85 General and Physical Chemistry. Refraction and Dispersion of Air Oxygen Nitrogen and Hydrogen and their Relations. CLIVE CUTHBERTSON and MAUDE CUTHBERTSON (Proc. Roy. Soc. 1910 83 A 151-171).-0n account of the discrepancies exhibited by existing data the authors have redeter- mined the refractivities of the four gases for the wave-lengths X=6563 5790 5461 and 4561. I n all cases it is found that Cauchy’s formula p - 1 =a( 1 + b/X2) is inadequate for the expression of the dispersion data the value of b increasing as the wave-length diminishes. Much better agreement is obtained when a formula of the Sellmeier type is employed to represent the results The constants involved i u the Sellmeier formula are calculated for each of the four gases examined and also for phosphorus sulphur and mercury.H. M. D. Refraction‘and Dispersion of Sulphur Dioxide and Hydrogen Sulphide and t h e i r Relation to those of their Constituents. CLIVE CUTHBERTSON and XAUDE CUTHBERTSON (Pmc. Roy. SOC. 1910 83 A 171-176).-Measurements of the dispersion of sulphur dioxide and hydrogen sulphide are recorded and the data are expressed in terms of Sellmeier’s formula for which the constants involved are calcul a t ed. The refraction data for sulphur dioxide acd hydrogen sulphide are compared with those of the component elements a n d . i t is shown that the refractivity does not follow an additive law. For sulphur dioxide the refractivity is IS% and for hydrogen sulphide 6% less than that required by the law of addition. Whereas the number of dispersion electrons in sulphur dioxide appears to be equal to the sum of the dispersion electrons in sulphur and oxygen this relation- ship does not hold in the case of hydrogen sulphide.H. M. D. Refraction and Dispersion of Neon. CL~VE CUTHBERTSON and MAUDE CUTHBERTSON (Proc. Zoy. Soc. 1910 83 A 149-llil).-The refractive index of neon a t 0’ and 760 mm. was found to be 1*00006716 for the green mercury line (X=5461). From this and values obtained for the red and blue lines of cadmium the dispersion has been calculated. Previous measurements having shown that the refractivities of the inert gases are very nearly in the ratio of whole numbers it was anticipated that the refractivity of neon would be exactly twice as large as that of helium. The experi- mental value is however less than this to the extent of 4%.H. M. D. Distribution of the Ultimate Rays in the Spectrum of Different Regions of the Sun. ANTOINE DE GRAMONT (Compt. reend. 1910 150 37-40. Compare Abstr. 1908 ii 645).-This paper contains a tabular statement of the wsve-lengths for the VOL. XCVIII. ii. 7ii. 86 ABSTRACTS OF CHEMICAL PAPERS. ultimate rays of great persistency found in the sun. The fact that metalloids such as tellurium phosphorus arsenic antimony and boron have never been recognised in the sun is probably to be explained by the absorption of their ultimate rays by the terrestrial atmosphere. The author considers that Lockyer's enhanced lines are not due to dissociation but that they are ultimate rays and that their appearance gives some indication as to the proportions in which the Critical Study of Spectral Series.I. The Alkalis Hydrogen and Helium. WILLIAM M. HICKS (Proc. Roy. Xoc. 1910 83 A 226-228*).-The experimental measurements of the spectral lines of the alkali metals hydrogen and helium have been analysed with the object of determining the relationships between the wave- length numbers. For any one series of lines the wave numbers can practically all be represented by a modified Rydberg formula. The relationships between certain constants for the series (lithium sodium potassium rubidium cmium) are expressible by the integers 1 2 4 5 and 6 and these integers are also involved when the atomic volumes of the respective alkali metals are compared. Line Spectrum of Calcium given by the Oxy-acetylene Burner.GUSTAVE A. HEMSALECH and CHARLES DE WATTEVILLE (Compt. rend. 1909 149 11 12-1 115).-Using the method described previously (Abstr. 1908 ii 336 445 547 745) the authors have investigated the spectrum of calcium in the oxy-acetylene flame. The flame contains a very brilliant blue cone which shows not only the bands which are obtained with R Burisen burner but also a series of supplementary bands which are distiibuted over the whole length of the spectrum. All the calcium lines do not exist a t the base of the flame but are only formed some distance above the orifice of the burner. Even the strongest lines although visible are only very faint at the base of the flame ; they become very intense just above the blue cone. A table IS given showing that with flames of different temperatures the number of lines between X 3900 and X 5000 increases with increas- ing temperature.It is also shown that the number of lines given by that portion of the flame which extends to the top of the blue cone decreases with rise in temperature. The Yellow Orange and Red Regions of t h e High Tempera- ture Flame Spectrum of Calcium. GUSTAVE A. HEMSALECH and CHARLES DE WATTEYILLE (Compt. rend. l909,149,1369-1372).-The authors compare the less refrangible part of the calcium spectrum obtained by them in the oxy-acetylene dame (preceding abstract) with the spectrum obtained by King ( A ~ o p h y s . Journ. 29 190) with the electric furnace. With the exception of the line A = 6708.18 which is present in the latter spectium and not in the former the characters of the spectra are the same and the causes which produce them are therefore probably thermal.I n order to produce the red lines in the * and Phzl. Trans. 1910 A 210 57-111. elements producing them are present. IV. 0. w. H. M. D. A full list of wave-lengths and intensities is given. T. S. P.ii. 87 GENERAL AND PHYSICAL CHEMISTRY calcium spectrum a high temperature such as that given with the oxy-acetylene flame is necessary. The line A = 6108.18 which is present in the spectrum of the electric furnace is probably due to lithium (A= 6705 OS).‘ This same line is found in the sun-spot spectrum and therefore lithium is present in the sun. T. S. P. Series Systems in the Spectra of Zinc Cadmium and Mercury. T. ROYDS (Ann. Phpsik 1909 [iv] 30 1024).-In refer- ence to Paschen’s paper (this vol.ii 3) the author poiots out that he has alreadymeasured the Zeeman effect for certain lines in the spectra of zinc and cadmium. I n a magnetic field the lines 6438.7 (cadmium) 6362.6 (zinc) 5528:7 and 4703.3 (magnesium) appear as symmetrical triplets. The magnetic displacement observed for these lines corresponds with values of elm in good agreement with the value for cathode rays. H. M. I). The Spectrum of Antimony. A KRETZER (Zeitsch. wiss. PAoto- graph. PhotophysiE. Yhotochem. 1910 8 45-79).-The spark arc and flame spectra of antimony have been investigated. The wave-length measurements of the lines and bands are recorded in detail and compared with the data of previous observers. H. M. D. Absorption Spectrum of Potassium Vapour.P. V. BEVAN (Phil. Hug. 1910 [vi] 19 195-200).-It has been found that t h e principal series lines of potassium appear in much greater numbers in the absorption spectrum of potassium vapour than in any form of emission spectrum which has as yet been examined. This result is analogous to that obtained by Wood in the case of sodium vapour. The method of experiment consisted in heating potassium in a steel tube the ends of which were closed by quartz plates. A beam of light was passed through the tube and the emergent beam was examined by means of a quartz spectrograph. I n this way the author was able to measure the wave-lengths of the principal lines of potass- ium up to the line corresponding with n=26 in the Kayeer and Runge formula. Hitherto the members of the series LIP to w = l l have been observed so that fifteen new lines have been added to the list.A comparison of the absorption spectra of sodium and potassium vapours shows a close correspondence between them and this is regarded as evidence that the mechanism involved in the two cases is the same. H. M. D. The wave-lengths of these range from 2928.0 to 2870.0. Absorption Spectra of Various Salts in Solution and the Effect of Temperature on Such Spectra. XXVI. HARRY C. JONES and W. W. STRONU (Amer. Chem. J. 1910 43 37-90).-A detailed account is given of an investigation of the absorption spectra of various potassium uranium and neodymium compounds by the methods employed by Jones and Uhler (Abstr. 1907 ii 147 211 212) and by Jones and Anderson (Abstr.1909 ii 197 359). The absorption epectra of potassium ferricyanide ferrocyanide chromate 7-2ii. 88 ABSTRACTS OF CHEMICAL PAPERS. and dichromate and of uranyl acetate bromide chloride nitrate and sulphate in aqueous solution of uranyl acetate nitrate and chloride in methyl alcohol and of uranyl nitrate and chloride in ethyl alcohol are recorded. The effect of dehydrating agents such as calcium and aluminium chlorides has been determined. Photographic records have been obtained of the absorption spectra of uranous chloride and sulphate and the absorption spectra of neodymium chloride in glycerol and in mixtures of glycerol and water have been studied. The effect of changes of temperature of solutions of various salts at different concentrations has been investigated and spectrograms for a given concentration of a salt have been made a t Oo 15O 30° 4 5 O 60° 75" and 90' for a layer of constant thickness.The absorption spectra of the uranyl salts contain a series of bands in the blue and violet which are usually diffuse. The position of the uranyl bands is not affected by dilution. I n methyl and ethyl alcohol the bands of each particular salt occupy different positions. A new set of fine bands in the green has been discovered in the spectrum of aqueous solutions of uranyl chloride; the presence of a small quantity of aluminium or calcium chloride causes them to disappear. The absorption spectra of the uranous salts are quite different from those of the uranyl compounds. The absorption spectrum of neodymium chloride in glycerol differs entirely from that of an aqueous solution.The intensity of the phosphorescence of the same uranyl saits obtained from different solvents by evaporation is found to vary greatly. Monochromatic stimulation fails to excite phosphorescent bands until the wave-lengths reach the region of the uranyl bands. The NO group has a great influence on the frequency of the uranyl and uranous absorption bands and of the uranyl phosphorescent bands. E. a. A Relation between Absorption and Phosphorescence. L. BRUNINGHAUS (C'ompt. rend. 1909 149 1124-1 127).-The absorption and phosphorescence spectra of the rare earths are dis- continuous consisting generally OF narrow bands. I n the absorption spectra groups of absorption bands are separated by regions of trans- parency whereas in the phosphorescence spectra the regions of emission are separated by dark regions.Taking the spectra of praseodymium erbium dysprosium terbium and samarium as examples the author shows (1) that the mean regions of emission are generally little separated from the mean regions of absorption and (2) that the regions of emission do not coincide with the regions of absorption but with the transparent regions (either those between the groups of absorption bands or those outside them). There is thus an alternation between the groups of absorption bands and those of phosphorescence. The results obtained with compounds of the common elements such as manganese chromium iron copper etc. confirm those given above. In general these substances phosphoresce with a colour which is little different in shade from their colour by reflected or transmitted light.For example chromium sesquioxirle when dissolved in alumina givesGENERAL AND PHYSICAL CIIEMISTRY. ii. 89 the ruby which phosphoresces with a red light; when dissolved in lime it is green and the phosphorescence is green. It seems as if the light emanates from “ phosphoragenic ” molecules in the interior of the phosphorescent substance. This light under- goes absorption in the superficial layers and the radiations observed at the surface are only those for which the “phosphorogenic” substance is relatively transparer,t. T. S. P. Theory of the Law of the Optimum of Phosphorescence L. BRUNINOHAUS (Compt. Tiend. 