ii. 40 General and Physical Chemistry. Refractivity of Unsaturated Compounds. I. GERVAISE LE BAS (Z’mns. Pnmday SOC. 1917 13 53-60).-A discussion of the influence on refractive power of the ethylene and acetylene linkings of the carbonyl group and of conjugated carbonyl groups and of the refractivities of nitrites nitrates and oximino-com- pounds. H. M. D. The Lke Spectrum of Titanium and its Applications. A. DE GRAMONT (Compt. rend. 1918 166 94-99).-A study of the line spectrum of titanium when fused in the form of its oxide with lithium sodium or potassium carbonate and submitted to the action of a condensed spark or with the direct spark between two fragments of steel cont,aining titanium. The spectrophoto- graphs are reproduced in the original together with tables show- ing the rays more sensitive to the eye than t o a photographic plate and those capable of being photographed.The author suggests that the approximate amount of titanium in a steel may be deter- mined by observing which of the titanium rays can be detected. I n a similar manner this method may be used for the examination of minerals. W. G. T h e Application of the Quantum Hypothesis to Photo- chemistry. E. WARBURG (Xatwwiss. 1917 5 489-494 ; from Chem. Zentr. 1917 ii 587-588).-Restricting the term photo- chemistry to such reactions as do not yield electric end-products it is probable that the photochemical process does not involve the separat’ion of electrons because gases undergoing a photochemical change do not exhibit conductivity. It is necessary to distinguish between chemical processes produced directly by the influence of light and secondary reactions.Only those rays which are absorbed exert photochemical action and the effect is proportional to the absorbed radiation ; the chemical change caused by an absorption of radiation equivalent t o one gram-calorie is termed the “specific photochemical effect.” The hypothesis that photochemical absorp- tion occurs by quanta explains the concentration of the effect on relatively few molecules and accounts for the greater activity of the shorter wavelengths the effect being due to an increase in temperature of individual molecules. From the quantum hypo- thesis it follows that a molecule for example of hydrogen bromide can be directly decomposed by radiation only if 2c/h > q where c is a constant h the wavelength and q the heat effect in gram- calories per molecular weight on recombination of the decomposi- tion products.The apparent contradiction that photolysis can occur with ammonia for wave-lengths of h=0’209 in which case 2 c / h < g is explained by the possibility that a molecule of ammonia after absorption of a quantum without chemical alkra- CXIV. ii. 5ii. 50 ABSTRACTS OF CHEMICAL PAPERS. Lion may subsequently meet a sccoiid molecule with tlie result that the change 2NH3= N + 3H2 occurs for which q possesses a smaller value than for the decomposition of a single molecule into its constituent atoms. Processes in which the production of a false equilibrium is accompanied by a decrease and an increase of the free energy of the system are described respectively as photochemical actions of the first and second class.The photochemical yield is the proportion of the absorbed radiation which undergoes con- version into chemical energy and in the case of the decomposition of ammonia by the wave-length 0.209 amounts to a% a notable parallelism appearing to exist between the yields and those of the silent discharge. If the quantum is greater than the heat change in the primary photochemical process the excess will be converted into heat and heat will aho be produced in the secondary pro- cesses. If in a photolytic process the highest possible photo- chemical yield is desired and a minimum heat effect it is necessary that the quantum should be greater than the heat change involved in the primary process but only greater to the extent required for the fission of the photolyte and the heat effect in the secondary processes should be as small as possible. True Photochemical Processes.FRITZ WEIGERT (Zeitsch. Elektrochem. 1917 23 357-368).-A theoretical paper in which after differentiating between ideal and real photochemical pro- cesses and surveying a number of real photochemical reactions the author propounds a theory of the mechanism of these processes. It is shown that a transformation of light energy into other forms of energy only occurs when the electrons rotating round the posi- tive nucleus travel in a distorted path and when the distortion is produced by neighbouring particles which are in irregular relative motion. There must be therefore an optical coupling in the sense of Stark’s intermolecular influence which is closely connected with the broadening of spectrum lines.According t o the Le Chatelier-van’t Hoff principle it is to be expected that a change in the system will be brought about by the absorption of light in the sense that the distortion will be as far as possible removed. This would occur most easily by the separation of the distorting particles from one another. The foregoing is in accord with Bohr’s work (A. 1913 ii 689 943 1045) and is sufficient to include all known real photochemical phenomena and the transformation of light into heat. The (‘ equivalent law ” of Einstein holds for ideal photochemical reactions and in its deduction simple processes without the mutual influences of neighbouring particles alone were considered.This law may be extended by the assumption that in real photochemical processes not a single molecule but’ the whole of the optically coupled particles take part in the absorp- tion of an energy quantum and that therefore changes must take place over the whole of this region. This region contains fewer particles the higher the frequency of the absorbed light and the more dilute the partition of the mass; the lower the temperature the more nearly the relationships approach those of an ideal D. F. T.GENERAL AND PHY SEAL CHBMk3T.R-Y. ii. 5i process. The iuom particles displaced by the absorpt'ion of itu energy quantum the Iarger is the fraction of the thermal energy lost. Hence in solid and stabilised systems changes brought about by the absorption of energy persist for long periods and are the cause of characteristic changes and after-eff ects.Changes brought about in other bands by the distortion indicate the cause of a number of characteristic colour changes effected by light. This theory differs from previous theories inasmuch as it does not presuppose either a primary electron emission or an electron loosening. The consideration of the fact that a relative displace- ment of small masses brings about a distortion of the electron orbit leads a t once to the meaning of luminescence phenomena. Shoiild the displacement be the resnlt of light absorption theii Iluorescence follows. dn coiinexion with the foregoing the author briefly considers the photoelectric effect and the action of Rontgen rays.It is shown in the first place that a connexion between photochemical fluorescence and photoelectric effects does not! of necessity exist but they can in individual cases appear simul- taneously. The only direct process iiecessarily connected with the absorption of light is the relative displacement of the individual optically coupled particles and therewith a change in the absorp- tion under consideration which shows itself most clearly by its dependence on the intensity of the incident rays. J. F. S. The Disintegration Constant of Radiothorium. B. WALTER (Physikal. Zeitsch. 1917 18 584-585).-Measurements on the decay of the radiation through 5 mm. of lead of two preparations of radiothorium extending over 500 days have given a mean value for the half-period of 1.876 years or 685 days or 989 days for the period of average life.This is less than the accepted half-period 2 years and is in agreement with the recent statement of Meyer and Paneth ( W i e n . Ber. 1916 Abt. I I . n 125 1253) that the value could not be greater than 1.9 years as a maximum. F. S. l 1 Spark-lengths " in Hydrocarbon Gases and Vapours. ROBERT WRIGHT (T. 1918 113 79-80. Compare A. 1917 ii 403) .-Comparative measurements of the spark-lengths have been made in hydrogen benzene toluene and paraffin hydro- carbons. I n the paraffin series the spark-length decreases with increase in the molecular weight4 of the hydrocarbon. The insu- lating power of isopentane is exceptionally high being greater than that of n-pentane and greater also than that of hexane.The results for benzene and toluene show that the insulating power of these is approximately the same as that of hexane. H. M. D. The Charge and Dimensions of Ions and Dispersoids. G. VON HEVESY (Ir'olloid Zeitsch. 1917 21 129-136. Compare A. 1916 ii 594).