105 General and Physical Chemistry Reflective Power of Ethyl Alcohol. HEINRICH RUBENS and ERIC LADENBURG (Sitxungsber. K. Akad. Wiss. Berlin 1908,1140-1143). -By means of a micro-radiometer the authors have measured the reflective power of ethyl alcohol for ultra-red radiation. In order to avoid errors resulting from the absorption of the radiation in the vapour above the liquid this was coded to a temperature of about - 20. The experimental data indicate the existence of selective reflection this being more pronounced than in similar experiments a t 18'. The maxima on the curve representing the reflective power as a function of the wave-length are all found to be displaced towards the visible spectrum by a lowering of the temperature from 18' to - 2 O O . Whether this is due to the diminution in the absorption of the alcohol vapour in contact with the liquid has not been established.H. M. D. Refraction and Dispersion of Krypton and Xenon and their Relation to those of Helium and Argon. CLIVE CUTHBERTSON and M. CUTHBERTSON (Yroc. Roy. Xoc. 1908 A 81 440-448).-The measurements were made by a method previously described (Cuthbert- son and Metcalfe Abstr. 1908 ii 545). The values of (p - 1).1OG for the green mercury line (h=5461) are for krypton and neon respectively 428.74 and 705.49. The dispersion for krypton is given by the formula p - 1 = 0.0004189(1+ 6*97/h2. loll) and that for xenon by p - 1 = 0.0006823(1+ 10*14/h2.1011) both gases being at normal temperature and pressure. When the refractivities for infinite wave-lengths of helium neon argon krypton and xenon are compared the numbers are found to be very nearly in the ratio 1 2 8 12 20.If the values of b in the equation p - 1 = a(1 + b/X2) are plotted against the refractivities for infinike wave-lengths a straight line can be drawn through the points representing helium krypton and xenon ; the corresponding point for argon is also quite close. On the other hand mercury does not fall into line with the other four monatomic elements in respect of the relationship between refractive and dispersive power. H. M. D. Spectroscopic Investigation of the Behaviour of Metallic Salts in Flames of Different Temperatures. HERBERT AUERBACH (Zeitsch. wks. Photograph. Photophysik. Photochem. 1909 7 30-39).- The apparatus used and details of the photographic method of regis- tration are described.For the production of spectra corresponding with different temperatures the Bunsen ordinary blowpipe and oxygen blow-pipe flames have been employed. No experimental results are given in this part of the paper. H. M. D. VOL. xcvl. ii. 8106 ABSTRACTS OF CHEMICAL PAPERS Radiation and Temperature of the Flame of the Buneen Burner. E. BAUER (Compt. rend. 1908 147 1397-1400).-The absorption and emission of certain rays by illuminating gas flames has been measured. The radiation from a source traversed a con; siderable thickness (3-5 cas) of a homogeneous flame it then passed through a hole in a screen was reflected three times from a polished surface of fluorspar and mas then examined with a micro-radiometer.The absorption which is due to the vapour of water is selective; a flame which absorbs only 11% of the rays emitted by a solid substance giving a continuous spectrum absorbs 28% of the radiation from anothet gas flame. From the emission and absorption measurements with the rays emitted by a solid substance and a knowledge of the emissive power of a black substance in the same position as the flame the temperature of the flame a little above the blue interior cone has been calculated by application of Kirchhoff’s law ; it amounts to 1760 & 50’. Extension of the Principal Series of the Sodium Spectrum. ROBERT W. WOOD (Phil. Nag. 1908 [vi] 16 945-947*).-The ultra-violet absorption spectrum of sodium vapour consists of a Balmer series of lines of which the author has photographed thirty this making the longest Balmer series known.In this series the D-lines appear as the third member whilst the end is a t X 2411. The photographs were obtained by passing the light from a cadmiumspark through a steel tube about one metre long containing sodium vapour at dull-red heat under nitrogen pressure of 15 cms. Fraunhofer Lima of Thallipm and an Error in Rowland’s New Table of Standard Wave-Lengths. CARL FHITSCH (Physihl. Zeittfcch. 1908 9 900-901).-The author points out that Rowland’s table of wave-lengths represents certain thallium lines as belonging to thorium. The spectral line measurements of thallium made by different observers are compared. Re-determination of the Wave-Lengths of the Iron Lines used for Comparison Purposes.A. H. PFUND (Zeitsch. Wit?& Photograph. Photophysik. Photochem. 1909,’? 1-16).-At the instance of the International Union for Co-operation in Solar Research the author has re-determined the wave-lengths of certain iron lines by the interference method of Fabry and Perot. Details are given of the apparatus used and of the method of mounting silvering and polishing the plates of the interferometer. The measurements are tabulated and compared with the results obtained by the same method by Fabry and Buisson and by Eversheim. The independent series of measurements are in very good agreement. Thermal and Chemical Absorption in Banded Spectra JOHARNES STARK (Physikd. Zeitsch. 1905 9 889-894).-The previous speculations of the author (compare Abstr. 1907 ii 147 417; 1908 ii 138 746 911) are supplemented by a consideration of the relationships between coupled bands of the nature of the double * aiid Phyez‘kal Zcitsch.1909 10 88-90. G. S. R. J. C. H. M. D. H. M. D.GENERAL AND PHYSICAL CHEMISTRY 107 and triple linkings in unsaturated compounds of the so-called oxide ’spectra of the metals and of the uitra-violet absorption of saturated valency electrons. In the second part of the paper application of the author’s ideas is made to photochemical phenomena. Primary and secondary photo- chemical changes are distinguished. Primary changes are characterised by the fact that certain valency electrons absorb more energy in the form of light than that corresponding with the energy of combination of the electrons with the atomic or molecular remainders.Three general relationships which hold for primary photochemical changes are deduced from the author’s hypothesis. Theory of Absorption in Gases. L ~ O N BLOCH (Compt. T e d . 1908 14’7 1284-1 286). -According to Drude’s theoryiof dispersion the motion of the electrons is affected by friction and the coefficient of friction T is assumed to be constant (independent of the wave-length). The author shows that in the case of absorption in gases c is a function of the wave-length. Employing the kinetic theory of gases the formula obtained leads to the conclusion that r varies aproximately as the square root of the absolute temperature a deduction in agree- Rotatory Power at Low Temperatures and the Relation between the Absorption of Light and Rotatory Polariaation in Crystale of Cinnabar.JEAN BECQUEREL (Compt. rend. 1908 147 1281-1284).-Cotton (Th&e 1896) has shown that there is a connexion between absorption of light and rotatory power in solution and the author now shows by experiments with well-formed crystals of cinnabar that there is a similar relationship in active crystals. The curve of rotatory dispersion at 18’ and -188O and the flutings in the spectrum at the same temperatures are shown side by side on the same figure. The rotatory power increases rapidly near the absorbed region. On cooling to - 18W the yellow and a part of the red appear and the great rotation appears no more in the orange but is thrown back into the green from which it is evident that the great rotatory dispersion is connected with the absorption.Cinnabar posseses considerable magnetic rotatory power and this property is being further investigated. Magnetic Rotatory Power of the Vapour of Calcium Fluoride and of Nitrous Oxide in the Neighbourhood of their Absorption Bands. A. DUFOUR (Compt. Tend. 1908 147 1471-1472).-1n the case of calcium fluoride the three bands X = 6037 X = 6050.8 and X = 6064.5 were examined. The last shows the normal Zeeman effect the other two the inverse phenomenon. The nitrous oxide absorption bands X = 5854.9 and X = 5846.8 show the ordinary whilst X = 5925.4 and X = 5856.9 exhibit the inverse effect. H. X. D. HUGO RAUFFWANN (Bey. 1908 41 4396-441 2).-Yellow nitroquinol di- H. M. D. ment with recent work by J. Becquerel. Q s. G. S. Nitroquinol Dimethyl Ether and Theory of Solution.S-2108 ABSTRACSS OF CHEMICAL PAPERS methyl ether yields coloured solutions in dissociating solvents and practically colourless solutions in non-dissociating solvents (Abstr. 1907 i 127). Hantzsch claims (Abstr. 1908 ii 447) that the ordinary yellow preparation cannot be the true nitroquinol dimethyl ether ; the latter is colourless and existent in solutions in feebly dissociating or non- dissociating solvents. His claim is based on the assumptions that all true nitrecompounds are colourless and that the molecular refraction of nitroquinol dimethyl ether in feebly dissociating solvents is only slightly greater than the calculated value whilst in solutions in strongly dissociating solvents a pronounced exaltation of the molecular refraction is observed.The author points out that Hantzsch contradicts himself since he states’(Abstr. 1907 i 513) that nitrobenzene shows absorption in the extreme violet which is only another way of saying that nitro- .benzene is not colourless. With regard to the second point it is known that substances containing auxochromes in the benzene nucleus exhi bit abnormally high molecular refractions. Therefore the true nitroquinol dimethyl ether must be that form existent in solutions in strongly dissociating solvents that is to say the yellow modification. The author however rejects the theory of the existence of several modifications of nitroquinol dimethyl ether and accounts for the varying colour of its solutions in different solvents by a novel concep- tion. The outlines of the theory are as follows Since the change of colour of a fluorescent substance in different solvents occurs in general without change of constitution (Abstr.1904 ii 52S) the conception of constitution alone is insufficient. A constitutively unchangeable substance must have some capacity of change within its molecule so that the molecule under different conditions can assume different states. The graphic representation of such states which necessitates the theory of the divisibility of valencies is called the conditioning formula. The substance has in general a single unchangeable constitutional formula and several conditioning formula The constitutional formula is an ideal state to which the conditioning formulae approximate more or less according to the conditions.Furthermore the author assumes that the divided valencies represent lines of force. A line of force starting from an atom of a molecule may terminate either in another atom of the same molecule or in R different molecule the latter case being important in the case of solutions. The molecules of a dissociating solvent have the power of attracting to themselves the lines of force issuing from the molecules of a solute. Non-dissociating solvents on the contrary tend to make the lines of force terminate in the molecule in which they originate. By this theory the author attempts to account for the change of the fluorescent colour of a substance in different solvents (Abstr. 1907 ii D ~ S I R ~ GERNEZ (Ann. Chim. PAY& 1 908 [viii] 15 516-556).