61 General and Physical Chemistry. Supposed Relationship between Molecular Size and Rota- tory Power in Solutions. THOMAS S. PATTERSON (Bey. 1905 38 4090-4101. Compare Walden Abstr. 1905 ii 130).-Walden states that there is a relationship between the molecular weight of a compound in solution and its rotatory power. The figures quoted for ethyl tartrate are invalidated by differences in the temperature at which experiments were made in different solvents and doubt is also cast on the validity of the molecular weight determinations on account of excessive concentration of the solutions. The most striking adverse evidence is found in the case of solutions of ethyl tartrate in benzene where in spite of a very great change of molecular weight the rotatory power is the same at 5 and at 25 per cent.concentrations ; in aqueous solutions on the other hand the molecular weight remains consbant but the rotatory power varies widely. Several other examples are quoted which go to show that the relationship postulated by Walden does not exist. T. M. L. Solubility and Specific Rotatory Power of Carbohydrates and certain Organic Acids and Bases in Pyridine and other Solvents. JOSEPH GE~CARD HOLTY (J. Pl~ysicul Chem. 1905 9 764-779).-The author’s experiments were performed chiefly with pyridine solutions for which the following solubilities are given laevulose 18.49 ; malic acid 14.6 ; dextrose 7.62 ; sucrose 6-45 ; galactose 5.45 ; erythritol 2.50 ; lactose 2-18 ; strychnine 1.23 ; mannitol 0.47. Propyl tartrate was found to be miscible with pyridine. The specific rotation of lacbose was less in pyridine than in aqueous solution and decreased with dilution.For dextrose and galactose the rotation was greater than for the equally concentrated aqueous solution. For lzevulose the laevorotntion was much less than that of the aqueous solution the values for a 7% per cent. solution being pyridine - 36.8 ; water - 87.3. The laevorotations of mannitol and malic acid are considerably greater than those of the aqueous solution values for a 7.6 per cent. solution being respectively -26.9 and -2.07 in the case of malic acid. Strychnine gave a lzvorotatory solution of higher specific rotation than the corresponding solution in chloroform. For a 6.45 per cent. solution of sucrose the specific rotation was 82.97 and the value increases slightly with dilution.From his inability to verify it experimentally the author considers that the theory of a greater solubility for small than for large particles is not true in the case of pyridine solutions of sucrose. L. M. J. Action of Alkaline Uranyl Salts on the Rotatory Power of Sugars and other Optically-active Hydroxyl Compounds. HERMANN GROSSMANN (Zeit. Fer. deut. 2ucker.-lnd. 1905 1058-1073). -Lzvulose and mannitol react readily with alkaline uranyl solutions VOL. XC ii 562 ABSTRACTS OF CHEMICAL PAPERS. forming intensely golden-yellow alkaline liquids the rotations of which differ considerably from the true rotations of the optically-active compounds. Complex compounds in which the alcoholic hydrogen atoms are replaced by the uranyl-residue are here formed.The addition of 1 mol. of uranyl salt t o 1 mol. of laevulose or mannitol changes the direction of the rotation. Dextrose galactose lactose and rhamnose also exhibit similar changes in rotation but the direction of the latter is not reversed. Sucrose is slowly transformed into l ~ v o - rotatory alkali-uranyl compounds but no change of the sugar molecule appears to take place here. The optical activity of mannitol is accentuated by a number of compounds for instance boric molybdic and uranic acids. The increase in the rotation of tartaric acid by uranyl nitrate and sodium hydroxide observed by Walden (Bey. 1897 30 28S9) is not a reaction occurring in alkaline solution but requires the presence of hydrogen ions so that the hydrogen atoms of the hydroxyl groups are probably not replaced by uranyl-residues.Saccharic acid exhibits on the whole similar behaviour its direction of rotation being reversed. T. H. P. Spectrum Analysis of the Light emitted by Crystals of Radium Bromide. F. HIMSTEDT and G. MEYER (Chenz. Centr. 1905 ii 1661 ; from P/~ysikaZ. Zeit. 6 688-689. Compare Sir William and Lady Huggins Abstr. 1904 ii 4).-In order to examine the spectra of the light emitted by radium bromide three crystals were arranged in front of the collimator tube of a quartz spectrograph in such a way that they did not touch each other. The whole apparatus was enclosed in an opaque air-tight case which was filled with dry air hydrogen or carbon dioxide. After an exposure of seven to ten days the photograph showed three continuous spectra extending from 460pp to 337pp.The blackest portion reached to 380pp and when the case was filled with air the nitrogen bands in this region were only faintly visible. Beyond 380pp however the nitrogen bands could be identified with certainty and were distinctly apparent in the spaces which separated the spectra. The nitrogen line 3 15*9pp extended with equal intensity from the bottom to the top of the plate. The nitrogen in the atmosphere near the crystals must therefore have been rendered luminous. The presence of traces of moisture was found to decrease the intensity of the light considerably. When an atmosphere of dry hydrogen or carbon dioxide was used the three continuous spectra extended to 310pp and did not show any traces of bands or lines in the intervening spaces.The strongest fluorescence was observed in an atmosphere of carbon dioxide. The wave-lengths of the nitrogen light are the same in the case of radium bromide and of radiotellurium. E. W. W. Some Phosphorescence Spectra indioating the Existence of New Elements. Sir WILLIAM CROOKES (Chem. News 1905 92 273-274).-8ub-fractionation has yielded earths in which certain groups of unassigned lines and bands of the phosphorescent spectra grow fainter or stronger independently of both other unassignedGENERAL AND PHYSICAL CHEMISTRY. 63 groups and of groups of lines assigned to known elements. provisionally regarded as indicating new elements. Radium in Sweden. JOHN LANDIN (Arkiv Kern Min. Geol. 1905 2 i No. 2 1-7).-The author has demonstrated the presence of radium in culm by preparing from it barium sulphate and testing the action of the latter on a photographic plate.The alum schists of Billinge answer t o the same test and like culm also contain uranium. The quantity of radium in these minerals is howevey too small to be of industrial importance. Hjelmite crystals slowly KNUT ANGSTROM (Chem. Cewty. 1905 ii 1575 ; from Yhysikal. &it. 6 685-688).-Since the quantity of heat emitted by radium does not decrease to an appreciable extent in a year and is not dependent on the nature of the surrounding medium the energy lost in the form of a- p- and y-rays can only be a very small proportion of the total energy (compare Yaschen Abstr. 