-Evidence is put forward in support of the view that there is a tendency for electrically charged particles to 5-2ii. 52 ABSTRACTS OF CHEMICAL PAPERS. c*oiubiiic with w:tt er iiioleciiles until the 1)otential of the particle is reduced to about 70 millivolts. Since the potential of a charged ion is given by r' -t:,'KR where 6 is the charge R the radius ol' the particle aut2 I< t hr rlielertzic. constant of the solvent medium ii; follows that It =2.8 x ' l 0 - s ctti.for i L iioimal univalent ioii. For a multivalent~ ioii the radius will be proportional to the charge. Since the rate of diffusion of an ion depends mainly on the radius the diffusion constants may be expected to depend on the valency. The available data for uni- bi- ter- and quadri-valent ions give average values for the reciprocals of the diffusion constants which are in the ratio 1 :1*99:3.03:3*88. This relation is cou- sidered to afford strong support for the theory of constant ionic potential. The combination which takes place when multivalent electrolytes are dissolved in water is supposed to be directly connected with the formation of aqueous envelopes in accordance with the above tendency. This combination is relatively small in the case of uni- valent electrolytes.The ionic mobility of a normal univalent ion for which e==4.7 x 10-10 R=2*8 x cm. and P=0*07 volt is 48 when expressed in terms of the ordinary units. I n the case of ions of large size such as complex organic ions there is no tendency to combine with water in that the potential of the anhydrous ion is already less than that which tends to be set up by the interaction. I n a certain sense these slowly moving ions are to be regarded as abnormal. The same tendency is supposed t o operate in the case of colloidal particles. On the assumption that these particles are characterisecl by an electrical double layer a t the surface of contact with the dispersive medium the potential of the particles may be calculated from the equation V=E(R - R1)/KRlR2? in which R is the radius of the colloidal particle and R,- R the thickness of the electrical double layer.This thickness has been found to be about 5 x 10-7 cm. Since the mobility of the colloidal particles is of the same order as the normal ionic mobility it is possible to utilise the above relation t o obtain the contiexion between the charge on a colloidal particle and its radius. It is thus found that the charge is approximately proportional to the radius in the case of very small particles but that the charge increases much more rapidly than the radius. I n concentrated solutions of electrolytes the ions are not entirely independent and by taking into account the electrical interaction it is inferred that the niobility of ions in aqueous solutions will increase with the concentration of the ions provided that a suit- able correction is introduced for the change in viscosity.In the case of fused salts the mutual interaction becomes much more pronounced and at the same time the proportion of neutral molecules with which the ions may combine is greatly reduced with the result that the normal potential cannot be set up. The elec-GENERAL AND PHYSICAL CHEMISTRY. ii. 53 trical properties of fused electrolytes and in particular the high electrical conductivity are to be explained in terms of these peculiarities which distinguish the fused salts from ordinary solu- tions of electrolytes. G. VON HEVESY (KolloicZ Zeitsch. 1917 21 136-138. Compare preceding abstract).-If the number size and mobility of the particles of a colloid are known it is possible to calculate the charge on the particles and also the conductivity of the colloidal solution.By calculating the conductivity of a 0.1% solution of colloidal gold on various assumptions relative to the radius of the particles and comparing the results with the observed conductivity it is possible to derive information with respect to the actual size of the particles and the charge which they carry. It is probable that the maximum conducting power of a solution of a colloidal metal is less than 0-5 x 10-5 mho and the size of the particles which corre- spond with this is represented by R=lO-7 cm. where R is the radius. Although other colloidal substances appear tlo conduct somewhat better than the colloidal metals the conductivity woulcl appear t o be in all cases less than that of a 0*0001A-salt solution.The small conducting power of the colloids is to be attributed entirely to the sinall number of the particles for the charge carried by the par- ticles is always greater than thO charge of an ion of the same size. The ratio of the charges is in fact. given by (R / d + 1). where 72 is the radius of the particle and d the thickness of the electrical double layer = 5 x 10-7 cm. Fallacy of Determining the Electric Charge of Colloids by Capillarity. A. W. THOMAS and I . D. GARARD (J. Amer. (7Jbem. Soc. 1918 40 101-106).-It has been suggested that posi- tive and negative colloids can be distinguished by the difference in the capillary effects which are observed when strips of filter paper are dipped into the colloidal solutions. If the particles are negatively charged they are said to ascend the strip readily whereas positively charged particles are coagulated and deposited OIL the filter paper within a short distance of the surface of the solution.Experiments made with colloidal solutions of ferric hydroxide. chromium hydroxide arsenious sulphide antimony sulphide and xnolybdenum lead t o the conclusion that the basis of the above method is illusory and that there is no relation between the sign of the electrical charge and the capillary behsviour. The ascent of the colloidal particles depends on the dilution of the sol 011 tI1e presence of electrolytes on the nature of the surrounding atmo- sphere a d on the nature and previous treatment of the filter- paper strip.It) is probable that the supposed depeiidencc of th(1 c-apillary behaviour on the charge of the particles owes i t s oligirl 1 C) the circunistaiice that the observations were iriade with relatively cvncentrat,ed solutions of I he positivc yolloidq. the iipgat,ivp cnlloid:ll solutions being relatively dilute. H. M. D. The Conductivity of Dispersoids. H. M. D. H. M. D.ii. 54 ABSTRACTS OF CHEMICAL PAPERS. Electrochemical Behaviour of Nickel. A. SMITS and C. A. LOBRY DE BRUYN (Proc. K . 14kad. 'GVetensch. Arnsterdum 1918 20 394-403) .-Certain anomalies exhibited by nickel in its electro- chemical behaviour are attributable to the slow rate a t which internal equilibrium is established. In contact with air or hydrogen the anomalous behaviour is intensified and this appears to be due to the circumstance t h a t oxygen and hydrogen exert a negative catalytic influence. This influence of hydrogen explains the fact that the potential of a nickel electrode in contact with an atmosphere of hydrogen is tlhe same as that of the hydrogen elec- trode.I n reference t o the normal calomel electrode this potential is - 0.640 volt whereas the true equilibrium potential measured in a vacuiim is -0.480 volt. To obtain this value the hydrogen- ion concentration should not exceed 10-3 gram ion per litre. H. M. D. A Lead Standard Cell and a Determination of the Potential of the Lead Electrode. W. E. HENDERSON and has been found t h a t lead amalgams containing 2.5 t o 6% of lead may be used for the attainment of constant and reproducible Toten- tial differences.The cell P b amalgam 1 PbSO I Na2S0,,10H20 I Hg2S041Hg has an E.M.F. which may be represented by the equation for temperatures between 1 8 O and 3OC. The E.M.P. is reproducible t o within threehundredths of a millivolt but a gradual diminution of the E.M.F. occurs when the cell is kept for an extended period of time. From the temperature coefficient of the cell the heat of the reaction 0.5PbzHg + Hg,SO =PbSO + 2*5Hg is found t o be 42139 cal. The value derived from thermochemical data is 41785 cal. Measurements of the E.M.F. of a cell of the above type with the amalgam replaced by pure lead gave 0.96973 a t 25O. The difference of 0.0051 volt is used in the derivation of the E.M.F. of the cell Pb I PbC1 IO.lNKC11 Hg2CI I'Hg.giving 0.5187 volt a t 25O. This is further employed in the calculation of the' potential of the norinal lead electrode using available data for the solubili'ty and degree of ioiiisation of lead chloride. The E.M.F. of the cell obtained by combination of the normal lead electrode with the WlN-calomel electrode a t 2 5 O is 0.4696 volt. [See J. Soc. Chem,. I??d. 1918 March.] H. M. D. S. R. MILNER (Phil. Hag. 1918 [vi] 35 214-220).-The author contends t h a t the true degree of ionisation of an electrolyte cannot be obtained from either osmotic or conductivity data. The failure of the law of mass action in its application to solutions of strong electrolytes i s said t o be such t h a t insuperable difficulties stand in the way of ally t,heory which ascribes the variations in conduc*tivity to rha11p~ i l l the llumber of the iow These difficulties are avoided if the GEBHARD STEGEMAN ( J .Amer. Chem. sot. 1918 40 84-89).-It E = 0.96463 + 0*000174(t - 25) + 0*00000038(t - 25)z Effect of Interionic Forces on Electrolytes.GENERAL AND PHYSICAL CHEMISTRY. ii. 55 variations in conductivity are attributed to the action of the elec- trical field on the ionic mobility. Equivalent Conductance of Electrolytes in Dilute Aqueous Solution. I. The Water Correction. EDWARD lv. WASHBURN ( J . Anter. Chena. Soc. 1918 40 106-122).-The introduction of improvements in the Kohlrausch method for the determination of the conductivity of electrolytes (compare A. 1917 ii 10) has made it possible to obtain increased accuracy in the experimental measurement and to investigate in particular the conductance of electrolytes in very dilute solution.I n this paper which lorms the first of a series in which the results of these investigations are described the author discusses the’ question of the magnitude and nature of the water correction the influence of carbonic acid and the products to which this gives rise by metathesis in dilute soln- tions of various kinds of electrolytes. The fact that the strong acids in very dilute solution appear to be abnormal in their con- ducting power even when the observed conductivity has been corrected for carbonic acid suggests strongly that basic or saline impurities are present and since these affect the magnitude of the carbonic acid correction it is considered that accurate data for the conductivity of dilute solutions can only be obtained by the use of ultra-pure water instead of water which is in equilibrium with the atmosphere.Equivalent Conductance of Electrolytes in Dilute Aqueous Solution. 