1909 140 1375-137‘i.)-Making use of the ideas put forward in a previous paper (preceding abstract) the author develops a relation connecting the intensity ( I ) of the radiatiou comprised between two wave-lengtbs X and h+dX which are very close together and the concentration (c) of the phosphorogenic sub- stance in a phosphor.The relation is I=kce-Bc where k and B are constants B depending on the phosphorescent substance. T. S. P. Electrolytic Conductivity of Fluorescent Solutions. A. RASSENFOSSE (Bull. Acod. roy. Belg. 1909 995-1 lO?).-According to the theory of do Heen the conductivity of a fluorescent solution should depend on the light to which it is exposed. Experiments on solutions of fluorescein and eosin show that the conductivity is a maximum when they are submitted to the action of light which is absorbed by them green in the case of fluorescein and yellow in the ’case of eosin.Blank experiments on solutions of potassium chloride proved that the conductivity does not vary under the action of light of different colours and also that the observed variation in the case of the fluorescent solutions is not a thermal effect. T. S. P. Luminescence of Crystals. ALFREDO POCHETTINO (Nuovo Cimento 1909 [v] 18 245-300. Compare Abstr. 1905 ii 430).- The phenomenon of luminescence has been investigated for 227 specimens of crystals representing 78 kinds of minerals. Several methods of exciting luminescence were used ; the more important results mere obtained with cathode Xuminescence but the anodic luminescence fluorescence phosphorescence and thermoluminescence of the crystals were also investigated when they occurred.In many cases the nature of the luminescence depends upon the method of excitation thus for distene the ordinary phosphoregcence and the cathode luminescence are of different colours and in the case of calcite the ordinary fluorescence is polarised and the cathode lumines- cence is not. The nature of the spectra of the cathode luminescence is in many cases independent of the Z.kf.F. applied to the vacuum tube but the intensity of the luminescence is generally greater the greater the applied E.M.F. provided the crystals undergo no permanent changes under the influence of the rays. There does not appear to be any definite connexioc between the luminescence of crystals and their mineralogical relationships ; even crystals of the same substance from different sources may show luminescence of different colours and some specimens of quartz show luminescence whilst others do not.ii.90 ABSTRACTS OF CHEMICAL PAPERS. Many observations have been made on the orientation of the plane of polarisation of the polarised part of the luminescence. The mineral milerite is dichroic greenocite shows double cathodic luminescence and the light emitted from the face of crystals of cassiterite parallel to the x axis is totally polarised. With one exception the lumines- cence emitted from crystals of the rhombohedric system is not rectilineally polarised. The greater the exhaustion of the vacuum tube the more completely is the luminescence polarised and in general all causes which tend to diminish the total intensity of the light emitted diminish the degree of polarisation of the light.Mauy minerals become brown under these conditions and on the face of a crystal directly exposed to the rays a stain made up of differently coloured concentric rings is sometimes observed. On continued exposure to cathode rays the capacity of a crystal to become luminous by excitation of the rays diminishes and this diminution may be temporary or permanent. Fewer substances are rendered luminous by anode rays than by cathode rays and every mineral so far examined which shows anodic luminescence can also be rendered luminous in some other way. I n all cases the luminescence due t o anode rays is weakest less durable and less polarised than that produced by cathode radiation under corresponding conditions.The effect of cathode rays on the minerals is discussed. G. S. Dependence of the Photo-electric Effect of the Alkali Metals in Polarised Light on the Wave-length. ROBERT POEIL (Ber. Deut. phpiknl. Ges. 1909 7 15-722).-Experiments with an alloy of potassium and sodium and ultra-violet radiation of short wave-length show that the photo-electric behaviour of the alkali metals in polarised light is the same as that observed in the case of other metals. H. M. D. Theory of the Ripening Process of the Silver Haloids. A. P. H. TRIVELLI (Zeitsch. wiss. Photocher~z. 1910 8 17-24).-As the result of a microscopic examination of ripened silver haloid plates of high sensitireness the author concludes tbat the ripening process is due to the development of definitely recognisabla crystalline structure.I n consequence of this structural change the silver haloid is in a condition of strain and in consequence is less stable than in the unripened condition. The smaller degree of stability is supposed t o be the cause oI the greater photo-sensitiveness. H. M. D. New Determinations of the Radioactivity of the Thermal Waters of Plombieres. ANDR~ BROCHET (Compt. rend. 1910,150 145-148).-The author has redetermined the radioactivity of the waters of Plombihres (compare Abstr. 1908 ii 143). The radio- activity of the gases spontaneously liberated from the various waters was measured and then the latter mere agitated with a n equal volume of air and the radioactivity of the latter measured in the apparatus of Chheveau-Laborde.The tabulated remlts give theGENERAL AND PHYSICAL CHEMISTRY. ii. 91 altitude of the spring the outflow per twenty-four hours mean temperatures in 1859-1861 Sept. 1908 and August 1909 the total solids per litre the radioactivities in milligram-minutes per 10 litres of the gases and the waters and the total radioactivity for twenty- four hours. The waters are strongly radioactive the radioactivity being due to radium emanation. The total radioactivity of the 22 springs is 74620 milligram-minutes for an average outflow of 67244 cubic metres of water per twenty-four hours. The average radioactivity is 1.11 milligram-minutes per 10 litres the Lambinet water being the most active (2.18). It is calculated that 55-60 mgms. of radium bromide are contained in the total output of water per minute (507 litres).This quantity is defined as the radioactive power of the Plombihres waters. T. 5. P. The Recoil of Radium-C from Radium-B. WALTER MAKOWER and SIDNEY Russ (Phil. Mag. 1910 [vi] 19 100-115. Compare Abstr. 1909 ii 455).-The active deposit of radium on a platinum plate was mounted opposite a metal disc in a n exhausted tube for some minutes so that the disintegration products recoiling from the plate would be received on the disc. The plate mas first freed from adhering emanation and from radium-8 by heating a t 360' in a vacuum for half an hour before use. It mas found that in general. both radium-B and -C were radiated to the disc but if three hours elapsed between the preparation of the active deposit and the recoil experiment only radium4 was obtained.They consider that the radiation of radium4 may not in all cases be a primary recoil effect but due to mechanical disturbance produced by the recoil of radium-D. The amount of radium4 recoiled varies with the same plate with the time in an unexplained manner and is always small compared with the amount of radium-B recoiled. Since radium-B gives only a /3-particle i t is t o be expected that radium4 will recoil with far less energy than radium-& which results from the recoil of the a-ray- expelling radium-A. But the energy of the recoiling radium-B atom is apparently far greater than theory would indicate. Its power oE penetrating air is about 1/40th of that of recoiling radium-B. Attempts to detect an electric charge on the recoiling radium4 atom have failed (compare Makower Abstr.1909 ii 456; Le Radium 1909 6 50). F. S. Disengagement of Emanation from Radium Salts. L. KOLOWRAT (Le Radium 1909 6 321. Compare Abstr. 1907 ii 729).-The paper is devoted to a detailed reconsideration of many of the points previously discussed. The anomaly encountered in that the disengagement of emanation from radiferous barium chloride decreases with rise of temperature from 830' to a minumum at 920° and then increases again quickly to the m. p. at about 950° is probably explained by Plato's observation of the existence of two forms of barium chloride with transformation point 34.4' below the m. p. (Abstr. 1907 ii 239). Similar behaviour of barium fluoride indicates a similar polymorphic transformation between 1000° and 11 00'.ii.93 ABSTltACTS OF CHEMICAL PAPERS. Additional salts studied comprise potassium nitrate silver chloride and czesium nitrate. Nineteen series of experiments have been made on the growth with time of the amount of emanation retained by the salts at varying temperatures after complete initial removal of the emanation by fusion both when the temperature is maintained and also when during part of the time the salt is not heated. The results are interpreted on the view that a definite temperature different for differ- ent molecules exists for each molecule of emanation formed within the salt below which it is retained. The amount of emanation retained at any temperature is the sum of all the molecules the temperatures of disengagement of which are above that temperature.No evidence of any want of homogeneity of the emanation disengaged at different temperatures was obtained (compare Rutherford AbGtr. 1909 ii 457 ; Debierne ibid. 534). F. S. Nature of the Ionisation of a Molecule by an a-Particle. R. D. KLEEMAN (Proc. Roy. Soc. 1910 83 A 195-199).-1f the energy of ionisation is derived from the kinetic energy of the a-particle the electrons ejected from molecules should on the whole possess a a motion in the direction of the a-particles. I n the experiment designed to test this a-particles from a plate covered with polonium were passed through a very thin aluminium foil and fell on a parallel aluminium plate in hydrogen and also in air a t a low pressure If the emergent electrons from the foil are more numerous than the incident electrons from the plate when an electric field is applied between the foil and plate the current should be larger with the foil negative than vice versa and this difference should be the more marked as the potentials are increased.Experiments in hydrogen at 0.8 mm. pressure show a well marked difference increasing rapidly as the voltage is increased above the point a t which the expelled electrons acquire sufficient velocity to ionise by collision. The experi- ments support the view that the energy of ionisation is derived from the ionising agency not by ‘‘ trigger ” action from an internal store in the molecule ionised. F. S. The Number of a-Particles Expelled from the Actinium and Thorium Emanations. H. GEIGER and E.MARSDEN (Physikccl. Zeitsch. 1910 11 7-11. Compare Bronson Abstr. 1908 ii 792).- By a further developnient of the method of counting a-particles by the scintillations produced i n zinc sulphide interesting confirmation and extension have been obtained of Bronson’s conclusion that the thorium emanation must give a t least four a-particles and the actinium emanation a t least two a-particles on disintegration. The number of scintillations produced by the emanation and active deposit together in equilibrium and by the active deposit alone after the supply of emanation has been stopped and the emanation allowed to decay is as 3 to 1 both for thorium and actinium showing that the emanation produces twice the number of a-pdrticles produced by the active deposit.The thorium active deposit is known to produce two a-particles. [A similar conclusion in the case of the actinium active deposit (compare Blanquies Abstr. 