-A resume of work already published (compare Abstr.1905 ii 365 430 431; 1908 ii 748) together with a claim far priority against Trautz (Abstr. 1905 ii 494 662). Contrary to Tschugaeff’s generalisation (Abstr. 1905 ii 132) d- I- and r-erythritol 519). c. s. Triboluminescence.GENERAL AND PHYSICAL CHEMISTRY. 109 all exhibit triboluminescence (Bertrand Abstr. 1900 i 424 ; Maquenne Abstr. 1902 i 131). M. A. W. Disintegrating Action of Light and Optical Sensitisation. JOHANNES STARK (Physikal. Zeitsch. 1908 9 894-900).-The dis- integrating effect which ultra-violet light has been found to exert on gold-leaf exposed to its action is shown to be in accordance with theoretical requirements. An experiment is described which appears to show that positively charged atoms are emitted in consequence of the absorption of ultra-violet light.The action of sensitisers in promoting certain chemical changes is discussed in terms of the author’s theory. H. M. D. Concerning the [Luminiferoue] Ether. W. A. KURBATOFF (J. Rzlse. Phys. Chem. Soc. 1908 40 1468-1471).-The author discusses the ether in relation to the atomic and electron theories. T. H. P. Action of Radium- and Rontgen-rays on the Colours of Precious Stones. CORNELIO DOELTEB (Monatsh. 1908 29 11 45-1 170).-A large number of observations are recorded relating to the changes in colour of precious stones and coloured glasses on exposure for different lengths of time to the action of radium- and Rontgen-rays. The colour changes are expressed in terms of the Radde scale. The two kicds of rays appear to act similarly on the coloured minerals.On account of the shorter periods of exposure in the case of the Rontgen-rays a comparison of the activity of the two types of radiation is not possible. I n certain cases the colour changes produced by tbe action of radium are reversed when the stones are heated to about 300° in oxygen or nitrogen. The nature of the colouring substances is discussed. Amethyst smoky quartz rose quartz and citrine are said to contain traces of iron thiocyanate yellow diamonds traces of iron and possibly cerium ruby traces of chromium and possibly iron. Chromium is also probably present as the chief colouring agent in yellow and green corundum and in certain sapphires. H. IN. D. Decay of Radium Emanation when Dissolved in Water.Josh MuRoz DEL CASTILLO (Anal. Fit?. &uim. 1908 6 488-489).-A claim for priority (compare Moore Abstr. 1908 ii 651). W. A. D. Radioactivity of the Atmosphere on Mountains. ALBERT GOCKEL and TH. WULF (Physikal. Zedsch. 1908 9 907-911).- Measurements have been made of the induced activity collected on wires suspended for periods of twenty-four to one hundred hours at various points in the neighbourhood of the Matterhorn. The decay curves indicate that a t a height of about 3000 metres the amount of induced thorium activity is relatively small varying from 0.5% to 10%. The activity of collected rain-water was found to be of the same order of magnitude as that collected in Freiburg (Swit8sTland) at a height of 650 metres.110 ABSTRACTS OF CHEMICAL PAPERS With a special type of electrometer the ionisation due to very penetrating rays mas examined.No influence of elevation on such ionisation could be detected and the authors surmise that this penetrating type of radiation emanates from the earth’s surface. H. M. D. Radioactivity of Italian Gaseous Emanations. RAFFAELO NASINI and MARIO G. LEVI (Atti R. Accad. Linmi 1908 [v] 17 ii 551-553).-The authors have measured the radioactivities of a number of gases of various Italian origins of which they also give certain analytical data. The greatest radioactivity was exhibited by the gas composed partly of hydrogen sulphide from the thermal springs of Abano (Padua). T. H. P G. COSTANZO and C. NEGRO (Jahrb. Radioaktiv. Elektronik 1908 5 395-402. Compare Abstr.1908 ii 551).-The activity of fresh rain-water collected during March and April 1908 was examined by bubbling air through it and leading the gas into the ionisation chamber of an electroscope. I n some cases the rate of discharge was greater when the electroscope was positively charged in others when the charge was negative I f the rain-water is perfectly fresh and especially if its fall is accompanied by atmospheric disturbances of an electrical nature the negative ions are in excess. Radioactivity of Waters from Alange. Josl Mu3oz DEL CASTILLO (Aml. Pis. Quirn. 1908 0 489).-The waters of Alange which have great value in the treatment of nervous diseases have a radioactivity of 1495.9 volts per hour litre. MAX BAMBERGER (Monatsh. 1908 29 1131-1 140).-The radioactivity of spring-water emanating from the South Bohemian granite region in the neighbourhood of Schloss Tannbach has been found to be very high 47-52 Mache units.The rate of decay of the induced activity indicates that the activity is due to disintegration products of radium Radioactivity of Roumanian Mineral Waters. DRAGOMIR HURMUZESCU and N. PATRICIU (Ann. Sci. Uniu. Jccssy 1908 5 159-1 65).-From measurements of the ionising power of the emana- tiop obtained from various mineral waters the authors have compared their radioactivities. These are of the same order of magnitude as the Swiss waters examined by Sury Sarasin Guye and Micheli if the highly radioactive waters of Dissentis and Lavey-les-Bains are excepted. H. X. D Radioactivity of Certain Spring-waters of the Semmering Province.MAX BAMBERGER (Mmatth. 1908 29 1141-1144).- Numbers representing the radioactivity of spring-waters are recorded. The springs emanating From quar tz-phyllite are the most radioactive whilst those from limestone exhibit the property very feebly. Ionisation Phenomena caused by Rain-water. H. M. D. W. A. D. Radioactivity of Certain Springs in Upper Austria. H. M. D. H. M. D.GENERAL AND PBYSICAL CHEMISTRY. 111 SpeqifIc Rdioaotivity of Thorium and ifs Produota. GEOBG~ C. ASHMAN (dmsr. J. SOi. 1909 97 65-72),-The specific aotivity a0 thorium has been deduced from experimental measurements of the activity of new and old thorium (1) freed from the easily separated products thorium=X emanation thorium-4 -23 and -C (2) after a period sufficiently long for the disintegration products to h v e aocumulated in the equilibrium quantities In order to separate thorium and radio-thorium from the other dis- integration products the authors have made use of m-nitrobenaoic acid After four precipitations of the thorium at 80' and at intervals of two hours the disintegration products are completely removed ; the precipitated thorium maitrobenaoate is then heated over a Bunses flame and finally for ten minutes over the blow-pipe.From the measurements specified above the author finds that the specific activity of thorium is 119 This represents 11% of the total activity of the thorium when the equilibrium condition has been reached. Of the remaining activity 20% is due to radio-thorium and 69% to thorium-X and its subsequent products.Heat Development due to Radioactivity of Thorium Oxide. G. B. PEGRAM and H. WEBB (Chem. Zentr. 1908 ii 1851 i from Le Radium 1908,5,271-276).-By means of a thermo-couple the temperature difference has been measured between 4 kilos. of thorium oxide in a Dewar vessel supported in a partly evaouated metal cylinder which was surounded with ice and the ice jacket. A similar measurement was made when the thorium oxide was heated by pass- ing a known current of electricity through a wire embedded in the mass. From these results it is found that the liberation of heat per second per gram is equal to 1.11 x gram- calories per hour per gram. T. H. LABY and G. W. C. KAYE (Phil. Mag. 1908 [vi] 16 879-889).-The authors have measured the ionisation produced by y-rays of radium in air hydrogen and carbon dioxide a t pressures from one to seventeen atmospheres.The y-rays passed through 7 mm. of aluminium and 10 mm. of brass into a flat chamber containing an electrode and filled with the gas under investigation. The ionisation directly induced by the y-rays is supposed to be small in amount and proportional to the pressure of the gas. The main effect is that of secondary p-rays from the alum- inium electrode ; this also varies as the pressure. Secondary P-rays excited in the gas which would produce an ionisation proportional to the square of pressure are not in evidence. The amount ,of ionisation produced in the case of hydrogen was approximately proportional to the pressure. In air and carbon dioxide the ionisation iucreased less rapidly than the pressure part- icularly a t high pressures.This phenomenon is attributable to a soft tertiary radiation produced in these gases by secondary p-rays. This would give a relatively larger effect at low pressures since at higher pressures the gas would be more opaque to it and the range of the prticles would be smaller. H. M. D. watts or 9.60 x J V. E. Gaseous Ionisation and Premure. R. J. C.112 ABSTRACTS OF CHEMICAL PAPERS. Experimental Investigation of the Nature of y-Rays. WILLIAM H. BRAGG and JOHN P. V. MADSEN (Phil. Mag. 1908 [vi] 16 918-939).-The authors adduce further experimental evidence in favour of the theory that y-rays are material and consist of p-particles combined with a positive charge (Abstr. 1908 ii 556).The y-rays from radium were purified as well as possible from p- rays by passing through a magnetic field and then allowed to impinge on plates of various elements. Measurements were made of ionisation on both sides of the plates that is of reflected secondary @-radiation and of emergent secondary P-radiation. The reflected secondary p-radiation depends very largely on the atomic weight of the material of the plate employed but the emergent p-radiation which is almost always greater is of the same order for all the elements. The production of @rays by hard y-rays appears to follow a density law independent of atomic weights but softer y-rays are affected by atomic considerations in that heavier atoms are the more absorbent and produce more secondary p-radiation.The speed and penetrating power of secondary p-radiation are governed by the corresponding properties of the primary y-radiation. The y-rays of radium produce secondary /3-rays which closely resemble the primary @rays of radium in speed. It is argued that the energy of the secondary p-ray does not come from disintegrated atoms but from the exciting primary y-ray. The authors' theory although it requires the existence of positive electrons is held to be simpler and more complete than that advocated by J. J. Thomson. The theory is also applicable to X-rays. R. 5. C. Law of the Optimum of Cathodic Phosphorescence in Binary Systems. GEORGES URBAIN (Compt. rend. 1908 147 1472-1 474).-The discovery of the general relationship in question is not due to the author but to Lecoq de Boisbaudran.I f instead of referring to the entire phosphorescent spectrum the component parts are taken into consideration it is possible to give the law of the optimum a more precise form. According to this if the relative proportions of the active substance and the diluent are varied each phosphorescent band passes through a maximum. Although the maxima for different bands do not necessarily correspond with the same composition of the binary system the proportion of the active eomponent is always very small. Minimum Quantity of Electricity. N. L. M~LLER (Jal~rb. Radioaktiv. Elektronik. 1908 5 402-408).