11. The Extrapolation of Conductivity Data to Zero Concentration. EDWARD W. WASHBURN ( J . Anzer. Chem. SOC. 1918 40 122-131).-The methods previously employed by Ilohlrausch Noyes Kraus and Bray and by Bates for the estima- tion of the limiting molecular conductivity are critically examined and rejectled as untrustworthy on the ground that. most of these involve the assumption t h a t the functional relation between the conductivity and the concentration which holds for the lowest measurable range of concentrations will also hold down to zero concentration.A new graphical method of extrapolation is described by which the author claims t o avoid the errors involved in the “arbitrary function ” methods and also those which attach to direct graphin4 extrapolation. The proposed method rests on two assumptions the first of which is t h a t with decreasing concentration ( c ) the value of ca2/(l - a) = k decreases and aqxoximates to a constant value ko a t extreme dilutions. According t o the second assumption the relation between c and Ic must be such t h a t deviations from the law of mass action do not increase with the dilution. The actual procedure in applying the method consists in plotting values of k against values of c for different assumed values of A and reject- ing those values which cause the curve to exhibit radical chsnqes in direction in the region of very dilute solution. I t is said to be possible t o determine A with a precision of 0.01% if the con- rluctivity data are of this order of acciiracy and extend t o I’ = 0-00002. H.M. D. H. M. D.ii. 56 ABSTRACTS OP CHEMICAL PAPERS. The method of extrapolation described by Bates (A 1913 ii 466) involves the assumption of the validity of the Storch equa- tion but otherwise resembles the method now proposed and when applied to the data for potassium chloride gives very nearly the samel result. H. M. D. Equivalent Conductance of Electrolytes in Dilute Aqueous Solution. 111. A Study of Dilute Solutions of Potassium Chloride. HENRY J. WEILAND ( J . ,4nzcr. Chenz. SOC. 1918 40 131-150).-A method for the preparation of “ultra-pure con- ductivity water ’’ is described in which ordinary conductivity water (k=0*6 to 0.8 x 106 mho) is heated to near its boiling point in a large quartz still a current of carefully purified air being passed through the water.The water vapour is condensed in a block-tin tuba and collected in a quartz receiver. The specific conductance of the water obtained in this way may vary from 0.05 to 0.07 x 10-6 mho a t 1B0 and has been used in the investigation of the Conductivity of very dilute solutions of potassium chloride. The conductivity cell of about 3 litres capacity is made of quartz and is provided with co-axial cylindrical platinum elec- trodes which are so constructed that the water does not come into contact with anything but.platinum or quartz The dilute solu- tions examined were prepared in the cell out of contact with the atmosphere by the successive introduction of small crystals of potassium chloride weighing about 0.005 gram. The error resulting from the adsorption of salt from these dilute solutions by the quartz surface of the cell has been examined and found to be negligible. The experimental data for solutions varying in concentration from about 0*000012\7 to 0.001N show that! the quantity k = c a 2 / l - a has a constant value for solutions for which the con- centration is less than 0.0001N. The results afford therefore direct experimental proof of the validity of the mass law in its application to the ionisation of potassium chloride a t very low concentrations.The limiting value of the equivalent conductance atl 1 8 O was found to be 129*64+0-02 (compare preceding abstract). The empirical relations suggested by Bates and by Kraus for use in extrapolating to zero concentration have been tested by refer- eiice t o the data for solutions between c-0 and c=O.005 and the conclusion is drawn t’hat these do not reproduce satisfactorily the experimental results 017er this range of concentrations. H. M. D. Equivalent Conductance of Electrolytes in Dilute Aqueous Solution. IV. Two Laws Governing the Ionisation Equilibrium of Strong Electrolytes in Dilute Solutions anda New Rule by means of which the Equivalent Con- ductance at Infinite Dilution can be Determined from a Single Conductance Measurement. EDWAWD W. WASHBURN (?7.A w f r . Cke/,i. Soc. 191s. 40 150--15S).--The behaviour ofGENERAL AND PIIY SICAL CHEMISTRY. ii. 51 eiectxolytes in very tiiliite solution as exemplified by the data for i'otassiuin chloride in the preceding paper and by the most accurate coiductivit y work of Kohlrausch arid his collaborators has led tho author to the following general conclusions. I n sufficiently dilute solutions (c < O*OOOl,V) all uni-univalent salts of strong acids and bases are ionised in accordance with the requirements of the law of mass action and the ionisation constant is the same for all. The values of k = ca2/ 1 - a for such salts are identical in sufficiently dilute solution and this identity persists up to concentrations which are higher the more closely the salts under comparison resemble each other.According to this second generalisation the identity in the ionisation relations extends beyond the concentration up to which the salts in question satisfy the requirements of the mass law. I n terms of the equation Ac,2C/A,(A,-Ac)=K the state- ment implies that K is independent of the nature of the salt and on simplification this equation leads to the relation Ao=Rc]a where a is independent of the nature of the salt. For C=O*OOOIAi\T this equation becomes A = 1.00475 A,. Evidence in support of the above conclusions is furnished by the dat'a for the salts of the alkali metals which are examined in detail. R. M. D. Electrolytic Deposition of Alloys and their Metallographic and Mechanical Investigation. VIII. Cathodic Deposits of Iron and Iron-Nickel Alloys obtained at the Ordinary Temperature under High Hydrogen Pressure.ROBERT KREMANN and HERMANN BREYMESSER (Monatsh. 1917 38 359-384. Compare A. 1915 ii 511).-In previous papers (Zoc. cit.) i t has been shown that electrolytic iron deposited at ordinary pressures and temperatures is charged with hydrogen and admixed with ferric hydroxide; it' is also brittle and hard. These condi- tions are shown to be due to the simultaneous liberation of hydrogen a t the cathode. It is shown theoretically that the simultaneous liberation of hydrogen can be prevented if the electrolysis is carried out under a high hydrogen pressure. To test this deduction iV-solutions of ferrous sulphate have been electrolysed under a pressure of 20 atm. of hydrogen with a C.D.of 0.25 amp./dm2 and 0.75 amp./dmz. A further series of experiments was carried out with solutions to which 10 grams of citric acid per litre were added. In the first two cases it is shown that the current efficiency is 99.33-99*46% whereas in the last case it is only 88.57%. The deposits obtained in these cases have been compared with those obtained under 1 atm. pressure of hydrogen and in the presence of 0*15N-sulphuric acid. It is shown that the material obtained under the higher pressure is composed of larger cryEtals than that under the lower pressure. The hydrogen content of the high- pressure specimens is much less than that of the low-pressure speci- mens whilst the hardness is very much reduced by the deposition under high hydrogen pressure The specimens obtained under the present conditions are much less brittle than those obtained under other conditions whereas the magnetic properties of the deposits 5"ii.58 ABSTRACTS OF CHEMICAL PAPERS. are iiot affected Ly chsiige iii the coiiditioiis of depositioii. Electro lytic iron produced under pressure when immersed in il'-ferrous sulphate solution shows a t once the equilibrium potential of -0.417 volt whilst iron deposited under other conditions only reaches this value after some considerable time. A further series of experiments was made under similar condi- tions with mixed solutions of ferrous sulphate 0-7N and nickel sulphate 0*3iV. I n this case the nickel-iron deposits did not show the improvement observed in the case of iron alone. A number of microphotographs of the deposits is appended t o the paper.[See also J . SOC. Chem. Ind. 1918 March.] J. F. S. Temperature Determinations by Eutectic Alloys. CHARLES P. STEINMETZ ( J . Ampr. Client. Soc. 1918 40 96-loo).- Eutectic points of alloys furnish a more satisfactory means for the determination of temperature than the melting points of pure metals. I n general the eutectic point is not dependent t o the same extent on the purity of the substance and the alloys afford a more numerous series of fixed points on the scale. The cooling curves of a number of alloys have been examined and twelve of these found to give satisfactory eutectic tempera- tures ranging from 6 9 0 5 ~ t o 194'0O. The approximate composition of these low-melting eutectic alloys is recorded in the paper.[See J . SOC. Chem. Ind. 1918 March.] H. M. D. Solubility of Sodium Sulphate as a means of Determining Temperatures. THEODORE W. RICHARDS and VICTOR YNGVE ( J . Amer. Chem. SOC. 1918 40 164-174).