1904 ii 798). Measurements have been made by an electrical compensation method in which an equal quantity of heat was generated electrically and balanced against a sensitive thermo-element.86.5 mg. of radium bromide were used. Lead copper and aluminium calorimeters were used the results differing by not more than 2 per cent. When about 1/7 of the y-rays which passed through the aluminium calorimeter were absorbed by a cylinder of lead the quantity of heat imparted to the cylinder was not equal to 1 per cent. of the quantity liberated in the calorimeter; the energy of the p- and y-rays citnnot therefore be equal to more than a few parts per 100 of the total energy emitted by the radium. The mean value of the quantity of heat given out by 1 gram of radium bromide per minute from September These are D.A. L. blacken a photographic plate. T. H. P. Emission of Heat by Radium. 1903 to January 1905 was 1.136 cals. E. w. w. Relative Absorption of the Rays of Radium and Polonium. E. RIECKE [with RETSCHINSKY and WIGGER] (Chena. Centr. 1905 ii 1574-1575 ; from Physikal. Zeit. 6 685-685),-The results of determinations of the coeilicients of absorption of the rays of radium and polonium should give a clue to the nature of the radiation. If for instance the radiation is effected by means of corpuscles then the absorption-coefficient cannot be constant since the absorption depends on the collision of the particles with the molecules of the absorbing substance whilst on the other hand constant values should be obtained if the energy is transmitted by waves. Although however the absorption-coefficient of the a-rays of radium in air has been found t o be constant the deflection of the rays in a magnetic or electric field appears to indicate a corpuscular structure.Experiments on the a-rays of polonium in which Marckwald's arrangement of a copper rod covered with radiotellurium was employed have shown that the absorption- coefficient is not constant. The velocity of the rays is decreased by collision of the particles with the molecules of the air whilst the abgorption is increased. The results of experiments on the absorption of the y-rays emitted by radium bromide in lead have proved that the 5-264 ABSTRACTS OF CHEMICAL PAPERS. rays are not homogeneous ; their nature is not affected by the absorp- tion. When a lead plate of sufficient thickness is used only those rays which have the lowest absorption-coefficient (0.25) remain.The theory of the identity of the y-rays with the Rontgen rays is supported by the fact that the absorption-coefficient of the ‘‘ residual ” rays in mercury lead copper iron zinc aluminium and sulphur is propor- tional to the sp. gr. of the substance (k/d = 0.0021). E. W. W. Radioactivity of Products of Etna. GIOVANNITROVATO CASTORINA (Nuovo Cim. 1905 [v] 10 198-202).-ln general the products of Etna are radioactive the following series being in order of increasing activity rock sand sandy tufa clay and clayey soil mud soil and pozzolane or soil coloured by the lava The activity of the soil is influenced by heating by chemical reactions or by absorption of wator.T. H. P. A. E. Nordenskiold’s Investigations on the Radioactivity of certain Swsdish and Norwegian Minerals. HJALMAR SJOGREN (Arkiv K e r n . illin. Geol. 1905,2 i No. 4 1-5).-The author gives an account of A. E. Nordenskiold’s researches on the radioactivity of Swedish and Norwegian minerals which in order OF their activity are as follows cleveite from Raade ; uraninite from Johanngeorgen- stadt ; oraugite and yellow thorite from Svinor ; nohlite from Noh1 ; fergusonite from Sandoen and orangite from Brevig ; euxenite from Morefjiir ; Eschynite from Hitter0 ; xenotime from Sandoen ; monazite from Raade and fergusonite from Ytterby ; malacon from Kammerfors. Columbite from Moss ixiolite from Skogbole and hjelrnite from Nya Kgrarfvet showed no action on a sensitive plate even after ninety-six hours’ exposure.The paper is accompanied by photographs illustrating the actions of the different minerals. T. H. P. Abnormal Anodic Polarisation produced by Fluorine Chlorine and Bromine Ions. ERICH MULLER and ALFRED SCHELLER (Zeit. anorg. Chem. 1905 48 112-128. Compare Abstr. 1904 ii 250 811 812. Skirrow Abstr. 1903 ii 69).-It has already been shown by Miiller and others that the addition of indifferent ions often causes increased polarisation at the anode (Zoc. c i t . ) ; in the present paper the effect of fluorine chlorine and bromine ions on the anode potential during the electrolysis of sulphuric acid is described. Experiments were made with a series of mixtures of normal sulphuric acid and normal solutions of the halogen acids with different current densities and a t various temperatures and the results are represented by curves the potentials being plotted as ordinates against the com- position of the mixtures as abscissae.With hydrochloric acid at the ordinary temperature the curves show maxima at about 2.5 per cent. of the acid; the greater the cur- rent density the greater is the rise in potential and the more distinct the maximum. With very small current densities the only effect of chlorine ions is to lower the potential. Increasing the temperatureGENERAT AND PHYSICAL CHEMISTRT. 65 has the same effect as lessening the current density ; thus the chlorine ion has a powerful polarising action a t 0" which quite disappears at 70". The bromine ion has a similar but less powerfnl action ; only mix- tures containing less than 10 per cent.of the acid show increased potential whereas stronger solutions lower i t very considerably. Hydrofluoric acid increases the anode potential in all concentrations ; in this case as well the effect decreases with rise of temperature. It is pointed out that the above results are very similar to those obtained in investigating the decomposition potentials of the halogen acids (compare Luther and Brislee Abstr. 1903 ii 708). The author suggests that the phenomena observed are due to the formation of the lower halogen oxyacids and shows that this theory accounts for the results obtained with chlorinc and bromine ions. G. S. Mechanism of the Production and the Nature of Cathodic Pulverisations. CH. MAURAIN (Compt.rend. 1905,141 1223-1225). -When a plate of glass or mica is placed opposite or inclined to the cathode inside a tube containing a rarefied gas the film of cathodic material which is deposited on it is most abundant at the edges that is at those parts of the plate farthest removed from the cathode. When the VrZcuiim tube is placed between the poles of an electro-magnet in such a position that the cathode is normal to the magnetic field the cathodic film deposited on the malls on the tube is most abundant in the trace of the luminescent plane and in a plane at right angles to it. Prom the results of these and similar experiments t,he author draws the conclusion that the cathodic pulverisations are formed of fairly large pnrticles torn from the cathode by the impact of the a-rays projected iii all directions and carrying variable electrical charges the e / m being illrich smaller than in the case of the particles forming the cathode rays.MI. A. W. Chemical Composition of the Nickel Oxide Electrode in the Jungner Edison Accumulator. JULIAN ZEDNER (Zed. EZekh*ociienz. 