fi. 69 General and Physical Chemistry. The Constitution of the Atom and the Properties of Band Spectra. H. DESLANDRES (Compt. rend. 1919 169 1365-1371. Compare A 1919 ii 206 310 441 486).-Further evidence is given in support of the views previously published (Zoc. c i t . ) . The author emphasises the fact that the formula given (A 1919 ii 206 310) is only a first approximation and offers some explanation of the divergencies found in certain cases. W. G. Intensity Relations in the Spectrum of Helium. T. R. MERTON and J. W. NICHOLSON ( P ? d . Trans. 1919 A 220 137-173).-The spectrum of helium has been examined by photo- graphing the radiation in front of an aluminium cathode in a tube containing helium of such a pressure that the thickness of the dark space was about 1 mm.A number of experiments were made with mixtures of helium and other gases particularly hydrogen. The most striking phenomenon observed relates to the difference in behaviour between the series of helium and parhelium for in the former lines belonging to a series maintain a practically constant intensity ratio a t every point whilst in the latter the relative intensity of any two lines of the same series variea with the distance from the cathode. I n the case of lines belonging to the principal series the seat of maximum emission is closer to the cathode and falls away with increasing distance from this point more rapidly than in the case of lines belonging to associated series. The diffuse series appear to preserve the most uniform intensity over a wide range of conditions. J.I?. S. F. PASCHEN (Ann. Physik 1919 Pv] 60 405-453).-An account is given of a very complete examma- tion of the spectrum of neon from A 9840.42 to A 2550’55. A long table is given of the wavelengths intensities and the spectrum combination of each line. The regularities between the lines of the principal and subsidiary series are investigated and discussed. Spectrum of Neon. J. F. S. Emission Spectra and the Chemical Reactions taking place in thesource of Radiation. ANGEL DEL CAYPO (Anal. Fis. Quim. 1919 17 247-27O).-Attention is directed to slight variations in the spectrum of the same element as described by different observers. The author regards these as being due to chemical reactions notably oxidation and reduction occurring within the radiant arc.These reactions may be localised according to circumstances in the immediate neighboyrhood of the electrodes or in the middle of the arc itself thus giving rise to a pol^ effect” in the photograph. This is illustrated in the w e of the VOL. OXVIII. ii. 3ii. 70 ABSTRACTS OF CHEMICAL PAPERS. silioa spectrum which shows variations it9 different sections of the arc between the carbon electrodes are examined. The view is expressed that the variations are due to the presence or absence of a reduction product the suboxide SiO. A similar phenomenon is observed with magnesium and is attributed to the suboxide Mg,O. I n studying the spectra of various specimens of aragonite three new lines 2554.6 2563.8 and 2565.0 A.U. were observed. These were given by all samples of naturally ocaurring calcium carbonate examined but not by pure hydroxide.In both cases the oxide formed is reduced to metal in contact with the carbon. This is immediately re-oxidised in the arc but in the former case the pro- tective action of the carbon dioxide liberated a t the same time ensures the persistence of the metallic vapour for a longer time and range so that the radiation from the metal is rendered perceptible. Further experiments confirmed the view that the lines in question belong to the calcium spectrum and are not due to impurities in the substances examined. W. S. M. The Emission of Positive Luminous Particles at High Temperatures by the Alkali Metals. G. A. HEMSALECH (Compt. rend. 1920 170 44-47).-Using the graphite plate with a layer of carborundum the plate being electrically heated as previously described (compare this vol.ii 1 2) it is shown that if the plate is first covered with a thin layer of the chloride carbon- ate or oxide of one of the alkali metals and this in turn covered with carborundum positively electrified particles are emitted a t varying temperatures. All the elements in the alkali group emit positive particles and for a given temperature the exten- sion and development of the luminous trajectories vary directly with the atomic weights and in consequence the critical tempera- ture or temperature a t which the phenomenon is apparent varies inversely with the atomic weight. For lithium the critical tempera- ture is 2700° and for caesium 1900O. Two hypotheses are put forward to explain this phenomenon but further work is necessary before a definite explanation can be given.[The Emission of Positive Luminous Particles at High Temperaturea by the Alkali Metals.] A. DE GRAMONT (Compt. rend. 1920 170 47).-The author states that Hemsalech used a monochromatic optical pyrometer for the temperature measure- men& in his work (preceding abstract). The instrument was Cali- brated and corrected up to 2500° above which temperature extra- polation was necessary. A t 2700O the temperature readings are accurate to *50°. W. G. The Direct Arc Spectra of Metals with Moderately High Melting Points A. DE GRAMONT (Compt. rend. 1920,170,31-%).- A comparison of the direct arc spectra with the spark spectra and the carbon arc spectra in the case of the metals zinc cadmium tin lead antimony bismuth magnesium and aluminium.By the photographic device employed the three spectra were obtained in W. G.GENERdL BND PHYSICAL CHEMISTRY. ii. 71 successive coincidence thus permitting of a direct comparison of the rays. By this means it has been possible to detect in the direct ray spectra on the one hand certain rays of the spark spectra considered by Lockyer as strengthened and which occur in the direct spectra with a marked intensity and on the other certain rays of the spark spectra also classed as strengthened but which are only fugitive in the direct arc spectra. The latter group of rays appears at the positive pole at the moment of striking the arc. They can most easily be seen by alternately making and breaking the arc.The detailed results are given for each of the metals and the general conclusions to be drawn are that the arc spectra must not be considered as invariable and always identical. They are subject to considerable variations not only in the intensi- ties of certain rays but also in the appearance of the rays these variations being closely connected with the intensity of the current producing the arc. W. G. New L i e s in the Arc Spectrum of Silver between X 4500 and X 2300. M. A. CATAL~N (Anal. Fis. QuinL. 1917 15 483486).-Measurements of eighty new lines were made with wave-lengths corresponding with similar lines in the spark spectrum described by Exner and Haschek (“Tabellen der Funkenlinien,” 1902). W. S. M. New Lines in the Arc Spectrum in Air of Iron between 2300 and 1980 A.U.S. PIBA DE RUBIES (Aizal. Fis. Quim. 1917 15 434444).-Measurements of more than 288 new lines in the arc spectrum of iron are given extending the spectrum from 2300 A.U. to 1994 A.U. W. S. M. New Lines in the Arc Spectra of Nickel and Cobalt between 2300 and 2000.&.U. S. PIRA DE RUBIES ( d i d Fis. Quim. 1918 16 338-350).-A total of 66 new lines f o r nickel and 165 for cobalt in the given range was measured. CARL RENZ (Helw. Chim. Acta 1919 2 704-717).-The action of light on thallous chloride has been investigated; the action was carried out with the dry substance and with thallous chloride under water hydro- chloric acid ammonia ethylamine ethyl alcohol glycerol toluene xylene and pyridine. I n all cases a darkening occurs which passes through the shades greyish-brown dark greyish-brown and blackish-brown.The change is due to the formation of a photo- thallous chloride and thallous-thallic chloride. Hydrochloric acid even in trams retards the formation of the photo-chloride and after prolonged illumination various yellow double thallous-thallic chlorides are formed. Nitric acid and sulphuric acid have a similar retarding action on the formation of the photo-chloride. Ammonia ethylamine and the abov*named organic substances act as semi. tisers in the reaction and many inorganic salts have a photo W. S. M. Photochemistry of Thallous Chloride. 3-2ii. 72 catalytic action. The sensitiveness of thallous chloride to light depends on its method of formation. Thallous chloride precipitated by metal chlorides is more sensitive than that precipitated with hydrochloric acid even though the acid has been thoroughly washed out. J.F. S. ABSTRACTS OF CHEMICAL PAPERS. Diffusion of Radium Emanation in Water. EVA RAMSTEDT (Medd. K . Vetenskapsakad. iVobel-lnst. 1919 5 No. 5 1-14).- The velocity of diffusion of radium emanation in water a t 14O has been determined by means of an apparatus containing a number of metal plates immersed in water. The coefficient of diffusion is found to be 0.820 cm. per day a value which is of the same dimensions (0,985 a t 18O) as that recently found by Rona (A. 1917 ii 286). Calculating from this value the product D& is found to be 12.2 and the molecular diameter of radium emanation 1'85 x 10-8 cm. J. F. S. Ionisation Potential of Helium.J. FRANCK and P. XNIPPING (Physikal. Zeitsch. 1919 20 481-488) .-Using as source of electrons a glowing wire in helium of pressure 1.5 mm. the reson- ance potential is found to have the value 20.5-10.25 volts and from this value the ionisation potential is calculated to 25.3k0.25 volts. Experimentally the ionisation potential of helium is found to be 25.41t0.25 volts and the resonance potential has the same value as before. The potential necessary for the removal of both electrons from the helium atom is also determined and the value 79.5rt0.3 volts obtained. The present results are in agreement with those of Horton and Davies (A 1919 ii 210) and indicate thaf. the helium atom is much less stable than is indicated by the atomic model of Bohr and LandQ.J. F. S. Conductivity. VI. Behaviour of Mixtures of Two Salts containing a Common Ion in Anhydrous Formic Acid Solution. H. I. SCHLESINGXR and F. H. REED (J. Amer. Chem. SOC. 1919 41 1921-1934. Compare A 1919 ii 91).-The con- ductivity of potassium f ormate in anhydrous formic acid solution has been determined a t ' 2 5 O over the range 0.1207N-0*3266hT. The degree of ionisation and the ionisation constant are caloulated in each case and the latter constant has the value 1.090. Similar measurements were also made f o r sodium formate and lithium formate; these salts have ionisation constants 0'810 and 0.557 respectively. Conductivity measurements were also made for mixed solutions of sodium and potassium formates lithium and potassium formates strontium and potassium formates and strontium and calcium formates all in anhydrous formic acid and a t 25O.A method of calculating from the ionisation constants the degree of ionisation of each of two salts containing a common ion when the two salts are both present in solution has been developed for the case in which both salts obey the law of mass action. It is found that in mixtures of sodiunl and potassium formates asGENERAL AND FHYSICAL CHEMISTRY. ii. 73 well as of lithium and potassium formates the mass law is obeyed by both of the highly ionised salts present. On the other hand solutions of mixtures containing as one or as both of the salts alkaline eart,h formates do not conform to the law although these uni-bivalent salts when in solution alone seem to follow the law over a certain range of concentration.This appears to make it quite certain that whenever the agreement of the salt with the law is an accidental one solutions of mixtures containing such a salt will not obey the law. Hence it may be definitely concluded that the agreement between the behaviour of the alkali met,al formates in anhydrous formic acid solution and the demands of the mass law is a real and not an accidental one. When the total concentration of mixed solutions becomes relatively great deviation from the mass law occurs also i n the solutions which contain uni- univalent formates. It has been found that this deviation begins when the concentration of the nan-ionised molecules of one of the salts reaches the same value as that a t which deviation begins in the solution of that salt by itself.This seems to be independent of the concentration of the other molecular species present and to indicate that in concentrated solutions it is the non-ionised mole- cules and not the ions which cause deviation from the mass law or a t least the non-ionised molecules cause deviation a t lower concentrations than do the ions. J. F. 5. Conductivity. VII. Transference Numbers of the Formates of Sodium Potassium and Calcium in Anhy- drous Formic Acid. H. I. SCHLESINGER and E. N. RUNTINQ ( J . Amer. Chem. SOC. 1919 41 1934-1945. Compare preceding abstract) .-The transport numbers of solutions of sodium potassium and calcium formate in anhydrous formic acid have been determined by the usual method a t 2 5 O for a number of con- centrations between 0.09N and 0.4N.The ionic conductivity of the formate ion is calculated t o be 51.5 that of sodium ion 14.6 and potassium 17.5. The transport number of the potassium ion changes with concentration. This is tentatively explained on the assumption that the ion is solvated and a method for calculating the extent of solvation from the transport numbers is suggested. The results indicate that if the assumptions made in the calcula- tion are correct each potassium ion is combined with from six to eight molecules of formic acid. The transport numbers obtained for calcium formate do not substantiate1 the view that calcium formate in formic acid solution ionises either entirely as a uni- univalent salt or ent$irely as a uni-bivalent salt.It is therefore possible that both methods of ionisation occur. Electric Conductivity of Weakly Ionised Neutral Salts. OLOF SVANBERG (Medd. K . Ventenslzapsakad. Nobel-Tnst. 1918 3 No. 26 1-7).