-The solubility of a sub- stance with a large temperature coefficient may be used for the accurate measurement of temperatures. Between 15O and 25O the solubility of sodium sulphate varies rapidly with the temperature and since it may easily be obtained in a pure condition and readily gives a saturated solution it has been examined with a view t o its application in the measurement of temperature. Measurements of the solubility were made a t accurately deter- mined temperatures in the neighbourhood of 15O 175O 20° and 25O. The results are represented very closely by the equation log s= 0.659970 + 0.02963889t + 0.0000688925t2 in which s is the solubility expressed as the number of grams of sodium sulphate per 100 grams of water. It is claimed that the solubility method described will permit of the determination of temperatures to within O*0lo and may be used for the standardisation of thermometers.H. M. D. The Transition Temperatures of Strontium Chloride and Strontium Bromide as Fixed Points in Thermometry. THEODORE W. RTCHARDS and VICTOR YNGVE ( J . Amer. Chem. Soc. 1918 40. 89-95).-The transition temperat'ures have been determined by methods described in previous papers (compare A. 1914 ii 244). The purification of strontium chloride can be effected by re-GENERAL AND PHYSICAL CHEMISTRY. ii. 59 c.ryslalliuiiig above arid afterwards below the transition tempera- ture.It is found that barium can be readily removed by recryst>allising below the transition temperature (61°) so as t o obtain the hexahydrate. The calcium cannot be removed in this way but recrystallisation above the transition temperature yields the dihydrate from which the calcium is rapidly eliminated. I n a similar way strontium nitrate may be purified by crystal- lising out under conditions which yield the anhydrous salt whereby the calcium is removed. The barium may then be removed sub- sequently by recrystallising the hexahydrated chloride prepared from the partially purified nitrate. The temperature at which the hexahydrate of strontium chloride is transformed into the dihydrate has been found to be 61.341O at atmospheric pressure on the hydrogen scale.Preliminary experiments show that hexahydrated strontium bromide is similarly transformed into the dihyclrate at aboiit 88*6F. !See J . SOC. Chem. 1?d. 1918 109~.] 13. M. D. Theory of Specific Heats. WALTHER JANKOWSKY (Zeitsch. Rlektrochem. 191 7 23 368-371).-A theoretical paper in which on the basis of the older kinetic theory it is shown that without making an assumption of an equal partition of energy it is possible to derive a formula for the specific heat' of gases and vapours. This formula has the form c=e,iz(l + i ) / m . n . i i n which c is the true specific heat that is the heat at constant volume em the mean energy change expressed in calories of a molecule when the temperature is raised lo i is the ratio o€ the change of molecular energy to that of the sum of the atomic energies i =eqn l e a .n and ?z is the number of atoms in the molecule. This formula is tested on a large number of gases and vapours and is found to give good agreement. The dependence of the specific heat on temperature is shown to be due to changes in the value of i. This is a direct con- tradiction of the theory of equal energy partition. This formula gives a simple theoretical basis t o the Law of Dulong-Petit and Joule and explains quantitatively the divergencies and also shows that a strict following of the law is impossible. A further formula is deduced for the calculation of the ratio of the specific heat a t constant pressure to that a t constant volume. This has the forms k=l+2C/3cVm and k=1/1-2Cf3cPrn in which C=2*98 k=cpfcV and m is the molecular weight.J. F. S. Calculation .of Gas Equilibria IV. D. TREADWELL (Zeitsch. Elektrochem. 1917 23 270-272).-A mathematical paper in which by a series of approximations the author shows that the value of a the chemical constant which is expressed by Planck as a = R 1Ogep - C'p logeT + r,/ T can on thermodynamic grounds be proved to have the value a = R log,p + r0/ T - C p log,T. I n both equations p is the vapour pressure C'p the specific heat at constant pressure of the vapour ro the heat of vaporisation a t absolute zero and C p the specific heat of the liquid phase. J. F. S. 5*-3ii. 60 ABSTRACTS OF CHEMICAL PAPERS. Specific Heat of Liquid Ammonia. NATH~N S. OSBOHNP and MILTON s.VAN TIUSEN (./. Anzrr. Chem. Soc. 1918 40 1 -13).-Measurements of the specific heat of liquid ainmovia havr been iiiatle according t,o two independent methocls. 1 II one ot 1 hese the change in temperature produced by a measured quantity of heat under saturation conditions was determined whilst in the other the calorimeter was kept full of liquid a t constant pressure greater than the saturation pressure the change in temperature produced by the added heat energy being corrected for the heat content of the expelled liquid. The interval of temperature covered by the measurements ranges from -45O to 45O and over this range the specific heat increases from 1.058 at - 4 5 O to 1.173 a t 4 5 O . The dependence of the specific heat on the temperature is represented by the empirical equation c = 0.7498 - 0*0001368 + 4.0263 2/ 133 - 8 in which c is expressed in terms of the 20° calorie and 0 is the temperature. H.M. D. Latent Heat of Vaporisation of Ammonia. NATHAN S. OSBORNE and MILTON S. VAN DUSEN ( J . Amer. Chenz. SOC. 1918 40 14-25) .-The calorimeter used in the determination of the specific heat of liquid ammonia (compare preceding abstract) was modified so as to render it suitable for the measurement of the heat' OF vaporisat'ion. A known quantity of heat developed and measured electrically is employed to vaporise a portion of the liquid ammonia contained in the calorimeter the ammonia vapour lwing withdrawn at measured temperature and pressure and its amount estimated. Corrections due to thermal leakage were reduced t o a minimum by special methods of manipulation.The results obtained show that the latent heat of vaporisation decreases from 333.0 cal. a t -42O to 252.6 cal. at 49O. The variation of t'he latent heat with the temperature - - - - may be expressed by the empirical equation L = 32.968 4 133 - 8 - 0*5895(133 - O ) in which 8 represents the actual temperature and 133 represents the critical temperature. From the latent heat of vaporisation and the specific heat of the liquid under the pressure of its saturated vapour the authors have calculated the specific heat of saturated ammonia vapour. Expressed in joules per gram per degree this varies in a con- tinuous manner from -4.42 a t - 4 5 O to -3.36 a t 45O. To reduce these numbers to 20° calories they must be divided by 4.163.H. M. D. The Vapour Pressure of Liquid Ammonia up to the Critical Temperature. 11. FREDERICK G. KEYES and R. B. BROWNLEE ( J . Amer. Clzem. SOC. 1918 40 25-45).-The vapour pressures of liquid ammonia have been measured between Oo and the critical temperature by a method involving the use of a piston gauge. The dependence of the vapour pressure on the tempera- ture may be expressed by the equation log p = - 196965 / T-I- 16.19785 - 0.04238582' + 5.4131 x 10-51'2-3-2715 x lO-RTT3. ThisGENERAL AND PHYSICAL CHEMISTRY. ii. G I equation holds satisfactorily for vapour pressures down to the freezing point ( - 7 7 O ) . The boiling point was found to be - 33-20 k 0.05'. 13. M. D. Formula giving the Saturated Vapour Pressure of a Monatomic Liquid. E. AHI~S (Conapt.mid. 1918 166 193-197). -The author deduces the formulae I1 = d 2 / ~ and where 7 is the reduced temperature and 11 is the reduced pressure of the saturated vapour and shows that the calculated results agree with the observed results in the cases of krypton xenon arid argon. W. G. Vapour Pressures of Liquid Metals. JOEL H. HILDEBRAND ( J . Amer. Chem. SOC. 1918 40 45-49).-1t has been shown previously (A. 1915 ii 416) that the heat of vaporisation divided by the absolute vaporisation temperature is the same for all normal liquids provided that comparison is made a t temperatures for which the concentrations of the saturated vapours have the same value. On the basis of this relation it is possible to superimpose the vapuur-pressure curve for one substance on that of another by means of a single constant a which expresses the ratio of the absolute temperatures referred t o above.By t,aking mercury as the standard liquid for which the vapour pressure may be represented by the equation log p = - 3140/ T + 7-85. i t is shown that' the vapour pressures of a number of other liquid rnet'als may be expressed by the equation log p= - 3140a/ T +- 7.85 +log u in which a varies from 1-74 for cadmium to 4-90 for iron. By meaiis of the tabulated values of u for the different metals it is possible t o calculate the vapour pressure a t any temperature and also the heat of vaporisation. H. M. D. Apparatus for the Determination of Boiling Points. ALFRED EDWARDS ( J . SOC. Ckent. Znd. 1918 37 38~).-A siinpfe form of boilirig-pointl apparatus is described which provides for tho complete immersion of the thermometer stem in the heated vapour.The still-head is traversed by a somewhat narrower inner tube with a hole near the top through which the vapour from the boiling liquid escapes into the outer tube passing therefrom to a condenser tube which is sealed into the still-head a t its lower end. The condensed liquid tends to seal the space between the inner s n c l outer tubes in its lower portion and thereby to secure a regular stream of vapour through t'he inner tube in which the therrrio- meter is supported. H. M. D. TheHeat of Formation of Liquid Water from its Ions. J . A. MULLER (Bull. Soc. chin,. 