1909 ii 634) is not referred to.]GEIVERAL AND PHYSICAL CHEMISTRY. ii. 93 Emanation was allowed to diffuse between two parallel zinc sulphide Bcreens placed close together film side inwards an6 the number of scintillations on exactly opposite portions of the two screens counted simultaneously by two observers with two microscopes. For actinium emanation a very large proportion of the scintillations up to 66% occurred in pairs showing that two particles are expelled simultan- eously or a least with less than 0.1 second between them With radium emanation and active deposit in equilibrium only 2 to 4% of pairs was observed.For thorium emanation it was established that the four a-particles expelled were not simultaneous. A large number of cases of successive scintillations in similar parts of the field were observed with a time interval between the two scintillations from half a second to an unmeasurably short time. This occurred even when extremely few scintillations were produced. The existence of at least one >short-lived a-ray product in the thorium series with a period of average life of about 0.2 second is indicated but full elucidation of this case is not yet arrived at. F. 5. The Absorption of Cathode Rays of Different Velocity in Helium. JAMES ROBINSON (Physikal. Zeitsch. 1910 11 11-13).- The cathode rays obtained by the action of ultra-violet light on a negatively charged plate were employed.The velocity of the rays was varied by varying the charge on the plate. With decreasing velocity of cathode rays the absorption by gases attains a maximum. In hydrogen the maximurn is reached more suddenly and at lower velocity than in other gases. I n helium the absorption with decreasing velocity increases a t first only very slowly down to far smaller velocities than for other gases. The absorption then rises with great abruptness to the maximum. Helium thus resembles hydrogen but the peculiarities shown by the latter gas are even more marked in helium. F. 5. Contact Electrification. ALBERT GRUMBACH (Compt. rend. 1909 149 846-848. Compare Gouy Abstr. 1906 ii 662).-According to Helmholtz the contact potential of a liquid with glass is proportional to pP/$Z where E is the E.H.F.acquired by the liquid of viscosity q- and resistivity p in passing through a glass capillary tube under pressure P. The author finds that in a N/lOOO-solution of potassium chloride in water E is strictly proportional to P. The addition of 5.6% of phenol to the solution increases the viscosity r] in the ratio 1 1.09 but does not alter the conductivity. The capillary E. M.F. however undergoes a marked reduction. In two series of experiments the addition of phenol altered the contact potential in the ratios 1 0.80 and 1 0.78 respectively. R. J. C. Dielectric Properties of the Elements. DIMITRI K. DOBROSERDOFF (J. Russ. Phys. Chem. Soc. 1909 41 llG4-1171).-As usually determined the dielectric capacity X refers to large wave-lengths and the index of refraction n to very small wave-lengths and in order to test Maxwell’s law n2 = K it is necessary to obtain values of n and K referring to comparable conditions.The value nm of n for very largeii. 94 ABSTRACTS OF CHEMICAL PAPERS. wave-lengths may be calculated by means of Cauchy’s formula r(r = A + BA-2 + CX-4 where m = A. But this process of extrapolation gives results agreeing with the experimental values only when X varies within narrow limits so that the values of n thus obtained must necessarily be somewhat inaccurate and it is found that (n2tro and K have identical values only in exceptional cases. A list is given of the values of d and K for such elements as have been previously investigated and in all cases the differences between the two are considerably less than those observed for complex liquids and solids.These results confirm Schmidt’s view (An%. Physik 1902 [iv] 9 919) that in the elements the molecules have an especially simple structure. The dielectric constants of the non-metals increase with the valency in any horizontal row of the periodic system and with the atomic weight in any vertical column. For helium ?a=1.0000375 or (n2= 1.000075 whilst K has the value 1.000074. For the other rare gases of the atmosphere the calculated values of K (n2) show a continuous increase with the atomic weight. Determinations of the dielectric constants of the metals are insufficient in number to allow of any similar regularity being observed. T. H. P. Quantitative Relations between the Dielectric Constants and other Properties of Substances.DIMITRI K. DOBROSERDOFF (J. Russ. Pl~ys. Chem. Xoc. 1909 41 1385-1406).-The author has investigated the validity of various relations between the dielectric constant and other physical constants which have been suggested by various investigators. It was shown by Obach (Abstr. 1892 158) that for the members of certain homologous series of organic compounds proportiouality exists between the dielectric constant and the latent hea% of evaporation or p/K=const. the actual magnitude of the ratio being different for different series. The author has calculated the value of the ratio for organic compounds belonging to nine homologous series the mean and limiting values found being as follows (1) propionic esters 15.20 (14-84-15.39); (2) butyric esters 15.41 (15*04-15.78) ; (3) valeric esters 15.56 (15.03-16.27) ; (4) ketones 6-00 (4.95-7-1) ; (5) alkyl derivatives of benzene 31.0 (29.73-33.02) ; (6) aliphatic acids 39.55 (39*06-40.05) ; (7) nitriles 5.65 (4.7-7.31) ; (8) amines 25.47 (25*0-26*0) ; (9) alkyl halogen compounds 6.61 (6.32-6-95).The values obtained by Obach were (10) formic esters 11-85 (I 1-03-12.78) ; (11) acetic esters 13.72 (13*26-14*30); (12) ethyl esters of fatty acids 13.34 (11*15-14*20) ; (13) monohydric alcohols and water 7.41 (6.22-7*91) the values for isopropyl (6.14) and cetyl (9.1) alcohols being of doubtful accuracy. The mean of the values obtained for groups (13) (7) (4) and (9) is 6.46; for groups (lo) (ll) (l) (2) (3) and ( l a ) 14.18; for group (S) 25.47 ; for group ( 5 ) 31 and for group (6) 39.55 these numbers 6.46 14.18 25.47 31 and 39.55 being approximately in the ratios of 1 2 4 5 and 6.When a11 the other compounds (inorganic as well) for which the values of p and K are known are included the values of the constant have the approximate proportionate magnitudes 0.5 1 2 3 4 5 6 . The chemical nature of any particular compoundGEKERAL AND PHYSICAL CHEMISTRY. ii. 05 seems to have no appreciable influence in determining t o which group it belongs members of one group being compounds completely different as regards their chemical characters. The relationship between dielectric constant and chemical constitu- tion of a dielectric given by Thwing (Abstr. 1894 ii 374)' and expressed by the equation K= d(a,K1 + a2K2 + .. .)/A4 (where d is the density and M the molecular weight of the compound a1,cc2 . . . the atoms or atomic groups of the same kind oomposing the molecule and Kl,K2 . . . . the dielectric constants of the atoms or groups) is not confirmed by the experimental numbers collected by the author. The relation discovered by Lang (Abstr. 1896 ii 144j for gases namely ( K - 1)106/S=const. = 123(116-145) (S being the sum of the valencies of the atoms constituting the molecule of the gas) holds only,as far as can be ascertained a t 0' and 760 mm. pressure for the six gases obeying Maxwell's law E=n2. T. H. I?. The Conductivity of Mixtures of Dilute Solutions. J. A. GARDINER (Trans. Roy. SOC. Canada 1908 [iii] 2 iii 37-52).-Burton has recently shown that a fall in the conductivity of the solution occurs when dilute hydrochloric acid is added to a colloidal solution of silver or to a dilute solution of silver nitrate. I f the ordinary law of electrolysis were followed a rise in the conductivity should occur.The author's experiments show that as a N/lO,OOO solution of hydrogen chloride is added to water there is a t first very little change in the conductivity of the solution but as the acid solution becomes more concentrated the conductivity gradually increases. On the addition of hydrochloric acid to silver nitrate of various dilutions a drop in the conductivity was observed. When however a 39.5 x 10-7 normal silver nitrate solution was reached this effect disappeared. A similar drop in the conductivity was observed when a dilute nitric acid solution was added to a dilute silver chloride solution. The conductivity of a silver nitrate solution steadily increased with the concentration. It is suggested that the abnormalities here recorded may be due to (1) the absorption of hydrogen by the platinum electrodes ; (2) the hydrogen ion attracting to itself the neutral silver chloride and becoming loaded so that its mobility falls below that of the silver ion which it displaces.E. J. R. Cadmium Chloride Concentration Cells. EUGEN VON BIROF and B. P. APHANASSIEFF (J. Buss. Phys. Chem. Soc. 1909 41 1175-1182).-The authors have measured the E.M.F. of cadmium chloride concentration cells with and without transport of the cation. In the first case silver chloride electrodes prepared by Jahn's method were used and one of the solutions had the same concentration and the other different concentrations for the different cells.The curve connecting the log. of the number of grams of salt per 1 gram of water in the variable solution with the E.H.F. calculated for 1 equivalent of salt per litre in tho constant solutioq exhibits a spread-out maximum for solutions containing 6-8 equivalents of salt. This maximum corresponds with solutions in which the transportii. 96 ABSTRACTS OF CHEMICAL PAPERS. number of the cation is zero. A line drawn parallel to the axis of log. concentrations cuts the curve in two concentrations which would give a cell having a zero E.M.F. although diffusion would occur. If the concentrations of the two 5olutions are below the maximum point on the curve the positive electricity is directed from the more con- centrated t o the more dilute solution; that is with the diffusion current whilst if the concentrations are greater than the maximum the reverse is the case.With cells without transport the curve connecting E.M.F. and log. of the concentration is approximately linear for solutions containing 1-3 equivalents of cadmium chloride but for higher concentrations it becomes mere complex. The curve does not however show any of the peculiar bends observed by Godlewski (Abstr. 1902 ii 445) whose observations are inaccurate owing to irreversible processes at the electrodes of which he took no account. The transport numbers for the anion arid cation are calculated the values for the former agreeing well with those obtained by Hittorf’s method (compare Abstr.1908 ii 145 250). Concentration cells with concentrated solutions hence show the same diffusion phenomena as are observed in those with dilute solutions and Nernst’s theory of diffusion of electrolytes is justified as well for concentrated as for dilute solutions (Zoc. cit.). T. H. P. Rapid Formation of Positive Lead Accumulator Plates. GERHARD JUST PAUL ASKENASY and B. MITROFANOFF (Zeitsch. EZektro- chem. 1909 15 872-892).-The effect of iepeatedly charging and discharging lead plates immersed in sulphuric acid and in mixtures of sulphuric and nitric acids is investigated. I n presence of nitric acid the attack on the plates is much more rapid lead sulphate being first formed.This is subsequently oxidised to lead peroxide. With smooth plates however it appears to be impossible to obtain a sufficiently adherent coating of lead peroxide ; with plates built up of a large number of small lead sheets so as to give a large surface good results were obtained. I n the solutions containing nitric acid lead nitrate is formed at the anode and this is converted into lead sulphate a t some distance from the surface of the plate. This distance increases as the concentration of the nitric acid increases and that of the sulphuric acid diminishes. When it is sufficiently small an adherenk deposit is obtained; when it is greater a loose deposit is formed and when it is greater still a precipitate of lend sulphate is produced in the solution.A solution containing 30 grams of potassium nitrate and 218 grams of sulphuric acid per litre appears to give the best results. T. E. Volatilisation of Cathodes. VI. VOLKMAR KOHLSCHUTTER (Zeitsch. Elektrochem. 