-The equivalent conductivity of potassium antimony1 tartrate and cupric acetate has been determined over the range N / 4 to N/1024 a t temperatures 0-65O. I n the case of potassium anti- rronyl tartrate the equivalent conductivity increases fram 64.4 at J. F. S.ii. 74 ABSTRACTS OF CREMICAL PAPERS. d = 4 to 134.5 a t d=1024 a t 25O; the more concentrated solutions (up to N/32) a t all temperatures follow the Ostwald dilution law but beyond this point the conductivity increases more rapidly than is demanded by any of the formuls for the extrapolation of the infinity value.There is no evidence that potassium antimony1 tartrate ionises as the potassium salt of antimonyl tartaric acid and it is also shown that the salt is not greatly hydrolysed in any of the solutions examined. I n the case of cupric acetate the equivalent conductivity a t 25O increases from 10.8 in N-solutions to 75.3 in N / 1024 solutions which indicates considerable hydrolysis. The equivalent conductivity a t infinite dilution has been obtained by extrapolation and the following values found potassium antimonyl tartrate Oo A = 76 ; 25O A = 154 ; SO0 A = 260 ; 65O A = 308 copper acetate 2 5 O A =90. J. F. S. Relation between the Electrical Phenomenon in Cloud- like Condensed Odorous Water Vapours and Smell- intensity. H. ZWAARDEMAKER and H. ZEEHUISEN (Proc.K. Akad. Wetensch. Amsterdam 1919 22 175-178).-0dorous organic sub- stances in dilute aqueous solution when sprayed yield electric charges. It is found on diluting these solutions to such an extent that the electrical phenomena are only just appreciable that the odour is also just appreciable. It is suggested that both these properties depend in a complicated way on (u) the volatility of the substances and (a) the lowering of the surface tension of the solvent (compare A. 1918 ii 351). J. F. S. Dielectric Constants considered in Connexion with the Theory of Dipolar Molecules. OSKAR RLEIN (Medd. K. Vetenskapsakad. Nobel-lnst. 1918 3 No. 24 1-48).-The dielee- tric constants have been determined for a number of solutions in organic liquids by means of a resonance method.The solutions examined consisted of (1) methyl alcohol ethyl alcohol propyl alcohol butyl alcohol amyl alcohol ally1 alcohol benzyl alcohol acetone and methyl ethyl ketone in benzene; (2) ethyl alcohol butyl alcohol and benzyl alcohol in chloroform. It is shown that the resonance method which is the method used technically for measuring capacity is well suited to the measurement of dielectric constants. Solutions of the alcohols in benzene or chloroform of the same normality have the same dielectric constant. Thus N-solutions of alcohols in benzene have a dielectric constant 2.63 in chloroform solution 5-68. Acetone and methyl ethyl ketone have the same value in benzene solution namely 3-09. J. J. Thornson’s rule that related substances have the same value for ( e - 1)M holds absolutely for mixtures of the alcohols and water.Some measurements of Pohrt on the dielectric constants of gases are examined and yield the temperaturecoefficient which is demanded by the dipolar theory. The formula of Debye ( e - 1) / ( e + 2) . v = A / T + B is changed t o the form ( e - l)(w - b ) = A /T in which for substances which belong t o a single group the relationship A M = L obtains ; I is a constant pro-GENERAL AND PHYSICAL CHEMISTRY. ii. 75 portional to the square of the molecular moment and M is the molecular weight. These formulz combine the above-named rules and are in keeping with the dipolar theory. The constant L has been calculated from several of the experiments and the values for alcohol (67,000) and acetone (180,000) obtained.The formula ( e - 1) = A / T(v - b ) is applied to the dependence of the dielectric constant on temperature in the case of water and found to be in keeping with fads. The formula K=L/T=M(u- b ) ( e o - l ) + p / [ l + k(e0- l)] holds for infinitely dilute solutions and together with (C - 1) = A / (v - b)T' furnishes an exact method for calculating the constants IC and b. The approximate formula R = p / [ ( l + k(eo-l)] is deduced from the above and allows the determination of R from a knowledge of the dielectric constant of solutions of the substances concerned. J. F. S. Dielectric Constants of Typical Aliphatic and Aromatic Hydrocarbons cycloHexane cycloHexanone and cyclo- Hexanol. THEODORE W. RICHARDS and J. W. SHIPLEY (J. Amer. Clienz. SOC. 1919 41 2002-2012).-The dielectric constants of a ncmber of hydrocarbons have been determined at 20° except in the case of cyclohexanol which was determined a t 25O by a modified Nernst type of apparatus.The percentage error of the method wa5 much decreased by the use of larger suitably balanced condenser troughs. The following values have been obtained as the mean of several concordant measurements hexane 1.876 ; heptane 1.973 ; obtane 1'962 ; n-nonane 1.967 ; B-methyloctane 1.967 ; 8-methyl- octane 1.967 ; B<-dimethylheptane 1.987 ; flB-dimethylheptane 1.89; Be-dimethylheptane 1.89; decane 1.956; xylene (com- mercial) 2.375 ; m-xylene 2.377 ; ethylbenzene 2.482 ; n-propyl- benzene 2.364; cumene 2.400; mesitylene 2.356; tert.-butyl- benzene 2.384 ; cyclohexane 2.055 ; cyclohexanone 18.2 ; and cyclo- hexanol 15.0.The present results are in moderate agreement with the values of previous observers although in many cases no previous measurement is on record. Relation between the Specific Inductive Capacity of an Electrolyte and the Electric Potential of a Metal placed in it. D. L. ULREY (Physical Rev. 1919 12 47-58).-Attention i 4 called to the inadequacy of the Nernst theory of the mechanism of galvanic current production. This theory omits an essential factor namely specific inductive capacity of the electrolyte and employs two hypothetical quantities metal solution pressure and osmotic pressure of the ions. Transfer of ions between metal and electrolyte is probably brought about by electrical forces the magnitude of which depends on the specific inductive capacity of the medium.The potential difference between electrodes of the same kind ,in a two-solution cell was measured for several different percentage mixtures of two liquids for four different cases and in each case was shown to be strictly proportional to the difference in the specific inductive capacity of the two solutions. The following systems were measured (u) ethyl alcohol-water (solution of cupric J. F. S.ii. 76 ABSTRACTS OF CHEMICAL. PAPERS. chloride) with copper electrodes (electrode in the aqueous solution electronegative) ; ( h ) ethyl alcohol-water with calomel electrodes (electrode in the water electropositive) ; (c) acetone water with calomel electrodes (electrode in the water electropositive) ; ( d ) methyl alcohol-water with calomel elee trodes (electrode in the water electropositive) ; and ( e ) solution of carbamide-water with platinum electrodes (electrode in the water electropositive).I n two cases investigated with copper elec- trodes one with platinum electrodes and six with calomel electrodes the direction of the E.M.P. of the cell is in accordance with the theory that the loss of ions from an electrode ia dependent on the specific inductive capacity of the electrolyte rather than on the concentration of those ions in the electrolyte and a solution tension of the electrode. Results obtained further substantiate the theory that the more electropositive metals have the higher specific inductive capacities. Changes of Potential in an Oxidising Agent by Ultra- violet Light.TORSTEN SWENSSON (Arkiv. Kern. Mzn. Geol. 1917 6 No. 12 1-32).-The effect of ultra-violet light on the potential of solutions of potassium dichromate in sulphuric acid has been examined. The souroe of light employed was a quartz mercury lamp and the solutions were maintained at 1 8 O during illumina- tion. The platinum electrode used in the measurements was bright since it was found that the same value was obtained with both bright and platinised electrodes but the former reached equiLi- brium more rapidly. It is shown that during the illumination with ultra-violet light of a solution of potassium dichromate and sul- phurio acid and also during the illumination of either sulphuric acid potassium dichromate or chromic acid alone a change in potential occurs.To achieve this result it is not necessary to illu- minate the platinum electrode. In the case of dichromate and sulphurio acid a speedy increase in the potential occurs ; whilst when sulphuric acid and dichromate are illuminated separately a decrease in potential occurs. The potential is also decreased f o r pure chromic acid and consequently it is shown that the positive effect does not depend on the liberation of this acid on the addition of sulphuric acid to the dichromate. The speedy change in E.M.F. is caused by the ultra-violet rays and is not observed when the action takes place in glass vessels. Both the increase and decrease of E.M.F. are of a photachemical nature. Over the rangeexamined the potential change is independent of dilution.The change of potential is in a high degree dependent on the composition of the solution and shows a maximum a t about 75 mol. % potassium dichromate. The decrease in E’.Jf.F. after the interruption of the illumination takes place gradually. It is also dependent on the com- position and is most rapid for pure sulphuric acid. It is further greatly dependent on the temperature and after a moment’s boiling the potential becomes normal again. Solutions which have been illuminated but have returned to their original potential show on renewed illumination changes in the velocity of the potential CHEMICAL ABSTRACTS.GENERAL AND PHYSICAL CHEMISTRY. ii. 77 increase and the maximum value. On very bright illumination a rapid increase i n E.il1.E’. was observed which on interruption of the light sank very quickly. I n this case the illumination of the electrode is essential.J. F. S. Polarisation Tensions of Iron in Solutions of its Complex Salts. Relations between these Tensions and the Dis- simulation of Analytical Characters of Ferric Ions. N. R. DHAFL and G. URBAIN (Compt. rend. 1919 169 1395-1397).- The electrolytic cell consists of two half elements joined by a solu- tion of potassium chloride. One half element consists of a metallic electrode i n a solution of one of its salts and the other is an elec- trode of mercury in a solution of calomel and potassium chloride. The polarisation tension is defined as the difference between the E.M.F. of the whole cell and the corresponding value as found i n Auerbach’s results (compare A.1912 ii 123) for the half element containing the mercury electrode. This tension should depend on the number of free metallic ions in the solution the tension diminishing as the number of ions increases. These measurements should therefore in the case of complex salts give an indication as to how f a r the metallic constituent of the complex ion is dissimu- lated. The results obtained with simple and complex iron salts are in agreement with those obtained by Pascal (A 1909 ii 487) as to their molecular magnetic susceptibilities. Hydrogen Overvoltage. 11. Applications of its Variation with Pressure to Reduction Metal Solution and Deposition. D. A. MACINNES and A. W. CONTIERI ( J . Amer. Chem. SOC. 1919 41 2013-2019. Compare A. 1919 ii 131).-The increase of hydrogen over-voltage with diminished pressure is shown t o follow i n a nearly quantitative manner from the theory of MacInnes and Adler (Zoc.cit.). The effect of a change in the gaseous pressure on several chemical processes involving the evolution of hydrogen has been studied. The changes i n rates of reaction and in reaction efficiencies were found in each case to be in the directions which follow from the change of hydrogen over-voltage with pressure; t h a t is a decrease of gaseous pressure produces (a) a decrease in the rate of solution of metals i n electrolytes ( b ) an increase i n the efficiency of reductions by metals and (c) an increased efficiency of metal deposition. The theory explaining the fluctuation of over- voltage accompanying the evolution of a single bubble of hydrogen from a platinum electrode is discussed. Determination of the Hydrogen Exponent.J. PINKHOF (Chem. Weekblad 1919 16 1168-1172).-1f in Poggendorff’s method the normal electrode is replaced by an electrode which does not differ in potential from t h a t of the hydrogen electrode then if the composition of the liquid is known and it.s relation t o the potential the potential of the hydrogen electrode and there- fore the hydrogen exponent can be determined. The results obtained with a silver electrode in a solution of silver cyanide in W. G. J. F. S. 3*ii. 78 ABSTRACTS OF UHEMICAL PAPERS. excess of potassium cyanide were not accurate. An electrode of cadmium amalgam in solutions of cadmium salt of various concen- trations was found to be suitable.A simple apparatus for the determination is described. W. J. W. Applicability of the Gas Law6 to the Strong Electrolytes. J. N. BRONSTED (Medd. R. Vetenskapsakd. Nohel-Inst. 1919 5 No. 25 1 4 9 ) . - A npmber of solubility and E.M.F. measurements have been carried out with the object of testing the hypothesis. "The gas laws hold for ions or salts when other salt solutions are employed as solvents the concentration of the latter being large when compared with the concentration of the dissolved salt." The E.M.F. of cadmium [ cadmium sulphate in magnesium sulphate solution has been measured. ( m n a l g % 3 * 1 % ) ' ~ ~ i ~ ~ ~ ~ ~ i o l . ! iMgso,(2 CdS04 - clMol* cl) mol. 1 (amalgam 3.1%) The value of c and c1 varied between 0'1M and 1/64OM and the temperature was 20° and 39.40.The potentials obtained are compared with those demanded by the gas laws and an excellent agreement found. The solubility of dinitrotetra-amminecobalt nitrate both a and B varieties has been determined at Oo and 20° in water and in solutions of various concentrations of potassium formate thiocyanate hydroxide and nitrate nitric acid scdium oxalate and sodium nitrate. The values obtained are in keeping with the demands of the gas laws. The Electroaffinity of Aluminium. I. The Ionisation and Hydrolysis of Aluminium Chloride. 11. The Aluminium Electrode JAROSLAV HEYROVSK~ (T. 1920 117 Occlusion of Hydrogen and Oxygen by Metal Electrodes. EDGAR NBWBERY (1. Amer. Chem. Sac. 1919 41 1887-1892 1895-1898); EARLE A. HARDING and DONALD P.SMITH (ibid. 1892-1894 1897-1898) .