1905 11 809-S13).-If nickelic oxide be prepared in any way and pasted on to a platinum wire the potential difference between the wire and zinc amalgam i n a solution of potassium hydroxide is the same 2s if the nickelic oxide were not present. A nickel oxide accumulator plate gives a larger potential difference and nickelic oxide deposited electrolytically in various ways on platinum gives the same result as the accumulator plate. The negative result obtained with the pasted oxide is probably due to its extremely smnll conductivity.Quantities of nickelic oxide sugcient for analysis were obtained by electrolysing a solution of nickel sulphate and sodium acetate with a rapidly rotating anode and adding dilute alkali hydroxide solution gradually so as to maintain the solution constantly neutral. The substances obtained had compositions varying between those represented by the formulae Ni20,.H20 and Ni203,11H,0 according to the method of drying. Exact measurements of the E.N.P. of the cell Zn I Zn(OK) + KO€€ I Ni,O showed that this oxide is ideri t ical with that formed in the nccumu1;itor. T. F A66 ABSTRACTS OF CHEMICAL PAPERS. Lead Peroxide as Anode in the Electrolytic Oxidation of Chromium Sulphate to Chromic Acid.ERICH MULLER and MAX SOLLER (Zeit. Elektrochem. 1905 11 863-872).-Chrome alum dissolved in 3-sulphuric acid is not appreciably oxidised t o chromic acid when an anode of smooth platinum is used. Traces of a salt of lead in the solution are precipitated a t the anode as lead peroxide and bring about considerable oxidation. ,Small quantities of chlorine ions also favour the change. The oxidation takes place a t a platinised platinum anode to about one-third the extent observed under similar conditions with a lead peroxide anode. With a lead peroxide anode the oxidation takes place almost quantitatively in fairly concentrated solutions of chrome alum when the current density is not too high (about 0.005 ampere per sq. cm. is suitable). The difference in the behaviour of the platinum and lead peroxide anodes is not due to the latter having a higher anodic potential (the reverse is rather the case) but appears to depend on a catalytic action of lead peroxide which is capable of producing the oxidation purely chemically.T. E. Conductivity and Viscosity of Solutions of certain Salts in Mixtures of Acetone with Methyl Alcohol Ethyl Alcohol and Water. HARRY C. JONES and EUGENE C. BINGHAM (Amer. Chern. J. 1905 34 481-554. Compare Jones and Carroll Abstr. 1905 ii 73).-A study has been made of the relations between the viscosities and conductivities of lithium nitrate potassium iodide and calcium nitrate in mixtures of acetone with methyl alcohol ethyl alcohol and water. Determinations have been made of the conductivity a t 0" and 25" of solutions of these salts of various concentrations in methyl alcohol acetone ethyl alcohol water and mixtures of acetone with methyl alcohol ethyl alcohol or water containing 25 50 or 75 per cent.of acetone. The results of the experiments are tabulated and plotted as curves. The temperature-coeflicients have been calculated in all cases. The viscosities a t 0' of acetone methyl alcohol ethyl alcohol and mixtures of methyl alcohol and acetone and of ethyl alcohol and acetone have been determined. 'Determinations have also been made of the viscosity a t Oo and 25' of water a mixture of water with 25 per cent. of acetone and solutions of calcium nitrate in the following solvents acetone methyl alcohol and mixtures of acetone with water methyl alcohol or ethyl alcohol containing 25 50 or 75 per cent.of acetone. The viscosity fluidity and density of the various solutions are given in tabular form. Com- parisons of the temperature-coefficients of conductivity and fluidity in these mixtures show that they vary in the same manner but that the former are uniformly smaller than the latter. The conductivities of the salt solutions in mixtures of acetone and water exhibited the minimum in conductivity which has been observed previously and which is shown to be intimately connected with the minimum in fluidity. I n the mixtures of acetone and alcohols the fluidities are in accordance with the law of averages and the fluidity curve is nearly a straight line whence it is concluded that acetone and these alcohols do not form complex molecular aggregates when mixed.The fluidities are also plotted as curves.GENERAL AND PHYSICAL CHEMISTRY. 67 The curves for the conductivity of potassium iodide in the various mixtures are nearly straight lines a t all dilutions and are very similar t o the corresponding fluidity curves. Lithium nitrate a.nd calcium nitrate in mixtures of acetone with methyl or ethyl alcohol show a deviation from the fluidity curves and exhibit a pronounced maximum in conductivity. After a consideration of t.he various factors which could influence the conductivity i t was concluded that the observed maximum is due to a change in the dimensions of the atmospheres about the ions (compare Kohlrausch Abstr. 1903 ii 403). The conclusion of Dutoit and Friderich (Abstr. 1899 ii 351) and of Jones and Carroll (Zoc.cit.) that conductivity is directly proportional to the dissociation and inversely proportional to the viscosity is incomplete since it does not allow for possible changes in the size of the ionic spheres. The conductivities of lithium nitrate and calcium nitrate in mixtures of acetone and water also shorn a tendency towards a maximum. This tendency to form a maximum in conductivity increases from potassium iodide through calcium nitrate to lithium nitrate and it is considered that this may be connected with the velocities of migration of the ions. E. G. Can an Element form both Positive and Negative Ions? MAX LE BLANC (Zeit. Elektrochem. 1905,11,813-818).-Both sulphur and selenium (in the form of a thin film on a platinum wire) when used as cathodes in potassium hydroxide solution go into solution in the form of polysulphide or polyselenide ions SZ” or Se,” the largest values of x observed being 5 and 4.4 respectively.Used as anodes no action was observed at all. Telluioium as cathode in a1 kaline solutions dissolves readily; in very concentrated solutions of potassium hydroxide the dust observed by Miiller and Lucas (Bbstr. 1905 ii 672) is not formed. Its formation is more probahly due to the dissociation of polytelluride ions T e l Z T e ” + Te in dilute solutions. At a tellu- rium anode in potassium hydroxide there is no vkible change but tellurium goes into solution w i t h the valency 3.9 ; it therefore gives off the ions Te”” which tben react with hyclroxyl ions to form the ions TeO,”.T. E. Experimental Study of the Three Parts of an Electrolytic Cell and their Relationships. G. HOSTELET (Zeit. Elektrochem. 