19U9 15 930--937).-Tn reply to the criticism of Starck and Fischer (Abstr. 1909 ii 718) the author says that between the value of the cathode fall of potential at which volatilisation begins and the higher value a t which it ceases to increase proportionally to the cathode fall the volatilisation is quite a definite reproducible quantity ; the relation of the quantity volatilisedGENERAL AND PHYSICAL CHEMISTRY. ii 97 to the atomic weight of the metal is too striking to be regarded as accidental. He inclines to regard Starck’s theory of the phenomena as a distinct advance.New experiments with gold and platinum in oxygen and with platinum in air are described. Satisfactory measurements could not be made in these gases with other metals because they react with the gas in circumstances which cannot be controlled. The chemical reaction causes the discharge to become intermittent which is indicated by a telephone in the circuit. In oxygen the regular volatilisation is observed between 440 and 1200 volts cathode fall for gold and between 500 and 900 volts for platinum. T. E. Electrode for Determining the Concentration of the CO,” Ion and the Condition of Silver Carbonate in Solution. JAMES F. SPENCER and MARGARET LE PLA (Zeitsch. anorg. Chem. 1909 65 lO-l5).-Pure silver carbonate is best prepared by adding a dilute solution of sodium hydrogen carbonate to a concentrated solution of silver nitrate stirring continuously.A silvered platinum wire covered with the freshly precipitated carbonate has a constant potential. With pure sodium carbonate solution as the electrolyte the value Ag.Ag2C0,.N/1-C0,” = + 0.7545 volt is found and the concentration of the CO,” ion c’ is found in any solution by tho relation E = 0.7545 -0.029 log. c’. The solubility of silver carbonate in water at 25” is thus found to be 1.16 x 10-7 mol. per litre (compare Abegg and Cox Abstr. 1904 ii 256) and the salt is almost completely hydrolgsed. C. H. D. Investigation of Electrolysis with the Ultramicroscope. J. J. KOSSONOGOFF (PhysikaZ. Zeitsch. 1909 10 976-986).-An arrangement is described by means of which the author has applied the ultramicroscope to the examination of solutions through which an electric current is passed.When a current is started through a solution of silver nitrate or copper sulphate a very considerable increase in the number of bright points in the field of view of the ultramicroscope is at once observed. The view is put forward that these are associated with the ionic carriers of the electric current. I n support of this it is found that ultramicroscopic effects of a special kind are observed when the potential difference between the electrodes is raised to that which corresponds with the tension of decomposition. I n the case of the above-mentioned solutions the attainment of the critical potential difference is accompanied by a very special dis- tribution of the bright points in the neighbourhood of the cathode.A t a short distance from this electrode and parallel to it a zone can be distinguished in which the bright points are very closely crowded together. Between this special zone and the electrode itself is a region which is almost entirely free from bright points. This is supposed to correspond with the dark cathode space in gaseous discharge. The appearance of this effect at a particular voltage affords an -optical method of determining decomposition tensions. Similar effects have also been observed at the anode in other cases. H. M. D.ii. 98 ABSTRACTS OF CHEMICAL PAPERS. Electrolytic Oxidation of Ammonium Carbonate. FRITZ FICHTER and HANS KAPPELEH (Zeitsch. Ekktroclzem.1909 15 937-943. Compare Brochet and Boiteau Abstr. 1909 ii 657).- A solution of ammonium carbonate containing 3.669 gram-molecules of ammonia and 2.494 molecules oE carbon dioxide per litre was electrolysed between platinum electrodes a t temperatures from 20' to 60' and with anodic current densities between 0.027 and 0.731 amperes per sq. cm. Ammonium nitrate is the principal product; neither carbamide nor nitrite could be detected. The experiments were always stopped when a comparatively small quantity of the ammonium carbonate was oxidised. The gases evolved contained carbon dioxide ammonia oxygen and hydrogen; nitrogen W ~ S not found. In an open vessel the yield increases with the current density and with the temperature. I n a closed vessel however very much worse yields were obtained; this is due to the fact that in an open vessel the solution loses ammonium carbonate rapidly and dilute solutions give better yields than more concentrated ones.The best results were finally obtained with a current density of 0.4 to 0.6 ampere per sq. cm. a temperature of 50-60° and a solution containing about 2 mols. of ammonia per litre the ratio COJNH being 0.27; in these cir- cumstances the current eficiency is from 82% to 97%. A solution of ammonium tetraborate gives very similar results. T. E. Electro-catalysis. D. ALEXJ~EFF (J. Buss. Phys. Chern. Soc. 1909 41 1155-116O).-By the electrolysis of ammonium sulphate with lead peroxide electrodes nitrogen oxygen and nitrous oxide are evolved but neither nitrite nor nitrate is formed in the solution.Hence the oxidation of the ammonia does not proceed further than the stage NH(OH) the anhydride of which is nitrous oxide hydroxylamine forming an intermediate oxidatib product. The evolution of nitrogen occurs according to the equation NH,*OH + KH(OH) =N + 3H,O. That the nitrogen and nitrous oxide developed do not arise by the formation and subsequent decomposition of ammonium nitrite and nitrate is shown by experiments on the oxidation of hydroxylamine (vide i..fi*u). With solutions of ammonium sulphate containing sulphuric acid electrolysis with lead peroxide electrodes yields only nitrous oxide and oxygen whilst if the ammonium sulphate solution contains ammonia pure nitrogen is obtained. I n the latter case the form- ation of NH,*OH predominates over that of (NOH) the reaction 2NH,*OH + (NOH) -+ 2N + 4H,O occurring in preference to (NOH) -+ 24-20 + H20.The reaction between lead peroxide and hydroxylamine proceeds according to one or the other of the two equations 2NH,-OH+ PbO = PbO + N + 3H20and2NH,*OH + 2Pb0 = 2Pb0 + N,O + 3H20 according to whether the hydroxylamine or the peroxide is in excess. I n practice as is shown by a study of the products obtained when the reaction takes place in absence of air a mixture of the two gases is always obtained; the amounts actually formed in two cases were (1) 22% N,O and 78% N and (2) 30% N20 and 70% N,. Thus lead peroxide is a typical catalyst with the peculiarity thatGENERAL AND PHYSICAL CHEMISTRY. ii. 99 one-half of its action is a chemical and the other a physical process The formation of nitrogen and nitrous oxide a t lead peroxide electrodes is hence a typical electro-catalytic process.T. H. P Manganese Aluminium and Copper. FRIEDRICH HEUSLER and FRANZ XICHARZ (Zeitsch. anorg. Chem. 1909 65 110-1 12. Compare Ross and Gray Abstr. 1909 ii 859).-The fact that certain manganese-aluminium bronzes heated above 200' and slowly cooled become strongly magnetic but have considerable hysteresis has been observed previously by Heiisler and by Asteroth. C. H. D. Magnetic Dichroism of Siderite in Liquids. GEORGES MESLIN (Comnpt. rend. 1909 149 855-S57. Compare Abstr. 1909 ii 529). -Siderite suspended in carbon disulphide or aniline exhibits magnetic dichroism t o such a high degree that the phenomena can be seen with an ordinary permanent magnet or the raidual magnetism of an electromagnet. The suspensions are also slightly dichroic spontaneously.A method of demonstrating dichroism by the optical lantern is described. R. J. C. Nagnetism of Solutions. PAUL DRAPIER (J. Chim. Php. 1909 7 385-404. Compare Pascal Abstr 19OS ii 927)-The author has examined the behaviour of a number of solutions when placed in a flat vertical cell between the pointed poles of a powerful electro- magnet. If the liquid has paramagnetic properties it tends to move radially in a plane perpendicular to the lines of force and thus forms a tumulus or convexity at the surface. This convexity is much accentuated in aqueous solutions when the surface-tension is lowered by addinga layer of ether or benzene. An aqueous solution of ferric alum or ferric chloride shows a convexity even when only 1% of ferric salt is present but n solution of the same concentratiorl in ether is unaffected.Ferric ammonium oxalate is but slightly affected and colloidal ferric hydroxide potassium ferrocyanide and potassium ferricyanide are unaffected . If a solution ol ferric chloride in ether is floated on water a striated layer is produced which in the field bends downwards into the water the striae being displnced horizont,ally. Displacements are also observable if precipitated ferric hydroxide or air bubbles are suspended in the p warnapetic fluid. Dilute manganese sulphate gives a marked convexity but potassium permanganate none. When ether is poured on the latter solution an intermediate layer is formed containing ether water manganese sulphate and precipitated manganese dioxide which is extremely sensitive to the magnetic field owing to manganese sulphate and perhaps to free oxygen occluded in the precipitate. Cobalt and nickel hydroxides precipitated by ammonia are also very sensitive to the magnetic field.Chromium sulphate cobaltous chloride and nickel nitrate areii. 100 ABSTRACTS OF CHEMICAL PAPERS. paramagnetic ; potassium dichromate potassium cobulticyanide titanium sulphate and platinum chloride are not. The author's observations confirm the conclusion arrived at by Pascal by a different method that as a magnetic metal becomes more and more removed from its normal (ionised) state it loses its para- magnetism. The magnetic capillary rise of solutions of ferric alum and ferric chloride of 5 7$ and 10% strengths was investigated.On increasing and decreasing the exciting currents between 0 and 10 amperes marked hysteresis was found in the capillary rise apart from the usual hysteresis of the magnet. This effect is supposed by the author to be true liquid hysteresis. R. J. C. Use of the Magnetic Field as a means of Determining Con- stitution in Organic Chemistry. 11. and 111. PAUL PASCAL (BUZZ. Xoc. clhim. 1909 [iv] 5 1110-1118 ; 1910,7 17-28. Compare Abstr. 1909 ii 487 788 859).-Part of this work has been published already. From comparisons of the magnetic susceptibility of oxygen in a series of oxygenated carbon compounds the conclusion is drawn that the value is - 48 x 10-7 where oxygen is joined to two different carbon atoms - 35 x 10-7 where it is doubly linked to a carbon atom the latter being itself joined to two oxygen atoms (as in carboxylic acids) and + 1s x 10-7 where a single oxygen atom is doubly lioked to carbon (as in aldehydes and ketones).A compara- tive list of (1) experimental molecular magnetic susceptibilities and of (2) values calculated from the data given above shows close con- cordance. The application of these rules to the case of paraldehyde lends support to the KekulB formula for this substance. Apart from the above effect due to the method of linking of oxygen the value of the magnetic susceptibility of the latter is also influenced by the general structure of the rest of the molecule and especially by the presence of (1) tertiary or quaternary carbon atoms (2) double linkings.The first of these effects is marked when the disturbing atom is in position a or y and more so in positions 6 and c but is very small in p t or 7 and ceases beyond position 8. The presence of a double linkage shows itself in an analogous manner and a table of corrections for the effect of double linkings in several positions is given. The second group of conclusions lends support to Bayer's strain hypothesis provided the carbon chain is regarded as having a roughly spiral form. The influence of the hexamethylene nucleus on the magnetic suscep- tibility is estimated at +31 x 10-7 and with this correction the calculated value for cineol according to Briihl and Wallach's formula agrees with that determined experimentally.and retains this in most of its organic derivatives but in thioacetic acid the atomic susceptibility of the oxygen atom is - 15 x lo-' as against - 35 x 10-7 for the same oxygen atom in acetic acid the greater effect in the former case being due to the presence of the sulphur atom. The normal value for nitrogen is - 58 x lW7 but in a cyanogen group Sulphur has the value - 156 xGENERAL AND PlXYSICAL CHEUISTRY. ii. 101 or where nitrogen is directly attached to a benzene nucleus the value becomes - 48 x I n closed chains containing carbon and nitrogen the latter has the same value as in aromatic amines and for purposes of calculation the value for one CH (or CK) group is replaced by t h a t for NH (or N) with the usual total correction for the influence of the nucleus.This rule is not applicable in the case of pyridine. T. A. H. Thermometers as Thermo - regulators. ERRARD GLASER (Biochem. Zeitsch. 