihydrol aiid the triliydrol molecules and account’ is taken of these in tlie author’s iiiterpretation of his observations.H. 31. D. The Viscosities of Solutions of Caesium Salts in Mixed Solvents. I?. B. DAVIS and HARRY C. JONES (J. Anzer. Cliem. SOC. 1915 37 2636-2G42).-Measuremeiits have been made of tlie viscosities of solutions of caesium chloride and caesium nitrabe in binary mixtures of water with methyl alcohol ethyl alcohol and acetone a t 1 5 O 2 5 O and 35O. Similar measurements have already been made for solutions of rubidium and potassium salts. The interest attaching to the salts of these three metals lies in the fact t h a t their atomic volumes are greater than those of all other metals. The atomic volume decreases in the order-czsium rubidium potassium and any peculiarities shown by solutions of rubidium and potassium salts may therefore be expected to be shown by czesium salt solutioiis in greater degree.Generally speaking the salts of these t h e e metals are the only ones whicli loner the viscosity of water. Tliis has been explained by reference t o tlie large volume of tlie cations resulting in a reduction of the frictional surface in comparison with that of pure water . As in the case of rubidium and potassium salts the transition from negative to positive viscosity as the proportion of non-aqueous solvent in the mixture increases is shifted towards the non-aqueous end of tlie series of mixtures aiid this effect is more marked with caesium than with rubidium. H. M. D. Willard Gibbs’s Adsorption Coefficient. ALFRED J\-. PORTER (Trrrns.Farcrdcry Soc. 1915 11 51-52).-A simple proof of Gibbs’s adsorption theorem obtained by taking the solut,iou in con- tact with its vapour aiid considering a trausf ormatioii in whicli an infinitesimal evaporation takes place so that tlie vapour pres- sure which is better known takes the place of the osmotic pres- sure. The formula is extended to the case of strong solutions for whicli it still holds good although tlie correcting term becomes relatively large. C. H. D. A Simple Method of Deriving the Gibbs Adsorption Formula. F. J. HARLOW and R. S. WILLOWS (Trans. Faraday SOC. 1915 11 53-54).-A simple thermodynamical proof of Gibbs’s adsorption formula is given avoiding the use of the chemical potential andii. 88 ABSTRACTS OF CHEMICAL PAPERS. obtaining the result in terms of osmotic pressure (compare pre- ceding abstract).(1. H. D. Determination of the Coefficient of Diffusion of Potassium Chloride by an Analytical Method. A. GRIFFITHS J. M. DICKSON and C. H. GRIFFITHS (Proc. Physical SOC. London 1916 28 73-80) .-The diff usometer described consists essentially of a series of parallel vertical tubes the lower ends of which communi- cate directly with a reservoir of large capacity containing the salt solution. The greater part of the reservoir is above the lower ends of the diffusion tubes and by the action of gravity the con- centration of the solution in this region is kept approximately constant. The upper ends of the diffusion tubes open into' a small chamber provided with inlet and outlet tubes. Pure water enters through the inlet tube and the dilute salt solution leaving by the outlet tube is collected and analysed.I n carrying out an experi- ment time is allowed for the attainment of- the steady state. This requires several days and the complete determination may take six weeks. The results obtained give 1.703 x 10-6 f o r the mean value of the diffusivity of potassium chloride a t 1 8 * 5 O in a solution of 3N con- centration. H. M. D. Velocity of Diffusion and Hydration in Solution. M. PADOA and PERNANDA CORSINI ( A t t i R. Accad. Lincei 1915 [v] 24 ii 461-467).-By means of an apparatus differing slightly from those employed by previous investigators the authors have studied the diffusion in aqueous solution of various non-electrolytes with the object of determining their molecular weights and consequently the extent to which they are hydrated (compare Herzog A.1911 ii 23). The measurements were made a t constant temperatures ranging from 20° to 2 4 O in the different cases alid the viscosity constants [ are calculated from Bingham and White's results (A. 1912 ii 1144). The results obtained are summarised in the follow- ing table II being the diffusion coefficient and the specific volume and the hydration being expressed in mols. of water per mol. of the compound /-<- Hydra- Mol. wt. k. V. & Calc. Found. tion. Methyl alcohol ............ 0.739 1-262 0.009172 32 66.6 1.9 Ethyl alcohol ............... 0.396 1.246 0.01006 46 319.0 15-0 Glycerol ..................... 0.555 0.793 0.009172 92 250.0 8.9 Phenol ........................ 0-525 0.932 0.009116 94 214.0 6.7 Sucrose ..................... 