-An explanation is given of the fact that the mass of the a-particle according to electro- magnetic measurements is only 2 whereas if the a-particle be a helium atom it should be 4.If Nis the number of ions in a gram equivalent then 2 x 96540,/N is the quantity of electricity necessary for the separation of a molecule of hydrogen by electrolysis. This is supposed to be taken up by the hydrogen ions at the negative pole. Denoting the quantity by 6 the process of formation of a hydrogen molecule is represented by the equation 2H' + e = H2r. Other molecules oan H. X. DGENERAL AND PHYSICAL CHEMISTRY. 113 be similarly formulated and in the case of helium the formula of the atom is Her whereas that of the a-particle is He and its mass is equal to 2. The direct measurements of ionic charges are discussed in terms of the author's hypothesis. H. M. D. Variation in the Electrical Conductivity at Constant Temperature of Mineral Waters Containing Radium.Jos& MuRoz DE CASTILLO and FAUSTINO DIAZ DE RADA (Anal. Pis. Quim. 1908 6 485-488).-Measurements are given of the electrical conductivity and radioactivity of a number of natural Spanish mineral waters. The authors are collecting data which will enable them to discuss the relationship existing between the changes of conductivity and radioactivity of natural waters. W. A. D. Relation of the Charge to the Mass of the Electron. Comparison of the Values Deduced from the Zeeman Effect and from Recent Measurements with Cathode Rays. A. COTTON and P. WEISS (Compt. rend. 1908 147 968-970).-From measurements of the distance between the side components of certain lines in the spectrum of zinc which become triplets in the magnetic field the value 2 x 1.767 x 10-7 is obtained for elm (e=electric charge m=mass of electron) on the basis of the theory of Lorente; a result practically double that obtained for t,he same ratio from measurements with slow cathode rays.From similar measurements on the helium Bpectrum however Lohmann (Abstr. 1908 ii 243) obtains e/m = 1.77 x 10-7 a value in good agreement with that obtained from observations on cathode rays. G. S. Electrical Resistance of the Alkali Metals Gallium and Tellurium. ANTOINE GUNTZ and WITOLD BRONIEWSKI (Compt. rend. 1908 147 1474-1 477).-The metals were introduced into capillary U-tubes the resistances being compared directly with that of mercury. Sodium potassium rubidium and caesium were obtained from the chlorides by the action of calcium and lithium by decomposition of the hydride.The specific resistances of these metals are recorded for a series of temperatures between - 18'7' and the melting points of the metals. At 0" the values are caesium 19.30 rubidium 12.80 potassium 7.01 sodium 4-30 lithium 8.55. Corresponding with the contraction which takes place on fusion the specific resistance of gallium is greater in the solid than i n the supercooled liquid condition. For the solid the value at 0" is 53.4. The specific resistance of tellurium at 0' is 0.102; it increases with the temperature attains a maximum at about 50° and then diminishes until the melting point is reached. H. M. D. E.M.F. of the Cadmium Normal Element at 0". I. ERNST COHEN and HUGO R. KRUYT (Zeitsch. physikal. Chew. 1909 65 359-370).-Jouaust (Compt.rend. 1908 147 42) has stated that whilst the behaviour of cadmium elements made up with an amalgam gostaining 12.5% of cadmium by weight is regular down to 0" the114 ABSTRACTS OF CHEMICAL PAPERS E.N.P. of different cells containing a 10% cadmium amalgam show9 deviations of several ten-thousandths and these cells show certain other irregularities. The authors now show that there is a small but probably real difference (since it is always in the same direction) not exceeding 0.2 millivolt between cells made up with 10% and 12pL amalgams respectively and consider that this small difference is due mainly to the difference in the concentration of the amalgams i n the two cells in accordance with the equilibrium relations discussed by By1 (Abstr.1901 ii 507). The differences however are not muah greater than those observed with '' identical " cells made up with the same amalgam which may also amount to 0.2 millivolt. The statement of Jouaust (Zoc. cit.) that at 0' the use of a 12.5% amalgam is to be preferred is therefore inaccurate. Certain other of Jouaust's statements are criticised but it is admitted that a further investigation of the electromotive behaviour of cadmium -amalgams betwe& 0' and 30' is desirable. G. s. Electrolytic Processes at the Surface of Electrodes. HANEI G. MOLLLR (Zeitsch. physikal Chm. 1908 65 226-254).-According to Nernst the phenomenon of supertension ( ' I uberspannung ") is connected with the capacity of the electrode material to occlude gases. The author considers that the power of occluding gases is connected with the molecular forces between electrode and gas and hence with surf ace tension phenomena.These theoretical deductions have been confirmed experimentally ; it is shown that there is parallelism between the supertension and the surface tension between gas and electrolyte and also between the supertension and the angle at the junction of gas and electrode. The supertension at different electrodes has been determined by constructing curves representing the variation of strength of current with B.N.F. and the precautions required to obtain accurate results are fully described. The angle 8 between gas and electrode was read off directly by means of a microscope capable of turning on its axis and a special method for obtaining conditions suitable for measurement is described.The relationship between supertension and the two capillarity factors mentioned above has been proved by experiments with mercury at temperatures between 0' and lo' and also with other electrodes the surfaces of which were of different degrees of roughness and in some cases were covered with a thin layer of fat. The relation between 8 and the supertension ?T is represented with great accuracy by the equation T = (eae - l ) / b which holds for all temperatures and all electrodes under any conditions of surface For hydrogen and N/lO sulphuric acid the constant a=O*O264 and b = 9 . It has not hitherto been possible to determine the absolute value of the surface tension at the junction of solid metals but the curve representing the variation of the angle 8 as a function of the polarisation must be parabolic and have a maximum at the game point as the electro-capillary curve.It is shown experimentally thatGENERAL AND PHYSICAL CHEMISTRY. 115 &be mtlximum is the same for all the metals and electrolytes examined provided the electrolyte contains metallic ions but no complex ions and that this value is at zero potential. If the metal is capable of forming complex salts the maximum is displaced. The important result is thus established that the potential difference between electrode and electrolyte disappears when the surface tension at the electrode attains a maximum value ; hence trustworthy results for absolute differences of potential are obtained by the use of the capillary electrometer.G. S. [Oxide Theory of the Oxygen Electrode.] EMIL BOSE (Zeitsch. EZektrochem. 1909 15 1 l-l2),-Polemical in reply to Lorenz (this vol. ii 15). T. E. Electromotive Behaviour of Mixtures of Uranyl and Uranous Salts. ROBERT LUTHER and ARTHUR C. MICHIE (Zeitsch. Elektrochem. 1908,14,826-829).-Uranous salts are strong reducing agents even in acid solutions ; the oxidation by atmospheric oxygen is accelerated by iron salts platinum black and especially by copper salts. The difference of potential between a platinised platinum electrode and a solution of uranous and uranyl salts in sulphuric acid is given by E = c0 + RT/2Plog[U02"][H* J4/[U""] where co is 0.419 volt more positive than the normal calomel electrode or 0.696 volt more positive tban the normal hydrogen electrode.The E.M.P. of mercury-mercurous sulphate electrodes in 0*1N to 1-0N-sulphuric acid is very little affected by the strength of the acid. This is shown to be in quantitative agreement with Luther's views of the dissociation of sulphuric acid (Abstr. 1907 ii 610). T. E. Minimum of Conductivity in the Titretion of Acids. ALFRED THIEL (Zeitsch. Elektrochern. 1909 15 1-4).-Bruni (Abstr. 1908 ii 935 1012) pointed out that the equation by means of which the author calculated the dissociation constants of weak acids from the position of the minimum conductivity of the partly-neutralised solutions (Abstr. 1908 i 787) was based on incorrect assumptions. I n the present paper the author gives the correct equation for the case that the acid is titrated with 0*1N sodium hydroxide (the volume of the solution increasing).The assumptions made are the same as those made by Bruni and the resulting equation becomes identical with his for the limiting case that the volume of the solution does not change during neutralisation. T. E. Formation of Salts and Basicity of Acids. 111. GIUSEPPE BRUNI and 0. SANDONNINI (Zeitsch. Elektrochem. 1908 14 823-825. Com- pare Abstr. 1907 ii 935,1012).-The theoretical equation previously obtained indicates that when an acid the conductivity of which is less than that of its neutral salt is progressively neutralised the con- ductivity must pass through a minimum which is displaced towards the salt end of the curve with increasing strength of acid and increasing dilution.The existence of this minimum is now proved for formic and acetic acids and p-nitrophenol. In all these cases it lies very close to the acid end of the c u m . T. E116 ABSTRACTS OF CHEMICAL PAPERS. Influence of Temperature on the I n t e r d Energy and the Free Energy of Electrolytic Dissociation of Weak Acide and Bases. HARALD LUND~N (J. China. phys. 1908,6 681-698).-The methods by which the heats of dissociation and of neutralisation may be calculated from the hydrolysis and dissociation constants are discussed. The relations between the heats of dissociation and neutralisation respectively and the absolute temperature !P are best expressed by formulae of the type U=f(a b T) for example U= a + bT (U= internal energy ; a and b are constants).The resulting formula are applied to calculate the changes of internal energy and of free energy obtainable in the electrolytic dissociation and neutralisa- tion of a number of weak acids and bases. The temperature-coe5cients of the internal energy dU/dT and of the free energy dA/dT and the inferior and superior limits of these coefficients have been obtained. Brom Nernst's theorem it is shown that dA/dT and dU/dT should have the same absolute value but be of opposite sign a deduction which is in moderate agreement with the facts. Magnetic Properties of Simple Substances. P. PASCAL (Compt. rend. 1908 147 1290-1292. Compare Abstr. 1908 ii 756 927).-It is shown that for diamagnetic substances of the same valency and of analogous chemical properties the atomic susceptibility x (the product of specific susceptibility and atomic weight) is an exponential function of the atomic weight a ; hence the formula logex=a+Pa applies where a and p are constants for the same group but vary especially p with the valency.The validity of the law is shown by observations on the halogen sulphur,and phosphorus groups. It holds independently of the temperature. G. S. Magnetic Dichroism of Calcite and Dolomite admixed with Liquids. GEORGES MESLIN (Compt. rend. 1908 147 1277-1279). -Calcite and dolomite form in association with a number of liquids systems which show magnetic dichroism but the sign of the effect is contrary throughout for the twp substances. I n both cases the change of sign takes place when the refractive index of the liquid exceeds 1.600 ; for calcite the change is from positive to negative at this point and conversely for dolomite Magnetism of the Rare Earths.B. URBAIN and GUSTAV JANTSCH (Compt. rend. 1908 147 1286-1 288).