1909,23,5-9).-1nto the thermometer which is used as thermo-regulator platinum wires are fused at certain definite points corresponding with temperatures the constancy of which it is desired to maintain. The lowest platinum wire is always in contact with the mercury when the thermometer is immersed in the apparatus the temperature of which is to be regulated. By means of this and another platinum wire corresponding with the temperature which is to be maintained in the thermostat an electric circuit is made with a coil in which is immerded a Hahn regulator. As soon as the mercury reaches the higher poict and the circuit is closed an iron core in the Hahn regulator is drawn down by the current and shuts off the supply of gas t o the burner heating the thermostat.As the latter cools the gas is automatically lighted again by means of a by-pass. The apparatus is figured in the paper. S. B. S. Krafft’s Boiling - point Estimations and his Theory of Volatilisation. C. VON RECHENBERG (J. pr. Chem. 1909 [ii] 80 547-555. Compare Abstr. 1909 ii 544).-Largely polemical in reply to Krafft (Abstr. 1909 ii 969 and Hausen ibid. 969). It is pointed out that the expression used by Krafft b. p./O mm. is a contradiction as if sufficient vapour is present for the temperature to be determined there must be a vapour pressure. The effects which are attributed by Krafft to the influence of gravity are regarded by the author as due to the condensation of vapour by external cooling.J. J. S. Preparation of a Mixture of Constant Boiling - point and Maximum Vapour Pressure by Distillation. D. D. GADASKIN and A. E. MAKOVETZKI ( J . Buss. P/ys. Chern. Soc. 1909 41 1160-1 163).-The authors describe experiments on the distillation of various aqueous solutions of the ether of methylene glycol under ordinary and reduced pressure (compare Abstr. 1908 i 753 ; 1909 ii 215). The resulbs obtained are discussed by Ma ovetzki (see following ex tract). Determination of the Composition of Constant Boiling-point Mixtures having Maximum Vapour Pressures and their Quantitative Separation by Distillation. A. E. MAKOVETZKI (J. h?uss. Phys. Chenz. Soc. 1909 41 1171-1175.Compare preceding abstract).-The author adduces further evidence in support of his view that a binary liquid mixture for which a maximum or minimum vltpour pressure exists may be regarded as consisting of two components one T. H. P. VOL. XCVIII. ii. 8ii. 102 ABSTRACTS OF CHEMICAL PAPERS. Deing the mixture of maximal or minimal vapour pressure and the other the component present in excess. By means of aqueous solutions of the ether of ethylene glycol it is shown that the mixture of maximum vapour pressure can be separated quantitatively by one distillation in a suitable fractionating apparatus. The quantity and composition of the mixture with maximum vapour pressure are not altered by the addition t o the solution of a nou-volatile substance which effects the separation of the liquid into layers but does not give a solid phase.T. H P. An Electrical Apparatus for the Direct Determination of the Water Value of a Calorimeter. W. ~WIFTOS~AWSKI (Bull. Acad. Sci. Cracow 1909 548-555).-The principle of the method consists in using two calorimeters which are heated by means of a n electric current. The one calorimeter (chief calorimeter) is filled with the solution the water value of which is required and the other (water calorimeter) with water. The heating is accomplished by means of a platinum wire spiral which is fused in between two concentric layers of glass. The two vessels are of tho same construction. The heat coefficient a is calculated from the formula a = A/l,(500 -t c1 + cl’)/AF2(500 + c2 + c2’) where AT and AT denote the increases in temperature c1 and c2 the water values of the calorimeter with stirrer and thermometer cl’ and c2’ the water values of the heating apparatus and where each calorimeter contains 500 grams of water. The sp.heat K of any liquid can then be calculated from the equation I<= aAT2/Al; where the increases in temperature of the two calorimeters are AT and AT;. Atomic Volume of Allotropic Modifications at Very Low Temperatures. EKNST COHEN and J. OLIE jun.* (Proc. K. Akud. Wetensch. Amsterdam 1909 12 437-445).-1n order to obtain information relating t o the densities of allotropic modifications at absolute zero measurements of the densities of diamond and graphite and of white and grey tin mere made at a. series of temperatures by a dilatometric method.The graphite was subjected t o pressures of 1000 to 5000 atmospheres until the sp. gr. remained constant after repeated compression. ‘l’he ratios of the specific gravities of diamond and graphite were found to be 1.585 1.583 and 1.582 at 1 8 O - 3S0 and - 164O respectively; those of white a n d grey tin 1.266 and 1.274 a t 78’ and - 164’. These numbers indicate that the specific volumes of the allotropic forms do not converge as the temperature falls. H. M. D. Associated Liquids. W. A. KURBATOFF and G. G. ELISEEFF (J. Russ. Plqs. Chem. SOC. 1909 41 1422-1425. Compare Abstr. 1909 ii 117 120).-In order to throw light on the abnormal values of the Ramsay-Shields constant given by certain apparently normal liquids the authors have examined acetic anhydride and ethyl malonate in this connexion.* also Zcitsch. yhysikcd. Cllem. 1910 Vl 385-400. J. J. S.GENERAL AND PHYSICAL CHEMISTRY. ii. 103 Acetic anhydride for which Trouton’s constant has the high value 22.9 is stated to give the normal value of the Ramsay-Shields constant K = d(yNw2’’)/dt = 2.129 corresponding with a non-associated liquid. This value of IT is confirmed by the authors’ measurements which give the mean result 2.12 +_ 3%’ in open and closed vessels. A whole series of esters are known which are normal as regards their thermal data and their values of Trouton’s constant but give abnormally high values for the Ramsay-Shields constant. This is also found to be the case with ethyl malonate which gives values of K varying from 2.20 to 2-56? The results obtained indicate that the value of the Ramsay-Shields constant varies for different homologous series.T. H. P. Association of Glycerol. G. G. ELISI~EFF and W. A. KURBATOFF (J. Rzcss. Phys. Chem. Soc. 1909 41 1426-1427).-The values of the Ramsay-Shields constant for glycerol at various temperatures are as follows 0.63 at 35*2-66-S0; 1-10 at 64%-74*3°; 1.50 at 74.3-101°40 snd 1.20 at 101.4-123.4” the alteration with tempera- ture being almost identical with that exhibited in the case of ethylene glycol. It hence appears that the degree of association of glycerol is not less than that of ethylene glycol. These results give no reason for expecting that the molecule of sugar is a simple one But if the sucrose molecules are associated as indeed all molecules containing hydroxyl groups appear to be then the laws of osmotic pressure derived from a study of sucrose solutions should be modified and there is exhibited also a possibility of an explanation of the coefficient i other than that given by Arrhenius.T. H. P. Adsorption of Ions. V. BOURNAT (Compt. Fend. 1909 149 1366-1368).-Dilute solutions of binary electrolytes excluding acids have a higher surface-tension than water; acids however lower the surface-tension. The action of acids is probably due to the accumula- tion of hydrions in the superficial layer giving the same effect as in the capillary electrometer. I n sopport of this explanation it is shown that the addition of a hundredth molecular weight of potassium ferro- cyanide t o a litre of NIS-nitric acid increases the surface-tension to a considerable extent whereas when added to a N/5-solution of a binary salt the increase is only very slight.Perrin has shown that multi- valent ions diminish considerably the charge in a double layer that is that they accumulate in the surface layer and in the experiment mentioned above they displace the hydrions and thus raise the surface- tension. Comparing equimolecular solutions the curve showing the relation between the molecular weights (abscissze) of binary salts and the diff ereiices in the surface-tension (ordinates) of their solutions from that of water is a straight line Monobasic acids also give a straight line lying below and parallel to that for the salts the difference in the ordinates being 0.35 absolute unit for N/lO-solutions.Sodium and potassium hydroxide raise the surface-tension of water but not t o the same extent as binary salts; in this case also a sbraight line is 8-2ii. 104 ABSTRACTS OF CHEMICAL PAPERS. obtained the difference in the ordinates being 0.25 unit. These differences are apparently proportional t o the molecular concentrations of the solutions. Assuming that the diminution in the surface-tension is due to the accumulation of ions in the surface layer it is shown that the number (n) of ions absorbed per unit surface is given by the formula ~ = u c ! where c is the concentration and p is a constant depending on the ion. This formula is analogous to the general adsorption formula of Freundlich. T. S. P. Adsorptive Power of Hydroxides of Silicon Aluminium and Iron.111. Adsorption by Clay. 11. PAUL ROHLAND (Zeitsch. anorg. Chern. 1909 65 108-109 ; Biochein. Zeitsch. 1909 23 278-280 Compare Abstr. 1909 ii 27 55l).-A property of a Fraustadt clay the analysis of which is given is to adsorb unsaturated hydrocarbons when it has imbibed its maximum quantity of water. It is however impermeable to saturated hydrocarbon. By means of this clay the unsaturated hydrocarbons can be separated from the saturated in American petroleum. The hydroxides of clays of this description (silicon aluminium iron and titanium) can adsorb organic substances containing oxygen such as alcohol and acetone but prevent the diffusion of organic substances such as carbon disulphide toluene etc. and hydrocarbons which do not contain oxygen with the exception of the unsaturated hydrocarbons. S.B. S. Chemical Dynamics and the Colloidal State. I 11 and 111. ALBERT REYCHLEK. (J. Chim. Phys. 1909 7 365-368 497-510. Compare Biltz Abstr. 1904 ii 324 393).