0.442 0.630 0.009616 342 530.0 10.5 Acetamide ..................0-709 0.870 0.009172 59 109.0 2.8 Hexamethylenetetramine 0.543 0.753 0-009838 140 221.0 4.4 Although these results are slightly incorrect owing t o the fact that ths calculated values of v do not apply exactly to the hydrated compounds a pronounced tendency t o undergo hydration is evident. Padoa and Matteucci (A. 1915 ii 676) have shown that inGENERAL AND PHYSICAL CHEMISTRY. ii. 89 benzene solution aromatic hydrocarbons exhibit no appreciable power of combiiiing with the solvent. This is borne out by the results of measurements of the diffusion of naphthalene diphenyl and dibenzyl in benzene solution ; the values of the molecular weight found in this way are 121 152 and 180 respectively the theoretical values being 128 154 and 182.T. H. P. Influence of Glycerol Dextrose Alkali Nitrates and Sul- phatee and Ammonium Salts on the Solvent Power of Water. JAMES CHARLES PHILIP and ARTHUR BRAMLEY (T. 1915 107 1831-1837>.-By the method described in a previous paper (com- pare A. 1915 ii 236) experiments have been made to determine the effect produced by the above-mentioned substances on the dis- tribution of ethyl acetate between water and benzene a t 20°. For all the dissolved substances examined excepting ammonium nitrate the equivalent relative lowering of the solvent power increases as the concentration decreases. From a comparison of the values of the lowering of the solvent power with those deduced from earlier observations on the influence of dissolved salts on the solubility of hydrogen oxygen ethylene and acetylene in water it is found that the authors’ values are greater except in the case of the nitrates Since the values are calculated on the assumption that the added substance has no solvent power of its own it follows that they must be regarded as minimum values.For this reason the equivalent lowerings yielded by the distribution method are probably nearer the truth and for the majority of the dissolved substances the specific character of the reference substance is of less significance in the method used by the authors than in other work of the same kind. H. M. D. The Electrical Synthesis of Colloids. H. T. BEANS and HERBERT E.EASTLACK ( J . Amer. Chem. SOC. 1915 37 2667-2683).-The experiments described were made with the object of throwing light on the nature of the electrical method of producing colloidal solutions. I n a preliminary calculation of the conductivity due t o colloidal particles of gold for which the size charge and mobility are known it is shown that the conductivity due to the colloid itself is negligibly small in comparison with the observed conductivities of colloidal gold solutions. Comparative experiments in which the increase in the con- ductivity due to the electrical dispersion of gold and platinum was determined showed that the average increase for sols of approxi- mately the same depth of colour is about eight times as large for platinum as for gold.The platinum sols were invariably more stable than the gold remaining apparently unchanged on keeping f o r several weeks whilst the gold sols coagulated in from two to twenty-four hours. The greater stability of the platinum sols is probably connected with the presence of small amounts of electro- lyte formed by oxidation of tbe metal. during the process of dis- persion.ii. 90 ABSTRACTS OF CHEMICAL PAPERS. I n accordance with this view it lias been found that certain anions have a stabilising influence on red gold sols and this property may be utilised in the preparation of stable colloidal solu- tions of the metal. Such anions are chlorine bromine iodine and hydroxyl and a iiiarked effect is observed in coiiceiitrations ranging from 0.00005S to 0.005,V.The stabilisiiig effect is not observed if the gold is dispersed in pure water and the electrolyte added subsequently. The lower limiting anion concentration appears to be independent of the concentration of the colloid forrned over a wide range and tlie upper limit is t o some extent iiifluenced by the coagulating power of the associated cation. Other anions appear t o have no influence 011 tlie stability of red gold sols negative results being obtained with solutions containing small quantities of fluorides nitrates chlorates and sulphates. I n all these solutions tlie colloidal particles are negatively charged and the stabilisiiig effect of certain anions would seem to be determined by the ability of these ions t o forin stable complexes with gold. Attempts to detect gold it1 the) filtrate after coagula- tion of the sols gave however negative result)s from whicli it is inferred tliat the airiouiit of the gold compound whicli is necessary to produce a stable colloid is extremely minute.The above observations lead to the conclusion that the electrical method of colloid synthesis involves a primary thernio-mechanical dispersion of tlie metals and the subsequent reaction of the highly disperse1 metals with certain ions present. in the surrounding medium resulting in the forniatioii of negatively charged colloidal complexes. H. M. D. Influence of Light on the Stability of Colloidal Solut'ions. H. NORDENSON (Zeitsch. physilial. Chem. 1915 90 603-626).