-The magnetic susceptibility of t h e pure sesquioxides of a number of rare earths has been deter- mined with the following results Element ............ Nd. Sa. Eu. Gd. Tb. Dy. Atomic weight ... 144'3 150.4 152.0 157.3 159.2 162.5 Magnetic power ... 33'5 6.5 38.5 161.0 237'0 290.0 G. S. G. S. The relative numbers in the third line are referred to the value for cobalt sulphate heptahydrate = 39.7. The above and other results indicate that there is maximum of magnetic power in the cerium group and a second in the yttrium group. As the magnetic power differs so greatly for closely-allied elements it may be used as a test far the camposition of mixtures.G. S.GENERAL AND PHYSICAL CHEMISTRY. 117 Measurement of Very Low Temperatures. XXII. The Thermo-element Gold-Silver at Liquid Hydrogen Tempera- tures. E KAMEBLINUH ONNES and J. CLAY (Proc. K. A h d . Wetemch. Anasterdam 1908 11 344-345).-The temperature-coefficient of the E.M.K of the gold-silver thermo-element increases rapidly with falling temperature. The great sensitiveness of the element at the temperatures of liquid hydrogen renders it suitable for temperature measurements in this region. H M. D. New Model of the Calorimeter of N. A. Hesehus and Determination by its means of the Specific Heat of Alloys of Tin and Bismuth. N. MARENIN [J. Buss. PAYS. Chem. Xoc. 1908 40 393-398 ; (Phys.Part) ].-This modification of Hesehus’ calori- meter (J. Phys. 1888,489) allows of the weight of water in the calori- meter tube being determined by weighing this tube alone instead of the whole calorimeter. The author has measured for alloys of bismuth and tin containing from 0 to 100% of bismuth (1) the temperatures of fusion ; (2) the specific heats which increase continuously with the content of bismuth and (3) the densities which diminish continuously as the percentage of bismuth increases. T. H. P. Specific Heat of Solid Substances. MAX THIESEN (Ber. Dezct. physikal. Ges. 19OS 6 947-952. Compare Abstr. 1908 ii 653 808 ; Griineisen Abstr. 1908 ii 653).-A consideration of specific heat data for elements in the solid condition leads the author to conclude that the coefficient of expansion and the specific heat vary with the temperature quite differently.With rise of temperature the coefficient of expansion increases the more rapidly. Although the formula given by the author can be used to calculate specific heat values for a wide range of temperatures the empirical relationship pointed out by Griineisen is incorrect. The remainder of the paper is devoted to a consideration of Einstein’s theory of the solid state of aggregation as furnishing a satisfactory basis for a theory of the specific heat of solids. H. M. D. Plait Point Temperatures of the System Water-Phenol. A. KEESINU (Proc. K. Acad. Wetensch. Amsterdam 1908,11 394-397). -Measurements ol critical temperatures for mixtures of water and phenol have been made by heating quartz tubes containing mixtures of known composition in the vapour of pure benzidine in contact with an atmosphere of nitrogen.A minimum critical temperature (357’) was found corresponding with 10% of phenol. The temperatures for pure water and phenol respectively are 3 6 4 ~ 7 ~ and 419.2O. The curve connecting the critical temperatures for mixtures between 10 and 100% of phenol exhibits a point of inflexion. H. M. D. Variation of Trouton’s Constant in a given Homologous System. W. A. KURBATOFF (J. Rws. Phys. Chem. Soc. 1908 40 1493-1501).-The author has collected or calculated the values of Trouton’s constant for members of the following series (1) elements ; (2) inorganic compounds ; (3) saturated hydrocarbons ; (4) poly-318 ABSTRACTS OF CHEMICAL PAPERS.methylene compounds j ( 5 ) homologues of benzene ; (6) ethers and (7) esters. The values for series (3) (Q) and (5) show that the value of the constant diminishes as the molecular weight and boiling point rise. In the first two of these series the constant exhibits a maximum value at b. p. 100-125O. With the ethers too the constant diminishes rapidly as the molecular weight increases. T. H. P. Melting Points. P. B. DALLTMORE (Phawn. J. 1908 [iv] 27 802).-A form of apparatus is described by means of which the m. p.'s of fatty acids waxes and similar substances can be readily determined. It consists of a glass U-tube one arm of which is longer than the other and open at the top. The upper end of the shorter arm is bent at right angles and drawn out to a fine point having a small aperture a t the end.The tube is completely filled with the melted substance and when cold is attached to a thermometer and heated in the usual way the open end of the longer arm being slightly above the level of the liquid in the bath. As soon a8 the m. p. is reached a drop of liquid is forced out of the aperture by the weight of the liquid in the longer arm. E. G. Hydrogen Polysulphides and Cryoscopy. EZIO PAT ERN^ (Atti R. Accad. Lincei 1908 [v] 17 ii 627-633).-1n criticising the conclusions drawn by Bruni and Borgo (Abstr. 1908 ii 102) concerning the formulix of hydrogen polysulphides the author points out that the cryoscopic constant of bromoform is variable and that no values obtained for the molecular weight of sulphur correspond exactly with the molecule S (compare Orloff J.Russ. l'hys. Chem. SOC. 1903 35 642). The calculation made by Bruni and Borgo of the apparent molecular weight when one of the molecules H,S H,S and H,S or a mixture of them exists in solution is faulty since the same results are obtained in all cases where the relation between the numbers of hydrogen and sulphur atoms is constant the total number of molecules also remaining unchanged. The uncertainty of the results obtained by these authors is confirmed by the wmk of Tanatar Choina and Kozireff (Abstr. 1895 ii 41) who found that the depressions of the freezing point of water caused by organic non- electrolytes are not altered sensibly by adding to the water 10 or 20% of methyl or ethyl alcohol whilst with salts the addition of alcohol increases the depressions obtained in water alone.The author has measured the freezing-point depressions of bromo- form produced by varying quantities and proportions of sulphur and carbon disulphide. The mean difference between the experimental results and the calculated ones is 6.8%. Further Bloch and Hohn (Abstr. 1908 ii 5'79 580) and Schenck and Falcke (Abstr. 1908 ii 762) have shown that crude hydrogen polysulphide is not a mixture of H,S H,S H2S7 and H,S but that it contains the two sulphides H,S and H2S3. The melting-point curves of mixtures of selenium and iodine obtained by Pellini and Pedrina (Abstr. 1908 ii 833) do not showGENERAL AND PHYSICAL CHEMISTRY. 219 that these dements do not form a compound but onIy that such compound decomposes before the eutectic temperature is reached.Olivari’s conclusions (this vol. ii 37 39) are also criticised. Law of Corresponding Boiling Points. ULRICH DUHRING (Zeit8cI.b. phgsikacl. Chem. 1909 65 338-340).-8 claim for priority with reference to the discovery of the law in question recently ascribed by Johnston (Abstr. 1908 ii 358) to Ramsay and Young. G. S. T. H. P. Relation between the Capillary Clonstants and the Latent Heat of Vaporisation of the Solvent. PAUL WALDEN (Zeitsch. physikal. Chm. 1909 65 257-288. Compare this vol. ii 122).- It is shown from the available data for over sixty liquids that there is a simple relationship between the heat of vaporisation A and the specific cohesion a2 (a2 = rh where ii is the height to which the liquid rises in a capillary tube of radius r mm.) both determined at the boiling point ; for non-associated liquids A/a2 = 17.9 ; for associated liquids A/az is greater than 17.9.Conversely the heat of vaporisation can be calculated from the above relationship when the specific cohesion at the boiling point is known and a more complex formula is deduced by which the value of X at the boiling point can be calculated from the value of a2 at lower temperatures. When liquids with very divergent boiling points are compared the relationship beween a2 and X is represented more satisfactorily by means of the formula Alas = 46*0/1ogT, where 2’ is the boiling point of the liquid on the absolute scale. By combining the last formula with Van’t Hoff’s formula for the molecular elevation 3 of the boiling point E - OS02P/W the relationship between specific cohesion at the boiling point and boiling-point constant is obtained in the form E = 0~00~43T’,,210g2‘u/u2 a formula which permits of the calculation of E with considerable accuracy.It is further shown that the molecular cohesion Ma2 the molecular heat of vaporisation iKA and the boiling point on the absolute scale are additive in character. For compounds containing only hydrogen oxygen (bivalent) nitrogen (tervalent) and carbon the formula Ma2 = 11*5Sn where 8n represents the sum of the valencies holds within very wide limits. The formula is not valid for compounds containing the halogens on any of the usual assumptions as t o their valehcy nor does it hold for water ammonia or methyl alcohol.The further relations I ’ m = 11 -2273 and MA = 2402n are also shown to hold ih rdany cases. Finally from the formula A/a2 = constant the conclusion is drawn t h a t the molecules of all non-associated liquids are approximately of Latent ‘Heat of Evaporation of isoAmyl Ether. W. A. KURBATOFF (J. Rugs. Phys. Chem. SOC. 1908 40 1471-1474).-The mean specific heat OF isoamyl ether between 24O and 170’ is 0.5721 &0*5% the latent heat of evaporation p being 53.1 cals. with an accuracy of about 1% and Trouton’s constant K 18.8. T. H. P. the same diameter at their respective boiling points. a. s.120 ABSTRACTS OF CHEMICAL PAPERS Latent Heat of Elvaporation and Specific Heat of Naphtha. lene. W. A. KURBATOFF (J. RUSH. Phgs. Chm. Xoc. 1908 40 1474-1477).-Taking the specific heat of naphthalene as 0.32 at 20' and 0.7 at 215O the latent heat of evaporation is found to be 74.7 cals.and Trouton's constant 19.3. The latter figure confirms the observation that in an homologous series Trouton's constant diminishes as the boiling point exceeds 100'. Latent Heat of Vaporisation of Benzoin &oButyl Ether. W A. KURBATOPF (J. Rzcss. Phys. Chem. Xoc. 1908 40,1477-1480). -The latent heat of vaporisation of benzoin isobutyl ether is found to have the mean value 55.7 cals. the greatest deviation from this mean being 1.2%. From this value Trouton's constant is calculated to be 19.3. The only substance having a high boiling point and a high Trouton *constant (21.5) is mercury which is not comparable with other substances owing to the extreme simplicity of its molecular structure T.H. P. Latent Heat of Vaporisation of Acetic Anhydride and Determination of the Association of Liquids. W. A. KURBATOFF (J. Russ. Phys. Chem. Xoc. 1908 40 1480-1487).-The author shows that the value of the Ramsay-Shields constant for determining whether a liquid is associated is greatly diminished by the observations (1) that the magnitude of the constant changes with the molecular weight and that (2) change oE the constant with rise of temperature occurs even with non-associated compounds. The author's measurements of the specific heat of acetic anhydride lead to the meau value 0.4551 f 0.33% for the range of temperature 20-138O. The latent heat of vaporisation is found to be 93.6 + 0.5% at 1384 Trouton's constant calcuated from these results is 22.9.Acetic anhydride is hence associated and is converted into simple molecules a t the moment of boiling. Lithium Strontium and Barium Oxides. ROBERT DE FORCRAND (Ann. Chim. Phys. 1908 [viii] 15 433-490).