1905 11 889-904).-In determining decomposition points by Glaser’s method (Abstr. 1899 ii 78) the author found difficulty in interpret- ing the results owing to variations in the resistance of the system and in the potential of the constant electrode. H e has therefore worked out a method in which the falls of potential between the anode arid the electrolyte between the cathode and the electrolyte and in the elec- trolyte itself are all separately and simultaneously determined. The method is applied to the study of the electrolysis of potassium chloride with a platinum anode and a cathode of mercury. T. E. Theory of Electrolytic Dissociation taking account of the Electrical Energy.R. MALMSTROM (Zeit. Elektrochem. 1905 11 797-809).-In addition to the ordinary attractions between the68 ARSTRACTS OF CHEMICAL PAPERS. various molecules in the solution of an electrolyte there are forces due t o the attractions and repulsions of the electric charges on the ions. Assuming these electrical forces t o be superposed on the ordinary attractive forces the free energy of a dissolved electrolyte is calculated and from this an equation for the equilibrium between the ions and undissociated molecules is obtained. This is found t o be in rather better agreement with measurements of conductivity than any of the equations proposed previously. It is suggested that owing to electro-striction the dielectric constant of the solution must be a function of the concentration and that a better agreement might be obta,ined by taking this into account.T. E. Theory of Electrolytic Dissociation. LOUIS KAHLENBERCr (f'hd. Compare Abstr. 1905 ii 139)- Mag. 1905 [vi] 10 662-664. A reply to Jones' criticism (Phil. Mag. 1906 [vi] 10 157). J. C. P. Accurate Measurement of Ionic Velocities. ROBERT B. DENISON and BERTRAM D. STEELE (Proc. Roy. Sac. 1905 76 A 556-557).-An apparatns has been devised for measuring the velocities of ions without using gelatin or other partitions during the experi- ment. The apparatus consists of two reservoirs each provided with a special electrode vessel and of a measuring tube of known cross-section in which the solution to be examined is placed.One of the reservoirs contains a solution of B salt with a less mobile cation than the cation under observation the other a solution of a salt with a less mobile anion than that of the salt under investigation. The measuring tube is provided at each end with parchment-paper partitions which facilitate the formation of a sharp electrolytic margin between the measured and the indicator solutions and these partitions are so arranged that they can be removed after the margins have been formed. Experiments have been made at dilutions down to 0.02 normal. The calculated transport numbers agree very well with those obtained by Hittorf's method and the measured average ionic velocities confirm in a remarkable manner those calculated from conductivity data.H. M. D. Mobility of Ions of Salt Vapours. GEORGES MOREAU (Compt. rend. 1905 141 1225-1227. Compare Abstr. 1903 ii 125 196 ; 1904 ii 536; 1905 ii 9).-The author has measured between the temperatures of 170' and 15" the mobility (k) of the ions in air which has been bubbled through solutions of potassium salts of different con- centrations and then passed through a tube heated to redness. From the values of k thus obtained the sizes of the ions for different tempera- tures and concentrations are calculated with the results that at the higher temperatures the ions have a diameter five to fifteen times that of the molecule and appear to consist of an electrically charged nucleus surrounded by two t o seven layers of molecules whilst at the l o ~ e r temperatures the number of layers is increased to ten t o twenty.M. A. W.GENERAL AND PHYSICAL CHEMISTRY. 69 Electrical Conductivity of Dilute Solutions of Sulphuric Acid. W. C. DAMPIER WHETRANI (Proc. Boy. SOC. 1905 76 A 577-583).-The equivalent conductivity of solutions of strong acids and alkali hydroxides increases with dilution but reaches a maximum a t 0.001 to 0.0005 gram-equivalent per litre and falls rapidly with further increase in dilution. The effect of traces of impurities on the equivalent conductivity of sulphuric acid in dilute solution has been examined by measuring the variation of the conductivity with the con- centration in four solvents (1) good quality redistilled water and the same containing (2) a trace of carbon dioxide (3) a trace of potassium chloride and (4) the same water freed as far as possible from volatile impurities by repeated boiling under diminished pressure.I n each case the conductivity of the solvent was subtracted from that of the solution. The results shorn that the conductivity of sulphuric acid in dilute solution is not affected by addition of a trace of potassium chloride or by boiling under diminished pressure but that it is appreciably diminished in presence of a litt,le carbon dioxide and in this case the maximum conductivity is reached at a higher concentru- tion. The abnormally low conductivity of dilute solutions cannot however be wholly attributed to the presence of carbon dioxide in the water used as solvent. H. M. D. Transit of-Ions in the Electric Arc. A. A. CAMPBELL SWINTON (Proc.Roy. SOC. 1905 76 A 553-556).-An experiment has been devised to show t h a t the current in the electric arc is cayried by positive and negative ions emitted by the electrodes. A small hole was bored axially through one of the carbon electrodes which was 6xed rigidly a t the end of a brass tube and an insulated Faraday cylinder w t s supported within the brass tube so that its aperture was in line with and immediately behind the aperture in the carbon electrode. With a current of three amperes and about 50 volts the Earaday tube remained unchanged in atmospheric air but when the apparatus was enclosed in a glass vessel and the pressure reduced the Faraday cylinder acquired the same charge as the opposite carbon electrode. The effect increased with the degree of exhaustion of the containing vessel ancl correspond- ing with the greater mobility of the negative ions the galvanometer deflection for a given degree of exhaustion was greater when the opposite carbon electrode was made the negative pole.H. M. D. Magnetisation and Magnetic Change of Length in Ferro- magnetic Metals and Alloys at Temperatures ranging from - 186' to + 1200". KOTARO HONDA and Su~ru SHIMIZU ( P l d . May. 1905 [vi] 10 642-661).-The papercontains n large number of data bearing on the magnetisation of Swedish iron nickel annealed cobalt and various nickel steels. The results obtained are in general harmony with those of earlier observers such as Osmond and Dumas. J. C. P. Coexistence of Paramagnetism and Diamagnetism in the same Crystal. GEORGES MESLIN (Conzpt.inend. 1905 141 1006-1008. Compare Compt. re?ziJ. 1905 140 i6S3).