- The influence of light from a quartz mercury lamp on a number of sols has been investigated spectrophotochemically and ultra- microscopically.It is shown that light has a slow coagulating influence on metal colloids which is very similar t o tlie influence exercised by a weak electrolyte. The quantity of light' necessary t o bring about a visible coagulation is very large which in all probability explains the negative results of other investigators. There is no chemical change occasioned in the disperse phase con- sequently so far as this phase is concerned the change is t o be regarded as purely physical. The photo-action does not bring about chemical changes in the dispersion medium and the action only takes place if the colloid is present during the action of the light in the dispersion medium. Various types of rays have an analogous action wliicli is iiidepeiident of tlie charge on tlie colloidal particles.Thus ultraviolet light and &rays act on both negative and positive colloids. The action cannot therefore be regarded as due t o a photo-electric emission of electrons o r t o the supply of certain charges by the light rays. I n all cases the photo-coagula- tion is accompanied by a decrease in the total charge of the colloid. This decrease in the total charge may be due to one of the follow- ing causes (i) a change in the dissociation of the adsorbed electro-GENERAL AND PHYSICAL CHEMISTRY. ii. 91 lytes; (ii) tlie setting up of potential differences after the manner of tlime in tlie Becquerel effect; and (iii) an adsdrption of hydrogen or hydroxyl ions from tlie dissociated water.I n any case there is a specific photo-chemical action wliicli occasions a disturbance of the adsorption equilibrium on the colloid and thus effects an increase in the size of the particles. This can then be continued after the removal of the light. J. F . S. Application of the Idea of Expansibility Tension to t h e Theoretical Study of Chemical Equilibria. L. GAY ( J . Chim. P?iys. 1915 13 303-311. Compare A 1915 ii 326).-A theoretical paper iu whicli the conception of expansibility tension is applied in the derivation of a general formula f o r the displace- ment of the equilibrium in a mixture resulting from a change in the volume pressure or temperature. H. M. D. The Ternary System Carbon Tetrachloride Alcohol and Water HARKY A. CuR'rIs and ESBON Y.TITUS ( J . PhysicaZ CheiiL. 1915 19 739 -752).- il large iiuniber of experiments have been carried out to determilie the critical solubility data of ternary mixtures of the above three coniponents. Further experiments have been made t o establish the relationship between the density composition and refractive index of the homogeneous ternary niix- tures. From tlie experimental data the equilibrium line between the one-liquid phase and the two-liquid phase systems has been located on the composition diagram for the temperature 19.75O. The general shape of the critical solubility surface in the com- position-temperature prism has been determined over a range of about 50°. It has been found that this surface rises abruptly from isotherm t o isotherm that is temperature has but little effect on the mutual solubility of the three components. The index of refraction has been determined f o r eight series of homogeneous ternary mixtures.It' has been found that as water is added to a mixture of carbon tetrachloride and alcohol rich in alcohol the refractive index a t first increases and then decreases. As mixtures richer in carbon tetrachloride are used the maximum refractive index is reached with less and less water until finally with mix- tures containing initially more than 35% of carbon tetrachloride the addition of water causes the refractive index t o decrease con- tinually. The densities have been determined for eight series of homogeneous ternary mixtures. It has been found that the addi- tion of water to a carbon tetrachloride-alcohol mixture rich in alcohol increases the density uniformly whilst the addition of water to a mixture rich in carbon tetrachloride lowers the density. The slope of the density curves does not however change sign a t the point where the density of the initial mixture is unity but a t a point somewhat higher tlian this showing t h a t the volume of the ternary mixture is less tlian tlie sum of tlie volumes of the com- ponents.It is shown that the determiiiatioii of two of the physical properties of a ternary niixture does not always fix its composition. J. F. S.ii. 92 ABSTRACTS OF CHEMICAL PAPERS Equilibria in the Systems of the Higher Alcohols Water and Salts. G. B. FRANKFORTER and STERLING TEMPLE ( J . Amer. Chem. SOC. 1915 37 2697-2716.Compare Frank- forter and Frary A. 1913 ii 685).