1918 [ivl. 23 8-13).-C~ing the method previously described (compare A. 1913 ii 115> theii.G2 ABSTRACTS OF CHEMICAL PAPERS. author has determined the heats of reaction of sulphuric aud hydrochloric acids with potassium hydroxide in aqueous solution a t infinite dilution and from his results has calculated the heat of formation of water from its ions. The values obtained were with sulphuric acid 13,966 cal. and with hydrochloric acid 14,003 cal. W. G . Thermochemical Studies. The Heat of Combustion of the Paraffins. DANIELAGERLOF ( J . p ~ . Chem. 1917 [ii] 96 123-124. Compare A. 1905 ii 76).-The heat of combustion of ,,-octane observed by direct measurement (Richards and Jesse A. 1910 ii 269) agrees closely with the value calculated with the aid of the author's hypothesis (loc. cit.) which thereby receives confirmation. D. F. T. The Heat of Ionisation in Aqueous Solution of Crystalline Barium Sulphate and the Solubility of this Salt in Water.J. A. MULLEI~ (Bull. Soc. c?&n. 1918 [iv] 23 13-16).-From measurements of the heats of reaction of barium chloride and sulphuric acid at increasing dilutions and determining the limits towards which these tend at three temperatures the following expression is found for the value of Q the heat of combination of the ions Ba and SO p = 105502.32 - 696*857T + 1.21187'2 T being the absolute temperature. The heat of ionisation of crystalline barium sulphate in aqueous solution is the inverse of this. Froin this it is possible to calculate the ratio of the solubilities of barium sulphate at different temperatures and the results agree with those of van't Hoff. [See also J.SOC. Chem. Znd. 1918 March.] W. G . Adsorption Compounds and Adsorption. 11. Replace- ment from the Surface. 1,. BERCZELLER and Sr. HETJ~NYI (Uiocl~e in. Zeitsch. 19 17 84 11 8-1 36) .-Stalagmometric measurements of the effect of addition of alcohols to solutions of various cryddloid and colloid substances. S. B. S. Utilisation of the Adsorptive Power of Fuller's Earth for Chemical Separations. ATHERTON SEIDELL (J. Amel.. Chewz. SOC. 1918 40 312-328).-A comparison has been made of the adsorptive capacities of thirty-six samples of fuller's earth and other similar clays by experiments on the adsorption of quinine bisulphate and methylene-blue. The adsorptive power of a given quantity of a particular sample increases with the concentration of the unadsorbed substance in the aqueous solution and also with the time of contact although the rate of adsorption gradually diminishes. The method of measurement consisted in mixing 1 gram of the fuller's earth sample with 10 C.C.of water and adding t o the mix- ture measured volumes of 1% quinine bisulphate or 0.5% methylene- blue solution. The tubes containing the mixtures were shaken for half an hour and the solutions then examined for the adsorbedGENERAL AND PHYSICAL CHEMISTRY. ii. 63 substances. From a series of such tests it was possible to obtain comparative numbers for the adsorption powers of the samples. I n the case of both the quinine salt and the methylene-blue the free base only is adsorbed. When both are present in the solution in equal amounts they are adsorbed to very nearly the same extent.If the fuller’s earth is first shaken with methylene-blue and then with quinine bisulphate a small amount of the latter is adsorbed and only a trace of the former liberated. If the order is reversed a much larger proportion of methylene-blue is adsorbed and a con- siderable amount of the quinine salt is displaced. The experiments show further the influence of dilution of the acidity of the solution and of the presence of ethyl alcohol and sucrose. [See also J . SOC. Chent. Ind. 1915 March.] H. &I. D. Preparation of Uniform Collodion Membranes for Dialysis. CHESTER 3. FARMER ( J . Biol. Chem. 1917 32 447-453. Com- pare Brown A. 1917 ii 362).-The membrane is prepared by filling a glass tube with collodion solution inverting and allowing to drain for one minute.The tube is then dried in a current of air for one minute and afterwards filled with cold water. After a few minutes the thin membrane can be removed from the walls of the glass tube with the aid of a pair of forceps. Convenient apparatus for performing these operations is described in detail in the original paper. H. W. B. The Colloidal Membrane its Properties and its Function in the Osmotic System. FRABK TINKER (Tranq. Fnrnday Soc. 1917 37 133-140).-Although i t is probable that the average kinetic energy of a molecule in the liquid state is the same as in the state of vapour this must not be taken t o imply that the average pressure of a solute molecule has the same value in the two states. In the solution a large proportion of the volume is occupied by the solvent molecules with the result that the free space is greatly reduced and the pressure which the solute mole- cules would exert on an imaginary flame is consequently much greater than the corresponding gas pressure.The internal bombardment pressure of the solute must therefore iiot be confused with the osmotic pressure which in the author’s opinion is an external mechanical pressure. The supposed analogy between osmotic and gas pressure is also considered to have no real foundation in that this analogy fails t o account for the funda- mental phenomena of diffusion. The mechanism involved in an osmotic system is said to be quite different from that which pro- duces gas pressure there being no pressure on the membrane unless the solution is compressed.This pressure is then exerted by the solution as a whole and not by the individual molecules whether of solute or solvent’. The similarity between a dilute solution and a gas is due to the fact that both experience no change in internal energy when theii 64 ABSTRACTS OF CHEMICAL PAPNRS. volume is varied. Equality in tlhis respect does not however afford any proof that the mechanism involved in gas and osmotic pressure is of the same nature. Reference is made to the importance of the study of the proper- ties of the membrane and a brief account is given of the author’s work on this subject. ALFRED FV. PORTER (Y’ram. Fayaday SOC. 1917 13 123-132).-The cause of osmotic pressure is discussed and it is pointed out that the kinetic theory is the only theory yet advanced which reproduces directly the values for the osmotic pressure which have been actually obtained in experiments with dilute solutions.The arguments against the kinetic theory which rest largely on the dissimilarity of the con- ditions in liquids and gases have been greatly weakened as the result of observations on the Brownian motion. These observations afford the experimental basis for a kinetic theory of liquids accord- ing to which solute and solvent molecules are in a state of rapid movement. The osmotic pressure represents the dynamical effect of this thermal motion of the solute molecules and in order to obtain a mental picture of the effect of the presence of the solute i t may be supposed that the molecular bombardmeat of the mole- cules of the solute on the boundary surface tends to enlarge the boundaries and thereby to relieve the total pressure on the solvent.In other words the kinetic pressure resulting from the thermal agitation of the solute molecules acts outwardly and diminishes the Laplacian pressure by an equivalenti amount. It is shown that the data for the osmotic pressures of sucrose solutions a t 20° can be represented satisfactorily by the equation P(v-b)=RT in which b is a constant which is great’er than the volume of the sucrose. On the assumption that this is due to the hydration of the sucrosel molecules it is found that about 5-3 mole- cules of water are associated with a molecule of sucrose. If this equation is applied t o the whole of Morse’s results and hydration values calculated for diff’erent concentrations and temperatures the numbers are not quite regular but show clearly that hydration diminishes with increasing concentration.The values for dilute solutions are surprisingly high but’ are considered to be quite plausible and it is suggested that the variation in solubility with the temperature may be due t o changes in the degree of hydration. According to the equation connecting osmotic pressure with the latent heat of dilution of the solution the latter quantity depends on the variation of PIT with the temperature T. I f these varia- tions derived from Morse’s values a t loo and 30° are compared with those calculated from measurements of the heatl of dilntion a t 20° the dBgree of correspondence is found t o be quite satis- factory.H. M. D. Osmotic Pressure in Relation to the Constitution of Water and the Hydration of the Solute. W. R. BOUSFIELD (T?.ans. E’arnday Sot.? 1917 13 141-155).