-A small70 ABSTRACTS OF CHEMICAL PAPERS. cube cut from a crystal of pyrrotite in such a way that its faces are parallel with or perpendicular to the magnetic plane exhibits para- magnetism when placed in a magnetic field in such a position that the magnetic plane passes through the axis of the electro-magnet whilst in a direction a t right angles to this it exhibits diamagnetism. By immersing the cube in a concentrated solution of ferric chloride its paramagnetic susceptibility is diminished and its diamagnetic sus- ceptibility is increased. M. A. W. Magnetic Compounds of Non-magnetic Elements. EDGAR WEDEKIND (Zeit. Elektrochem. 1905 11 850-861).-The manganese borides previously described (Abstr.1905 ii 322) are found to be strongly magnetic ; they also possess considerable permanent magnetism. Compounds of manganese with antimony and bismuth are also magnetic ; manganese arsenide however is not magnetic but becomes magnetic when heated in the air. The phosphides carbides and sulphides of manganese prepared by the alumino- thermic process are also magnetisable whilst the silicide is not so. T. E. Specific Heats of Gases. SILVIO LUSSANA (Nuovo Cirn. 1905 [v] 10 192-195).-The author points out that although in their paper on the specific heats of gases a t high temperatures Holborn and Austin (Abstr. 1905 ii 228) affirm that it is impossible to state with certainty that the specific heats of the simple gases increase as the temperature rises yet all their results exhibit such an increase the mean value of which is 3 per cent.whilst the accuracy of their measurements is about +_ 1 per cent. The author's investigations on this subject extending over a number of years lead to the conclusions that for all gases the specific heat a t constant pressure (1) increases with the pressure up to a. certain limiting value and then diminishes the maximum corresponding with a pressure which is lower as the gas deviates more from Boyle's law and (2) increases or diminishes with the temperature according 5s the latter is near t o or far from the temperature of liquefaction of the gas ; the temperature variation of the specific heat increases with the pressure. T. H.. P. Latent Heat of Fusion of Ice. ANATOLE LEDUC (Compt.rend. 1906 142 46-48).-There is a difference of nearly 1 per cent. between Bunsen's value (S0.03) for the latent heat of fusion of ice and the value (79.25) found by Laprovostaye and Desains and confirmed by Regnault and the difference is still of the same order if the numbers are corrected for the variation of the specific heat of water taking 1.004 as the specific heat of water a t 0" according to the results obtained by Callendar and Barnes. If however the author's value 0-9176 for the specific volume of ice a t 0' is substituted for the value found by Bunsen 0.9 1674 Bunsen's value for the latent heat of fusion of ice becomes 79.15 and the author considers that 79.3 is the most probable value for t h i s constant. M. A. IT.GENERAL AND PHYSICAL CHEMISTRY.71 Connection between Depression of Freezing Point and Latent Heat of Fusion. KARL DRUCKER (Zeit. Elektrochem. 1905 11 904-905).-Goebel (Abstr. 1905 ii 679) has calculated the latent heat of fusion of ice a t temperatures below zero from the vapour pressures of ice and water. The author points out that a very small inaccuracy in the formula representing the vapour pressures makes a very large error in the heats of fusion calculated. Goebel’s results differ enormously from Pettersson’s direct determinations. If the latter are used to calculate the vapour pressures of ice numbers are obtained in good agreement with the experimental numbers. Hence the correction of the freezing-point depressions given by Goebel is much too large. He has also erroneously taken the concentrations in gram-molecules per litre of solution instead of per kilogram of water which partially compensates the first error.Solvent and Ionising Properties of Ethylene Cyanide (Succinonitrile). GIUSEPPE ‘ BRUNI and ANTONIO MANUELLI (Zeit. Elektrochem. 1905 11 860-862).-The molecular depression of the freezing point of succinonitrile is 182.6 from which the molecular latent heat of fusion is calculated to be 0.94 cal. which is the smallest value yet observed for any organic compound. Ethylene bromide and chloride give abnormally small depressions of freezing point when dissolved in succinonitrile which is doubtless due to the formation of solid solutions. Compounds containing amino- or hydroxyl groups give small depressions probably owing to molecular association.The high dielectric constant of succinonitrile points to considerable dis- sociating power. The freezing points of solutions of potassium and tetraethylammonium iodides were determined and values of i calcu- lated which lie between 1-26 and 1.3 for potassium iodide and 1.78 and 1-89 for tetraethylammonium iodide. The latter solutions conduct well. The anhydrous chlorides of copper manganese cobalt and nickel are insoluble in succinonitrile. Critical Temperature of Solutions in Liquid Carbon Dioxide. ERNST H. EUCIINER (Chem. Centr. 1905 ii 1618-1619 ; from Chem. Weekblad 2 691-694).-The results of experiments on solutions of p-dichlorobenzene tribromomethane p-dibromobenzene o-chloronitro- benzene naphthalene and 1 3-dichloro-5-nitrobenzene in liquid carbon dioxide have shown that van’t Hoff’s law holds approximately in these cases the molecular rise of the critical temperature being almost constant.The results may however be more accurately expressed by means of van Laar’s hypothesis in which the rise per gram-molecule per cent. is not only dependent on the critical tempera- ture of the solvent but also on therelationship of the critical temperature of the dissolved substance to that of the solvent and on the ratio T of the critical pressures. Since as a rule T = 1 6 = O(O - a) in which 8 = l/Tk.(dT,/d,) and 5“k =critical temperature of the solvent. T. E. T. E. E. W. W. Relative Vapour Tensions of the Three Modifications of Carbon ANDREAS SMITS (Ber. 1905 38 4027-4033. Compare Schenck and Heller Abstr.1905 ii 619).-The equilibrium constant72 ABSTRACTS OF CHEMICAL PAPERS. of the reaction CO2+C,2CO is X = p ~ ~ ~ / p c o p c where p c is the partial vapour pressure of the carbon that is the maximum vapour pressure of the solid carbon at the given temperature. As this increases so do the values of K,pc and p2co/pco in the equation Klpe =p‘co/pco,. The equilibrium constant for the reaction Fe + CO t FeO + CO is AT= ( ~ ~ ~ ~ . ~ ~ ~ o ) / ( p ~ o / p ~ ~ ) ; if the differences in the solubilities of the three modifications of carbon in iron are neglected may be considered as constant and the equation may be written K2=pco/pco,. If the sum of the partial vapour pressures CO + C02 = P then P = k,pc( 1 + k2)/k,2 which differs from Schenck and Heller’s equation by the presence of pc (Zoc.cit.) and the greater the vapour pressure of the modification of carbon the greater is the equilibrium pressure P. If pc p’c a n d p ” ~ are the ‘a our pressures of the three modifications,then P P’ P”=pc :p’c :p”c. h e author discusses the work of Schenck and Heller from a theoretical point of view and concludes that the equilibrium pressures measured by them are not trustworthy. If however these are assumed to be correct the vapour pressures of the diamond amorphous carbon and graphite are in the proportions pc :p‘c :pNc = 3.7 1.8 1 at 500”; 4.1 2.4 1 a t 550O; 5.2 4.1 1 at 600”; and 5-4 4.2 1 a t 641O. G. Y. Distillation in High Vacua by means of Dewar’s Air Absorption Method and a Shortened MacLeod Vacuum Gauge. ALFRED WOHL and M.S. LOSANITSCH (Rer. 1905 38 4149-4154. Compare Pischer and Harries Abstr. 