-8 detailed account of work already published (compare Abstr. 1907 ii 615 683 928; 1908 ii 155 493 764). Using the corrected value for the heat of solution of strontium oxide (Abstr. 19OS ii 765) the heat of formation of strontium oxide is 138.64 Cal. and not 137.6 Cal. as stated (Abstr. 1908 ii 155). The heats of formation of the peroxides of calcium strontium and barium are 157.37 153.07 and 145.58 Gal. respectively and not 157.33 152.10 and 145.71 Cal. as stated (Zoc. cit.) WILLEM P. JORISSEN (Chern.Weekblad 1908 5 8'75-888).-A summary of the literature on heat of hydration and a criticism of the results obtained. A. J. W. T. H. P. T. H. P. M. A. W. Heat of Hydration. Heats of Dissociation of some Electrolytes in Organio Solvents. PAUL DUTOIT and H. DUPERTHUIS (J. Chim. Phys. 1908,8 699 -725).-'I!he electrical conductivity and hence the degree of ionisa- tion of sodium iodide in ethyl propyl isobutyl and isoamyl alcohols,GENERAL AND PHYSICAL CHEMISTRY. 121 and in acetone and pyridine has been determined at intervals of 10' between 0" and 80" (in the case of acetone up to 40") and similar measurements have been made with potassium iodide and sodium and potassium thiocyanates in pyridine. The dilutions employed were from 1000 to 20,000 litres per mol. of salt within which limits Ostwald'b dilution law applies in most cases so that the degree of dissociation can be accurately determined.From the results the heat of dissocia- tion has been calculated by Van't Hoff's formula. Contrary to Walden's statement (Abstr. 1907 ii 437) the heat of dissociation of a salt has not the same value in different solvents ; for sodium iodide as solute it is positive in ethyl alcohol zero in acetone and negative in the other alcohols and in pyridine. Further the heat of ionisation also varies greatly with the temperature in most cases ; in isoamyl alcohol it is - 2008 cal. at 15' and - 16,760 cal. at 75'. On the other hand the heats of ionisation in pyridine of the four salts above mentioned are nearly equal and do not vary much with temperature The fact that for what is presumably the same reaction NaI = Na' + 1' the thermal effect in different solvents is very different indicates that the process in question is not simple; it is probable that combination between ions and solvent is of importance. The differences between the authors' results and those of Walden are ascribed to the fact that the latter observer used erroneous values for the molecular conductivity at infinite dilution obtained by extrapolation from relatively con- centrated solutions G.S. The Density of Liquids Below Zero. JEAN TIMYERMANS (Bull. Xoc. chim. Belg. 1908 22 427-439).-It has been shown previously (Abstr. 1908 ii 85) that none of the liquids studied has a maximuui densityanalogoiis to that of water at 4O. The original experiments which were accurate to the extent of 1 in 1000 have now been repeated more exactly the error not exceeding 1 in 10,000.The present paper relates to the apparatus and experimental methods employed. The measurements were made by means of the dilatometer. A series of the latter were employed in order t o avoid capillary tubes of unwieldy length to cover the range of temperature considered and were calibrated with an accuracy of 1 in 20,000. The apparent internal volume of the dilatometers at 0" was not altered by two hours' immersion in liquid air. The errors due to the meniscus were corrected in the case of mercury as calculated by Schalkwyk and by Gackel and in the case of liquids which '' met " glass by assuming the surface of the liquid in the capillary to be hemispherical.This correction amounts to about 1/10,00Oth of the volume to be measured. With very volatile liquids a correction amounting to about 1/10,00Oth of the total weight had to be made for the weight of saturated vapour in the dilatometer above the surface of the liquid. The densities of the liquids at 0' compared with that of water at 4' were taken as standards and were determined by means of an Ostwald-Perkin pyknometer. The apparatus for maintain- ing the dilatometer at constant and uniform temperatures below 0" consists of a tube to contain a pure liquid (for cooling the dilatometer) solidifying at 8 known temperature below Oo surrounded by a Dewat YOL. XCVI. ii 9-1 22 ABSTRACTS OF OHEMICAL PAPIRS tube provided with a tubulure and stopcock which again is surrounded by another Dewar vessel the walls of thelatter being silvered except over a small space left transparent for reading the volumes in the dilatometer.The outer vessel must be maintained (by means of liquid air or carbon dioxide) at least 15-20° below the solidification point of the cooling liquid. Air is kept between the walls of the inner Dewar vessel until the cooling liquid (which must be kept continually stirred) solidifies (this should occur rapidly throughout the entire liquid) when it is pumped out until the internal pressure is 60-200 mm. according to the difference of temperature of the inner tube and the outer Dewar vessel. This prevents further cooling of the cooling-bath and enables the latter to be kept at a known constant and uniform temperature for some hours.Thus the volume of liquid in the dilatometer can be measured at fixed uniform temperatures below 0'. The temperatures are determined by a Callendar resistance thermometer in conjunction with a Callendar-Griffin bridge by means of which resistances can be measured t o 1/10,00Oth of an ohm. The exact evaluation of the temperature is made according to the calculations of Travers and Gwyer (Abstr. 1905 ii 372). The sole systematic errors of any importance in the work are (1) that in the calibration of the dilatometers equal to 1/20,00Oth of the density (2) that in reducing the scale of the platinum resistance thermometer to the scale of the gas thermometer equal to 1/10,00Oth of the density. The results obtained for ethyl acetate show that the density of this liquid varies according to the expresson Dt = 0.92540 - 0*0011863t Dt being the density at -to t the temperature interval below O' and 0.92540 the density at 0".The greatest difference between the observed densities and those calculated by this formula is only 0*00014. E. H. Expansion Coefficisnt Speciflc Gohesion Surface Tension and Molecular Weight of Solvents. PAUL WALDEN (Zeitsch. physihl. Chem. 1908 65 129-225).-The densities of a number of carefully purified liquids which have considerable solvent power have been determined a t Oo 25" and 50°. From these results and the available experimental data on the density specific cohesion and surface tension of a large number of liquids and the variation of these properties with the temperature certain relationships are deduced between these properties themselves and between the physical properties and molecular weights of the solvents in question.The general formula Vt = Vo/(l - I$) proposed by Mendelkeff to represent the variation of the density of liquids with temperature has been tested by application to the data for forty-nine different liquids including some liquefied gases. It is valid for liquids boiling below loo" but only mQderately so for liquids boiling a t bigher temperatures k varies for different liquids between 0*000178 and 0.00316. For non-associated substances the '' expansion modulus," k is inversely proportional to the critical temperature t on the centigrade scale and in many cases kt,=0*34. For certain liquefied gases and other liquids with low critical temperatures the value of kZ\=GENERAL AND PHYSICAL CHEMISTRY.123 constant = 0.49 the critical temperature being measured on the absolute scale. There is also parallelism but not proportionality between the expansion modulus and the compressibility of liquids. The data a6 to the surface tension y for over fifty liquids are taken from the results of Ramsay Guye and others ; those for the '' specific cohesion," a% are mostly calculated from the values of y by means of the equation uz= 2~/9*81cl where d is the density. In some cases the y values are calculated from the experimental values of a2 by means of the same formula. The variations% of the surface tension and the specific cohesion respectively with the temperature are repre- sented through a wide range of temperature by the equations yt = yo( 1 -@ and at2 = a,2(1- k,t) where p and k are constants for any one liquid but vary with the nature of the liquid.It is sbown that for non-associated liquids constant = 0.94. From these equations the critical temperature can be calculated when /3 or k are known and it is shown that the calculated and observed results are in good agreement. Other methods which have been suggested for calculating critical temperatures are discussed. Further the quotient of the latent heat of evaporation at the boiling point A by the specific cohesion at the bame temperature is constant and -17.9. Combining this with the mathematical expression for Trouton's rule JlAu/Tu=20*7 where M is the molecular weight of the liquid the equation MC~~~/Y' = 1.16 is obtained and is shown to be valid for non-associated liquids (compare Kistiakowsky Abstr.1906 ii 655). For associated liquids this expression is no longer constant but is smaller the more the liquid is associated. We have thus a new method for determining the molecular weight of liquids which appears to give more consistent results than any other method so far suggested. In the first instance the above equation only applies to observations at the boiling poiat but a rather more complicated equation is derived which admits of the calculation of the molecular weight from capillary observations at any temperature. When allowance is made for the fact that Trouton's rule is only valid over a limited range of temperature a more accurate formula M= 0-4351'JogTc/cc2 is obtained and is applied to the determination of the molecular weight of liquefied gases and other liquids.Of the ten liquefied gases considered only hydrogen chloride is considerably associated (association factor = 1-32). The association factor for cadmium is rather less than unity; it is therefore monatomic in the liquid state. When instead of the specific cohesion the surface tension y is used the expression $A/Yyu=3*64 is obtained where Y is the molecular volume and the other factors have the same significance as above. Combining this with Trouton's rule we obtain YyU/Tu = 5.67. These expressions are shown to be valid for a large number of non-associated liquids but for associated liquids the quotients are less than 3.64 and 5.67 respectively and from.the observed values the degree of association can be calculated. According to Eiitvcis-Ramsag-Shields the temperature-coefficient of = constant = 1 *16 and that k,t 9-2124 ABSTRACTS OF CHEMICAL PAPERS the molecular surface energy " should be constant for non-associated liquids that is d(yVj)/dt = constant = 2.12. The above considerations however lead to the conclusion that d(yV)/dt should be constant and it is shown by application to twenty-four typical liquids that the latter expression varies less than the former ; the mean value is 8.80. For associated liquids the value is less and from the above equation the degfee of association can be calculated as in the case of the Eotvos equation. The results obtained by the two methods are in satisfactory agreement.G. S. Liquefaction and Compressibility of Gaseous Mixtures. A Case in which Combination Occurs. E. BRINER and E. CARDOSO ( J . C'hi7n. Phya 1908 6 641-680).