-The aiithor’s previous H. M. D. Kinetic Theory of Osmotic Pressure.GENERAL AND PHYSICAL CHEMISTRY. ii. ti5 work 011 the properties of solutions is considered with reference to the kinetic interpretation of osmotic pressure. This interpretatioii rejects the idea that the molecules of the solute are directly responsible for the osmotic pressure and a modified gas theory is put forward which involves the assumption that water is a mix- ture of three kinds of molecules vapour liquid and ice molecules represented by H,O (H20)a and (H,O) respectively and attributes the osmotic pressure to the thermal agitation of the vapour mole- cules.The validity of the ideal gas equation for dilute solutioiis is supposed to indicate that these hydrol molscules comport them- selves towards changes of pressure and temperature in the same way as the molecules of a gas. The addition of a solute to water is said to be accompanied by a reduction in the proportion of both the vapour and ice molecules in the equilibrium mixture resulting in a diminution of the vapour pressure and a lowering of the freezing point. Reduction of the molecular interspace by external pressure raises the vapour pressure and the osmotic pressure is defined as the liquid pressure under which the external vapour pressure of a solution is equal t o the internal vapour pressure of the pure solvent.The relations between the osmotic pressure the lowering of the vapour pressure and the freezing point are discussed in reference tlo the above theory and i t is claimed that the various osmotic data are brought into line with ot’her properties by t’hs assumptioil that’ the active hydrol molecules enter into combination with the solute molecules. It is said that different properties lead to the same value for the degree of hydration of the solute. H. M. D. Solubility and Internal Pressure. JOEL H. HILUEBRANU ( J . Amer. Chenz. SOC. 1918 40 198).-Corrections to a previous paper (this vol. ii 36). H. 04. D. Changes in Volume during Solution. 111. GBEGORY YAur BAXTER ( J . Amer. Chem. SOC. 1918 40 192-193).-1f A l l A f B f *lB’ and A’H represent the four salts formed by the ions A U Ll’ and B’ then the apparent volume in solution of one of these may I)e obtained if the apparent volumes in solution of the three others are known.The calculation depends on the fact that the sum of the apparent volumes of LIB and A’B’ is equal to the sum of the apparent volumes of AB’ and A’B. The apparent volume of dissolved calcium carbonate obtained in this way from the apparent volumes of calcium chloride sodium carbonate and sodium chloride is 3.0 C.C. Since the molecular volume of solid calcium carbonate is 36.9 c.c. the change in volume on dissolution is 3.0 - 36.9 = - 33.9 C.C. per mol. The contraction is thus more than 90% of the volume of the solid salt. IT. M. I). The Structure in Steps in certain Anisotropic Liquids.F. GRANDJEAN (Coinpt. r e n d . 1918 166 165-167).-This struc- ture already found in ethyl azoxybenzoate and cinnamate (com- pare Bull Soc. franc. Min. 39 167) has been found also in theii. 66 ABSTRACTS OF CHEMICAL PAPERS. oleates aiid the positive phase of cholesteryl decoate. It is described in detail. This structure in steps aiid particularly the existence of steps of extremely slight thickness separated from steps infinitely near by abrupt lateral surfaces reveals a discontinuous property of the liquid which is not observed in the group of azoxyphenetole. W. G . Precipitation of Colloidal Gold and Platinum on Metallic Surfaces. ELLWOOD B. S P ~ A R and KENrjETlf n. I ~ A H N (J. Awzer. C’hem. SOC. 1918 40 181-184).-The precipitation of colloidal gold and platinum solutions when brought into contact’ with polished plates of zinc steel nickel lead tin copper or platinum has been examined with results which indicate thatl the rate of coagu- lation increases with the electro-positive character of the metal.The effectl also depends on the nature of the metal surface in that the rate of coagulation for a given metal decreases if the surface is roughened or if the metal is employed in a finely divided condition. It is suggested that ions of the active metal are formed and that these are adsorbed by the colloidal particles thereby neutralising their negative charge. I n support of this view it has been found that copper does not bring about! coagulation if the colloidal gold or platinum solution is freed from air by the passage of a current of purified hydrogen.[See J . SOC. Chein. Tnd. 1918 March.] H. M. D. Laws of Chemical Equilibrium. ERSKINE D. WILLIAMSON and GEORGE TV. MOREY ( J . Anzer. Cherri. SOC. 1918 40 49-59). -A theoretical paper in which the authors derive general equa- tions for the equilibrium in heterogeneous chemical systems by methods which are essentially based on the work of Gibbs. Special forms of equations applicable to systems of simple type are specifically ref erred to. Pressure-Temperature Curves in Univariant Systems. GEORGE ’CV. MOREY and ERSKINE D. WILLIAMSON ( J . Amer. Chenr. ,Sot. 1918 40 59-84. Compare preceding abstract).-The pressure-temperature curves for univariant systems are discussed in reference to Gibbs’s equations. The conditions under which different pressure-temperature curves become coincident are ex- amined and a method is developed by which the order of succession of the pressure-temperature curves intersecting a t an invariant point can be determined.The applicability of the method is shown by reference to the five curves which melt in the quintuple points characteristic of the ternary system H,O-K,SiO,-SiO,. H. M. D. H. M. D. Univariant Equilibria in the Ternary System-Water Sodium Sulphate Ammonium Sulphate. C. MATIGNC N and F. MEYER (Compt. rend. 1917 165 787-789).-The experi- mental data recorded show the composition of solutions which areGENERAL AND PHYSICAL CHEMISTRY. ii. 67 ia equilibrium with two solid substances. The two sulphates combine to form the double sulphate NaNH4S0,,2H,O antl numbers are given for solutions saturated with respect to the double salt and Glauber's salt between -13O and 25'5O double salt and anhydrous sodium sulphate between 29O and 5S0 double salt and ammonium sulphate between -19O and 58O and with respect t o anhydrous sodium sulphate and ammonium sulphate betweei 1 62'5O and logo.A solution saturated with sodium sulphate boils a t 1 0 2 O and contains 2.10 mols. per 1000 grams of solution. Similarly the b. p. of saturated ammonium sulphate solution is 108'9O and it cont'ains 3.922 mols. per 1000 grains whilst a solution saturated with respect to the two sulphates boils at 1 1 1 O and contains 1.125 mols. Na,SO apd 3.175 niols. (NH,),SO per 1000 grams of solution. [See further J . s o c . Chem. T ? d .1918 29~.] Heterogeneous Equilibria between Aqueous and Metallic Solutions. G. McP. SMITH and 8. A. BRALEY (J. Anzer. C h m . Sot. 1918 40 197).-A correction of results recorded in a previous paper (A. 1917 ii 455).-The error necessitates a further in- vestigation of the ionisation relations in mixtures of sodium and strontium chlorides. H. M. D. Invariant Equilibria in the Ternary System Water- Sodium Sulphate-Ammonium Sulphate. C. MATIGNON and F. MEYER (Compt. rend. 1918 166 115-119).-A study of the equilibrium of the solution in the presence of the various com- binations of three solid phases the cooling curves being plotted. I n a trilinear diagram with co-ordinates giving respectively the temperature the concentration of anhydrous sodium sulphate antl the concentration of ammonium sulphate are shown the surfaces corresponding with the states of equilibrium of the solution with respect to one solid phase.[See also J . SOC. Chem. Ind. 1918 March .] W. G. Effect of Hydrogen Chloride on the Nitrogen-Hydrogen Equilibrium. E. B. LUDLAM (TTCL~S. E'nrnday SOC. 19',7 13 43-52) .-The observations made by Deville suggest that the stability of ammonia a t high temperatures is increased very con- siderably by the presence of hydrogen chloride and lead to thc supposition that the equilibrium between nitrogen hydrogen and ammonia will be displaced in favour of the ammonia i f hydrogen chloride is added to the mixture. Experiments in which a mixture containing equivalent quanti- ties of nitrogen hydrogen and hydrogen chloride was subjected t o the action of an electrically heated platinum wire or carbon rod stretched along the axis of a water-cooled tube afforded no evidence in support of the above hypothesis.The soaking of the carbon rod in solutions of sodium calcium or magnesium chloride made no difference to the result. Other experiments in which a mixture of nitrogen and hydrogen was passed slowly through a quartz tube containing sugar charcoal H. M. D.ii. GS ABSTRACTS OF CHEMICAL PAPERS. a t about 800° gave results which seemed to show thatl t'he amillonis formed was increased when hydrogen chloride was added to the nitrogen-hydrogen mixture although the effect was much smaller than that calculated from the mass action equation. On the assumption that' the smallness of the effect was due to the slowness of the reaction attempts were t'hen made t o approach the equilibrium condition by starting with ammonium chloride.Il'eighed quantities of this were accordingly heated in an evacuated quartz tube in presence of gold silver copper and iron with results which seemed to show that ammonium chloride is not nearly so stable a t high temperatures as Deville's observations would suggest. The evidence afforded by these experiments pointed to iron as the most active catalyst and further observations were therefore made in which nitrogen hydrogen and hydrogen zhloride were passed through a layer of iron asbest'os heated a t about 450O. Even a t this low temperature ferrous chloride is formed and sublimes and the volatility of this substance would evidently be a serious obstacle to the use of iron in practice even if the catalytic activity of the metal were very considerable.H. M. D. Equilibrium Data on the Polybromides and Polyiodides of Potassium. G. A. L r N H A w ( J . ilnzer. Chenz. SOC. 1918 40 155-163).-0n the assumption that KBr and KBr are present in aqueous solutdons which contain potassium bromide and bromine the constitution of the solution is determined by the equations [Brg']/[BrfJ[Br2J = K' and [Br,]j[Br3][Br,J=Kf'. By reference to Worley's data for 26*5O it is shown that' K' remains very nearly constantl= 15.9 i f it is assumed that li" =1*2. A t Oo Rf = 19.6 and Xff=2.08 and at' 3 2 . 