1902 ii 491 ; Erdmann 1904 ii 20).-The apparatus t o which is attached an absorption vessel containing 20-30 grams of extracted blood charcoal is evacuated by means of an ordinary water pump and the receiver and absorption vessel are then cooled by liquid air. Pressures of 0.006-0.008 mm. can thus be attained within thirty minutes. It is not necessary that the apparatus shall be absolutely air-tight as small amounts of air which enter are absorbed by the charcoal. The same cooled charcoal (25-30 grams) may be used for several distil- lations provided the apparatus is exhausted before being attached to the absorption vessel. For measuring high vacua a MacLeod apparatus shortened in much the same manner as suggested by Stlock (Abstr.1905 ii 514) for the Topler mercury pump is recommended. J. J. S. Distillation of Liquids which are Mutually Insoluble. C. VON RECHENBERG and W. WEIPSWANGE (J. prakt. Chem. 1905 [ ii] 72 47S-4SS).-Experiments on the weights of the two components of a non-miscible liquid mixture which distil over are not in accord with the conclusions of Charabot and Kocherolles (Abstr. 1904 ii 234). The authors’ results are however in accord with their deductions that the weights of the components are proportional to the products of their molecular weights and partial pressures. If the vapour mixture is saturated with respect to both components then the partial pressure for each component is the normal vapour pressure corresponding with the temperature of the vapour.I f the vapour mixture is unsaturatedGENERAL AND PHYSICAL CHEMISTRY. 73 with respect to one component the unsaturated vapour has the pressure equal to the difference between the pressure of the mixed vapour and the vapour pressure of the saturated component a t the temperature of the mixed vapour. Progressive Dissociation of Dibasic Acids. 11. RUDOLF WEGSCHEIDER (Moncctsh. 1905 26 1235-1239. Compare Abstr. 1902 ii 643).-In the case of sparingly soluble dibasic acids or of those which have a high value for the dissociation constant s of the second hydrogen atom it is necessary to calculate both dissociation constants from the results of the same experiments which may be done if approximate values for the two constants are known or can be guessed.A maximum value for k the dissociation constant of the first hydrogen atom is obtained according to Ostwald’s law of dilution from the conductivity in the most concentrated solution. is substituted for g1 in the equations L. Ill. J. If the value (I) - 2g2(zK+ Po0 (1) and (2) previously given then (11) - 2SVVK + 17.4) + ?( 1 - 1/1 - 8SVP(L + 4 . 4 - P ’) g = 4(Z + I - p a ) L2SV(l + Z2J + p] ’ in which g is a function oh the known constants I and pa of the experimentally determined values v and p and of the unknown constant s and (111) F(kls) = Kvp \‘pa - p + g2[2(E! + ZZA) - per 1) - [ p - 2g2(Zli + EZa4)][p - 2g,(E + E - pa ) J = 0 where F is a function of k and s. If the approximate value of k and the values of g calculated by equation (11) from the results of a number of experiments with assistance of the approximate value of s are substituted in equation (HI) as many values of 3’ are obtained and the values of k and s must be altered until in each case P-0.I f k+6k s+Ss are the correct values for the two constants k and s then (IV) O = P 6P SF ( k + 6k,s + 6s) = F(k,s) + -as + -6k ; from these are obtained two 6s 6k equations by means of which 6k and 6s may be calculated. Equations are given also for the adjustment of s and k in case of more than two observations. The method of calculation is illustrated in the case of 4-acetyl- aminoisophthalic acid. G. Y. Dissociation of Electrolytes. CARL HENSGEN ( J . prakt. Chem. 1905 [ii] 72 345-477).-1t has been shown previously that con- centration differences exist in a salt solution not entirely a t the same temperature.This case is first investigated for a solution of copper sulphate the upper layer of which was kept a t a temperature of about 98” whilst the lower was maintained at the ordinary temperature. It was found that not only was the concentration different for the two layers but also that the ratio CuO SO was not unity and was greatest for the highest layer. Numerous expressions have at different times been proposed to connect the sp. gr. of a solution with its chemical composition; these are fully considered and found by the author to be unsatisfactory for solutions of copper sulphate and it is74 ABSTRACTS OF CHEMICAL PAPERS. found that the ratio CuO SO varies within wide limits.Numerous tables showing this ratio a t different temperatures and ooneentrations are given in the paper and fullF discussed. The author considers that his results indicate that solutions of normal salts are the liquid phases of a heterogeneous equilibrium system brought about by the chemical action of the water and dependent on temperature &c. The general idea associated with the term solution of a normal salt namely the complete passage of the solid normal salt into the solution main- taining throughout the same proportion of its components regarded as base and acid is not consistent with the actual conditions. The molecular contents of normal salts in solution are therefore not calculable. Crystallisation of normal salts takes place only in solutions which do not contain acid and base in the same proportions as those of the precipitated salt.It is evident that these conclusions are not in accord with ordinary chemical views of solution and they and their consequences are discussed a t considerahle length by the author in the paper (compare Abstr. 1901 ii 540). KAKL DRUCKER (Zeit. Elektrochem. 1905 11,845).-The author assumes that the surface tension of a binary mixture may be expressed on the analogy of van der Waals' molecular attraction constant by l / y =L l/y,x2 + 1/y2( 1 - x ) ~ + l / ~ . ~ ( 1 - x)x. The formula is applicable to mixtures of non-associating substances ; yIm2 becomes constant in mixtures of acetic acid and benzene or carbon tetrachloride when the molecular weight of acetic acid is taken as double its formula weight ; in mixtures of water with alcohols or fatty acids constancy of yl.a is obtained by trebling or quadrupling the molecular weight of water.T. E. L. M. J. Molecular Weight of the Solvent in Binary Mixtures. Origin of the Formation of Layers in Solutions observed by A. Sinding-Larsen. C. CHRISTIANSEN (Chem. Centr. 