-The gases were liquefied by means of liquid air or solid carbon dioxide and ether and purified by fractional distillation. The compression was effected with a Cailletet pump the compression tube contained a stirrer worked by a magnet and the pressure was measured on a nitrogen manometer. The compressibilities and vapour tensions of certain binary mixtures and of the components separately have been determined at different temperatures. A mixture of sulphur dioxide and methyl ether in equivalent pro- portions liquefies with very little alteration of pressure and the vapour- pressure curve of the system shows a minimum a t the same composition from which the conclusion is drawn that a compound SO,,Me,O exists.The critical temperature of a mixture of this composition is 152-5" between those of the compouents and the critical pressure 62 atmo- spheres. From the variation of the pressure of the binary mixture with temperature it is calculated by means of Van't Hoff's equation that the heat given out by the reaction Me,O(gas)+SO,(gas)= Me,G,SO,(solid) is 11.8 cal. between 50° and 100' and 8.7 cal. between 13s' and 147'. No evidence of combination has been obtained in the following binary systems methyl ether and methane methyl ether and carbon dioxide and methyl ether and nitric oxide. G. S. Relation between Surface Tension and Speoiflc Volume of Non-associated Liquids. REGINALD 0.HERZOCI (Zeitsch. E'kktro- cherra. 1908 14 830-831 ).-By combining Avenarius' equation for the dependence of specific volume on temperature with Van der Waals' equation for surface tension and temperature the author obtains logy + pv = c where y and v are surface tension and specific volume a t the same temperature and p and r specific constants of the substance. The relation is shown to hold good for six liquids. The relation p = 5.28 times the critical density is also trua T. E. Supertension and Viscosity. CHARLEEI MARIE (Cornpt. rend. 1908 147 1400-1402).-The effect of variations in the viscosity of the electrolyte on the excess voltage required to liberate hydrogen at different electrodes dipping in N/l-sulphuric acid has been examined.When the supertension is small the effect of increased viscosity is slight thus 0.5% of gelatin raises the supertension a t platinisedGENERAL AND PHYSICAL CHEMISTRY 125 platinum electrodes at most by 0.001 volt but for metals with high supertension the effect is much greater thus 0.1% of gelatin raises the supertension a t lead electrodes by 0.05 volt. G. S. Viscosities and Limiting Conductivities. PAUL DUTOIT and H. DUPERTHUTS (J. Chim. Phys. 190S,6 736-731).-Walden (Abstr. 1906 ii 335) has shown that for tetraethylammonium iodide i n organic solvents the product of the viscosity 7 of the solvent and the molecular conductivity at infinite dilution pa is independent of the nature of the solvent and of the temperature i y .~ ~ = 0.7 approximately. The authors have determined the viscosities at intervals of loo between OOand 40' (in some cases up t o SOo) of the seven solvents already mentioned (see this vol. ii 120) and show that Walden's rule is not of general validity. The product qpm is not the same for the different solvents nor does it tend towards a fixed value as the temperature rises. A t 0' the extreme values are 0.374 (in isoamyl alcohol) and 0,573 (in pyridine) and at 80' the extreme values are 0.225 and 0.559 (in the same solvents). Further 7pw varies with the temperature; in four of the solvents it diminishes and in one it hcreases with increase of. temperature. The variations are consider- able in some solvents ; from SOo to Oo qpa increases by 47% in isoamyl alcohol and by 25% in isobutyl alcohol.G. S. Adsorption. LEONOR MICHAELIS and PETER RONA (Biochem. Zeitsch. 1908 15 196-21 6).-Determinations were made of the adsorption of acetone by charcoal and the results were found to agree with the theory of Gibbs and Freundlich. The adsorbabilities of acetone in the presence of acetic acid or acetic acid in the presence of acetone by charcoal were also determined and it was found that the simultaneous presence of two adsorbable substances diminished the adsorbability of either. The phenomenon of the diminution of adsorbability by the presence of a second adsorbable substance was not observed however in the case of proteins and certain dyes. I n the case of the former substances there is a further peculiarity in that the adsorption process is not reversible.This was thought t o be due to the conversion of the protein into an insoluble form such as takes place on coagulation but the same phenomenon takes place when albumoses are adsorbed and it was experimentally shown that in this case the presence of acetone does not diminish its adsorbability by charcoal. The conclusion is drawn that the adsorption of proteins is not a simple mechanical process but depends probably on the electrical properties of the surfaces. s. B. s. Coefficient of Diffusion. BASIL W. CLACK (Phil. Mag. 1908 [vi] 16 863-879).-The author has endeavoured to elaborate an accurate method of measuring the diffusion coefficients of dissolved salts and to determine to what extent diffusion depends on concen- tration.The solutions under investigation were enclosed in a glass bulb into which a diffusion tube 4 cm. long was sealed in such a way that the end projected 2 CIP. ioto the bulb The bulb was suspended is126 ABSTRACTS OF CHEMICAL PAPERS. distilled water beneath the pan of a delicate balance SO that the weight of its contents could be neasured automatically from time t o time. As salt diffused out of the bulb tha solutions contracted and water entered to make up for this contraction at nearly constant rates so long as the concentration was not appreciably diminished. A steady distribution of concentration being established in the diffusion tube an equation could be deduced connecting the rate of loss of weight of the bulb with the coefficient of diffusion. The results obtained are held to show that the method is capable oE yielding accurate values when certain conditions of experiment are observed.The following diffusion coefficients a t 0’ are given KNO in 10% solution 0,844 in 5% solution 0.870; KCl in 20% solution 0.972 in 10% solution 0.954. R. J. C. Osmosis of the Alkali Haloids. FRANZ WIEMERS (Ann. Physik 19OS [iv] 27 1081-109S).-Measurements have been made of the relative changes in concentration which take place in solutions of the halogen salts of the alkali metals when the solutions are separated from water by a membrane of pig’s bladder. The data are used to calculate the so-called endosmotic equivalent the quantity of water which in a given time replaces unit quantity of the dissolved substance. The endosmotic equivalent decreases as the molecular weight of the dissolved haloid salt increases.It is independent of the temperature although the absolute magnitude of the two diffusion currents increases with rise of temperature in consequence of the greater velocity of the molecules. With increasing concentration of the solution the diffusion currents incream in magnitude but since the flow of water increases more rapidly than that of the salt solution the endosmotic equivalent is greater the more concentrated the solution. H. M. D. Osmotic Pressures of Aqueous Solutions of Calcium Ferro- cyanide. I. Concentrated Solutions. EARL OF BERKELEY ERNALD G. J. HARTLEY and C. V. BURTON (Phil. Frans. 19OS A 209 177-203).-Porter’s equation taking account of the effect of compressibility of solvent and solution on the relation of osmotic to vapour pressures (Abstr.1907 ii 743; 1908 ii 670) can be modified and applied to determinations of vapour pressure made in air a t atmospheric pressure instead of under vacuum conditions. Whilst Porter’s equation yields results differing from the authors’ experiments by 2 or 30/ the modified equation is in close agree- ment with experiment. The authors have endeavoured to increase the accuracy of their vapour-pressure determinations i n order to render the correction for compressibility significant. Determinations were made a t 0’ of osmotic pressure vapour pressure and compressibility of concentrated solutions of calcium ferrocyanide. The purest salt recrystallised to eliminate potassium consisted of yellow crystals containing 11 H,O. I n previoiis measurements of vapour pressure by the bubbling method an unexplained loss of weight occurred in the whole apparatusGENERAL AND PHYSICAL CHEMISTRY.127 as well as in the particular bulb containing the solution. The complete elimination of rubber bubing from the apparatus and the substitution of mercury seals although an advantage was not entirely satisfactory because a ponderable amount of mercury now volatilised. Finally plat inurn-irid ium connecting tubes with ground-in junctions. were used throughout and no appreciable source of error was found in the method so modified. Determinations of osmotic pressure were carried out as previously described (Abstr. 1906 ii 599). The amount of calcium ferrocyanide leaking through the copper ferrocyanide membrane was found to be practically nil A special form of piezometer was employed for compressibility determinations.The compressibility tube was made of Jena glass the compression coefficient of which has been determined by Straubel to be 228 x 10-8. A check experiment made by the authors gave the value 223 x loTs. Possible sources of error particularly those due to temperature were carefully guarded against. The concordance of observed osmotic equilibrium pressures and those calculated from vapoiir pressures and compressibilities by the modified Porter’s equation is held to indicate a satisfactory agreement between experiment and thermodynamic reasoning R. J. C. Contact Pyrogenetio Oxidation of Eydrogen and Carbon Monoxide by Air E.I. ORLOFF (J. Buss. Phys. Cham. Soc. 1908 40 1590-1595),-The author has studied the oxidation of excess of hydrogen and of carbon monoxide by air in presence of copper or platinum gauze under such conditions that the contact material is heated by the oxidation process. The results are given in detail and do not admit of condensation (compare following abstract). T. IT. P. General Conditions of Contact Oxidations accompanied by Auto-heating of the Contact Layer. E. I. OBLOFF (J. Russ. P~TJB. Chem. Soc. 1908 40 1596-1603. Compare preceding abstract),- These contact oxidations are divided into two classes (I) reversible reactions such as 2H2+02 t 2H,O 2SO,+O2 ZZ 2SO and 4HCl+ 0 dH20 + 2C1 ; (11) reactions which are not reversible under the experimental conditions such as 2CO + 0 = 2C0 and 4NH + 30 = 2N2 + 6H,O.Class I.-In this case the velocity of flow of the mixed gaseg is without influence on the magnitude of P.C,/C,!a which has the con- stant value 0.55 C and C representing the initial concentrations of the hydrogen and oxygen respectively and C the concentration of the water vapour. In contact with the heated layer of copper or platinum gauze the gases occupy a volume Vm greater than the volume V they occupy after leaving this layer and cooling. If T is the ratio of the initial speed to the mean speed of the gases and 1p and Kl the concentrations of the hydrogen and oxygen respectively in the final gases (C2.Cl/r2)v2 = (K2E,)”. Inserting the values of C2C,/r2 and P . K l given by the four experiments previously described (see preceding abstract) the quotient Vm/V is found to be 6 6-87 6.87,128 ABSTRACTS OF CHEMICAL PAPERS.qnd 6.8. If to represents the fall of the temperature of the gases on leaving the heated contact layer Vm/V= 273/(273 - t ) from which to = 227O 233O 226O and 230'; using a contact layer 8-12 cm. in length the actual difference of temperature between the heated and unheated parts was found to be approximately 230'. Class II.