6 O Rf=15.5 and 1</f=1*06. From the values of the constants at the two lower temperatures the author calculates the heat of the reactions Brf + Br2(aq) = Br,' - 1290 cal.and Br3/ + Br,(aq) =B:,/ - 3390 cal. When the value K/f=1*2 for 2 6 . 5 O is applied t o solutions which are saturated with bromine the calculated value of lif is appreci- ably higher than that indicated above and it is suggested that this may be due to the formation of KBr7. Measurements of the ratio of distribution of iodine between carbon tetrachloride and water a t 25O show that the ratio of the concentrations expressed in mols. per 1000 grams of solvent is constant= 57.7. The constitution of iodine-potassiuni iodide solutions is also dis- cussed briefly. H. M. D. Influence of Substitution in the Components on the Equilibrium in Binary Solutions. X. Equilibria in Binary Solutions of p-Toluidine and Carbamide respectively with Nitro-derivatives of Benzene.ROBERT KREMANN and BRUNO PET~RITSCHEK (Hotzutsh. 1917 38 385-404. Compare A. 1905 ii 307; 1906 ii 268; 1912 ii 1151).-The binary systems ;rr-toluidine with the three dinitrobenzenes 2 4-dinitrotoluene and the three nitrophenols respectively and carbamide with the threeGENERAL AND PHYSICAL CHEMISTRY. ii. 60 cliiiitrobenzenes aud I! 4-dinitrotoluerte respectively have bee t i iiivestigated by means of time-cooling curves. It is shown in the cease of p-toluidiiiei with the three dinitrobenzenes am1 with 2 4-dinit rotolnene gives no compounds but only simple eutectics ; with m- and p-nitrophenols a cornpound is formed in each case. With 7t~-nitrophenol and ptoluidine the compound produced consists of one molecule of each constituent whilst with p-nitro- phenol the compound consists of two molecules of p-nitrophenol to one molecule of ptoluidine.In the case of o-nitrophenol and p-toluidine no compounds are formed but simply a eutectic. Carbamide does not form any compounds in any of the mixtures examined and in all cases there are large gaps in the mixture series. As a result of the experiments the authors state that the tendency to cornpound formation with the dinitrobenzenes is deter- mined by the residual affinity of the benzene nucleus and not by the affinity of the amino-group whereas in the case of the nitro- phenols the amino-group is the determining factor. Influence of Substitution in the Components on the Equilibrium in Binary Solutions.XI. Binary Solution Equilibria between Phenol and the Three Isomeric Nitro- phenols respectively with the Three Isomeric Phenylene Diamines. ROBERT KREHANN and BRUNO PETRITSCHEK (Moizntsh. 1917 38 405-444. Compare preceding abstract) .-By means of cooling curves the authors have investigated the twelve possible binary systems formed between phenol and the three nitrophenols on the one hand and the three phenylenediamines on the other. The system phenol-p-phenylenediamine gives rise to the compound consisting of one molecule of diamine t o two molecules of phenol. This compound forms a eutectic with phenol a t 40° and with pphenylenediamine a t 94O. The system phenol-m-phenylene- diamine gives rise to a compound made up of three molecules of phenol and two molecules of the diamine; this compound forms a eutectk with phenol at' 24O and with diamine a t 41° and has m.p. 52'6O. I n the case of the system phenol-o-phenylenediamine two compounds appear ; these consist respectively of four molecules of phenol and one molecule of the diamine and one molecule of each component. The eutectics in the case of the first compound lie atl 2 8 O with phenol and 2 9 O with the second compound. The system pnitrophenol-o-phenylenediamine forms a single cornpound composed of two molecules of nitrophenol and one molecule of the diamine (m. p. 87*9O). This compound with nitrophenol has a eutectic a t 85.5O and with diamine a t 7 8 O . A compound of similar composition is formed in the system p-nitrophenol-nz-phenylene- diamine; this melts at 1 1 9 * 9 O and its eutectics lie atl 1 0 2 O with nitrophenol and 5 2 ' 4 O with the diamine.In the system p-nitro- phenol-p-phenylenediamine two compounds are found ; these have compositions (a) four molecudes of nitrophenol t o one molecule of the diamine and ( 6 ) one molecule of nitrophenol to one mole- cule of the diamine respectively. The eutectics lie a t 109.5O for p-nitrophenol and compound n 117*li0 for compound n and com- J. F. S.ii. 70 ABSTRACTS OF CHEMICAL PAPERS. pou~id I) a i d 107' for the couipotmd b aiid tche diamine. The systems o-nitrophenol and the three phenylenediamines do not give rise to compounds; the eutectick in these cases lie a t 4 2 . 5 O for p-phenylenediainine 3 3 ' 5 O for m-phenylenediamine and 38*S0 for 0-p hen y 1 ene diamine .The system m-nitr op henol-m-phen ylen e- diamine gives rise t o two compounds composed of two molecules of nitrophenol and one molecule of the diamine and one molecule of nitrophenol and one molecule of the diamine respectively. In the system m-nitrophenol-o-phenylenediamine two compounds are found; these are two molecules of nitrophenol with one molecule of o-phenylenediamine and an equimolecular compound. Similar rela ti onships are found in the system m-ni t rophenol-p-ph en ylen e- diamine. J. F. S. A Complete Review of Solutions of Oceanic Salts. 111. ERNST JANECKE (Zeitsch. anorg. Chem. 1918 102 41-65. Compare A. 1917 ii 527).-In the previous paper a graphic representation was devised for the doubled ternary system (Na,,K,,Ng) (Cl,,SO,) in presence of sodium chloride as a constant solid phase.The diagram took the form of a triangle for any particular temperature the corners of the triangle corresponding with the three salts MgCl 2KC1 Na,SO,. The temperature co- ordinate being perpendicular to the plane o€ the triangle the com- plete diagram took the form of a three-sided prism. Although sodium chloride is always present' as a solid phase the quantity in solution or the quantity of water corresponding with saturatlion with salt a t different temperatures has been hitherto neglected. In the present paper this new factor is taken into account. A t first the new variable is considered in connexion with the simple salts represented by the corners of the triangle and later with respect to the systems associated with the sides.For an interpretation of the numerous diagrams given the original paper must be referred to. [See further J . SOC. Chew?. Ind. 1918 March.] E. H. R. Chemical Kinetics. U. PRATOLONGO (Atti R. Accad. Lincei 1917 [v] 26 ii 182-190).-The author bases on Marcelin's work (Contribution 3 1'8tude de la cine'tique physico-chimique Thesis Paris 1914) the essentials of a new thermodynamics of irreversible phenomena. T. H. P. Effect of Temperature and of Pressure on the Limits of Inflammability of Mixtures of Methane and Air. WALTER MASON and RICHARD VERNON WHEELER (T. 1918 113 45-57).- Theoretical considerations indicate that the effect of increasing the initial temperature of mixtures of inflammable gases with air should be to widen the difference in the composition of the mix- tures which correspond with the upper and lower limits of inflam- mabili ty .Experiments made with mixtures of methane and air show that the percentage of methane corresponding with the lower limit,GENERAL AND PHYSICAL CHEMISTRY. ii. 71 cleweases f r m i 6-00'$) of iuetliane when the initial teluperaturtl of the mixture is 20° to 3.25% for an initial temperature of 700'. In the case of the higher limit mixture the percentage of methane increases slowly with the initial temperature of the mixture up to about 600° after which there is a considerable augmentation of the rate of increase of the methane content. I t is suggested that this is probably due to the disturbing influence of surface com- bustion of the methane during the interval which elapses between the introduction of the gas mixture into the heated explosion vessel and the passing of the igniting spark.I n general the results obtained for the influence of the initial temperature on the limits of inflammability agree closely with those obtained by Taffanel (Contpt. rend. 1915 157 593). Experiments made to determine the influence of pressure on the composition of the limit mixtures show that the percentage of methane increases with pressure for both the lower and upper mixtures. This result is in agreement with previous observations made by Terres and Plenz ( J . Gasbeleucht. 1914 57 990 1001 1016 1025). The smaller effect obtained by these authors in the case of the upper limit mixtures is presumed to be due to the circumstance that they did not make use of a sufficiently powerful source of ignition in order to obtain strictly comparable results.The lowestl pressure at which self-propagation of flame occurs in mixtures of methane and air was found to be 120 nim. when the initial temperature was atmospheric. In similar experiments Burrell and Robertson (U.S. Bureau of Mines Technical Paper No. 121 1916) obtained a limiting pressure of 300 mm. The differ- ence between the two results is explicable on the asumption that the igniting source employed by these authors was not sufficiently powerful. I n these circumstances their results give merely the limiting pressure for ignition by a spark discharge of particular intensity. The mixtures which correspond with the lowest limiting pres- sures contain between 8.75 and 9.40% of methane these numbers being derived from observations in which the initial temperatures were 20° 250° and 500O.11. J. P. TREUB (Proc. K. Akud. TVetensch. Amsterdam 1918 20 343-357. Compare A. 1917 ii 528).