-In extending thO earlier investigation (Zoc. cit.) the authors have selected the systems n-propyl alcohol-sodium carbonate-water isopropyl alcohol- potassium fluoride-water isopropyl alcohol-potassium carbonate- water allyl alcohol-potassium fluoride-water allyl alcohol- potassium carb ona t e-wa ter allyl alco hol-sodium carbona te-wa ter and allyl alcohol-sodium chloride-water. The method of represent- ing such systems by triangular co-ordinates is described but the results are represented in part by rectangular co-ordinates with the percentage of alcohol in the solvent plotted against the per- centage of salt in the solvent. The n-propyl alcohol used was purified by successive dehydration with potassium fluoride quicklime magnesium amalgam sodium amalgam and again quicklime; it then had b.p. 97*3-97.5O/ 752 mm. D20 0.8032 nH 1.39023 whilst isopropyl alcohol purified by successive treatment with potassium fluoride quicklime sodium amalgam and quicklime had b. p. 82'5O D20 0.7881 nga 1.37960. Potassium fluoride and potassium carbonate were found t o be satisfactory dehydrating agents for the above-mentioned akohols. No evidence was forthcoming as to any relation between the salt- ing-out effect and the chemical constitution of the salts o r of the alcohols and it would appear that the effect is a function of the solubility of the salt in water and in the alcohol concerned together with the amount of water with which the salt unites to form its lowest hydrate and with the ability of the alcohol to replace the water of hydration.Relative Affinity of Metals in Non-aqueous Solutions aDd their Reactivity in Ineulating Media. I. JNANENDRA CHANDRA GHOSH (J. Plzysikal. Chem. 1915 19 720-733).-The author has studied the action of a number of metals on solutions of salts in a series of nonconducting solvents with the object of ascertaining whether chemical replacements take place in solutions of salts which are not ionised. Solutions of ferric chloride mercuric chloride mercuric cyanide cadmium iodide and lead oleate in benzene toluene nitrobenzene aniline chloroform amyl alcohol amyl acetate and ethyl acetate were shaken in sealed glass tubes a t 21° with the metals mercury copper lead iron tin nickel zinc and magnesium. The specific resistance of each solution was measured by the replacement method.I n every case the resistance was very high the value varying between 45 x 1O1O and 0.25 x 106 ohms. It is shown that magnesium zinc and lead replace iron in ferric chloride in all the solvents used; tin replaces iron in most cases; copper replaces iron in aniline solution and probably in amyl alcohol solution. Mercury does not replace iron in any solvent but reduces the ferric chloride t o ferrous chloride in all the solvents except aniline in which case there is no action. Magnesium zinc and copper replace mercury in mercuric chloride in all cases. Iron D. F. T.GENERAL AND PHYSICAL CHEMISTRY. ii. 93 replaces mercury in mercuric chloride in ethyl acetate amyl acetate and amyl alcohol solutions.Lead has no action on mercuric chloride in chloroform solution but in all other cases it reduces mercuric chloride to mercurous chloride. Copper replaces mercury in mercuric cyanide in all solutions. Iron replaces mercury in nitrobenzene and amyl alcohol solutions magnesium replaces mercury only in amyl alcohol solution and lead replaces mercury in the cyanide in amyl alcohol and amyl acetate solutions. I n other cases there was no action. Magnesium replaces cadmium in cadmium iodide in all solutions whereas zinc does not replace it in any; iron replaces cadmium in aniline chloroform and amyl alcohol solutions and copper replaces cadmium in aniline chloro- form amyl alcohol and amyl acetate solutions. Magnesium copper and zinc replace lead from lead oleate solutions in all the solvents; iron has no action in benzene chloroform nitrobenzene and ethyl acetate solutions of lead oleate and mercury has no action on any of the solutions of lead oleate.It is thus shown that the replacement series is not at all a characteristic of the replacing metal but is dependent on the mutual relationship of the metals the acid radicles and the solvent present in a given com- bination. A series of experiments was carried out with solutions of mercuric iodide and mercuric chloride in ethyl ether ethyl acetate glycerol carbon tetrachloride chloroform benzene toluene and carbon disulphide. These solutions were shaken with silver and in every case mercury was deposited thus showing that the order of replacement is the same in all cases as in water.J. F. S. Dynamic Researches on the Formation of Some Aromatic Sulphones Under the Influence of Aluminium Chloride in a Medium of Benzenesulphonyl Chloride. S. C. J. OLIVIER (Rec. trav. chim. 1915 35 166-179. Compare A. 1914 i 676 818; ii 846).