-The experi- ments made by Biltz on the removal of arsenious acid fromits solution by shaking with colloidal ferric hydroxide led to the conclusion that the amount of acid removed (x) was related to the amount remaining in solution (a - x) by the equation x5 =]<(a - x). The phenomena were attributed to adsorption but were not further investigated mathematically. The amounts of ferric oxide used by Biltz in all his experiments were sufficient to form a normal arsenite with 1.4 grams of arsenious oxide whereas the amount of arsenious oxide actually adsorbed never exceeded 0.824 gram even when almost four equivalents were available.If a normal arsenite is produced its concentration would be represented by xl1.4 that of the free ferric hydroxide by (1.4 - x)/1.4 and of the free arsenious oxide (a - 2)/3. Assuming that normal ferric arsenite is hydrolysed in the usual manner for a salt of a weak acid and a weak base the eqoilibrium will be (~/1*4)~ = X3((1.4 - ~)/1*4)~(a - x)/3,whence If on the other hand only two of the basicities of arsenious acid are exercised the constant is 1C2 = Jx"(a - x) It is shown that K satisfactorily expresses Biltz's values up to the point where about one-half of the theoretically possible arseriious oxide is combined and that above this K2 gives a constant pointing to the formation of some acid arsenice.K,= ~Z'/(CG-Z) + (1.4-x). (2 x 1.4 - x).GENERAL AND PHYSICAL CHEMISTRY. ii. 105 If the equivalent amount of arsenious oxide in Biltz’s experiments is taken as 3.2 instead of 1.4 an even better constaat (K3) is obtained. It is not certain whether alumina has a similar a6nity for arsenious oxide. The author suggests that all so-called adsorption phenomena regu- lated by equations of the form d / ( u - x) = K may eventually be made amenable to the ordinary laws of chemical dynamics. The experiments of Freundlich on the adsorption of various acids etc. by blood-charcoal (Abstr. 1907 ii 155 939) can be considered as cases of chemical combination. Since however the basicity of charcoal is unknown i t is necessary to assume that 1 gram of charcoal is capable of combining with n-milli-equivalents of acid m-Grams of charcoal can therefore combine with mn-milli-molecules of a monobasic acid. If x milligram-molecules of acid be adsorbed the degree of saturation of the charcoal is elinn and the unsaturated charcoal is (mn - x)/mn.If cc is the initial concentration in the solu- tion (a - x) is the final concentration. Assuming that the adsorption compound is hydrolysed like a salt of a weak acid with a weak base (x/mn)2= K,(rnn - x)/rnn x (a - x) whence Kl = ( ~ / r n ) ~ / ( a - x)(n - x/m). By choosing a suitable value for n a very satisfactory constant K can be obtained. The values are Acetic acid ............... Propionic acid ........... Dichloroacetic acid.. ... Formic acid ............... Butyric acid ............Chloracetic acid ......... Benzoic acid ............ Sulphanilic acid ......... n=6 n = 5 n=6 n=S 71 = 6 n=4 12=4 n=5 K~ = 0.043 irl = aeoza K,= 0.30 K1=0*139 Kl = 0.21 Kl =0*33 K1 = 24 Kl = 0 -21 The strong acids trichloroacetic acid and benzenesulphonic acid do not give satisfactory constants ; it is supposed t h a t their adsorpticin compounds are not hydrolysed according to the same law. I n the case of dibnsic acids with only one active valency R= ( ~ / m ) ~ / ( a - 2)(2n - x/m) whereas if both valencies be active X2 = J / ( ~ / r n ) ~ / ( a - x) 4 (n - z/m). With succinic acid both valencies are active n= 10 and K2=0.32. Citric acid acts as a tribasic acid R = 0.16 when n = 15. Bromine appears to follow the same adsorption lnw as monobasic acids whereas methy lamine follows a simple partition law.Adhesion dissolution etc. mas in many cases superpose their effects on chemical adsorption. The author’s hypothesis is based on the theory that a large number of colloidal solutions may be considered as strongly basic or acidic salts. For instance the small proportion of hydrochloric acid which stabilises a solution of ferric hydroxide acts by combining with all the hydroxyls in turn thus preventing tbe hydroxide from forming large complexes which on dehydration would be precipitated. For every degree of dilution of the colloid sol there is a correspond- ing minimum of acid to prevent precipitation. All the known agentsii. 106 ABSTRACTS OF CHEMICAL PAPERS. for precipitating ferric hydroxide can be explained to act by disturbing the chemical equilibrium of acid water and ferric hydroxide. Colloidal silica can be considered as a very acidic.sodium silicate and its properties can all be explained on chemical grounds. The phenomena of cataphoresis and anaphoresis the precipitation of colloids byelectrolysis may be due to the transfer of water to anode or cathode respectively the colloid appearing to travel in the reverse direct ion. R. J. C. Adsorption of Arsenious Acid by Ferric Hydroxide. WILHELM BILTZ '(L Chim. Phys. 1909 '7 570-574).-1t is shown that Reychler's explanation (preceding abstract) of the author's observations relating to the absorption of arsenious acid by ferric hydroxide is untenable. This explanation is based on the assumption of the formation and hydrolytic decomposition of ferric arsenite.New experiments have been made in which varying quantities of the hydroxide were shaken up with the same volume of a solution of arsenious acid of determined concentration. IF x denotes the quantity of arsenious oxide taken up by rn grams of the hydrogel and x is the quantity which remains in solution the observed results can be satisfactorily represented by the equation log z/m = 0937 log x + log k. From this the author concludes that the removal of arsenious acid from the solution by the ferric hydroxide is a pure adsorption phenomenon. H. M. D. Thermodynamics of the Capillary Layer. GERRIT BAKKER (Zeitsch. physikal. Chem. 1909 68 684-692).-8 mathematical paper. The author indicates certain errors very often committed in applying thermodynamical considerations to the capillary layer.When the capillary layer is considered by itself instead of the usual equation d& = d e - Hds (where dQ is the heat absorbed in varying the surface d e is an energy difference H is the surface-tension and ds the change of surface) the equation d& = d e +pNdv - Hds must be used where p is the vapour pressure and v the specific volume of the capiilary layer. The last equation is the correct expression for the specific heat of the capillary layer per unit of mass. Nothing is known as to the variation of this speci6c heat with temperature. The energy equation obtained when a vessel filled with liquid vapour and the capillary layer as transition layer is considered differs from that deduced for a thin sheet.G. 5. Relationship between Physical Properties of Solutions. I. Density aad Electrical Conductivity of Aqueous Solutions of Salts. ADOLF HEYDWEILLER (Aim. Physik 1909 [iv] 30,873-904). -From an examination of the data for a large number of aqueous solutions of electrolytes it is shown that a connexion exists b$ween the density of a solution and that of the solvent which can be expressed by the equation A = B + ( A - B)i. I n this equation A denotes the percentage change in density per gram-equivalent of the dissolvedGENERAL AND PHYSICAL CHEMISTRY ii. 107 electrolyte i is the ratio of the equivalent conductivity of the given solution to the conductivity at infinite dilution and A and B are constants. A and B represent respectively the percentage changes in density which are caused by one gram-equivalent of ionised and non- ionised electrolyte ; A - B represents the influence of ionisation on the density of the solution. The values of A and B are tabulated for a number of electrolytes.For certain electrolytes the values of 3 indicate that the volume of the undissociated electrolyte in solution is the same as that of the salt in the solid State. I n other cases changes in volume take place on solution. The contractions which are found in the case of salts which form hydrates in the solid state indicate that hydrated molecules are also present in the aqueous solutions. Ionisation of the electrolyte is always accompanied by an increase in density and this is found to become greater as the sum of the mobilities of the constituent ions increases.The observed contrac- tion is shown to be probably due to a diminution in the volume of the water. The values of A exhibit additive relationships and ionic moduli are calculated by means of which it is possible t o calculate the influence of the ionised portion of any electrolyte on the density of its aqueous solution. Certain electrolytes are abnormal in t h a t they do not agree with the relationship A = B + ( A - B)i. The anomalous behaviour is traced in some cases to the formation of complex ions in the more concentrated solutions and to the large affinity of the dissolved salts for water. H. M. D. Condition of Equilibrium between a Dilute Solution and the Pure Solvent Separated by a Semi-permeable Diaphragm or by the Vapour of the Solvent.GIOVANNI GUGLIELMO (Atti R. Accad. Lincei 1909 [v] 18 ii 536-544).-Making use of two relations which mere obtained by van der Waale (see Die Continuitat des Jussigen und gasformigert Zustccndes) and which express the condition of equilibrium of a large number of molecules (considered as material points) in perpetual motion and attracting one another the author derives (1) in two forms the condition of equilibrium of the molecules of a chemically homogeneous liquid and (2) the equilibrium conditions for a chemically heterogeneous liquid-solvent and solute ; (3) the condition of equilibrium between pure solvent and solution separated by a semi-permeable surface. With the aid of the results thus obtained the following questions are discussed independence of the molecular attraction on the mass of the molecules and a hypo- thesis on the nature of this attraction; causes of the lower vapour pressure of solutions compared with the solvents and of the equality of vapour pressure for equimolecular solutions ; influence of the curvature of the surface of a liquid on its vapour pressure.T. H. P. Remarks on THOMAS S. PATTERSON (Zeitsch. physihl. Chem. Compare Dolezalek Abstr. 1909 ii 22).-It is Binary Mixtures and Concentrated Solutions. Doleartlek’s Paper. 1909 67 572-574.ii. 108 ABSTRACTS OF CHEMICAL PAPERS. shown that the theoretical densities of mixtures of chloroform and acetone calculated by Dolezalek are incorrect as they were obtained by multiplying the respective densities of the components by the molecular fraction of the component in the mixture instead of by the usual method The observed density of n particular mixture of chloro- form and acetone only differs very slightly from the theoretical value ; the deviation is only one-ninth of that given by Dolezalek and the conclusions of the latter investigator therefore require revision.G. S. Existence and Properties of Disperse Systems in the Region Separating Colloidal and Crystalloidal Solutions. THE SVEDBERG (Zeitsch. Chern. Ind. Kolloide 1909 6 3 18-325. Compare Abstr. 1909 ii 389).-Experiments are described which show that the absorption of light by a gold hydrosol increases as the size of the colloidal particles increases. By raising the temperature of a ruby- red hydrosol or by the addition of electrolytes the particles were caused to coagulate and measurements of the colour intensity showed a gradual increase in the absorptive capacity of the hydrosol.The addition of a non-electrolyte was found to be without influence on the absorption. By the reduction of a solution of gold chloride by means of hydrazine in presence of gelatin (free from electrolytes) as protective colloid gold hydrosols consisting of extremely small particles can be obtained. The intensity of the colour of the hydrosol obtained in this way is much smaller than that of the hydrosol obtained under similar conditions in the absence of the protective colloid. The action of this consists in reducing the rate of coagulation of the hydrosol and it is shown that the activity of the gelatin is approximately pro- portional to its concentration.Gold hydrosols prepared in different ways exhibit considerable differences in respect of the position and the intensity of the absorp- tion maximum. This is found to depend on the size of the colloidal particles. As the size diminishes the absorption maximum shifts to- wards the region of smaller wave-lengths. For the most highly disperse hydrosols this maximum is in the ultra-violet and approxi- mates to the position of the maximum for a solution of gold chloride. H. M. D. Theory of Colloids. JACQUES DUCLAUX (J. Chim. phys. 1909 [vii] 405-446. Compare Duclaux Abstr. 