-Polemical. In the first paper Newbery gives a theoretical discussion of the paper put forward by Harding and Smith (A. 1918 ii 424) and a criticism of the theory put forward as to the condition of the occluded hydrogen in palladium. The following papers cont.ain replies and counter-replies by Harding and Smith and Newbery. J. F. S. Electro- and Thermo-chemical Investigation of the ,Cells Cu or Cu Amalgam I CuS0,-Hg,S04 I Hg. L. W. OHOLM (Medd. K . Vetemskapsakad. Nobel-Inst. 1919 5 No. 4 1-20). -The E.M.F. of the elements Cul N-CuSO,,Hg,SO,(sat.) 1 Hg and Cu(12% amalgam) I N-CuS0,,Hg2S04(sat.) 1 Hg have been measured a t 100 15O 17O 200 250 and 30° daily for a period of two months. The copper amalgam cells are fairly constant and easily repro- ducible and the variation of the E.M.F.with temperature is represented by the formula The cells with a copper electrode had a slightly higher (0*004-0~005 The element is made up Cd J. F. S. 11-26 27-36). Et= 0'35030 - 0*00064(t - 20) - 0'0000025(t- 20)'.GENERAL AND PHYSICAL CHEMISTRY. ii. 79 volt) E.M.P. than the amalgam cells and it is nothing like so constant. The chemical energy of the copper cell is calculated and the value Q =24,860 cal. obtained. The quantity Q is regarded as made up of three quantities ql x2 and q3 q1 being the heat change when 1 gram- atom of copper is withdrawn from the amalgam qz the difference in the heat of formation of copper and mercury sulphate and q3 the heat change accompanying the withdrawal of water from the solu- tion by the newly formed copper sulphate.These three values are calculated to pi= - 1297 Gal. at 20° q,=7600 Cal. and q,=18.527 Cal . J. F. S. HORACE G. BYERS and CURTIS W. KING (1. A m e r . Chem. SOC. 1919 41 1902-1908).-When cobalt is used as anode in the electrolysis of 0.02N-sulphuric acid or sodium sulphate a t Oo it becomes passive if a high current density is employed but with low current density it remains active; if potassium dichromate is present it readily assumes the passive con- dition in all circumstances. The passive state is indicated by an increased drop in potential a decreased current by the evolution of oxygen and by the failure of the cobalt t o pass into solutiou. The potential measurements of a cobalt-platinum cell with various electrolytes and a comparison with similar results with iron and nickcl show that when the cobalt becomes passive t.here is a marked increase in voltage across the cell.The potential measurements show that this is due to a change in the potential of cobalt as it changes from the active to the passive condition. Cobalt may therefore be classified with the passive metals since it exhibits all the characteristics of iron and nickel when they are passive. The essential difference between cobalt and the other passive metals lies in the fact that cobalt when used as an anode will not become passive a t the low current densities required by nickel and iron. If cobalt has once assumed the passive condition it will remain so even though the current density is reduced.I n ths absence of the anodic relation cobalt becomes active in acid solution more readily than nickel. Activity Coefficient for Ions. NIELS BJERRUM (&?kid. K. Vetenskapsakad. Nobel-lnst. 1919 5 No. 16 1-21).-A theoretical paper in which activity coefficients are considered with the object of bringing strong electrolytes into line with the ionic hypothesis. The activity coefficient is defined as the effect of the interionic forces on the activity of the ions the conductivity coefficient as the influence of the interionic forces on the con- ductivity and the osmotic coefficient as the influence of the inter- ionic forces on the osmotic pressure. Various relationships between these and other similar coefficients are evolved mathematically. E =0*3542 and dEjdT = - 0.000'72 volt.Passivity of Cobalt. J. F. S. J. F. S. Comparative Electrolysis of Various Alkali Chlorides. E. BRINER (MLLE.) A. TYROCINER and B. ALFIMOFF (Helw. Chim. -4cta 1919 2 666-672).-Solutions of the chlorides of sodium 3'-2ii. 80 ABSTRACTS OF CHEMICAL PAPERS. lithium and potassium of various concentrations have been electro- lysed and the relative current yields of alkali hydroxide compared. After the passage of 50,000 coulombs through 3'lN-solutions of these salts it is shown that the total current yields are respectively for lithium sodium and potassium 7556 82*5% and 87%. I n the case of lithium chloride two cases are considered (u) Where the initial concentration of lithium chloride lies on the ascending branch of the conductivity cusve (concentrations below 5N) ; here the fraction of the current carried by the lithium chloride diminishes due to the presence of lithium hydroxide and to the reduction of the salt concentration.( b ) Where the initial con- centration lies on the descending branch of the conductivity curve (concentrated solutions) ; here the yield is diminished by the lithium hydroxide present but it is increased by the fact that the reduc- tion of the lithium chloride concentration brings with it an improve- ment of the conductivity of the salt The Transport Number of the Ions of Cadmium Iodide. GEORGES HEYM (Ann. Physique 1919 [ix] 12 443-454).-From a series of measurements i t is shown that the transport number of iodine ions varies from 0.55 to 1.0 for solutions of cadmium iodide in which the concentration varies from 0-007 to 0.07 gram of iodine per litre.W. G . Electrochemistry of Uranium and the Single Potentials of Some Oxides of Uranium. CHESTER A. PIERL~ (J. Physical (:hem. 1919 23 517-553).-The electrolysis of uranium com- pounds in various solvents and under many conditions has been investigated. It is shown that in aqueous solutions with low current density uranyl salts deposit in the first place hydrated uranic oxide U03,H,0 which is changed as the electrolysis pro- ceeds to a black oxide of varying composition. With higher current density uranyl sulphate is reduced to uranous sulphate but in the presence of free acid the deposit obtained is small in amount and poorly adherent although metallic in appearance. I n neutral or alkaline solutions the deposit formed is a mixture of black and yellow oxides.The use of a porous cup diaphragm does not change the character of the deposits. The deposit obtained when alkaline uranyl tartrate or citrate solutions art? electrdysed is an oxide much richer in uranium than that deposited from solutions acidified with t.artaric or citric acid. The con- ductivity of non-aqueous solutions of uranyl salts is a function of the water present and the deposite formed are oxides contaminated with organic matter Anhydrous pyridine dissolves anhydrous uranium tetrachloride to form conducting solutions ; these on electrolysis deposit a compound containing uranium and pyridine on the cathode.. Solutions of uranium tetrachloride in acetone do not yield metallic uranium on electrolysis; the solution is a good conductor of electricity. The deposit obtained replaces mercury from mercurous sulphate.Uranium tetrachloride reacts with anhydrous acetone forming PP-dichloropropane. During electro- J. F. s.GENERAL AND PHYSICAL CHEMISTRY. ii. 81 lysis hydrogen is evolved. Solutions of potassium uranyl fluoride whether acid alkaline or neutral form a deposit containing fluorine; in acid solution the deposit is UF4,6H,O; neutral and alkaline solutions give a deposit containing uranium tetrafluoride and uranium oxide. Deposits obtained from neutral solutions of potassium uranyl cyanide consist of pure potassium uranate. When acidified with hydrocyanic acid the deposit is the yellow hydrated oxide contaminated with a little of the black oxide.The single potentials of the metal and the oxides have been measured by pasting the finely powdered material on a platinum electrode with gelatin. The followin potentials of the more stable oxides have been obtained U,OjUO,(NO,) 14.3 grams per litre// = 0.776 volt; UO,,H,O I UO,(NO,),[I = -0.860; black oxide from aqueous uranyl salts 1 U02(N0,) I/ = - 0'6872 volt ; uranium 91.49% I UO2(NO3),I/ = - 0.093 volt. Uranous oxide UO giv? a single potential identical with that obtained for the green oxide U,O,. The black deposit formed when uranium salts are electro- lysed is not U,08,2H,0 as stated by Smith (Amer. Chem. J . 1879 1 329) but a compound U,0,,,2H20 and it has a different potential to that of U,O,. Johan Gadolin's Electrochemical Theory its Origin and Development.H. G. SODERBAUM (Medd. K. Vetenskapsakad. Nobel- I n s t . 1919 5 No. 9 l-l4).-Historical. Diamagnetic Phenomenon in Luminous Nitrogen and the Magnetic Behaviour of its Band Spectrum. W. STEUBINQ (Physikal. Zeitsch. 1919 20 512-519).-The intensity changes of the nitrogen bands in a magnetic field have been examined with the magnet in various positions. It is found that if a flat bulb 33 mm. in diameter is blown in the middle of the tube and the tube filled as for an ordinary spectrum observation the light passes across the bulb in the ordinary way except that the positive column is slightly broadened. When the field is made active the middle of the bulb is filled with a blue luminous sheath which extenas up to the walls and is at right angles to the lines of force of the field.This sheath has an identical spectrum with the negative luminescence and is probably to be accounted for by the diamagnetic molecules setting themselves a t right angles to the magnetic lines. Magnetic Properties of some Rare Earth Oxides at Low Temperatures. E. H. WILLIANS (Chem. News 1919 119 287-288) .-The magnetic susceptibility of the oxides of dyspros- ium erbium gadolinium samarium neodymium lanthanum and yttrium has been determined a t a series of temperatures from 2OC to - 140O. The following values are given Dysprosium oxide 20° 384'2; -120° 430.2; -140' 490.0. Erbium oxide 20° 188.6; - 140° 402.8. Gadolinium oxide 20° 129.7; Oo 138.2; -40° J. F. S. T. S. P. J. F. S. X x lo6 233.3'; O' 250'0; -40° 291'3'; -80° 347'4; -loo' O' 201.7; -40° 234.8; -goo 282'3; -looo 314'8; -120° 355'0; 160.7; -80° 194'3; -looo 217.0; -120' 244'6; - 140° 279'0.ii.82 ABSTRACTS OF CHEMICAL PAPERS. I n all cases the oxides were purer than 99.5%. I n the case of samarium oxide the susceptibility a t -140° was about 10% larger than at 20°. Yttrium oxide increases in susceptibility with decrease in temperature but very slightly and since the susceptibility is less than 1 x 10-6 the experimental error is relatively large. It is shown that the product of the susceptibility and the absolute temperature is not constant but decreases with falling temperature to a slight extent. J. F. S. The Paramagnetism of Solid Salts and the Theory of the Magneton. B. CABRERA (Anal. Fis. Quim. 1918 16 436-449).-A mathematical discussion of the Curie-Langevin law when the mutual actions of the paramagnetic atoms are no longer negligible.W. S. M. WALTER P. WHITE and LEASON H. ADAMS (Physical R e g . 1919 14 .4448).-By making the heating coil of an electric furnace one arm of a Wheatstone bridge and combining this wit,h a galvanometer regulator thus keep- ing the resistance of the coil constant the temperature of electric furnaces may be kept constant. This device is effective regardless of variations in the current supply and requires no attention particu- larly in the case of furnaces which are not directly influenced by the temperature of the room or where the surrounding air is kept constant. The arrangement operates as follows Changes in the temperature of the furnace and consequently in the resistance of the heating coil operate a boom which either hits or misses a contact-maker which controls a suitable relay.This relay operates a larger magnet which controls the main current. The power available in this regulator is verv large; nothing has to be inserted into the furnace cavity and the lag is practically non-existent. The regulator is often almost a t its best under conditions most unfavourable t o other regulators. Using this regulator the authors have kept a small furnace constant to within 0’lc for hours a t temperatures from 500° to 1400O. An Analysis of the Radiation Emitted in Gaseous Explosions. W. T. DAVID (Phil. Mag. 1920 vi 39 84-95)- I n explosions of coal-gas and air and of hydrogen and air the ratio of the energy in the 2.8 p radiation emitted to that in the radiation of longer wavelength decreases as the temperature decreases. In the neighbourhood of 1200O abs.the 2.8 p radiation decreases very rapidly with the temperature and is negligible at 1000° abs. Radia- tion of longer wavelength is emitted after the temperature has fallen to 900° abs. but a t this temperature the emission is small. The ratio of the energies in the 2.8 p and the 4.4 p radiation in coal- gas and air mixtures appears to depend on the composition of the mixture and the temperature. The loss of heat by radiation expressed as a percentage of the heat of combustion of hydrogen and air mixtures between the limits of composition of 10 and 25.4% of combustible gas decreases very rapidly as the latter increases.Furnace Temperature Regulator. J. F. S.GENERAL AND PHYSICAL CHEMISTRY. ii. 83 The variation for mixtures of coal-gas and air between 9.8 and 15% is small. The author attempts a theoretical explanation of these results on the following lines. The intra-molecular energy acquired on combustion by the freshly-formed molecules of carbon dioxide and water is not equally partitioned over the various internal degrees of freedom of the molecules. When the combustion is gentle the intra-molecular energy is concentrated in the rotational degrees of freedom and in such very low frequency vibrations as the mole- cules may be capable of executing. As the combustion becomes more violent the higher frequency vibrations share in this energy and it is possible that during combustion of extreme violenoe equi- partitioning of energy amongst all the internal degrees of freedom of the molecules may be approached momentarily.This is believed to be capable of explaining many phenomena of explosive combus- tion such as the pre-pressure period. Isothermals of Monatomic Substances and their Binary Mixtures. XX. Isothermals of Neon from 20' to - 217O. C. A. CROMMELIN J. PALACIOS MARTINEZ and H. KAMMERLINGH ONNEE (PTOC. R. Akad. Wetensch. Amsterdam 1919 22 108-118. Compare A. 1917 ii 407; 1918 ii 9).