-From a mathematical discussion of the ideal case in which the saponification of a triglyceride takes place in solution the ester groups being equivalent and no complications arising it is shown that the relative concentration of free glycerol a t any moment is equal to the third power of the relative concentration of the free fatty acid in the case of stagewise saponification that is passage through the di- and mono-glycerides. The same holds good for the ideal case of esterification. The experimental results obtained when trilaurin is saponified with strong sulphuric acid show only very slight deviations from this rule the deviations being due to the fact that the three ester groups are not perfectly equi- valent.Rather greater deviations are found in the esterification of H. M. D. The Saponification of Fats.ii. 7 2 ABSTRACTS OF CHBMTCAL PAPERS. 1au t*ic acid slit1 glycerol by strong sulphuric wi(1. [See also J . Dynamics of Nitrile Formation from Acid Anhydrides and b i d e s . I. Investigation of the Reaction O(COPh),+ COPh.NH -+ 2Ph*CO,H+PhCN by Methods based on the Phase Rule. ROBERT KREMANN and MAX WENZING (i%?onats?L. 1917 38 445-456).-In a series of experiments designed for the investigation of the binary mixture benzoic anhydride-benzamide the authors found that the temperature of the primary crystal- lisation varied with the time during which the mixture had been heated and also with the temperature. This they were able to show was due to the reaction O(COPh),+COPh*NH,+ 2Ph*C02H + PhCN.To follow this reaction mixt’ures of the four substances taking part were made in a number of different propor- tions commencing with 100% of the substances on the right hand side of the equation and ending with 100% of the substances on the left hand side but always so that the molecular ratio of the benzoic acid to the benzamide was 2 l and that of the benzoic anhydride to the benzamide was 1:1. These mixtures were rapidly heated to 98O and then allowed to cool and the tempera- ture of primary crystallisstion noted. The crystallisation tem- peratures were plotted against the composition and an analytical curve produced which on the assumption that no chemical change had occurred gives the composition of any mixture in which the proper ratios of the two sides of the equation are maintained directly from the temperature of its primary cryst’allisation.The reaction was then studied quantities of benzoic anhydride and benzamide in molecular proportions were mixed in a closed vessel heated to 9 8 O or 123O and the temperature of primary crystallisa- tion measured at stated intervals of time (one to one hundred hours) and from the analytical curve the progress of the reaction ascertained. It is shown that the reaction is bimolecular and at 9So has a value 1;=0*053 whilst a t 123O k=0*24. The tempera- ture-coefficient is therefore 1.8 for loo.It is shown also that the reverse reaction does not take place to the extent of more than 1%. J. F. S. Influence of Carbon Monoxide on the Velocity of Catalytic Hydrogenation. EDWARD BRADFORD MAXTED (Trans. Farada;ll ,Sot. 1917 13 36-42).-Measurements have been made of the rate of absorption of hydrogen by olive oil a t 180° in presence of small quantities of carbon monoxide. The absorption vessel con- taining the oil and a nickel catalyst was connected to the gas- measuring tube by rubber tubing and was adjusted so that the contents could be continuously and thoroughly shaken during the progress of the absorption. Comparative experiments with pure hydrogen and with hydrogen containing from 0.25 t o 2% of carbon monoxide show that the rate of hydrogenation is very considerably reduced by these quan- tities of carbon monoxide.The curve obtained by plotting the hydrogen absorption for a given interval of time against the per- ,\‘or. ( ‘ h ~ n ? . 7nd. 191 8 March.] w. G.GENERAL AND PHYSICAL CHEMISTRY. ii. 73 centage of carbon monoxide in the hydrogen is convex towards the origin indicating that the retarding influence of successive increments in th0 carbon monoxide content decreases with increase in the quantity of the poisonous gas. Apart from the poisoning effect of the carbon monoxide the admixture of this gas dilutes the hydrogen but this effect can be readily calculated and allowed for. H. M. D. Hydrogenation under the Influence of Colloidal Catalysts and how to account for this Process.J. BOESEKEN and H. W HOFSTEDE (Proc. K . Akad. Wetensch. Amsterdam 1918 20 424-434) .-The question of the mechanism of colloidal metal catalysts in hydrogenation processes is discussed and the results of observations are recorded on the rate a t which hydrogen is ab- sorbed by solutions of cinnamic acid cinnamic esters and ethyl undecenoate in presence of colloidal palladium. These results are not sufficiently regular t o admit of mathematical treatment and the authors infer that the normal course of the reaction is dis- turbed by impurities in the hydrogen by coagulation of the cata- lyst and by other unknown factors. [See J . SOC. C'hem. Znd. 1918 March.] H. M. D. The Fundamental Values of the Quantities 71 and & for different Elements in Connexion with the Periodic System.V. The Elements of the Carbon and Titanium Groups. J. J. VAN LAAR (Proc. K . Akad. Wetensch. Amsterdam 1918 20 492-504. Compare A. 1916 ii 386 610; 1917 ii 67).-The methods previously described have been applied to the calculation of the values of b and . \ l a f o r the elements of the carbon and titanium groups. The critical data for these elements are of course not known but the values of 6 can be obtained from the compounds for which in certain cases the requisite data are avail- able. The chief result to which the author's calculations lead is that the value of J a must be very large ranging from 0.32 for carbon to 0.40 for lead. This is supposed to indicate that the attractive forces measured by 4% are those of the free atonis.The estimated values of .\ln and b and also of the critical tmi- perature and pressure are recorded in tables. H. M. D. The Fundamental Values of the Quantities b and Jcu for different Elements in Connexion with the Periodic System. VI. The Alkali Metals. J. J. VAN LAAR (Proc. K. A k d . Il'etetzsch. A msterdam. 1918 20 505-519. Compare preceding abstract).-An attempt is made to estimate the values of 7 and \/. for the alkali metals. the approximate value of the critical 1 emyerature required in the calculations being derived froin the irielting points and boiling points. The physical properties of tlhc alkali metals necessitate the assumption of high valnes for tIho tritraction constant and this is assumed to be connected with th(h existence of these elements in the atomic condition.The estiiriatcdii. 74 ABSTRACTS OF CHEMICAL PAPERS. values of the critical data and of b and J a are recorded in tabular form . H. M. D. Considerations on the Nature of Chemical Minity and of the Valency of Atoms G. CIAMICIAN and M. YADOA (Atti R. nccad. Lincei 1917 [v] 26 ii 165-173).-The bearing of the results of recent work on the structure of the atom and on valency is discussed. T. H. P. Vacuum Balance Cases. BERntAiv BLOUNT and WILLIAM H. WOODCOCK (T. 1918 113 81-84).-Attempts have been made to construct a vacuum balance case of gun-metal. On account of the porosity of the metal it was not found possible to reduce the rate of leakage much below that represented by 0.01 mm. per hour. Better results were obtained with a glass case consisting of a large bell-jar with a side tubulus through which the rider is con- trolled by an arrangement which is commonly made use of in the ordinary glass hypodermic syringe. A balance case constructed on these lines has been found t o give satisfactory results. By means of a Gaede pump it can be exhausted to 0*001 mm. and the vacuum can be maintained for more than an hour. H. M. D. Met.hod for Preventing Salts from Creeping over the Sides of Evaporating Dishes. W. 0. ROBINSOW ( J . Amer. Chem. SOC. 1918 40 197).-Creeping of salts can be prevented by paint- ing a strip about 6-7 mm. wide round the inner rim of the dish with collodion. The film cont,ains no non-volatile residue and can be easily burnt off. Lecture Experiment on the Vapour Pressure of Solutions. H. S. VAN KLOOSTER ( J . Amer. Chesn. SOC. 1918 40 193-195). -A simple apparatus is described for demonstrating the lowering of the vapour pressure of a volatile liquid on the addition of a foreign non-volatile substancs. It consists of an outer glass tube in which the pure liquid is boiled and an inner tube containing the solution. The inner tube is constricted somewhat about the middle of i€s length and at' its lower end is sealed on to a narrow tube which is bent round t o form a U with the wider tube and is used ,as a gauge tube. The constriction serves t o close the inner tube when all the air has been removed and this is conveniently effected by a rubber cork attached t o a glass rod. When the outer tube is closed and the liquid boils vapour passes through the solution vi2 the gauge tube and when the air has been displaced the inner tube is closed by the rubber stopper. The level of the solution in the gauge tube is then found t o be less than that in the wider tube. H. M. D. H. M. D. Preparation of Argon as a Lecture Experiment. W. P. .TORISSEW (Chem. TVeekbZad 1917 14 1151-1153).-A descrip- toin of an apparatus for demonstrating the extraction of argon.. INORGANIC CHEMISTRY. 11. 75 from air the oxygen being absorbed by phosphorus and the nitro- gen by a mixture of magnesium-powder fresh quick-lime and sodium. A. J. W.