-The author has determined the velocity of forma- tion of sulphones from benzene and chlorobenzene on adding them to a solution of aluminium chloride in benzenesulphonyl chloride a t 30°. The velocity constant K was calculated assuming that the compound C,H,*SO,C1,A1Cl3 reacts with the benzene or chloro- benzene the solvent not being able to remove any of the aluminium chloride from its combination. The results show that one molecule of aluminium chloride cannot transform more than one molecule of benzenesulphonyl chloride and that for low concentrations of aluminium chloride the values of K tend to decrease during the reaction.A t high concentrations the formation of sulphone is a true bimolecular reaction. An increase in the concentration of the aluminium chloride or of the chlorobenzene causes a diminu- tion in the value of IZ. The molecular weight of the combination C,H,*SO,Cl,AlCl in solution in benzenesulphonyl chIoride was determined by cryoscopic methods and gave no indication of association with increase in concentration of the aluminium chloride. On combinations of sulphones with aluminium chloride the above solvent exerts aii. 94 ABSTRACTS OF CHEMICAL PAPERS. greater or less dissociating action. The results show that for the same solvent the relative velocities of introduction of a secoiid substituent are shown by C,H > C,H,Cl.W. G. Determination of Ionisation Constants and the Titration of Weak Bases by the Conductivity Method. CH. DE ROIIDEN ( J . Chim. Phys. 1915 13 261-301).-The methods usually employed in the determination of the ionisation constants of weak acids and bases depend on the measurement of the degree of hydro- lysis of their salts in aqueous solution. The equilibrium relations in solutions of such salts are discussed and attention is drawn to the essential difference between tlie hydrolysis constants (1) when the acid and base are both weak and (2) when the acid is strong and the base weak or vice vers2. The difference gives rise t o little ambiguity so long as the salt falls into one o r other of these limit- ing groups but the hydrolysis constant of a salt which occupies an intermediate position is a quantity which cannot be defined by either of the ordinary formulze.The magnitude of the hydrolysis constant is determined by the ionisation constants of the acid base and water and on this account hydrolysis constants are not of primary significance. A detailed description is given of a method for the determina- tion of the ionisation constants of weak bases from measurements of the electrical conductivity of solutions after the addition of successive quantities of a standard solution of hydrochloric acid. This addition was effected without aiiy appreciable change in the volume of the solution. The results obtained for aniline and f o r 0- 712- and pchloro- aniline are recorded.H. M. D. Reaction Velocity in a Viscous (Eeterogeneous) Medium. RAPHAEL H. CALLOW (Trnns. Fnmday Soc. 1915 11 55-59).- The effect' of the presence of gelatin on tlie velocity of hydrolysis of methyl acetate has been determined. Results cannot be obtained with sodium hydroxide on account of its reaction with the gelatin but with hydrochloric acid the change in the velocity constant is small in comparison with that of tho viscosity being only 10% below the normal even in tlie case of a jelly which is completely set. The form of the relation between viscosity and velocity has not been fully determined. C. H. D. Aspirin. 111. Anqmalies of the Decomposition of Aspirin by Water. TI. E. TSAKALOTOS and S. HORSCII (BzLZZ. Soc. c l i i m 1915 [iv] 17 401-406. Compare A.1915 ii 47 433).-A quantitative study of the decoiriposition of aspirin in aqueous solu- tion a t 23O 50° and 60° the rate of decomposition being measured by removing 25 C.C. of the solution a t given intervals and titrating this against N/50-barium hydroxide. At 50° tlie reaction was also followed by electrical conductivity measurements. I n each case the velocity curve shows two minima the times a t which these occur varying with the temperature. W. G .GENERAL A N 0 PHYSICAL CHEMISTRY. ii. 95 The Indirect Determination of Velocity of Hydrolysis by the Polarimetric Method. JAMES CODRINGTON CROCKER (T. 1915 107 1762-1765).