-The isothermals of neon have been determined over the range 20° to - 2 1 7 O up to a pressure of 90 atms. by the method previously described. From the results of the experiments the authors have calculated the virial-coefficients from the equation of state.It is shown that the pv values obtained in the present work agree in a very satisfactory manner with the older measurements. Only in the isothermal -217'52O is any marked divergence visible and here the divergence does not exceed 0'5%) whilst for the isothermal -200'08O they do not reach 0.1%. J. F. S. A New Improvement of the Equation of State of Fluids. E. ARIBS (Compt. rend. 1919 169 1140-1143).-A mathemati- cal disoussion of the subject. Equation of Condition. FRED. G . EDWARDS (Chem. News 1920 120 4-5).-Using Nernst's value in the equation for the mean molecular heat of gases Cv=a+ b t it is shown that these linear equations are chords to the curve mz= k log T where m= 5 . 9 5 1 ~ or 4 - (2~12.975) below or above y=2.975 from which is obtained the general equation m2= kllog T = 1 / v log T and hence pv = mZT mz being a constant for monatomic gases.Corrected van der Waals's Equation of Condition for the Quasi-diminution of the Molecule. E. A. HOLM (Medd. R. Vetenskapsakad. Nobel-Znst. 1919 5 No. 27 1-33).-1n the cor- rected van der Waals's equation an infinite series of b l v appears as the correction factor of the volume. This has been inserted by Boltzmann i n an approximated form as 1 - b / v + 0*375b2/va- 0'0369bSlvS. This cannot be true for very small volumes. The equation has been tested on Amagat's oxygen isotherms and it is J. R. P. W. G. W. G.ii. 84 ABSTRACTS OF CHEMICAL TAPERS. found to be strictly applicable at pressures 260 300 and 450 atm. and a t temperatures Oo 15.63O 99.50° and 1 9 9 ' 5 O .The negative divergence found for very high pressures is due entirely to the mathematical incompleteness of the formula J. F. S. Determination of Chemical Constants. ALFRED C. EGERTON (Phil. B a g . 1920 [vi] 39 1-20).-The author emphasises the fact t h a t the usual formula for the calculation of chemical con- stants from vapour pressures due to Nernst cannot give results having any theoretical significance because the assumptions as to the specific heats a t low temperatures which are contained i n it have been shown to be incorrect by recent experiments. H e there- fore modifies the formula by substituting for these assumptions the more exact expressions for the specific heats of solids a t low tem- peratures and assumes t h a t the atomic heats of gases remain con- stant and equal t o 5 1 2 .R a t the lowest temperatures. Thevaluesof the chemical constants calculated by the new formula agree within the limits of experimental error with those given by the formula - C',= C - 1.5 log M where C is a universal constant and M is the atomic weight. The values of C obtained from the vapour-pressure formula are given in the second column below; those from the above formula in the third column C from vap. press. Mercury ............ 1.820+_0.03:! 1.845 Cadmium ... 1.65 k0.31 1.468 Zinc . . . . . . . 1.23 f 0.26 1.115 The latent heats of vaporisation are calculated from an expression previously used (Phd. Mag. 1917 [vi] 33 193) and found to agree with the experimental values as well as those calculated by the equation X = T"8.5 log T due to Nernst.The value of Stefan's constant u is calculated and found t o be5.27. erg. cm.-2 deg.-4 which agrees with the value adopted by Planck but is distinctly lower than the more recently determined values. Significance of the Chemical Constant and its Relation to the Behaviour of Gases at Low Temperatures. F. A. LINDEMANN (Phil. ,&fag. 1920 [vi] 39 21-25).-It is shown t h a t the chemica.1 constant has the dimensions of the logarithm of a pressure if the atomic heat of a monatomic gas becomes zero at the absolute zero. I n this case it should be of the form X + 312 . log A + 5 / 2 . log 8 where 8 is a characteristic constant of the substance. If the atomic heat of a monatomic gas remains 5 / 2 . R down to the absolute zero the chemical constant has the dimensions of a pres- sure divided by a temperature to the power 512 and is of the form X + 3/2 .log A . Experimental determinations show that the latter form is true within the limits of error. It follows either that 6 is very nearly equal to lo for all substances which seems improbable or t h a t the atomic heat remains constant down t o the lowest tem- peratures. It is further shown that the chemical constant may be eliminated and the vapour pressure expressed i n terms of the pressure of full radiation. It is therefore suggested that the C from formula. The value of C is found to be -1.622. J. R. P.GENERAL AND PHYSICAL CHEMISTRY. ii. 85 chemical constant may express the interaction of matter and full radiation rather than requiring that a gas can assume only a finite number of microphases from the point of view of statistical mechanics a t a given temperature pressure and volume.If the value of the chemical constant could as suggested be derived from the radiation pressure the quantum assumption would be avoided iii the case of gases although it would be necessary for t.he deduction of the law of full radiation. Latent Heat and Surface Energy. 11. D. L. HAMMICK (Phil. Mag. 1920 [vil 39 32-46. Compare A. 1919 ii 389).- On the assumption that van der Waals’s constant a varies with the temperature a relation is derived between the surface energy p the molecular volume I’ and a namely a,=6pV/d where d is the molecular diameter. This is valid a t low temperatures only. By assuming that aT diminishes linearly with the temperature to the critical value a,.latent heats are calculated by Bakker’s formula A=aT(l/ul-l/u2) for several liquids.The results are in good agreement with experimental values. The ratio a T / a c is found to be the same for many liquids a t the boiling point and equal to 1.4. This leads to the value of Trouton’s constant. The empirical rela- tionship between latent heat and surface energy due to Walden and the Eotvos-Ramsay law can also be deduced. Heats of Fusion Velocities of Crystallisation and Chemical Affinities in Crystals. M. PADOA (Atti R. kccad. Lincei 1919 [v] 28 ii 239-243. Compare A. 1919 ii 51 96). -From the considerations previously developed the conclusion is drawn that the affinity acting in the process of crystallisation should be expressed by or a t least included in the heat of fusion of the compound so that under similar conditions as t o molecular magni- tude and structure those compounds exhibiting the greatest veloci- ties of crystallisation should also have the highest heats of fusion.The melting points heate of fusion and velocities of crystallisa- tion are tabulated for a number of pairs of isomorphous organic compounds such as naphthalene and dihydronaphthalene benzene and cyclohexane etc. I n each pair the greater heat of fusion corresponds with the greater velocity of crystallisation character- istic of the compound containing double linkings. Further un- saturated compounds in general melt a t higher temperatures than the corresponding saturated compounds this indicating the greater stability of the crystals of the former.As regards heats of fusion the values f o r different series of compounds cannot be compared the value for the saturated ethyl succinate for instance exceeding that for benzene; thus the heat of fusion may be the resultant of various thermal effects and not merely the effect of the affinity inherent to the union of the molecules in the crystal in the same way as the heat of solution represents the algebraic sum of the various heats of ionisation hydration etc. With a compound containing a triple linking such as tolane the velocity of crystallisation is less than that for the corresponding J. R. P. J. R . P.ii. 86 ABSTRAOTS OF CHEMICAL PAPERS. isomorphous compound with a double linking (stilbene) and approaches that for the corresponding compound with a single linking (dibenzyl).The parallelism between velocity of crystal- lisation and heat of fusion is maintained also in this case; the melt- ing points of compounds with triple linkings are likewise lower than those of the compounds with double linkings. The high velocities of crystallisation shown by compounds with double linkings are explainable on the assumption that change of form on crystallisation is unnecessary with these compounds which in the liquid state are wholly of the fumaric or wholly of the maleio type and with the help of the latent valencies crystallise with great rapidity; on the other hand the compounds with single or triple linkings undergo in part a t least a preliminary transposition which results in retardation of the crystallisation. As an instance of the influence of the molecular configuration it may be noted that the velocity of crystallisation of dibenzyl with which equil- ibrium between the cis- and trans-forms is possible in solution is 580 mm.per minute whereas that of dihydrophenanthrene which differs from dibenzyl only in the closure of the third ring and with which no equilibrium between isomeric forms is possible is 1200 mm. per minute. It may be however that the results obtained with compounds containing the three different linkings are to be explained by the assumption that the latent valencies of the triple linking exert no action in the formation of crystals. Thomsen’s thermochemical calculations (“Thermochemische Untersuchunpen.” 1906.310) i \ i I 3cm 4 show that the thermal value of :simple linking in . -3 .- the lower members of the paraffin seriea is 14.71 cal. whereas that of the double linking in the correspond- ing olefines is 13.27 cal. and that of the triple linking in acetylene etc. approaches zero. W c m . T. H. P. Apparatus for Determining the Melting Point of Very Hygroscopic Substances. H. J. BACKER (Chem. W e e k h l a d 1919 16 1564-1565).- The apparatus consists of a narrow thin-walled glass tube into which the material under examination is placed the tube being then drawn out and sealed. It communicates by means of a small curved latera? branch with a wider tube in which is placed about 2 C.C. of phosphoric oxide after which this tube is drawn out into a capillary exhausted by means of a water-pump and sealed The material is thus con- tained in a vacuum desiccator during the determin- ation of its melting point.w. J. w. Freezing-point Lowerings in Mixtures of Two Electro- lytes. K. G. DERNBY (Medd. K. Vetemkapsakad. Nobel-Inst. 1918 3 NO. 18 l-lO).-The depression of the freezing point has been determined for aqueous solutions of mixtures of two electrolytes. The following pairs were examined bydro.chloric acid and the,GENERAL AND PHYSICAL CHEMISTRY. ii. 87 chlorides of sodium potassium and magnesium respectively potassium and magnesium chloride nitric acid with potassium and magnesium nitrates respectively. The concentrations examined varied between ON and 0 * 8 N . It is shown that the freezing-point lowerings of mixtures of electrolytes with a common negative ion are always greater than the sum of the individual lowerings.For mixtures of binary electrolytes such as KCl,HCl NaCI,HCl KNO,,HNO the lowering of the freezing point is approximately proportional to the concentration of the salt and of the acid. For mixtures of a ternary electrolyte and a binary electrolyte such as RIgCl,,HCl MgCI,,KCl and Mg(NO&HNO the lowering of the freezing point increases faster than the concentration. Mixtures of magnesium chloride and hydrochloric acid lower the freezing point more than mixtures of potassium chloride and magnesium chloride of the same concentration. J. F. S. CalcuIation of the Neutral Salt Action from the Depres- sion of the Freezing Point of Aqueous Solutions. SVANTE ARRRENIUS and ERIK ANDERSSON (Medd.R. Vetenskapsakad. Nobel-lnst. 1918 3 No. 25 1-9).-A theoretical paper in which on the basis of the Arrhenius hypothesis of the mechanism of neutral salt action an attempt is made to calculate the neutral salt action from freezing-point depressions. The osmotic pressure of a solution is strongly influenced by other substances particularly salts; this is also true of the osmotic pressure of hydrogen ions. Since the velocity of reaction is proportional to the osmotic pressure of the surrounding substance (sucrose or ethyl acetate) and further the osmotic pressure of the catalyst (hydrogen ions) increases pro- portionally the neutral salt action can be calculated as scan as the corresponding changes of the two osmotic pressures are known.These for the case under investigation are known from the freez- ing-point measurements of Dernby (preceding abstract) and from the hydrogen-ion activity determinations of Harned (A. 1916 ii 8). The calculation shows that the method is one capable of furnishing the desired result Simplification of the Inverse-rate Method for Thermal Analysis. P. D. MERICA (B~11. Bureau Standards 1919 No. 336 101-104).-1n plotting and recording cooling curves the author recommends the following method using a thermocouple a direct- reading potentiometer and galvanometer and two stop-watches. The watches are mounted in a single frame and held in one hand ; the potentiometer is set a t the desired point. When the time- temperature readings are being made the stem of both watches is pressed a t the moment the galvanometer reaches the zero thus stopping one watch and so recording the time interval and start- ing the other watch on the next interval which is recorded in the same way.This method saves the time necessary in plotting and reading chronograph records and does away with the use of ex- pensive chronographs. Specimen curves f o r the cooling of iron made by this method are given in the paper. J. F. S. J. F. S.ii. 88 ABSTRACTS OF CHEMICAL PAPERS. Exact Formula for the Saturation Tension of Water Vapour between 0' and 50'. PAUL SCHREIBER (Physikal. Zeitsch. 1919 20 521-523).-The author gives the following formulze for calculating the saturation tension of water vapour a t temperatures between Oo and 50° log s= log s1 + log f ( T ) in which logsl=p+p logT and logf(T) is a correcting factor s is the corrected saturation tension.The values calculated by this equa- tion are practically identical with the values given in the inter- national meteorological tables. The values log sl= - 7.0814 + 17.8 log (10-22')) logf(T) =0*0115 -0*0000494(2"-298)2 and are used in the calculations. F. P. SOEBEL (Science 1919 50 49-50).