-In this case not only the concentrations but also the speed with which themixed gases meet the contact layer is of import- ance and unless the velocity of flow reaches a certain value in each particular case no heating of the contact layer takes place. For the oxidation of carbon monoxide and methyl and pt-opyl alcohols the con- stant C2C,/P2 has the values 0 *022-0-031 0.00103 and 0-000103 respectively C and C being the initial volume concentrations of the carbon monoxide etc and of the oxygen respectively and Y the mean velority of the gaseous mixture.From the equation (C2Cl/r2)" = (XZKJy it is calciilated by means of the experimental numbers obtained (see preceding abstract) that X / Y = 2 * 0 4 3.21 and 3.1. It is found experimentally that the ratio of tbe absolute temperature of the heated portion of the contact layer to that of the cool portion T1/Tz has the approximate value 2 from which it is deduced that Ti/T2= X/Y. Replacing X and Y in the equation given above by T1 and Y\ respectively an expressiou is obtained showing that increase of r decrease of V and increase of Tl/T2 all cause increase of the reaction between carbon monoxide and oxygen. With methyl and propyl alcohols the primary oxidation is accompanied by decomposition of the alcohol and oxidation of the decomposition products whilst with ammonia part of the nitrogen originally formed is oxidised to oxides of nitrogen.T. H. P. Polyiodides. P. OLIVARI ( A t t i R. Accad. Lincei 1908 [v] 17 ii 584-590).-The author has studied the influence of the solvent on the relation z/y = 8 between the concentration x in gram-mols. of iodine dissolved and the concentration of an insoluble iodide (KI) passing into solution to form polyiodides (K12n+l). The results show that all the alcohols acids esters ketones cyauo- and nitro-derivatives on which iodine has no chemical action allow of the formation of polyiodides.The solubility of potassium iodide in presence of iodine increases approximately with the dielectric constant of the solvent. With solvents having analogous chemical functions the ratio 0 and also x increase and consequently the absorption of potassium iodide diminishes with increase of the molecular weight of the solvent that is with increase of its anomalising power and with decrease of its dielectric constant. Comparison of the results obtained with solvents having different chemical functions shows that the absorption of potassium iodide depends on the anomalising property as well as on the dissociating power of the solvent. The solvents which increase the ratio 8 do so in the following order acetic acid (9*7) ethyl acetate (6.1 l) ethyl alcohol (25-8) nitrobenzene (36*4) ethyl bromide (9) hydrocarbons (Z) the numbers in brackets representing the respective dielectric constants.As regards anomalising power the series CO,H CO,R OH NO Br 1GENERAL AKD PHYSICAL CHEMISTRY. 129 Me is in very good agreement with the series given by Auwers (Abstr. 1903 ii 268) considering that these groups are not uuited t o the same radicle in the solvents investigated. With different iodides and the same solvent it is found t h a t when the solutions of iodine are dilute almost identical values of 6 are given by the iodides of any one group of the periodic system With concentrated solutions however deviations occur in a definite sense which seems to indicate that they bear a relation to the electropositive character of the metal.Diminution of the latter character is accom- panied by a diminution in the absorption of the iodide by a nitro- benzene. solution of iodine. T. H. P. Kinetics of Alkylation. HEINRICH GOLDSCHMIDT (Zeitsch EZek trochem. 1909 15 10-11).-The experiments of Acree and his upils (Abstr. 1908 ii 919) do not prove the existence of tautomeric r orms in solution as they are equally well explained by the supposition that the alkyl halide is added on t o the anion of the salt in two different ways the two complex ions yielding the 0- and N-ethers tespectively . T. E. Researches on the Formation of Eaters. HEINBICH GOLD- SCHMIDT [with M. ASRIEL v. KOREN LUND and OLAF UDBY] (Zeitsch. Elektrochent. 1909 15 4-lO).-Kailan (Abstr. 1908 ii 936) finds that the rate of esterification of trichloroacetic acid without a catalyst is not in agreement with the theory of hydrogen ion catalysis.The author therefore gives a rAsum6 of the unpublished results which he has obtained during the past few years. The rate of esterification of tri- and di-chloroacetic acids is increased slightly by the addition of a salt of the acid. that of formic acid is much diminished. With di- and tri-chloroacetic acids the reaction is of the second order ; with formic acid the second order constants rapidly diminish but are constant in presence of a formate ; the aromatic sulphinic acids give a reaction lying between the first and second orders. Probably in these cases the hydrogen ion catalysis is of secondary importance some other reaction accounting for the main part of the ester formation.The esterification of formic acid using picric trichloroacetic tri- chlorobutyric and dichloroacetic acids as catalysts was studied in alcohol dried with calcium. The addition of the piperidine salt of the acid diminishes the rate of change the effect being larger with the weaker acids. The catalytic action of hydrogen chloride is not affected by the addition of a chloride. The conductivity of the alcoholic solutions of the acids was also measured and the catalytic action is found to be almost proportional to the conductivity of the solutions. The rates of esterification of formic acid by picric acid in different concen- trations are also nearly proportional to the conductivities of the solutions. With weaker acids the agreement was not so good possibly the non-dissociated complex of catalytic acid and alcohol reacts slowly with the acid to form the ester the effect being superposed on that of the hydrogen ion.The addition of water to a solution of hydrogen chloride in absolute alcohol diminishes its130 ABSTRACTS OF CHEMICAL PAPERS. conductivity to a minimum; for all other acids the conductivity increases steadily. It follows from the theory of hydrogen ion catalysis that the ratio of the catalytic accelerations produced by two acids must be independent of the nature of the acid undergoing esterification. The ratio of the rates of esterification of acetic acid in presence of hydrochloric and picric acids is 13.5 ; for propionic acid it is 13.1 ; for m-butyric acid,;l3*l ; for phenylacetic acid 13.8 ; and for phenylpropionic acid 13.0.The ratio of the accelerations of hydro- chloric and sulphosalicylic acids is also found to be the same with five acids. The formation of isobutyl and methyl esters in presence of hydro- chloric acid is studied. The equation obtained by the aiithor and Udby (Abstr. 1907 ii 852) applies in these cases as well as in that of the ethyl esters. The rates of esterification of acetic and phenylacetic acids are determined at different temperaturep and the temperature- coefficients found to be the same All these results are in good agreement with the theory of hydrogen ion oatalysis developed by the author (Abstr. 1907 ii 853). The self-esterification of an acid however appears to be mainly due t o some other reaction.T E. History of the Discovery of the Rules of the so-called (( Steric Hindrance.” FRIEDRICH KEHRMANN (Ber. 1908 41 4357-4358).-8 claim for priority in the statement of the phenomanon of steric hindrance. The author foreshadowed this in 1888 (Abstr. 1889 243) and called attention (inter alz’a) to the difficulty of esterifging o-substituted carboxylic acids in 1890 (Abstr. 1890 484) whilst V. Meyer’s first paper on this subject was published in 1894 (Abstr. 1894 i 243). J. C. 0 Physico-chemical Investigation of the Reactions between Egg-albumin and Acetic Acid. LUDWIK ZOJA (Zeitsoh. Chem. Ind. Kolloide 1908 3 249-269).-On the addition oE acetic acid to dialysed egg-albumin the viscosity of the resulting solution increases with time and under certain circumstances the end result is a solid jelly The process of gelatinisation has been studied by measurements of viscosity electrical conductivity and density.The variaCion of the viscosity with time depends very much on the concentration of the acetic acid. Below 3.22 mols. per litre the viscosity remains constant ; with increasing acetic acid concentration the rate of alteration of the viscosity with time increases attains a maximum at 7.01 mols. per litre and then decreases as the acid concentration is further increased. On the other hand the electrical conductivity does not change with time and the curves which represent the dependence of the con- ductivity on the acetic acid concentration are closely similar to the corresponding curves for acetic acid-water mixtures.Both curves exhibit a maximum a t about 3.2 mols. acid per litre. The change of viscosity with time is supposed to be due to an action of hydrogen ions on the particles of the colloid. I n agreement with this it is Found that the rate of gelatinisation is greater when the acetic acid is replaced by formic acid aud less when replaced by propionic acid.GENERAL AND PHYSlCAL CXEMISTRY. 131 Observations relating to the influence of calcium acetate indicate that the reaction is catalytically accelerated by calcium ions. The increase in viscosity takes place more rapidly at 25’ than at higher or lower temperatures when the other conditions are maintained c o ~ s tant. H. H. D. Hydration in Solution as the cause of certain Solubility Influences. C.S. HUDSON (Zeitsch. EZektrocham. 19OS 14 831-823 ; J. Amr. Chem. Soc. 1909 31 63-66).-The solubility of anhydrous sodium sulphate in water is diminished by the addition of carbamide whilst that of Glauber’s salt is increased. The author supposes that the solution contains molecules of the anhydrous salt and molecules of the hydrate ; in presence of the solid anhydride the concentration of the former is a that of the latter is h. The total solubility of the anhydrous salt is therefore s=cl+h. Tf p is the vapour pressure of the solution h/a=const. pn. The addition of a foreign substance to the solution diminishes p and therefore diminishes h also since a (in presence of the solid anhydrous salt) is constant. The solubility of the anhydrous salt is therefore diminished by the addition of a foreign substance to the solution.Precisely the same argument applied t o the hydrated salt shows that its solubility must be increased. T. E. Theory of Colloidal Solutions. LUIGT ROLLA (Atti R. Accad. Lincei 1908 [vJ 17 ii 650-654).-The author has carried out a series of experiments on metallic colloids with the object of determin- ing the charge of the granules and the relation between charge and mass making use of Stokes’s formula. The colloids examined were Eredig’s platinum and various preparations made by Zsigmondy ’s method. Using a P.D. of 13 volts between electrodes i cm. apart the time required to traverse uniformly a path of 0.125 mm. was found to be four seconds for the gold granules and 4.3 seconds for those of platinum.The mobility or velocity for a field of 1 volt per cm. is hence 26.0 x 10-5 crns. per sec. for gold and 24.0 x loe5 crns. per sec. for platinum. Burton’s measurements (Abstir 1906 ii 275) gave the values 21.6 x crns. per sec. for gold and platinum respectively. Electric transport measurements made on red violet and blue colloidal gold solutions prepared according to Zsigmondy’s method (Abstr. 