1909 ii 303 ; Malfitano ibid. 473 ; Pappada ibid. 473)-The author develops the theoretical ideas already put forward by him into a complete theory of colloids.The physical theory which postulates that the stabilising ions in a colloidal solution are permanently combined with the colloid particles and that the whole osmotic pressure is due t o colloid particles acting as molecular units fails to explain the difference in properties when one stabilising ion is substituted for another. According to the chemical theory put forward by the author colloid particles are very large multivalent ions forming salts with the stabilising ions which surround them but which are capable of super-adding their osmoticGENERAL AND PHYSICAL CHEMISTRY. ii. 109 pressures conductivities etc. t o that of the nucleus granules or micella (compare Reychler this vol. ii 105). The degree of ionisation will vary with the nature of the stabilising ion.For instance ferric hydroxide stabilised with sulphuric acid has a much lower oemotic pressure than the same colloid stabilised with hydrochloric acid. It is shown that if the colloidal micella is assumed to be exactly comparable to an ionisable salt the osmotic pressures calculated from the conductivities and ionic mobilities of the solutions are about double the experimental osmotic pressures. The measurements mere all made on solutions with a very pure intergranular liquid (compare however Malfitano Zoc. cit.) and a small correction was subtracted for the con- ductivity of the intergranular liquid. The ionic mobilities of the colloid granules were determined by passing a direct current through the solution and afterwards analysing the liquids in the anode and cathode chambers it being assumed that no transference of water had occurred.The ultramicroscope is not available here because colloids with an appreciable osmotic pressure are so small as to be almost invisible. When a colloidal solution is dilute the micella are so widely separated that the ions surrounding each one are never attractad from it but the whole comprises a stable almost neutral sphere. On establishing an electric field the ions all crowd to one side of the sphere and the parent granule to the other. Since the voltages required for electrolysis are small it follows that only a few of the ions become detached. When the solution is concentrated the micella approach each other and ultimately their neutral spheres intersect.Each micella will then facilitate the ionisation of its neighbours a kind of Grotthus’s chain being set up. It follows that as a colloidal solution is concentrated its ‘‘ molecular ” conductivity increases. This is showri to be the case with ferric hydroxide and gum arabic. The degree of ionisation of a colloid does not mean the proportion of granules ionised but the average extent to which each granule is ionised. Knowing the number of stabilising ions the degree of micella ionisation can be calculated. It varies in the author’s experi- ments from 0-008 in copper forrocyanide to 0.88 in gum arabic. If the intergranular liquid contains electrolyte this must have an ion in common with the micella and will influence its ionisation. When the degree of ionisation of the micella the ionic velocity and the viscosity are known the radius of the micella can be calculated with the aid of Stokes’ theorem.This varies from 0 . 5 5 ~ ~ in tungstic acid to 5 . 2 ~ ~ in Prussian-blue the number of free ions per micella being 2.9 and 24 respectively. These values are in accord with the relative retoention of the colloids by collodion acd are also of the same order as the radius < l p p found by Zsigmondy for colloidal gold particles. The molecular weights of the colloids calculated from their micella radii are tungstic acid 1900 thorium hydroxide 7000 gum arabic 16,000 ferric hydroxide 115,000 copper ferrocyanide 700,000 Prussian-blue 1,000,000. The neutral sphere probably has a radius about ten times that ofii.110 ABSTRACTS OF CHEMICAL PAPERS. the micella that is to say on an average an ion does not move further than 10pp from the micella. Direct measurements have been made of the osmotic pressures of ferric hydroxide copper ferrocyanide Prussian-blue thorium hydr- oxide gum nrabic and caramel with collodion membranes. The results confirm those obtained by the filtration method (Abstr. 1909 ii 303) and clearly show that the osmotic pressure of a colloid increases more rapidly than the concentration. In very dilute solution each micella with its ions complete acts osinotically as one molecular unit. As the concentration increases it may attain by splitting off of ions an osmotic value of S or more. R. J. C. General Equation of State. KARL DRUCKER (Zeitsch. physikal. Chem.1909 68 616-636).-A theoretical paper. A general equation of state is deduced on the assumption that the gas laws hold for gases and pure liquids in general and that the deviations are to be accounted for on purely chemical grounds that is on the formation of complex molecules (polymerides). On this basis the general gas equa- tion pv = RT.8n leads to the equation p/RT= 8n/v = Sc = el + c2 + c3 + . . . . (l) where 8 c = c1 + k,c12 + k,cI3 + . . . . I n this equation cl c2 etc. represent the respective partial concentrations of the simplex and complex molecules ; k k etc. represent the respective equilibrium constants and the other symbols have the usual signi- ficance. It is assumed in deducing this equation that equilibrium between simple molecules and their polymerides is established instantaneously.I n equation (1) the coefficients are necessariIy positive and it appears at first sight as if it applies only to cases where the com- pressibility is too great. Although this difficulty can theoretically be got over it has been considered advisable to insert a volume correction (analogous to that of vim der Waals) in equation (1). The equation is then tested by application to the data for ethyl ether methyl alcohol and other vapours given by ltamsmy and Young and others and simplified forms of it are found to give satisfactory results. The application of these considerations to liquids leads to the con- clusion t h a t a liquid under ordinary conditions consists of a dilute solution of unimolecules in complex molecules.The fact that this result is in apparent conflict with the method of determining the molecular complexity of liquids due to E6tvos-Ramsay-Shields based on surface-tension measurements is not regarded as an insuperable objection as the method in question has no purely thermodynamic basis. The author considers that liquids such as water usually regarded as complex are really comparatively simple. Provisional suggestions are made for determining the partial concentrations in liquids. G. S. Demonstration of the Phase Rule. R . BOULOUCH (Compt. rend. 1909 149 1377. Compare Abstr. 1909 ii 802).-The author maintains his criticisms of Miiller's demonstration of the phase rule,GENERAL AND PHYSICAL CHEMISTRY. ii. 111 pointing out that Muller has confused three things which are essenti- ally distinct namely (1) the actual changes taking place in a system not in equilibrium; (2) the virtual changes which one imagines to take place in a system in equilibrium ; (3) the atomic interchanges occurring according to the atomic theory in a system in equilibrium.T. S. P. Invariant Systems and the Regularity of Composition of Certain Eutectics. ALEX. GORBOFF (J. Buss. Phys. Chem. Xoc. 1909 41 1241-1300).-The author discusses the phase rule and the composition of eutectic mixtures a large number of examples from the work of various investigators being considered. The principal results arrived at are as follows. One of the fundamental propositions of chemical mechanics is that the laws ‘by which material systems are characterised are determined not by the number of their components but by the number of effective degrees of freedom.Examination of the compositions of eutectics formed by the elements and by chemical ‘‘ molecular,” and ‘‘ complex ” compounds shows that the composition of any eutectic corresponds with a chemical compound formed by the elements occurring in the eutectic which may hence be expressed by a chemical formula with rational indices. So that eutectics obey not only the law of constant composition but also the law of multiple proportions. Not only may the composition of a eutectic formed by two independent components capable of giving chemical compounds melting without decomposing be expressed by a chemical formula but this formula is often constructed according to a simple rule-equal masses of one of the two components being distributed in both solid phases of the eutectic.Thus if the two components A and B form the compounds A + xB and A + yB melting without decomposing then the eutectic between A and A +xB is expressed by the formula 2 A +xB and that between A + x B and A + y B by the formula ( g A + x y B ) + (d + q B ) or ( x + g ) A + 2xgB and so on. Excluding the limiting eutectics answering to the general formula 2A +xB in all the others expressed for example by (x + y)A + 2xyB in both solid phases that independent component is distributed in equal masses which possesses the more basic chemical character. The latter in the euiectics formed by crystallo-hydrates of sulphur dioxide hydrogen chloride hydrogen icdide nitrogen pentoxide copper nitrate magnesium chloride ferrous nitrate and ferric chloride is water and in the metallic eutectics the more alkaline metal.I n every case where a chemical compound of two components does not melt without decomposing but exhibits a transition point below the melting point its solubility in one of the components is lower sometimes very considerably lower than is required by the above rule; in such cases this component consequently occurs in excess in the eutectic. For the large numbers of experimental data from which these conclusions are drawn and for references given the original must be consulted. T. H. P.ii. 112 ABSTRACTS OF CHEMICAL PAPERS. Influence of Centrifugal Force on the Equilibrium of Chemical Systems. A.V. DUMANSKY (J Buss.Phys. Chem. Soc. 1909 41 1306-1308).-When a concentrated solution of cadmium iodide is subjected to centrifugalisation in a tube closed with a cork a brown precipitate containing cadmium and iodine is deposited whilst aqueous hydrogen iodide under similar conditions gives a deposit of iodine. The author's results seem to indicate that the cork acts as a catalyst. Also centrifugalisation of solutions of ferric chloride and mercurous nitrate produces a marked increase in the electrical conductivity of the solutions the increased values persisting after removal of the centrifu- gating force and mixing of the liquids. Since all the compounds used in his experiments are readily decomposed by water the author suggests that such decomposition may occur to a slight extent and that one of the products of the decomposition being the heavier may be readily removed from the sphere of action by the centrifugalisation.Thus with ferric chloride the ferric hydroxide formed by the hydrolytic dissociation would pnss to the periphery of the centrifuge and thus permit of the hydrolysis of further quantities of ferric chloride ; the hydrochloric acid formed by the hydrolysis would cause the increased conductivity observed. Colloidal solutions of antimony sulphide and ferric hydroxide deposit precipitates when subjected to centrifugalisation. T. H. P. Chemical Affinity. 111. Solution-affinity of Binary Systems. 11. Sulphuric Acid and Water. J. N. BRONSTED (Zeitsch. physikal. Chem. 1909 68 693-725).-The theoretical conclusions discussed in the previous paper (compare Abstr. 1909 ii 29) are now tested by application to the system sulphuric acid-water. The heat of admixture of sulphuric acid and water has been determined over the whole range of concentrations and as large amounts of the substances mere used the results are probably very accurate ; they are represented in tabular form in various ways.The heat of formation of 1 mol. of monohydrate is 6710 cal. The m. p. of pure sulphuric acid on the hydrogen scale is 10.49". The results obtained are in good agreement with those of Pfaundler and of Pickering (Trans. 