-Assuming that at the b. p. the energy of vibration of the individual molecules of a liquid and of its vapour must be equal the author deduces the equation T,=mp,v,/1.49 for a pure liquid,.where m is the molecular weight and p and v the pressure and the volume of the vapour a t the absolute b. p. T,. For ordinary liquids containing impurities the equat.ion becomes T,= (p,v,-C)m/1~49 where C is a constant characteristic of each liquid.I n the case of water (m = 18 ; C = 8.7) the calculated values of T a t 273O 313O and 473O (abs.) are 270° 313*3O and 473.5O respectively. Similar agreement is found in the case of other liquids in which association does not occur. Vapour Pressure and Free Energies of the Hydrogen Haloids in Aqueous Solution. The Free Energy of Formation of Hydrogen Chloride. STUART J. BATES and H. DARWIN KIRSCHMAN (1. Amer. Chem. Soc. 1919 41 1991-2001). -Determinations have been made of the vapour pressures of hydrogen chloride hydrogen bromide and hydrogen iodide above their aqueous solutions between the concentrations 3.2 and l O . O M 5.8 and ll*ON and 6.0 and 9.7N respectively a t 25O and of hydrogen chloride between 5.5 and 9.2N a t 30°. By the method employed which consisted in determining the hydrogen haloid contained in a given quantity of air or nitrogen in equilibrium with its solution and comparing this with the amount of aqueous vapour which the same air contained when in equilibrium with pure water a t the same temperature partial pressures as small as 0.001 mm.were determined with an accuracy of a few per cent. The vapour- pressure measurements of hydrochloric acid solutions between 3'2N and 7.15N are in good agreement with the E.M.F. data for hydro- chloric acid concentration cells. The free energy of formation of hydrogen chloride a t 25O is -22,700 cal. The free energies of formation of the halogen acids a t various concentrations 0'1-1 1.0 mols.per litre in aqueous solution are given in a table in the paper. Conatant Temperature Still-head for Light Oil Frac- tionation. FREDERICK &I. WASBBURN (J. Id. Eng. Chem. 1920 12 73-77).-The apparatus described consists of two essential log S = - 7.0699 + 17.8 log (10-2T) - 0*0000494(T -298)' J. F. S. Boiling Point of Liquids. CHEMICAL ABSTRACTS. J. F. S.GENERAL AND PHYSICAL CHEMISTRY. ii. 8 parts namely a Hempel column and a constant temperature still- head of the type suggested by F. A. Brown (T. 1880 37 49). This still-head is a spiral of about 6 inches diameter made from 12 feet of $ inch iron tube; it is surrounded by an oil-bath which is provided with a stirrer and maintained at the required temperature by a coil of resistance wire thernio-regulator relay etc.The lower end of the spiral is connected with the side-tube of the Hempel column whilst the upper end is fitted with a thermometer pocket and a side-tube leading t o an ordinary condenser. Heat of Formation Calculated from the Wave-length of Absorption Bands. A. L. BERNOUILLI (Helw. Chim. Acta 1919 2 720-728).-The author has deduced an expression whereby from the absorption bands corresponding with characteristic ultra-violet electron vibrations the heat of formation of a compound from its elements may be calculated. This expression for a binary com- pound - - - has - Q = 0~01128[(~~@l/avl~ - / c % / ~ w ~ $ ) - d/O'/(a+@-6)v'$] i n which Q is the heat of formation q w2 and d the atomic volumes of the constituent elements and the molecular volume of the compound respectively; @ a and a' are the melt ing points i n absolute degrees of the elements and the compound respectively a and j3 are the electron numbers or valencies of the elements and 6 is the number of charges given up i n the formation of the molecule.This formula is tested in the case of a number of binary compounds and the results compared with the experimental values. The two sets of results are remarkablv close for example. W. P. S. the - form A carbon dioxide calc. 108.3 obs. 106.0 ; silverdchloride calc. 29.05 obs. 29.0. J. F. S. Heat of Reaction of Ammonia Oxidation. GUY B. TAYLOR (1. Znd. Eng. Chem. 1919 11 1121-1123).-The temperature most favourable to the catalytic oxidation of ammonia by means of air is 800O. In order to determine the amount of external energy required t o maintain the catalyst a t the optimum temperature a formula has been based on the equations (1) 4NH3+5O2=4NO+ 6H20+ 214200 cal.( 2 ) 4NH + 30 = 2N + 6H,O + 300600 cal. The temperature rise is expressed by the formula t = Q/C where Q represents the heat of reaction i n calories and C the specific heat of the products of the reaction. From Oo to 800° water-vapour is the only gaseous product with an appreciable temperature-coeffi- cient and this may be taken as 8.34. I n the following formula t0=(75150x- 21600y)/(7~~08+3*41a-O~25y+ 8.34VIB- V ) V re- presents the partial pressure of water-vapour B the total pressure of the air-ammonia-water vapour mixture 5 the molecular fraction of ammonia i n the mixture and y the molecular fraction of nitric oxide produced in the oxidation.For maintaining the catalyst a t the right temperature preheating the mixture electric heating or enriching the current of ammonia and air with oxygen have been shown to give equally satisfactory results. C. A. M.ii. 90 ABSTRACTS OR CHEMICAL PAPERS. Critical Densities of Hydrogen Helium and Neon. J. J. VAN LAAR (Chem. Weekblad 1919 16 1557-1564).-1n thecases oi hydrogen and helium the extrapolated values for Do obtained from the curve B(Dl+D2)=f(T) are found to be too high. Thus the figure 0*0310 so obtained for hydrogen would correspond to y =0.46 which is an impossible value. The author calculates that for hydro- gen D,=0.0287 and for helium D,=0*0598 instead of 0.066 as given by the curve. I n the case of neon there is agreement between the extrapolated value and the calculated figure U being 0.456.The ‘‘ Density Numbers ” of Groshans. W. P. JORISSEN (Chem. Weekblad 1917 14 1066-1071).-The volume in C.C. of 1 gram of a vapocised substaiicc a t tho boiling point is given by the expression 82T / M where llf is the molecular weight. Elimination of Tb,p. between this expression and the boiling point formula of Groshans Tbp = 27*8iMdi/n gives the vapour volume in C.C. of 1 gram 2280dxln. x is a constant for the class of substances con- sidered and n is the sum of the density numbers of the elements involved (Ann. Phys. Chcm. 1849 78 116). For the elements carbon hydrogen and oxygen the density numbers are unity and n is therefore the number of atoms in the molecule.For other elements n is greater than unity. For unassociated liquids the expansion in the transition from liquid to vapour is 570&. By comparing this formula with the expansion determined experimen- tally by Masson (A 1891 379) for methyl ethyl propyl phenyl butyl and amyl chlorides x is determined for each and the density number for chlorine is determined from the boiling-point formula. The values found range from 3’40 to 4.44. The value given by Groshans is 4. Determination of Avogadro’s Normal Volume V and of the Atomic Weights of Hydrogen Helium and Argon. J. J. VAN LAAR (Chena. Veelcblad 1919 16 1243-125O).-A con- tinuation of it previous communication (A. 1919 ii 461). The most probable value of VIA is considered to be 224153 C.C. The atomic weights of hydrogen helium and argon are stated to be 1.007697 (= 1’00770) 3.9998 (=4’000) and 39.95 respectively.W. J. W. W. S. M. W. J. W. Molecular Attraction. 111. The Characteristic Equation. K. K. JARVINEN (Ann. Acad. Sci. Fennicae 1919 [ A ] 12 Reprint 44 pp.).-A continuation of the theoretical discussion on molecular attraction (compare A. 1913 ii 293; A. 1915 ii 251). The internal prmsure pa is calculated by the equation p = R TI V . f -pa on the basis of the law of molecular attraction F=lem2/r5. It is found that pa=a/vi for monatomic substances and pa=a/ { v + ( v ~ - 0.242b*)5} for polyatomic substances where a and b are the cbn- stants of van der Waals’s equation. It is thus much smaller than that calculated from the latter equation. The critical pressure p h = R T / V .f is also calculated and thence a table of values of fGENERAL AND PHYSICAL CHEMISTRY. ii. 91 deduced which may be used to find the deviations from the ideal state. Values of f are also calculated from the equations f = p - p a / RT . v and f = d p / d t . v / R the latter obtained by differentiation of the characteristic equation. These agree approximately with the theoretical values. The differences are taken to mean that the theory is yet incomplete. It is also found that b decreases with temperature but this need iiot in reality be the case. The pressure the critical data and the values of vk/bo RT/pvk d p l d t . v / R and d p / d t . T / p may be obtained with fair approximation. Only approximate accuracy is claimed. Effects of Acids and Bases on the Surface Energy Relations of PP‘-Dichloroethyl Sulphide (Mustard Gas).WILLIAM D. HARKINS and D. T. EWING ( J . Amer. Chem. SOC. 1919 41 1977-1980).-The surface tension between &3’-dichloro- ethyl sulphide and a number of liquids has been determined with the object of finding a suitable emulsifying agent for the prepara- tion of aqueous emulsions of this substance. The following values of the surface tension in dynes per cm. have been found between &3’-dichloroethyl sulphide and the liquids named water 28.36 ; O.1N-hydrochloric acid 28.90 ; O’lN-sodium hydroxide 12-78 ; O’liV-sodium carbonate 18.82; 1% solution of turkey-red oil 14.47 ; 1% solution of turkey-red oil in O‘lN-sodium carbonate 8.35; ls’o Twitchell’s solution 12.32 ; 1% Twitchell’s solution in O’lN-sodium carbonate 12.89; 1% maize oil solution 12.94; and 1% maize oil in O‘IN-sodium carbonate 10.91.The surface tension of several organic liquids has also been measured aa-diphenylpropane 20° 37.15 25O 36.64; aa-diphenylethane 20° 37.67 25O 37.20; diphenylmethane 20° 37.56 ; ditolylmethane 20° 35.51 25O 34.80; propylbenzene ZOO 32.22 25O 31.30. All determinations were made by the drop-weight method. J. R. P. J. F. S. Viscosity of Pure Liquids. SVANTE ARRHENIUS (Xedd. K. Vetenskapsakad. Nobel-lnst. 1918 3 No. 20 140).-A theoreti- cal paper in which a large volume of work on the viscosity of liquids is discussed and correlated. It is shown that the ratio q b x 105 d F i s approximately constant for non-associated organic liquids. The symbols have the significance qb is the viscosity of a liquid a t the boiling point and s is the density and the constant varies between 230 and 320.The constant is not obtained in the case of a few liquids which are characterised by the fact that the expression Kl = P ( d log 7 id) / d t does nof show a minimum below the boiling point. These liquids are ethylene propylene and iso- butylene bromides benzene and carbon tetrachloride. The presence of negative atoms such as the halogens sulphur or oxygen increases the value of the ratio q b .\/; The iduence of pressure 011 the viscosity is nearly represented by a linear equation log qp= logqo+Z(p-po) where qo is the viscosity a t the external pressure zero (practically atmospheric pressure) p is the total pressure (external and internal) calculated according t o van der Waals’sii.92 ABSTRACTS OF CHEMICAL PAPERS. theory and p the internal pressure a t zcro external pressure. The influence of temperature on viscosity may for non-associated liquids be expressed by the formula dlog q . d / d t = R / T 2 . K is nearly proportional t o the absolute boiling point Tb so t h a t K Tb if ordinary logarithms are used for calculatiiig Kl does not change for normal organic liquids between greater limits than 1-1.2. Benzene and the liquids named above are exceptions. For associated liquids the ratio R Th possesses higher values between 1‘25 and 4.32. There is a pronounced parallelism between the values of the ratios q b d/sand R Tb. I n an homologous series li’ Tb gener- ally increases with the boiling point.There is a great analogy between the effect of temperature on the vapour pressure and on the product of the viscosity and the specific volume. J. F. S. Investigations concerning the Viscosity of Binary Liquid Mixtures. HANS EGINLR (Medd. K. Vetenskapsakad. Nobel-Inst. 1918 3 No. 22 1-13).-A4 theoretical paper in which a large number of viscosity determinations of various authors are collected and examined in connexion with the modified logarithmic formula of Arrhenius log i i = n log q1 + n2 log q2 where and v2 are the viscosities of the components of the mixtures and n and n2 the molecular concentrations. It is shown that this formula represents the variation of the viscosity in binary liquid mixtures more exactly than any other formula.J. F. S. Theoretical Signiflcance of Viscosity Measurements of Colloidal Solutions. SVASTE ARRHENIUS (Medd. K. Vetens- Lapyakad. h T o b e l - l m t . 1918 3 No. 21 1-22).-A theoretical paper i n which a large number of viscosity determinations of colloidal solutions are discussed. It is shown t h a t Einstein’s viscosity formula is fully established by Bancelin’s results (A. 1911 ii 586 1067) as soon as the logarithmic formula is employed. The proposed extension of Einstein’s formula put forward by Smoluchowski is not confirmed by Oden’s measurements on colloidal sulphur (A. 1913 ii 485). Viscosimeter for Measuring Viscosities and Fluiditiee. GEORGES BAUME and HENRI VIGNERON (Ann. Chim. and. 1919 [ii] 1,379-383).-The viscosimeter tube consists of a capillary pro- vided with a bulb at? its upper end and a short length of wide tube a t its lower end; this tube is fixed vertically by means of a cork in a test-tube containing 20 C.C.of the oil under examination the lower end of the tube being immersed i n the oil up t o a mark just below the capillary. A thermometer passes through a T-piece carried by the cork the bulb of the thermometer dipping into the oil. The test-tube is inserted in a boiling bulb provided with a reflux apparatus This boiling bulb may contain ether acetone. benzene or water according t o the temperature a t which the viscosity is t o be determined ; when the required temperature has iogr-iogqp=e+ J. F. S.GENERAL AND PHYSICAL CHEMISTRY. ii. 93 been reached the oil is forced upwards into the viscosimeter tube by a pressure ball attached to a branch of the T-piece; the pressure is then released and the time taken f o r the level of the oil to fall from a mark just above the bulb (on the viscosimeter tube) to one Emulsification by Adsorption at an Oil-Water Interface.S. E. SHEPPARD (J. Physical Chem. 1919 23 634-639).