1898 ii 522) show that the method of preparation does not influence the nature of the colloidal metallic solution obtained the charge of the granules remaining practically constant. The radii of the gold granules measured by means of a hmnato- meter were found to be red gold 192 x low5 crns. ; violet gold 1.7 x crns. ; blue gold 2.0 x cms. If the granules carrying a charge e move in an electric field of intensity + they are subjected to an electric force e+ and their velocity is given by k+ where k is their mobility.Owing to the very great dilution (0.00006 gram of gold per c.c.) the internal friction 7 will be sensibly thb same as if the granules moved in pure water. According to Stokes’s law the velocity of spheres of radius r iq 8 and 20.3 x132 ABSTRACTS OF CHEMICAL PAPERS. medium having the internal friction 7 under the influence of a force f=e+ is given by the expression v = e+/6nr)r and since v=k+ e = 6 ~ 7 & r . The experimental determinations give the value 7 = 0*01782/(1 + at). The charge of Bredig's gold granules is hence e= 12.19 x electrostatic units the values for the other colloids being Zsigmondy's red gold 9.75 x ; violet 12.19 x lo-'"; blue 16.2 x lo-''); Bredig's colloidal platinum 11.32 x 10-10. Tbe charges on the granules are hence of the same order of magnitude as the charges carried by the ions.T. H. P. Behaviour of Colloidal Suspensions with Immiscible Solvents. W. LASH MILLER and R. H. MCPHERSON (J. Physical Chem. 1908 12 706-71 6).-The experiments were undertaken to ascertain if colloidal suspensions distribute themselves between t w o immiscible solvents on shaking. No such distribution was observed in the case of two liquids ; the colloid usually coagulated on shaking (compare Winkelblech Abstr. 1907 ii 17). More definite results were obtained with ternary systems which form two layers in the neighbourhood of the " plait point )' (the point at which the layers become almost identical in composition).The binodal curve tie-lines and plait point have been determined directly at room temperature for a number of ternary systems such as water-chloroform-alcohol water-chloroform-acetone and water- alcohol-ether by a simple method. No distribution was observed with colloidal arsenious sulphide in the first two systems but in the latter systgm and in certain other systems distribution was observed in the neighbourhood of the plait point. It is however only for relatively clear solutions which approach true solutions in character that distribution has so far been obtained. G. S. Has the Crystallising Force any Influence on Properties in the Amorphous Condition. Latent Heat of Vaporisation of p-Toluidine. W. A . KURBATOFF (J.Buss. Phys. Chem. SOC. 1908 40 1487-1492).-The latent heat of vaporisation of o-toluidine m. p. - 40° b. p. 198" is 95.1 and Trouton's constant 21-6 (compare Luginin Abstr. 1903 ii 7). The author's measurements of the latent heat of vaporisation of p-toluidine m. p. 45O b. p 199*5O give the value 9 1-8 Trouton's constant being consequently 20.6 the normal value. p-Toluidine is hence non-associated and o-toluidine associated. The author regards this difference as due to the difference in the crystallising powers of the two compounds. The crystallising power hence seems to be a molecular force and to influence the properties of the compound in the amorphous condition. The values of the Ramsay-Shields constant are different for the two toluidines being 1.72 for the p-compound between 107' and 210*6O and for the o-compound 2.04 a t 8.1-108.4° and 2-06 at 108.4-208*5°.T. H. P. Forms of Matter. 11. P. P. VON WETMARN (Zeitsch. Cham. Ind. I1TbZZoide 1908 3 283-304).-An account is given of methods which have been used in the preparation of crystalliue substances in so-calledGENERAL AND PEYSICAL CHEMISTRY. 1.38 colloidal amorphous forms. According to the author’s theory (crystallisation theory of colloids) the amorphous condition has no real existence ; “ amorphous ’’ substances represent sponge-like forms- tions consisting of extremely minute crystals In a large number of cases it has been shown that crystalline substances can be obtained i n “ colloidal ” form and conversely “colloidal” substances i n crystalline form by suitable variation of the conditions under which separation from solution takes place.The character of the precipitate which separates from a solution as the result of a chemical change (for example the double decomposition of two salts) depends on the rate of formation and the solubility of the solid phase the degree of association of the components of the solution which react to form the insoluble substance and on the viscosity of the !iquid medium. The fact that the rate of formation of the solid and the association of the components of the solution both increase when the concentration of the solution increases enables the process of crystallisation to be regulated by varying the concentration of the solution. Starting out from solutions just saturated with respect to the slightly soluble substance and progressively increasing the con- centration of the reacting solutions it is possible to distinguish five concentration intervals according to the crystalline character of the the substance which separates out.The character of the barium sulphate precipitate which is obtained on mixing differently concentrated solutions of barium thiocyanate and manganous sulphate is described and illustrated by photographs. The following t intervals of concentration are distinguished. Between X/ZO,OOO and N/7000 no precipitation takes place even when the solution is left for years. Between N/7000 and N/600 well-developed crystals separate out whilst between N/600 and 0.75N the precipitate consists of star-shaped and needle-like aggregates.From 0*75N to 3N ‘‘ amorphous ” precipitates are obtained distinguishable under the microscope as spherical granules whilst between 3N and 7rV ‘* amorphous ” precipitate separate which cannot be differentiated either by the aid of the microscope or the ultra-microscope. It is shown that the ‘‘ amorphous ” forms obtained in the fourth and fifth concentration intervals are actually crystalline. The minuteness of the crystals which result from these concentrated solutions is probably due for the most part to the very high velocity of formation. The conditions for the preparation of a stable suspension of barium sulphate are considered and it is shown that such a stable suspension can be obtained by mixing alcoholic solutions (N/50O-N/lO) of barium thiocyanate and cobalt sulphate.This is due to the increase in the number of centres of crystallisation consequent upon the diminished solubility of the barium sulphate. H. M. D [Forms of Matter.] MAX M~NDEN (Zeitsch. Chem. Ind. Kolloide 1908 3 304-505).-A claim of priority in respect of P. P. von Weimarn’s paper (preceding abstract). The author points out that he has previously demonstrated by microscopic observations that all forms of matter organic and inorganic are fundamentally identical H. M. D.134 ABSTRACTS OF CHEMICAL PAPERS. [Form of Matter.] 9. P. VON WEIMARN (Zeitsch. Chsm. $ad. Xolloids 1908 3 305).-In reply to Munden (preceding abstract) the author states that the question of the identity of living and dead matter has not been considered by him.Conservation of Weight. CONSTANTIN ZENGELIS (Zeitsch. physikal. Chern. 1909 65 341-358. Compare Abstr. 1906 ii 831). -It is shown by a number of different methods that many gases and vapours of solids can pass through glass even at the ordinary temperature. I n one set of experiments substances such as stannic sulphide and iodic acid mere supported inside a glass beaker inverted on a glass plate the junction between plate and beaker being closed by para&n. On the outside of the beaker at the same height as the material inside tt piece of silver foil was fixed and the whole arrangement was surrounded by a large inverted beaker also closed air-tight with paraffin. After some days the silver foil showed evidence of the action of the solid which must therefore have passed in the form of vapour through the glass.I n another set of experiments chlorine bromine and iodine were sealed up in separate flasks on the outside of which were fixed strips of silver foil and the flasks were then enclosed in separate air-tight vessels. The silver outside the chlorine vessel showed signs of attack after fifty days that of the bromine vessel in about the same time but in the case of iodine there was distinct action in three days. The effect was much more rapid when the outer vessel was exhausted. The rate of action is not proportional to the volatility of the substance and in some cases at least is not markedly accelerated by a moderate rise of temperature. The effect is the more rapid the thinner the glass walls. The author considers that Landolt's experiments contrary to the latest views of Landolt himself (Abstr. 1908 ii 366) afford evidence of a real loss of weight in certain chemical reactions and that the loss of weight is due to the passage of vapours through the glass walls. The fact that there is no loss of weight when the inner surface of the vessel is coated with paraffin (Landolt) is in favour of this view. H. &I. D. G. S. Are the Stoicheiometric Laws Intelligible without the Atomic Hypothesis? FHANZ WALD (Chem. Zeit. 1908 32 1249-1250 1276-1279. Compare Abstr. 1908 ii 367).-Polemical. A furthsr reply to Kuhn (Abstr. 1908 ii S26). Combining Power of Metallic Atoms with Atoms of the Same Kind. LEOPOLD RUGHEIMER (Anlzalen 1908 364 5 1-631.- I n a previous paper (compare Abstr. 1905 ii 576) the author has pointed out that the tendency of atoms of metals to link together is very small even in non-ionised compounds ; the only definite exception is tin triethyl which according to Ladenburg has the formula Sn,Et (from vapour density determinations). The author has re- determined the molecular weight of this compound by the boiling-point method ethyl ether being used as solvent. I n the moat dilute solution W. H. G.INORGANIC CHEMISTRY 135 the value found is not much greater than that corresponding with the formula SnEt but it increases rapidly with the concentration so that the substance shows a considerable tendency t o associate. The remainder of the paper is occupied with a discussion on the nature of valency. G. 8. Physical and Chemical Properties as Functions of the Magnitude of the Qranules (Number of Molecules in the granule) of the Solid and Liquid Phases. P. P. VON WEIMARN (J. Izuss. Phys. Chem. ~ o c . 1908 40 1434-1447).-The author gives a rdaumd of the literature on this subject and bhows that some of Pawloff's conclusions (Abstr. 1908 ii 92'7) are in entire disagreement with the results of previous investigators (compare also Abstr. 1908 ii 90 263 1023). Method to Avoid Cracking Vacuum Vessels whilst Manipulating Liquid Air. ENRIQUE HAUSER (Anal. Pis. Quim. 1908 6 428-429).-8 tube of paper rolled several times on itself is introduced into the mouth of the vacuum vessel so as to fit it closely and to pass halfway to the bottom of the vessel and project an equal distance outside. The liquid air can then be poured out through this tube without undergoing violent ebullition in the neck and without risk of cracking the vacuum vessel. T. H. P. W. A. D.