1890 57 94) but not with those of Thomsen (Thermochemische Untersuchungen) ; i t is probable that Thomson's "pure" acid contained a little water. From the results the differential curves for the heats of admixture are determined by means of the equations given in the earlier paper.The solution-affinity of the components throughout the whole range of concentrations is then determined from the combined results of i7.M.X measurements of vapour-tension measurements and of freezing-point determinations. The reaction the E.M.F. of which has been determined is the formation of sulphuric acid by the reduction of mercurous sulphate by hydrogen the cell being built up as follows H I H,SO I $g,SO I Hg. Measurements have been made between 15" and 80" with varying proportions of acid and the affinity is calculated from the results by means of the Helmholtz equation in the usual way.GENERAL AND PHYSICAL CHEMISTRY. ii. 113 As regards vapour-pressure measurements the results of Tammann (Zeitsch.physihl. Chern. 1888 2 42) at loo" and of Dieterici (Ann. Phys. Chern. 1893 [ii] 50 47) at O" have been supplemented by measurements at 20' and 30" with a special form of apparatus. In connexion with the calculation of solution-afinity from freezing- point determinations with the help of thermal constants it has been found that the heat of fusion of sulphuric acid is 2485 + 6 - l t calories per mol. and that of the monohydrate 4290 + 18% calories. From these results the differential solution-affinity curves for water and acid respectively have been obtained and for comparison are plotted on the same diagram with the corresponding curves for the heat of admixture. The forms of the curves are very different from the ideal type due to chemical reaction between water and acid.As regards the sulphuric acid curve A the solution-affinity is greater than U the heat of admixture for x= 1 (x is the molar pro- portion of acid in the mixture) ; the curves intersect at x = 0.53 where A = U= 1400 cal. beyond which the A-curve is lower than the U-curve until they again intersect at x = 0*006,~~tvhen A = U= 17,000 cal. FOP water the affinity and heat of reaction curves practically coincide for concentrations between 0.1 and 0.3. The values found are throughout in accord with the general thermodynamic equation A - U= T x dA/dT (where the symbols have the usual significance). G. S. PAUL TH. MULLER (J. Chim. Phys. 1909 7 534-539).-The theorem recently published by Nernst (Abstr. 1907 ii 153) correlating affinity with temperatlure in condensed systems is applied to the affinity of sodium phosphate for water.The affinity of the hydrate Na2HP0,,12H,0 for its water is given by the formula A = 1*985T'log,f/f' cd. where J' andf' are the vapour pressures of pure water and hydrate respectively a t temperature 1'. Nernst's equation gives A = Qo - UP - P P / 2 where Qo is the heat of hydration at absolute zero. The values of A in the first formula are calculated from Frowein's measurements of the vapour pressure of the hydrate between 6.8" and 27*00°. I f Qo= 1200'58 cal. a = 0.01 19827 and /3 = 0 Nernst's formula gives values agreeing with Frowein's within 0.35%. The heat of hydration at any temperature is equal to Qo+uT2+pl'3. Hence the heat of hydration of sodium phosphate at 18" should be 2215.3 cal.whereas Thomsen and Pfaundler obtained the values 2234 cal. and 2244 cal. by direct measurements. When the affinity for water ( A ) is zero (@-aT2)=0 whence 5!'=316.5O abs. It follows that at 43.5O the phosphate becomes anhydrous. Extrapolation of Frowein's results indicates that at about this temperature the vapour pressure of the phosphate begins to exceed that of water OTTO SACKUH. (Zeitsch. Elektrochem. 1909 15 865. Compare Trautz Abstr. 1909 li 651).-Trautz'a equations for the velocity of a reaction are obtained essentially by dividing the well known equation d log K/dI'= Q/RP into the two equations d log k1/d2'= q1/12Tz and d log k,/dT= q2/&P where k and k are the Affinity of Sodium Phosphate for Water. R. J. C. Chemical Kinetics.ii.114 ABSTRACTS OF CHEMICAL PAPERS. I. 11. Calcium platinichloride ........ 1 2 6 Zinc platinichloride ........... 11 7 Copper platinichloride ......... 18 6 Cadmium platinichloride ...... 18 6 Cobalt platinichloride ........... 12 10 Nickel platinichloride ......... 12 10 Sodium platinochloride ......... 4 3 Sodium platinibromide ......... t i 5 Barium platinichloride ......... 6 5 Manganese platinichloride ...... 11 7 I. 11. Ammonium palladiochloridc 5 4 Palladious iodide ........... 6 (at 0”) 2 Rhodium chloride ........... 4 3 Ruthenium chloride ......... 3 . 3 Ruthenium bromide ......... 3 2 Cuprous chloiide.. ............. 3 1 Cuprous iodide .............. 3 0 Silver nitrate .................. 3 2 Chronious chloride ............ 6 3 Uranium tetrachloride ......3 3GENERAL AND PHYSICAL CHEMISTRY. ii. 115 Neither hydrogen chloride nor phosphine combines with any dry inorganic salt. Acetylene is absorbed only by cuprous chloride. Ethylene gives no additive products; its absorption by ferrous and platinous chloride in ethereal or hydrochloric acid solution only takes place when these chlorides are formed by reduction of the correspond- ing ferric and platinic chlorides. Carbon monoxide was not absorbed by any of the salts investigated under the particular conditions of experiment . T. s. P. Eder’s Solution. I. CHR. WINTHER (Zeitsclh. wiss. Yhotochern. 1909 7 409-441).-The rate of the photochemical reaction between mercuric chloride and ammonium oxalate in aqueous solution is increased by ceric salts and potassium ferricyanide and decreased by cupric salts potassium tin chloride and many organic colouring matters.Potassium iodide in small quantity accelerates the reaction but when this is present in excess the velocity‘of the reaction diminishes. I n order to obtain information relative to the nature of the catalytic effect the author has examined the behaviour of chlorine potassium permanganate and more especially ferric salts. I n the case of chlorine the catalytic phenomenon is traced to the inducing effect of the reaction between chlorine and ammonium oxalate on that between the mercuric salt and the oxalate. The action of potassium permanganate is found to be accompanied by a period of induction during which the permanganate is reduced to a manganic salt which then accelerates the photochemical change.The mode of action of this is in all probability similar to that of ferric salts. The numerous experiments made on solutions containing iron salts show that the catgltic effect is very largely dependent on the asmount of oxygen which is present. I n the absence of oxygen the catalytic process can be resolved into two stages in one of which the rapid photochernicaJ reduction of ferric oxalate is involved whereas the other consists in the inducing effect of the oxidation of the ferrous oxalate formed in the photochemical reduction process. I n presence of oxygen the process is complicated by reason of the action of the oxygen on the reduced ferric salt. In support of this view it is found that the rate at which mercurous chloride is precipitated from a solution containing a given amount of ferrous salt increases as the amount of oxygen in the solution diminishes.For a given quantity of ferrous salt the total amount of mercurous chloride precipitated increases as the oxygen concentration diminishes. Ferric salts diminish the rate of the reaction. For a given ratio between ferric and ferrous salts the retarding effect increases rapidly with the total amount of iron in the solution. These observations are in accord with the fact that maximum photo-sensi- tiveness is obtained for a particular iron concentration. Since the retarding action of ferric salts increases when the amount of oxygen in the solution diminishes the iron concentration corresponding with maximuq sensitiveness diminishes with the oxygen concentration.The rate of precipitation of mercurous chloride from Eder’s solutionii. 116 ABSTRACTS OF CHEMICAL PAPERS. is recommended as a means of estimating small quantities of dissolved oxygen. H. M. D. Vacuum Correction of Weighings Applied to Atomic Weight Determinations. PHILIPPE A. GUYE and N. ZACHARIAD~S (Compt. rend. 1909 149 1122-1123. Compare Abstr. 1909 ii 989).-The authors have determined previously the magnitude of the error made in reducing weighings to vacuum values through the presence of condensed air on the surface of the substance. These calculations have now been revised after taking into account another source of error. The new numbers together with the results of fresh determinations for other common substances are given in tabular form.The results were obtained by weighing a flask in air (u) exhausted of air ( b ) full of air (c) containing air and the substance ( d ) containing the substance only. The apparent weight of the salt in air is reduced to the vacuum value by the usual method and compared with the “actual weight in vacuum” given by ( d - cc) - p where p is the loss of weight in the air of the standard weights. A further correction should be made for the air condensed on the surface of the standard weights. The “actual weight in vacuum” was in each case found to be higher than the calculated value the difference ranging from 1 mg. in the case of silver bromide to 25 mg. in the case of sodium chloride per 100 grams. I n the case of silver however the numbers were the same. w. 0. w. The Fundamental Constant of Atomic Vibration and the Nature of Dielectric Capacity. WiLLIAM SUTHERLAND (Phil. Mag. 1910 [vi] 19 1-25).-0n the assumption that positive and negative electrons are associated in pairs and are revolving round one another in such a way that each pair has an average electric moment it follows that if these moments are similarly directed the atom as a whole mill have an electric moment and can be investigated as a uniformly electrised sphere. It is shown that the internal electric fields cause atomic vibrations and that the atomic vibrator can be regarded as the single electron which is involved in the explanation of the Zeeman effect. The common constant which appears in Rydberg’s formuh for the series lines of many elements is discussed in terms of this conception of the atomic vibrator. An explanation of dielectric capacity in terms of the electron theory is given and it is shown that Balmer’s formula relating to spectral structure can be interpreted on a kinematical basis. H. M. D. Molecular Diameters. WILLIAM SUTHERLAND (Phil. Mug. 19 10 [vi] 19 25- 26).-0n the basis of the value 2.77 x obtained by Rutherford for the number of molecules in 1 C.C. of a gas under standard conditions the author has recalculated a series of molecular diameters with the following result H2 2.17 He 1.02 CO 2.74 C,H 3.31 N 2.95 NO 2.59 0 2.71 A 2-66 CO 2.90 N,O 3-33 C1 3.76 x lo-* em. H. M. D.INORGANIC CHEMISTRY. ii. 117 Liquid Extraction with the Aid of Soxhlet’s Apparatus. TADASU SAIKI (J. Biol. Chem. 1909 7 21-22).-9 modification of Soxhlet’s apparatus is described and figured for the extraction of liquid material with ether. W. D. H. Apparatus for Evaporating Ethereal Solutions. GILBERT P. GIRDWOOD (Analyst 1909 35 16)-The one end of an inverted siphon has the shape of a funnel and the other end projects several inches below the level of the ethereal solution to be evaporated. The watch-glass containing this is placed under the funnel-shaped opening and by applying suction to the long arm the ether vapour siphons over and the residue is finally deposited in a small space in the centre of the watch-glass. The same means may be adopted for concentrating an ethereal solution in a test-tube or beaker by gradually lowering the funnel of the siphon as the ether evaporates. L. DE I<.