- Emulsions of nitrobenzene in sulphuric acid and hydrochloric acid may be readily prepared by making the acid of the same density as the nitrobenzene. Similar moderately stable emulsions were prepared by saturating the acid (sulphuric) with lead sulphate or calcium sulphate or hydrochloric acid with sodium chloride or lead chloride in both cases the acid being of the same density as the nitrobenzene. It therefore appears that in accordance with Ban- croft’s general theory of peptisation adsorption a t a liquid-liquid interface is capable of effecting emulsification.The systems thus produced on’ ageing pass into a condition approaching Pickering’s emulsions but with a great diminution of the dispersity. just below i t is noted. w. P. s. J. F. S . Nature of Osmotic Pressure MITSUJI KOSAKAI (Proc. SOC. Ezpt. Biol. Med. 1919 16 118-119).-The hnmolytic effects of formaldehyde and carbamide were found to be like that of boric acid the result of osmotic pressure. The rates of diffusion of boric acid formaldehyde and carbamide are 90 sec. 30 sec. and less than 5 sec. respectively. These differences correspond with the differ- ences in the hamolytic action of the three substances and confirm the view that osmotic pressure is not a direct property of a solute but is merely water pressure developed by the process of diffusion.CHEMICAL ABSTRACTS. Osmotic Pressure. 11. The Nature of Osmotic Pressure MITSUJI KOSAKAI ( J . Immunology 1919 4 49-65).-The same degree of osmotic hEmolysis is not produced by identical concentra- tions of boric acid formaldehyde and carbamide or by a corre- sponding lowering in the medium of suspension of the treated corpuscles. As the treating concentration of the three hnmolytic substances is correspondingly diminished the ratio between that concentration and the final concentration in the hnmolytic experi- ment increases disproportionately with the different substances. The osmotic hzemolysis of corpuscles which have been treated with the three hzmolytic substances in the same osmotic concentration is not inhibited by identical concentrations of sodium chloride or of the hnmolytic substances themselves. All these facts contradict the assumption that osmotic pressure is exerted directly by the solute.Osmotic pressure is not a direct property of a solute but is solely the pressure exerted by water which has passed by the unexplained process of diffusion through a semipermeable mem- brane to the side of the higher osmotic concentration. CHEMICAL ABSTRACTS.ii. 94 ABSTRACTS OF CHEMICAL PAPERS. Osmotic Pressure of an Electrolyte. OSKAR KLEIN (Medd. K Vetenskapsakad. Nobel-lnst. 1919 5 No. 6 1-9).-A mathe- matical paper in which an expression is deduced showing the con- nexion between osmotic pressure concentration temperature and dielectric constant.The reasoning is based on thermodynamical and mechanical (kinetic) principles and is simpler than that used by Milner (A 1913 ii 481) in a similar problem. T. S. P. The Diffusion of Electrolytes into Jellies. 11. The Dependence of the Diffusion on the Mobility of the Ions and the Hydration and Polymerisation of the Molecules. OTTO VON FURTH and FRANZ BUBANOVIO (Biochem. Zeitsch. 1918 92 139-170. Compare A. 1919 ii 13).-Diffusion in jellies differs in most cases from free diffusion in aqueous solution and is dependent on the character of the jelly. The velocity with which acids bases and polymerised salts penetrate jellies is smaller than would be expected from the mobility of their ions. Weakly hydrated salts show normal diffusion (diffusion similar to that in aqueous solution) in jellies.St,rongly hydrated salts on the other hand contrary to their behaviour in aqueous solution diffuse relatively faster than in the previous case so that their diffusion velocity approximates to the theoretical value obtained from the mobility of the ions. s. s. z. Influence of the Concentration of Electrolytes on the Electrification and the Rate of Diffusion of Water through Collodion Membranes. JACQUES LOEB (2. Gem Physiol. 1919 20 173-199).-When an aqueous solution is separated from pure water by a collodion membrane the initial rate of diffusion of water into the solution is influenced in an entirely different way by solutions of electrolytes and of non-electrolytes.The latter influence the rate of diffusion in direct proportion to their concen- tration and this effect is termed the gas-pressure effect. Solutions of electrolytes show the gas-pressure effect also but. it commences a t a somewhat higher concentration than in the previous case. If the concentration of the electrolvte is below that a t which the gas- pressure effect is observed the solutions have a specific influence on the initial rate of diffusion which is not found in the case of the solutions of non-electrolytes and is due to the diffusion of the water in an electrified condition the sign of the charge depend- ing on the nature of the electrolyte in solution according to the theory which has been advanced by the author (A. 1919 ii 497). I n these lower concentrations the curves representing the influence of the concentration of the electrolyte on the initial rate of diffusion of water into the solution show that within a range of concentra- tions between MI256 and MI16 or more (according to the nature of the electrolyte) the reverse of what should be expected on the basis of van’t Hoff’s law is noticed namely that t.he attraction of a solution of an electrolyte for water diminishes with an increase in concentration Whilst no definite assumption concerning theGENERAL AND PHYSICAL CHEMISTRY.ii. 95 origin of the electrification of water and the mechanism by which the ions influonce the rate of diffusion of water particles through collodion membranes is made it is suggested that in the lowest concentrations attraction of the electrified water particles by the ions with the opposite charge prevails over the repulsion of the water particles by ions with the same sign of charge whilst beyond a certain critical concentration the repelling action of the ion with the same charge as that of the water particles on the latter increases more rapidly with increasing concentration of the solute than the attractive action of the ion with the opposite charge.It is shown that negative osmosis is due to the repulsion of the electrified particles of water by the iun with the same charge as that borne by the water. J. C. D. The Effect of Strain on Solubility. J. C. HOSTETTER (Science 1919 50 25) .-It is possible that fluctuating tempera- ture and perhaps some indirect effects brought about by pressure may account for the solidification of crystals compressed in contact with their solution by loosely fitting pistons as found by James Thomson Le Chatelier and Spring without the necessity of postu- lating large increases in solubility due to non-uniform pressure.I n preliminary experimenta individual crystals were subjected to stress a t constant temperature by direct loading and the effect on the concentration of the surrounding solution was studied by measuring the electrical conductivity. No change in concentration was found. The test was sufficiently sensitive to indicate that the effect of non-uniform pressure is much less than that produced by the same pressure acting uniformly. However in another series of experiments in which an unloaded crystal was placed alongside a loaded crystal the former grew at the expense of the latter show- ilig that a very slight increase of solubility was produced by the stress.The method of loading the crystals has a large influence on the effects found thus indicating the importance of the stress distribution. The experiments of Becker and Day on the linear force of growing crystals are cited as indicating the stability of a crystal in its solution even when subjected to pressure. I n their experiments loaded crystals were found to lift the load during growth although the pressures on the supporting edges of the crystals were finally of the order of magnitude of the crushing strength of the crystal. The evidence so far obtained indicates that the effect of strain on solubility is a second-order effect.CHEMICAL ABSTRACWS. Influence of Electrolytes on Solubility. EBBE LINDE (Arkiw. Kern. Min. Geol. 1917 6 No. 20 1-17).-The solubility of ethyl ether and ethyl acetoacetate has been determined in water and in solutions of certain electrolytes. In the case of ethyl ether it is found that 7.88 grams dissolve in 100 C.C. of water a t 1 8 O . The influence of mixtures of two electrolytes on the solubility of ether in water is determined for the pairs of electrolytes sodiumii. 96 ABSTRACTS OB CHEMICAL PAPERS sulphatesulphuric acid sodium sulphate-sodium hydroxide and sodium chloride-sodium acetate. I n the first two pairs the total electrolyte concentration was 0.5N and in the last pair A'. The concentrations of the two electrolytes were varied within these limits.For the pair of electrolytes sodium sulphate-sulphuric acid it is shown that the solubility is 5.52 grams per 100 C.C. of 0.5N-sodium sulphate and rises regularly with decreasing sodium sulphate con- centration and increasing sulphurio acid concentration to 7.72 grams per 100 C.C. of 0.5N-sulphuric acid. I n 0.5N-mixtures of sodium sulphate and sodium hydroxide the solubility is constant 5.53 grams per 100 C.C. of solution whilst in the case of sodium chloride and sodium acetate the solubility is 4.62 grams in 100 C.C. of ;2i-sodium chloride and falls regularly to 3.44 grams in 100 C.C. of N-sodium acetate. The electrolytes are seen therefore to have an additive effect. The solubility is also determined in solutions of sodium acetate and sodium chloride of varying concentration and in both cases shown to increase steadily with increasing dilu- tion but in neither case is the depression of the solubility in keep- ing with Steiner's expression.Both sets of solubility values are fairly in keeping with the expression of Hoffmann and Langbeck (A. 1905 ii 374). The solubility of ethyl acetoacetate has been determined in water and solutions of sodium chloride and sodium nitrate of various concentrations a t 16-16*5°. It is shown that 100 C.C. of water dissolve 12.5 grams of ethyl acetoacetate a t 16-16.5O. I n N-sodium chloride 8 . 4 grams dissolve in 100 c.c. and this value increases with decreasing concentration of sodium chloride but the solubility change is in keeping with neither of the above expressions. I n sodium nitrate solution 11.4 grams of ethyl acetoacetate dissolve in 100 C.C. of W-sodium nitrate and the solubility increases in accordance with both the above-named ex- pressions on decreasing the nitrate concentration.The addition Df alcohol to solutions of sodium chloride and nitrate increases the solubility of ethyl acetoacetate in these solvents. * J. F. S. Contrasting Effects of Chlorides and Sulphates on the Hydrogen-ion Concentration of Acid Solutions. ARTHUR W. THOMAS and MABEL E. BALDWIN ( J . Amer. Chem. Soc. 1919 41 1981-1990) .-The hydrogen-ion concentration of a chrome tanning liquor has been determined in the presence of various concentra- tions of sodium chloride ammonium chloride sodium sulphate. ammonium sulphate and magnesium sulphate.The effect of the addition of chlorides is to increase the hydrogen-ion concentration whilst that of sulphates is to decrease it although on keeping the concentration increases somewhat Similar experiments were carried out with pure chromium sulphate solution chromium chloride solution sulphuric acid (0*0005N) hydrochloric acid (0.004A7) sulphuric acid (0.1N). and hydrochloric acid (O.lN) using a number of chlorides and sulphates. In every CWB the chlorides are found to increase the hydrogen-ion concentration andGENERAL AND PHXSICAL CHEMISTRY. ii. 97 The power of increasing the hydrogen- the sulphates to reduce it. ion concentration follows the order MgCl2>BaCl2>LiC1,>NaC1>NH4Cl = KC1. J. F. S. Degree of Ionisation of Very Dilute Electrolytes.GILBERT N. LEWIS and GEORGE A. LINHART ( J . Amer. Chem. SOC. 1919,41 1951-1960) .-A theoretical paper in which the authors deduce a general equation for the freezing-point lowering of dilute strong electrolytes. This equat,ion has the form n h - e / c = p c ~ in which n is the number of dissociated parts h the degree of dissociation 8 the lowering of the freezing point c the concentration and p and a characteristic constants. The calculation and observed values of 0 for fourteen salts a t a series of concentrations are compared and shown to be in good agreement. The thermodynamic degree of dissociation is calculated for the same salts over a range of con- centration 10-1-10-~iV. The constants a and B of the above equation can be obtained from freezing-point determinations.The calculated degree of dissociation diverges extraordinarily from t,he value obtained by the usual method. The Influence of the Dielectric Constant of the Solvent and of the Electric Energy of the Ions on Electrolytic Dissociation MARIO BASTO WAGNER (Anal. Fis. Quim. 1919 16 229-257) .-A purely mathematical paper. Mechanism of Electrolytic Dissociation. B. CABRERA (.4naZ. Fis. Quinz. 1918 16 186-225).-A mathematical paper developing the theory that the ions remlting from the solution of a crystalline substance are already present as such in the crystal the formation of neutral molecules being thus subsequent to the act of solution or fusion. J. F. S. W. S. M. W. S. M. Atomic Constitution of a Crystal Surface. E. MADELUNG (Physikal.Zeitsch. 1919 20 494-496) ,-A mathematical paper in which the author considers the condition of the surface layer of molecules in crystals of the sodium chloride type. It is shown that in crystals of the regular type consisting of binary compounds the atoms of one kind on the surface are displaced in directions at right angles to the surface with respect to the atoms of the other kind. The displacement decreases in the body of the crystal according to a simple exponential equation. It is also shown that since the atoms of electrolytes carry electric charges an electric double layer must in consequence of the displacement surround the crystal surf ace. J. F. S. A New Method of Electrical Synthesis of Colloids. THE. SVEDBERG (Medd. K . Vetensbapsakad. Nobel-lnst.1919 5 NO. 10 l-l8).