-A mixture of sucrose formamide and hydro- chloric acid was made up and the change in the rotation of the solution a t constant temperature followed in the polarimeter.The acid liydrolyses the formamide and is partly used up in the process. It also liydrolyses tlie sucrose remaining unchanged in amount by this reaction. The rate of hydrolysis of the sucrose being dependent on the amount of acid present gradually diminishes in proportion as the formamide is decomposed. From tlie rate a t which the velocity of inversion of the sucrose decreases the velocity of hydrolysis of the formamide can be arrived a t . For the method of calculation the original must be consulted. The rota- tion of the sucrose solution is not appreciably affected by form- amide. Data are given showing t h a t a satisfactory value for the velocity constant for the hydrolysis of formamide can be obtained by this method the principle of the metho-d being applicable to any reaction during which the hydrolyst changes i n concentration and a t the same time can act as a simple catalyst on a third substance present in the solution.T. S. PA. Velocity of Crystallisation of Supersaturated Solutions. ERNESTO CABALLERO Y L ~ P E Z (Anal. Fis. Qzcim. 1915 13 350-375).-An investigation of the effect of mechanical physical physico-chemical and chemical influences on the velocity of crystal- lisation of supersaturated solutions of tlie sulphates of sodium potassium copper zinc ferrous iron magnesium and potassium aluminium. A. J. W. The Catalytic Hardening of Fats by Means of Nickel Oxidee. W. MEIGEN ( J . pr. CIzem. 1916 [ii] 92 390-411. Compare Meigen and Bartels A. 1914 i 482).-Jn reply to- Bedford and Erdmanii (A.1913 i '701) and Siegmund and Suida (A. 1915 ii 626) the author submits the evidence as to the activity of an unisolated nickel suboxide during tlie catalytic reduction of fats t o a critical survey and confirms the opinion already expressed by Meigen and Bartels (Zoc. c i t . ; Normann and Pungs A. 1915 ii 159; Bosshard and Fischli A. 1915 ii 788) t h a t oxides of nickel exhibit activity as catalysts only when they contain a certain amount of the free metal. Further experimental evidence with other oils is adduced in support of this view and stress is laid on the u n t r u stw o r t hi ness of specific gravity or e 1 ec t r i c a 1 conductivity data as evidence of the coinposition of the catalyst after a reduc- tion experiment.Such results must be dependent on the nature and amount of the organic impurities included in the finely divided nickel and these indeed also can affect the sensitiveness of the carbon monoxide test for the presence of free nickel. The free nickel can be removed fairly satisfactorily from a used nickel oxide catalyst by a magnet and in this way the author obtained frac- tions which contained more nickel than could be present' even inii. 96 ABSTRACTS OF CHEMICAL PAPERS. it suboxide of the formula Ni,O so that the presence of free nicker can scarcely be denied. From the facts that metallic nickel is known t o be an effective accelerator and that it can be found in nickel oxide which has been used as a catalyst the author is unable to accept the suggestion that nickel oxide owes its apparent catalytic activity to the forma- tion of a suboxide the presence of which is not proved and the catalytic properties of which are merely postulated. D. F. T. [Determination of] Molecular Weights of Gases by an Evaporation Method. H. L. TRUMBULL ( J . Amev. Chem Xoc. 1915 37 2662-2667).-Measurements have been made of the rate a t which water evaporates in contact with dry air carbon dioxide and hydrogen at atmospheric pressure. The entire apparatus was immersed in a thermostat maintained a t constant temperature (25'07O k O*0lo). The rate of evaporation (dwldt) multiplied by the barometric pressure ( B ) is constant for a given gas. The values of B . dw/dt f o r the three gases %re approximately in the inverse ratio of the square roots of the densitdes of thO gases indicating that the rate of vaporisation is determined by the rate of diffusion. H. M. D. The Valency Theoryof J. Stark from a Chemical Point of View. DOROTHY A. HAHN and MARY E. HOLMES ( J . Amer. Chenz' SOC. 1915,37,2611-2626).-An account is givemn of Stark's electron theory of valency in its application to the interpretation of chemical phenomena. The account represents an abstract of a treatise by P. Ruggli (" Die Valenzhypothese von J. Stark von Chemischen S tandpunkt," Stuttgart 19 1 2). H. M. D. An Improved Form of Gas-washing Bottle. F. R,. VON BIGHOWSKY and H. STORCH ( J . Amer. Chem. Soc. 1915 37 2695-2696).-In its simplest form the apparatus consists of an ordinary wash-bottle which in operation is laid on its side. The tube which reaches t o the bottom of the bottle is slightly con- stricted below the cork and given a very slight inclination to the horizontal. The gas to bO washed enters the bottle through this tube and forms bubbles a t the constriction opposite t o which and on the lower side of the tube is a sufficiently wide aperture t o admit freely the entrance of the liquid in the bottle. H. M. D. An Automatic Vacuum Pump. OTTO MAASS ( J . Amer. Chenz. SOC. 1915 37 2654-2656).-A form of Topler pump which works automatically in conjunction with an aspirator. No mechanical spring or valve is involved and experience has shown that it can be allowed to run with perfect safety for an indefinite period. H M. D.