1905 ii 161'7 ; from Overs. K. Danske Vidensk. SeZsk. Forh. 1905 307-315).- Sinding-Larsen (Ann. Fiqsik 1903 [ iv] 9 1186) observed that when solutions of sodium chloride in which the concentration was not uniform were heated sharply defined layers were formed and attri- buted the cause to surface tension and the formation of hydrates. Experiments on water water and toluene and with various solutions have shown however that layers may be formed by heating the liquids on one side and cooling them on the other.The formation of layers is due to movement of the liquid and to differences of temperature and is in no way dependent on the formation of hydrates. The surface of contact between two layers of different concentrations can only remain. sharply defined when the liquid is in motion and although the mixture of the layers is thereby delayed it is not entirely prevented. An electrolytic method of causing two layers of sulphuric acid solu- tions t o circulate is also described in the original paper. E. w. w. Limiting States of some Dissolved Chromic Salts. ALBERT COLSON (Compt. rend. 1905 141 1024-1027. Compare Abstr. 1905 ii 94 255 460 592).-When a solution of the normal green chromicGENERAL AND PHYSICAL CHEMISTRY.75 sulphate Cr,(S0J3 is exposed to sunlight the colour becomes turquoise-blue the density increases (compare Boisbaudran Abstr. 1875,427,730) and the heat of decomposition by potassium hydroxide diminishes t o a limiting value reached more rapidly in dilute than in more concentrated solutions as is shown in the following table Heat of decomposition [Cr,(SO,),:GKOH]. Duration of r- h . to sunlight. 1/10 ntol. per litre. 3/10 mol. per litre. exposure Solntions of Solutions of 0 57,300 cal. 57,200 cal. 10 days (May) - 56,100 28 9 9 48 7 50,400 84 7 50,100 52,200 165 9 50,100 to 50,000 cal. 51,800 51,600 - I The change in density corresponds with a contraction of 1 mol. H20 per 1 mol. Cr2(S04)3 and the turquoise-blue solution probably contains the hydrated sulphate S04<dr:so cr(0H)'s04H of which the heat of decomposition by potassium hydroxide is 50,000 cal.and not an equilibrium mixture of the green and violet chromic sulphates of which the heat of decomposition would be (57,200 + 45,000)/2 cal. or 51,100 cal. A solution of the violet chromic acetate OCr,(OAc) in excess of acetic acid becomes green on exposure to sunlight and cryoscopic and thermochemical measurements indicate that the tetra-acetate has combined with acetic acid to form the penta-acetate OH-Cr,(OAc) which however cannot be isolated for on evaporating the green solution to dryness it decomposes into the two salts Cr(OAc) and OH* Cr(OAc),. M. A. W. Abnormal Increase of Solubility with Organic Substances. DANIEL STROMHOLM ( A ~ k i v Kern.Hin. Geol. 1905,2 i No. 7 1-9).- The author has examined the influence of the addition of small quantities of water on the solubility in ether of benzoic acid p-nitro- benzoic acid picric acid acetanilide quinol phloretin methyl terephthalate benzil P-naphthaquinone phthalimide a-nitro- naphthalene p-nitroaniline P-naphthylamine benzyl sulphoxide sulphonal benzy 1 sulphide azobenzene and phenanthrene. The presence of water in excess causes the solubility in ether to increase by the following percentage proportions benzoic acid 80 ; quinol 296 ; a-nitronaphthalene 10-5 ; methyl terephthalate 16.7 ; benzil 19 ; P-naphthaquinone 43; P-naphthylamine 33.3; p-nitrobenzoic acid 161 ; picric acid 573; acetanilide 5 17; p-nitroaniline 127; phthalimide 47.7; benzyl sulphoxide 92.2 ; and sulphonal 30.3.With benzyl sulphide azo- benzene and phenanthrene the increases in solubility are inappreciable. Augmentation of the solubility hence occurs with all the acid substances examined and with all non-acid compounds containing amino-groups. Hydroxyl compounds carboxylic acids and phenols apptar to be very greatly influenced.76 ABSTRACTS OF CHEMICAL PAPERS. I n the cases of mercuric chloride and iodine the author has previously shown (Abstr. 1903 ii 547 and 644) that the increase of solubility caused by addition of water is due to the existence of hydrates of these substances in solution. This is probably also the case with the above-mentioned compounds. T. H. P. Partition of some Organic Acids between Two Solvents. WALTER HERZ and MARTIN LEWY (Zeit.EZektrochern. 1905 11 $1 S-S20).-The partition of acetic chloroacetic and picric acids between water on the one hand and chloroform bromoform carbon disulphide and carbon tetrachloride on the other hand is studied. The results point to the existence of double molecules in the more dilate solutions and of more complex molecules in the more concentrated solutions. T. E. Passivity of Metals. WOLF J. MULLER (Zeit. E’lektrochem. 1905 11,8a3-824). -The author’s theory of passivity was incorrectly stated by Ruer (Abstr. 1905 ii 755). T. E. Passivity. CARL FREDENHAGEN (Zeit. Elektrochem. 1 905 1 I 857-S60).-With special reference to Rner’s recent work (Abstr. 1905 ii 795) the author maintains the view that passivity is better explained by assuming a protecting layer of gas (Abstr.1903 ii 353) than by assuming a skin of oxide. The gas theory is capable of accounting for cathodic as well as anodic passivity. The protecting layer of gas may exist as a solid solution as a surface film or in some intermediate condition. T. E. Rate of the Reaction between Arsenious Acid and Iodine in Acid Solution ; Rate of the Reverse Reaction ; and the Equilibrium between them. (11.) J. R. ROEBUCK (J. Physical Chem. 1905 9 727-763. See Abstr. 1903 ii 14).-It has been shown previously that the velocity of the reverse reaction is given by dx/dt = (arsenic acid)(pot. iodide)a(sulphuric acid)@ where a and /3 approach unity in dilute solution. The investigation of this was extended and the values a and /3 found to increase considerably with increasing concentration values as high as 2.4 and 3.7 being obtained. Where the concentration of iodide (C) and acid (D) considerably exceeds that of the arsenic acid the expression may be written dzldt=K5CDV-3(E-x) and it was found that the value of K5 increases as equilibrium approaches and also when clxldt is decreased by decreasing (E - x).The author considers that these anomalous results may be explained by the supposition that the reaction pro- ceeds in two stages namely (1) H + I + H3As0 = H3As0,HI and (2) H,AsO,HI = H,AsO + HIO the reaction of the first proceeding with a measurable velocity. The author extends his determination of the direct rate to near the equilibrium point and from the reaction velocities in the neighbourhood of the equilibrium calculates the eq ui- librium constant which agrees well with that directly determined the two values being 1.4 x lo5 and 1.5 x lo5.The temperature-coefficient + -GENERAL AND PHYSICAL CHER.IlSTR1’. 