-The author has prepared gold and silver sols of high dis-ii. 98 ABSTRACTS OF CHER1ICA.L PAPERS. persion by a modified method. The apparatus consists of a quartz tube in which a small hole about 1 mm. diameter is bored. The metal electrodes gold or silver pass down this tube and are so placed that the arc may be formed opposite the hole. The quartz tube is so arranged that a current of nitrogen passes through it both from the top and the bottom. The quartz tube is placed in an outer glass jacket containing about 30 C.C. of the dispersion medium (alcohol) and this jacket is surrounded by a further jacket containing a cooling agent (ice and salt or solid carbon dioxide and alcohol). An electromagnet is placed with its polw on either side of the small hole in the quartz tube.The current (I amp. 220 volts) is switched on and an arc of the usual type appears for about a second then owing to a melting of the lower electrode (anode) the quartz tube becomes somewhat stopped up and so the electrode is protected from the dispersion medium ; the arc becomes a sharp-pointed flame and this is drawn through the hole in the tube by the action of the magnet. Metallic clouds appear and are absorbed by the dispersion medium. The appearance of the arc has been examined spectrographically in both conditions and tthe differences noted. The author holds that the condensation of metal vapour is the cause of the sol formation. Other forms of apparatus are described in the paper. J. F. S. Colloid Metal Reactions. Spectrum Analysis and Blood Colouring Matter.EDUARD RICHTER (Kolloid. Zeitsch. 1919 25 208-211).-A number of colloid react8ions between 1% gold chloride solution and 1 1000 solutions of adrenaline alloxan t+annic acid and p-phenylenedimethyldiamine are described. I n each case colloidal gold is produced but of different degrees of dispersity. From the expesriments i t is deduced that certain diamines possess a particularly powerful reducing action and are probably to be classified with the reducing substances of the amino- group which occur in the animal organism. Certain hydroxy- substances also show a similar reducing action. J. F. S. Colour of Colloids. IX. WILDER D. BANCROPT (J. Physical Chem. 1919 23 554-571. Compare A. 1919 ii 500).-A con- tinuation of the previous discussion.It is shown that with very small particles the light which is ordinarily reflected selectively is transmitted by resonance whilst the light which is ordinarily trans- mitted is scattered. Massive gold is red by multIple reflection and thin films are green by transmitted light. Very small particles reflect green and transmit red light. Massive gold reflects yellow when compact and brown to black when porous. Particles which do not resonate are yellow or brown by reflected light and transmit blue light. Silver is yellow by multiple reflection and thin films are blue to green by transmitted light. Very small particles reflect blue and transmit yellow. Particles which do not resonate transmit blue light and reflect red. Colloidal indigo solutions transmit red light and the surface colour of indigo is red.Sodium fog scattersGENERAL AND PHYSICAL CHEMISTRY. ii. 99 blue light and transmits the yellow which the vapour absorbs. Iodine fog scatters red light. J. F. S. Colours of Colloids. X. WILDER D. BANCROFT ( J . Physical Chem. 1919 23 603-633. Compare A. 1919 ii 500).-A con- tinuation of the previous discussion. In the present paper the colours of glasses and glazes are considered. It is shown that in glasses and glazes gold silver copper platinum iridium oxide selenium tellurium sulphur lead antimonate carbon magnetite ferric oxide stannic oxide zirconium oxide arsenious oxide titan- ium oxide and calcium phosphate occur usually as a second phase. Chromium occurs in some form as a second phase in chrome Aven- turine glass and copper in Egyptian blue.Some glasses coloured by iron chromium manganese and cobalt are optically empty. In enamels the substance causing the colour is probably chiefly adsorbed by the material causing the opacity. J. F. S. Colours of Colloids. XI. WILDER D. BANCROFT ( J . Physical Chem. 1919 23 640-644. Compare preceding abstract).-A con- tinuation of the previous discussion. I n the present paper the colours of gem stones are considered. J. F. S. Coagulation. ARNE WESTGREN (Arkiv. Kern. Min. Geol. 19 18 7 No. 6 1-30).-The coagulation of gold sols has been examined with the object of ascertaining the influence of the size of the colloid particles the nature of the coagulating ions and the temperature on the velocity of the slow coagulation.Forty C.C. of the sol were mixed with 10 C.C. of an electrolyte and after definite intervals of time had elapsed 5 C.C. of the mixture were withdrawn and run into 25 C.C. of 0.5% gelatin solution and the particle concentration deter- mined. Using soh with particles of radius 120 pp 77 p p and 49 pp respectively it is shown that the velocity of coagulation is indepen- dent of the size of the particles. Using the following electrolytes to coagulate the sols (hydrochloric acid sodium lithium potassium and rubidium chlorides sodium hydroxide and sodium iodate) it is shown that the ionic conductivities of the ions are determinative of the coagulating power. Thus hydrochloric acid the cation of which has the greatest ionic conductivity is the most active coagulator of all the chlorides whereas lithium chloride with it3 slow-moving cation is the least active.Potassium chloride and rubidium chloride which have cations of about the same ionic con- ductivity have about the same coagulating power. Since gold sols are negative colloids it follows that when they are coagulated by cations the anions must exercise a stabilising influence and this is shown in the coagulation by sodium derivatives for sodium hydr- oxide with its rapidly moving hydroxyl ion is the least efficient coagulator whilst sodium iodate with its slow-moving anion is the most efficient coagulator. Using hydrochloric acid and sodium chloride at various temperatures it is shown that the velocity of coagulation does not depend on the temperature only because theii.100 ABSTRACTS OF CHEMICAL PAPERS. Brownian movement of the particles depends on it but also because t,he specific coagulation power of the electrolytes changes with the temperature. J. F. S. Effect of the Wall of Vessels on the Velocity of Gaseous Reactions. ELICHI YAMAZAKI (J. Tokyo Chem. Xoc. 1919 40 606-608).-This is a theoretical paper attempting to give an explanation for Kooij's work (A. 1893 ii 569) in which the velocity of dissociation of phosphine was shown to be affected by the nature of the vessel the ratio of the velocity in the new vessel to that in the old being 1 2.25. According to the author the decomposition of phosphine in a vessel should be considered to be a heterogeneous chemical reaction that is (1) a formation of the diffused zone between gas and vessel (2) chemical reaction. I n the new vessel the wall becomes the " diffusion zone," but in the old vessel the decomposition product (phosphorus) adheres t o the surface of the vessel thus creating a larger surface which accelerates the velocity of the diffusion.Ordinarily the velocity of the d3u- sion of the gas over the wall is so fast that only the velocity of the chemical reaction is measured as the total reaction velocity. The author believes that the vessel acts as a catalyst in the same manner as platinum black or sponge by removing the product of the reaction thus maintaining general equilibrium. More detailed investigation and explanation are promised. CHEMICAL ABSTRACTS. An Unsolved Problem in the Application of the Quantum Theory to Chemical Reactions W.C. M. LEWIS (Phil. Mug. 1920 [vi] 39 26-31).-0n applying the quantum theory to a uni- molecular reaction i t is shown that very different results are obtained according as it is assumed that the absorption of radiation is continuous or discontinuous. A very large discrepancy exists in both cases between the calculated and observed velocity con- stants which is much greater on the discontinuous view than on the continuous. This discrepancy is always in the sense that the observed velocity constant is many times greater than the calcu- lated ; on the hypothesis of continuous absorption the observed con- stant is of the order of 107 times greater than the calculafed and also appears to be independent of temperature for different reac- tions.The explanation of this discrepancy is expected to throw light on tthe theory of physicochemical processes. J. R. P. The Propagation of Flame in Mixtures of Methane and Air. I. Horizontal Propagation. WALTER MASON and RICHARD VERNON WHEELER (T. 1920 117 36-47). The Propagation of Flame in Complex Gaseous Mixtures. IV. The Uniform Movement of Flame in Mixtures of Methane Oxygen and Nitrogen. id.Maximum-speed Mixture0" of Methane and Hydrogen in Air. WILLIAM PAYMAN (T. 1920 117 48-58),GENERAL AND PHYSICAL CHEMISTRY. ii. 101 Hydrolysis of Iodoacetic Acid. BROR HOLMBERQ (Medd. E. Vetenskapsakad. Nobel-lnst. 1919 5 No. 11 l-l2).-The velocity of replacement of iodine by hydroxyl in the sodium and barium salts of iodoacetic acid by the action of sodium and barium hydroxidm has been studied; the influence of neutral salts (sodium chloride and iodide and barium chloride) was also investigated.The reaction is bimolecular its velocity being intermediate between those observed for the corresponding bromo- and chloro-acids (com- pare Johansson A. 1912 ii 544). Contrary to expectation there- fore iodoacetic acid is more stable than bromoacetic acid (compare Drushel and Simpson A. 1918 i 57). The effect of the concen- tration of the metal ions on the velocity constant C is given by the equations C=O-lOG[Na*].O*1 and C= 0-220[ba]0*2 where ba=+Ba. The reaction between iodoacetic acid or its salts and silver nitrate was also studied. The formation of silver iodide takes place much more quickly than the production of acid and the author concludm that the most important reactions are represented by the equations ZCH&*CO,’ + Ag’ = CH,I*CO,*CH,*CO,’ + AgI and CH,I*CO,*CA,*CO,‘ + Ag’ = O < ~ ~ z ~ ~ ~ > O + A gI.If the solutions are not too dilute the formation of iodoacetylgly- collic acid takes place much more quickly t*han t’he formation of glycollide. Glycollic acid is then formed indirectly or it may also be formed directly according to the equation CH,I*CO,’ + Ag’ + H,O = OH-CH,-CO,’ + H’ + AgI. T. S. P. Effect of Chlorine on Periodic Precipitation. [Miss] A. W. FOSTER ( J . Physical Chem. 1919 23 645-655).-When silver nitrate is allowed to diffuse into gelatin films cont?aining potassium chromate a series of concentric rings is formed the spacing of which decreases with the distance from the centre.The appearance and spacing of these rings does not depend on the time which Eas elapsed between the formation of the gelatin film and the addition of the silver nitrate. I f however a trace of chloride is present in the gelatin the appearance is altered. Making up the gelatin chromate solution with tap water is sufficient to produce the change. I f such a film is allowed to harden for three hours and then treated with silver nitrate the rings are spiral in form and in groups of three. If a little calcium hypochlorite is added to the gelatin and the film allowed to harden for an hour narrow rings close together are formed by two precipitations. Some Problems in Contact Catalysis. WILDER .D. BANCROFT (Trans. Amer. Electrochem. Soc. 1919 36 reprint).-A number of specific cases of catalytic reactions are cited where a satisfactory explanation of the mechanism of catalysis is not available and further study of these is suggested.Thus the reaction COCl,+ H20 = CO + 2HC1 is so far as is known irreversible. In the pres- ence of excess of water the reaction goes to completion but in the presence of concentrated hydrochloric acid the rate of hydrolysis is J. F. S. VOL. cxvar. ii. 4ii. 102 ABSTRACTS OF CHEMICAL PAPERS negligible. A theory is suggested but study is necessary to prove it. Also trichloromethyl chloroformate (superpalite) reacts as follows in the presence of alumina CCI,*O*COCl=CO,+CCl and in the presence of ferric oxide CCl,*O*COCl= 2COC1,. The reverse reac- tion has never been made to take place to any extent.Yet when some superpalite and ferric oxide were placed in a tube the reaction sGon came to an apparent end. When the temperature was raised a little the reaction went further and did not reverse on cooling. It is suggested that this phenomenon was due to poisoning of the ferric oxide. Finally some experiments by Lind are cited which may if desired be regarded as the displacement of an equilibrium by a catalytic agent. If radium emanation is placed in water in the liquid phase hydrogen and oxygen are formed and escape into the vapour phase; if the emanation is moved up into the vapour phase oxygen and hydrogen are caused to combine. Additional examples are cited and some interesting lilies of experiments suggested. CHEMICAL ABSTRACTS. Catalytic Decomposition of Hydrogen Peroxide.GOSTA PHRAGM~N (Medd. K . Vetenskapubud. Nobel-Inst. 1919 5 No. 22 1-13).-The rate of decomposition of hydrogen peroxide in alkali phosphate or sodium hydroxide solutions has been determined a t 17-18O. I n the phosphate mixtures it is shown that the reac- tion velocity reaches a maximum at P,=11-8. I n alkaline hydr- oxide solutions the velocity is not easily reproducible and some- times gives a large value and sometimes a small value. Since the velocity decreases with both increasing and decreasing hydrogen- ion concentration it is assumed that hydrogen peroxide forms a salt with sodium hydroxide which is verv stable and that only the undissociated hydrogen peroxide molecules undergo decomposition. The decomposition of hydrocen peroxide by yeast extract in the presence of a phosphate buffer mixture has been determined at 17-18O.It is shown that fresh yeast decomposes dilute hydrogen peroxide without sending a soluble enzyme into the surrounding liquid. The reaction is of the first order between certain limits and the reaction coefficient increases proportionally to the quantitp of yeast. The catalytic action per cell or per gram can he increaFed by treating the yeast with sugar solution before use. Ths reaction constant is no criterion of the quantity of cat-alase in the cells. J. F. S. Catalytic Actions at Solid Surfaces. 11. Transference of Hydrogen from Saturated to Unsaturated Organic Compounds in the Liquid Staye in Presence of Metallic Nickel E. F. ARMSTRONG and T. P. HILDITCH (Proc.Roy. SOC. 1919 [ A ] 96 322-329. Compare A. 1919 ii 403).-The evidence that metallic catalysts during the hydrogenation process interact primarily with the unsaturated organic compound together with the resemblance of the whole process t o enzyme action (Zoc. c i t . ) led to the consideration whether the catalytic action like thatGENERAL AND PHYSICAL CHEMISTRY. ii. 103 of certain enzymes might not be reversible. Evidence in support of this view has no’w been obtained. A t 180° an equimolecular mixture of cyclohexanol and methyl cinnamate may in the presence of nickel be transformed into one in which about 10% of the cinnamic ester has become hydrogenated to methyl P-phenyl- propionate the cyclohexanol being transformed into cyclohexanone. The experiments indicate that it is necessary that both components of the system should be present in the liquid state. A similar action took place to a certain extent at 230° when the cyclohexanol was replaced either by dimethylcyclohexane or bv dihydropinene.Similarly a mixture of ethyl stearate aqd methyl cinnamate when heated at 230° with catalytic nickel gave a product containing small quantities of methyl /3-phenylpropionate and ethyl oleate. An important point to note is that simultaneous dehydrogenation and hydrogenation have now been effecte,d at temperaturw not far removed from the general optimum hydrogenation range (170-180°) and that hydrogen has been transferred from one compoun’d to another instead of from one molecule to another of the same species as was observed by Zelinski and Glinka (compare A 1911 i 870).There is no absolute proof that the mechanism of the change is not dependent on the production of hydrides of nickel but the aut-hors prefer to regard it as a further case of the catalytic equilibria previously discussed which may be indicated as follows +Ni [ c::EZ:Ni] Saturated compound Unsaturated compound Ni Unsaturated + Ni + H,. [ hydrogen 1 - compound The ultimate equilibrium will depend on the resultant of the vary- ing affinity for nickel of the saturated and unsaturated compounds involved. This view of catalytic hydrogenation and dehydrogenation affords some explanation of the products obtained (compare Moore J . Sac. Chem. Tnd. 1919 38 3 2 0 ~ ) during the hydrogenation of un- saturated glycerides where partial isomerisation occurs.It is suggested that in the hydrogenation of ethyl oleate a portion of the freshly produced ethyl stearate in contact wit.h the nickel under- goes dehydrogenation the hydrogen liberated being transferred to more ethyl oleate the ‘‘ dehydrogenated ethyl stearate ” formed being the ethyl AX-oleate isolated by Moore (Zoc. cit.). Constitution and Structure of the Chemical Element. RAWKSWORTH COLLINS (Chem. News 1919 119 295-296).- Single electropositive charges emanate from masses of 1 (hydrogen) 7 (lithium) 23 (sodium) and 39 (potassium) also two electro- positive charges emanate from 4 (helium) of which one has been shown to emanate from a mass 3 (H3) hence single electropositive charges emanate from masses 1 3 7 23 39. Taking the first twenty-six elements it is shown that twenty of these may have their W.G . 4 *ii. 104 ABSTRACTS OF CHEMICAL PAPERS. atomic weights split up into the above-named numbers in such a way that the number of parts is the same as the maximum valency. All the parts are odd numbers and the non-metallic elements are characterised from the metallic by the presence of one or more portions (1 + 3) (helium). Elect>ropositive forces are distinguished from electronegative forces by the following rule an electropositive force emanates from each of the masses 1 3 7 23 39 except when an element has one or more portions (1 + 3) in which cases electro- positive forces emanate from these portions whilst electronegative forces emanate from each of the remaining portions of the element.Hydrogen forms the connecting link in the structure of these elements. From tables given in the paper it is shown that the longer the chain of -H-H the more volatile the element and the stronger the electronegative forces from the metallic portions. On this basis the structure of the sulphur atom is given as I I I * Na-H-H3-H-H,-H I I in which Na represents 23 H represents 1 and H 3. The thick lines represent electropositive charges the curly lines electronegative charges and the thin lines forces which are not chemically evident but which have to be overcome before the element can be dis- integrated. The atomic number of sodium is 11 that of H and H unity which makes the atomic number of sulphur 16 which is in keeping with fact. The formula therefore shows the connexion between atomic weight atomic number and maximum valency and distinguishes between electroposit,ive and electronegative charges.The molecule K,SO may therefore be represented OK OK 0 /\ Na-€3-H,-H-H,-H. \/ 0 J. F. S. A Model of Radioactive Atoms. TORAHIKO TERADA (Proc. Phys. Math. SOC. Japan 1919 [3] 1 185-195).-According to the recent view of atomic structure an atom consists of a positive nucleus about which a number of electrons are revolving. The conception of the positive nucleus is slightly extended by consider- ing the case of two heavy nuclei with opposite charges revolving about the common centre of mass. Assuming a primary with the larger mass having a positive resultant charge and the secondary of smaller mass charged negatively as a whole and leaving out of consideration the ring electrons i t is evident that the two members of the nuclear system will exert on each other a mutual action analogous to the tidal action in the case of the gravitating planetary system.Both members are considered to consist of a number of positive and negativel elementary charges bound together by some unknown forces; also most of the positive charges are assumed toGENERAL AND PHYSICAL CHEMISTRY. ii. 105 consist of two elementary charges as in the case of a-part-icles. When the tidal action exceeds a certain limit the secondary may become unstable and when the elementary charges in the two substances happen to take a definite configuration one of the charges may escape and be projected from the nuclear system with the momentum possessed during the orbitual motion.If the negative secondary loses a positive charge by the expulsion of an a-particle the attraction of the primary will increase and the orbitual velocity be accslsrated. I n actual radioactive transformations the veloci- ties of the expulsion of the a-particles generally increase with successive transformations although with some exceptions. The tidal action will increase with the decreasing distance bet<ween the two nuclei and consequently the chance of disruption will increase because the chance of the a-particle stepping out of the critical threshold will increase. The result will be the shortening of the life of the atom in question. This seems to suggest a possible way of explaining Geiger and Nuttall's law concerning the relation between the average lives of radio-elements belonging to a dis- integration series and the velocities of the a-particles emitted from these elemeats.If the two members of the nuclear system gradu- ally approach each other by successive emission a stage may be attained a t which the two substances are amalgamated into one pro- vided the resulting configuration is a stable one. When this final stage is reached the disintegration will be stopped and the end- product of the radioactive transformation will be reached. One of the simplest systems conceivable is proposed as a provisional work- ing model and the consequences tested in the light of the available experimental data. CHEMICAL ABSTRACTS. Metals and Non-metals.A. SMITS (Proc. I<. Akad. Wetensch. Amsterdam 1919 22 119-125).-The author explains the posi- tive charge of the chlorine electrode by assuming that the chlorine atom has the power of splitting off and absorbing electrons and further that these two processes can take place side by side. This is represented by the equations XCl dX(31,' + 2XOG and YCl,,+2YO S 2YClG' in which X and P indicate the frac- tions of the chlorine molecules which have undergone positive and negative ionisation respectively. As the electrons which are absorbed in the second action proceed from the first it is clear that Y Z X but chlorine being a non-metal both fractions must be extremely small. If then i t is further assumed that the negative ions almost exclusively pass into solution the electrode will then possess a positive charge.Extending the above hypo- thesis the author shows that t-here is only a difference of degree between metals and non-metals and not one of kind. If it be assumed that all elements can both split off and absorb electrons then t-he factor X is comparatively large for metals and small for non-metals and for non-metals P is exceedingly small. This view is in keeping with the large electric conductivity of metals and the small cmductivity of non-metals. In the case of the metals the 4"-2ii. 106 ABSTRACTS OF CHEMICAL PAPERS. positive ions have the greater solubility and the negative ions in the case of the non-metals. This is in keeping with the different electromotive behaviour of metals and non-metals. The difference in solubility between the two ions is so great as to justify t-he assumption that only one kind is present in solution.Amphoteric elements probably send appreciable quantities of both positive and negative ions into solution. J. F. S. Atomic Structure and the Periodic Law. TYCHO E:SON AURBN (Medd. K. Vetenskapakad. Nobel-Inst. 1919 5 No. 18 1-7) .-A theoretical paper. The author’s experiments on the absorption of Rontgen rays (details not yet published) indicate the possibility of differentiating between the inner and outer elect,rons in the atom the outer electrons being identical with t.he so-called valency electrons and playing a totally different r61e from the inner electrons. I n accordance with this view Kossel’s table (A. 1916 ii 243) for the first twenty-three elements of the periodic system is modified._The number of electrons in the outer ring increases from one in the case of hydrogen to seven in the case of nitrogen; it does not become eight in the case of oxygen but four electrons go from the outer ring to tvhe inner system; the number of outer electrons then increases to seven again for sodium whilst for mag- nesium four more electrons go from the outer to the inner system and so on. The consequences of these conceptions are discussed special reference being made to the valencies of the elements in the various groups. For example it is argued that sulphur should be bivalent; the existence of SO and SF is no definite proof of the sexavalency of sulphur since the former can be represented by a ring structure and the latter as :>F.S.F<g.T. S. P. Atomic Numbers HAWKSWORTH COLLINS (Chem. News 1919 1 19 285-287) .-The author considers the atomic weights (nearest whole number) atomic numbers and valencies of the first twenty-eight elements taken in order of their atomic weights. It is shown that with four exceptions (glucinum nitrogen scandium and cobalt) when the atomic weight is represented by an even number the maximum valency and the atomic number are both represented by even iiumbers and when the atomic weight is repre- 3ented by an odd number the valency and the atomic number are also represented by odd numbers. Further the elements with the exceptions named follow one anot-her alternately odd and even. The probability that this condition of things has happened accident- ally is 1 :42*. The reason for the odd and even rule is shown to be as follows. If the atomic weight of an element be split up into 3’s and 1’s alternately always commencing with a 3 the number of portions obtained gives the atomic number and should the atomic number be an even number the atomic weight must of necessity be an even number but if the atomic number is odd the atomic weight must also be odd. The atomic number is exactlyGENERAL AND PHYSICAL CHEMISTRY. E 107 obtained as follows. If the atomic weight is an even number the atomic number is onehalf of the atomic weight but if the atomic weight is an odd number the atomic number is onehalf of the atomic weight from which 1 has been subtracted. J. F. S. The Mathematical Possibility of Increaaing the Yield or of Reducing the Proportion of the Reacting Substances in certain Chemical Reactions. ANGEL PBREZ HERNANDEZ (Anal. Fis. Qzcim. 1918 16 302-317) .-An ele,mentary mathe- matical method is illustrated for the constz-uction of chemical equa- tions in which two or more reactions are simultaneously involved. W. S. M. Some Biographical Notes on Hermannus Follinus. W. P. JORISSEN (Chem. JVeekbZad 1919 16 947-951).-The biographer gives a very brief sketxh of Hermannus Follinus who was born in Friesland circa 1590 and joined the medical faculty in Cologne where he died of the plague in 1622. He was the author of “Den Nederlandtsche Sleutel van’t Secreet der Philosophie ” (Key to the Secrets of Philosophy) “ Physiognomie,” and “ Simonides ofte die Memori-const ” (Science of Memory). W. J. W. An Apparatus for Preparing in a Very Short Time Homogeneous Liquid Mixtures. PIERRE JOLIROIS (Compt. rand. 1919 169 1095-1098).-A simple apparatus is described consistr ing essentially of two containers one for each of the liquids to be mixed arranged so as to deliver into the two limbs of a Y-tube of the pattern shown the liquids in the requisite proportion. The diameter of the Y-tube is 6.2 mm. and the M lower limb has a restriction as shown 1.5 mm. in diameter and 2 cm. long. With such an ap- paratus a homogeneous mixture may be obtained U from two miscible liquids in 1/100th of a second. W. G . Eykman’s Suction Pump. L. TH. REICHER (Chem. Weekblad. 1919 16 951-956).-This apparatus embodies the principle of Geissler’s suction pump but the operation of filling the vessel in communioation with the apparatus which has to be exhausted by lifting a second vessel periodically is avoided. Special features of Eykman’s pump are a “ vacuum reservoir,” containing air-free concentrated sulphuric acid and a drying device also filled with that liquid. A sketch of the complete apparatus is given and its mode of operation described. W. J. W.