77 between 10’ and 0” for the equilibrium constant was found to be 1 41 and that calculated from the coefficients of the direct and reverse velocities was 1-43. Experiments were also made with hydriodic acid in p!ace of sulphuric acid and results in accord with theory were also obtained. L. M. J. Affinity Constants of Amino-acids. RUDOLF WEGSCHEIDER (Monatsh. 1905 26 1365-1276. Compare Abstr. 1902 ii 494 ; Walker Abstr. 1904 ii 309 ; 1905 ii 1 3S).-The observed ‘‘ acid ” dissociation constant K of an amino-acid is less than the constant k Calculated from simple stoichiometric relations for those acids which obey Ostwald’s law of dilution or by means of Walker’s constant for the ‘‘ acid ” dissociation of amphoteric electrolytes which do not obey the law of dilution if an intramolecular salt is formed or if the amino- group forms with water an ammonium hydroxide which is less negative but is greater if the ammonium hydroxide is more negative than the amino-group. I f A’ k” .. . are the true constants of the “acid ” dissociation of different acids having an amino-group in the same position relatively to the carboxyl and x’ x” . . . . are the dissociation constants of the corresponding acids with hydrogen in the place of the amino-group then (1) kl’/xl’ = ~iN/xl“ = . . . . -f wheref is the factor representing the influence of the amino-group. I n a dibasic acid which has x1 and x2 the dissociation constants of the two carboxyl groups where fl andf are the factors representing the influence of the two arnino-groups which are usually in different positions.As the observed constant (3) li’=k,P for monobasic acids and it follows from (1) and (3) that (4) -F=fp’=fp”=. . . dibasic acids A’= Flzl + F2x2. (2) kl =fix1 +A?% r / x t = T / X ” . . = F k1P =flPIxl +f2Paxs =!fix1 +fiX,)P. For From (Z) (3) and (4) it follows that As in this equation f and x can be realised only if p = p = p2 p must be the same for all amino-acids the affinity constants of which can be calculated by means of the same factor for the influence of the amino- group irrespective of the position of the amino-group relatively to the carboxyl. observed for acetylaminoterephthalic acid 0.098 f o r 1-methyl hydrogen 3-acetylnminoterephthalate 0.07 and for 4-acetyl- aminoisophthalic acid 0.079 agree moderately well with the constants calculated 0,084 0.066 and 0.069 respectively with the aid of Ostwald’s factors for the influence of the group NHAc* in o- m- and p a c e tylam inoben zoic acids.The P-anilino-fatty acids which are weak acids give no constant factors whereas the stronger a-anilino-acids give only small variations from the mean factors for the anilino-group 2.1 for the p-toluidino- group 0.63 and for the o-toluidino-group 3.3. For the introduction of the acetylanilino-group -NPhAc into the a-position of acetic proyionic and butyric acids the factors are 14 9.4 and 7.3 ; for the introduction of the o-acetyltoluidino-group The constants VOL.XC. ii. G78 ABSTRACT8 OF CHEMICAL PAPERS. into the a-position of the same acids the factors are 12 7.8 and 6.2 ; and for the introduction of the p-acetyltoluidino-group into the a-position of acetic propionic butyric and isobutyric acids the factors are 12 7.8 7.2 and 6.6 respectively. These factors are highest for the introduction of the groups into a methyl lowest for introduction into a methenyl group. The factors for the influence of the amino-group in the ortho- meta- and para-positions obtained from the three aminobenzoic acids 0.17 0.27 and 0.20 respectively are much smaller than the factors obtained from 5-nitro-3-aminobenzoic acid which gives f(NH2,m) = 0.6 1 from 1 -methyl hydrogen 3-aminoterephthalate which gives and from aminoterephthalic acid which gives f(NH,,o) = 0.33 and f(NH,,m) = 1-38.The amino-group appears in some cases as a nega- tive substituent and it may be that in all other cases the smallness of the factor is due to salt formation. f( NH,,o) = 0.33 G. Y. Theory of Dyeing. IV. Lakes. WILHELM BILTZ [with KURT UTESCHER] (Bey. 1905 38 4143-4149. Compare Abstr. 1904 ii 392 ; 1905 ii 807 808).-Experiments have been made by dyeing precipitated metallic hydroxides with alizarin and similar dyes. The results are arranged as curves in which the concentration of the dye- bath is plotted against the number of grams of dye taken up by one gram of hydroxide. With ferric hydroxide and an alkaline solution of alizarin the curve points to the formittion of a definite compound of 1 mol.of hydroxide with 3 mols. of dye. The curve also indicates a certain amount of a d sorption. With chromic hydroxide and alizarin-red SW the curve indicates adsorption and not chemical combination. I n the case of ferric hydroxide the quantity of dye taken up depends to a large extent on the nature of the hydroxide; when gelatinous some ten times as much dye is absorbed as when in the granular s t ate. Experiments with ferric hydroxide and alizarin in ammoniacal or alcoholic solution with the same hydroxide and aqueous solutions of acid alizarin-blue and with aluminium hydroxide and alcoholic gallein solution gave no very definite results. [Construction of the Periodic System.] AUGUSTO PICCINI (G'uzzettu 1905 35 ii 417-42 l).-The author criticises adversely the periodic system of the elements deduced by Werner (Abstr. 1905 ii 308). T. H. P. J. J. S. An Automatic Steam Generator and Superheater. JOHANNES THIELE (Zeit. ancd. Chem. 1905 44 767-768).-The apparatus consists of a vertical boiler having a conical opening down its centre. In this opening is placed a cylindrical vessel. The water is boiled by a buruer under the boiler and the steam passes through a pipe into the upper part of the cylindrical vessel. The burner is now placed under the latter from which the superheated steam'is withdrawn by a pipeINORGANIC CHEMISTRY. 79 reaching to the bottom of the vessel. The burner is regulated by a specially constructed thermostat which by means of a aide-tube is not brought into action until the water in the boiler is at a boiling temperature. w. P. s. New Apparatus for the Extraction of Liquids. EFISIO llrlAmLr (Chenz. Centr. 1905 ii 1569 ; from La Sccrdegm Scmituria 1). -An apparatus is described which only differs from a Soxhlet extraction cyliuder in the following respects. The tube which conveys the vapour of the solvent from the flask is not sealed on to the side of the cylinder but is connected with another tube which passes through the cork and reaches to the bottom of the cylinder. The siphon tube is lengthened YO that it terminates near the bottom of the flask. I n another form a small tubulns is attached to the bottom of the cylinder through which a tube passes t o the bottom of the flask reaching to a convenieut height in the cylinder and being straight or bent in the upper portion into a U form. The vaponr of the solvent passes through a tube fitted in the cork of the flask and coonected with the extraction cylinder as described above. E. w. w.