pressure is 212.6 cm. of mercury. From the sulphur dioxide partial pressures at equilibrium the mean value of the heat of dissociation into magnesia and sulphur trioxide over the range 1000-1100° has been calculated as 66.07 cal., whilst that obtained from thermochemical measurements is 65-6 cal. Beryllium sulphate dissociates appreciably at 565° and the dissociation follows the reaction 5BeSO, SO3,5 Be0+ 4SO. Up to 785°, the sulphur trioxide equilibrium pressure is represented closely by the equation log Pso,-14.907/T-14.10 log T+57-97, whence the calculated values for the heat of decomposition over the range 700-800° has a mean value of 41 cal. compared with 49.8 cal. given by thermochemical measurements for the direct decomposition to the oxide. The fact that over the temperature range 590-775° the dissociation pressure of aluminium sulphate is considerably higher than that of beryllium sulphate suggests a possible method for the separation of the two metals. The double sulphate of potassium and magnesium melts at 750°, but does not commence to dissociate appreciably until 895°; this agrees with the thermodynamic deduction that its dissociation pressure should be considerably lower than that of magnesium sulphate. The double salt of potassium and beryllium melts at 900°. Dissociation can be detected at 700-710° and the dissociation curve is normal up to about 975°, but at higher temperatures, owing to the solution of potassium sulphate in the fused double salt, the pressure is diminished. The calculated mean value of the heat of decomposition up to the m. p. is 65.2 cal., above which temperature it is 58.8 cal. The value obtained from thermochemical measurements is 58.5 cal., from which it follows that the value 6-4 cal. does not necessarily correspond with the heat of fusion of the salt. A. E. MITCHELL.

Thermochemistry of beryllium. C. MATIG NON C. MATIG NON and (MLLE.) G. MARCHAL (Compt. rend., 1925, 181, 859-861).—The heats of solution of some beryllium compounds in water and in solutions of hydrogen chloride, hydrogen fluoride, and sodium hydroxide are recorded. The following heats of formation are given beryllium oxide, 137-4 cal.; hydroxide, 209-3 cal.; sulphate, 276-9 cal. The results obtained illustrate the close chemical analogy between beryllium and aluminium. S. K. TWEEDY.

Heats of combustion of normal substances. P. E. VERKADE and J. COOPS (Z. physikal. Chem., 1925, 118, 123-128).-A reply to the criticisms of Jaeger and von Steinwehr (A., 1925, ii, 126) that undue weight has been given to Dickinson's value for the heat of combustion of benzoic acid in view of the results obtained by other observers. Recent accurate determinations (Verkade and Coops, A., 1923, ii, 294; Schläpfer and Fioroni, A., 1923, ii, 832; Swientoslawski and Starczewska, A., 1922, ii, 616) are in agreement with Dickinson's value, and it is claimed that the results quoted by Jaeger and von Steinwehr are less trustworthy. In particular, the electrical method of determining heat values involves a systematic error.

J. S. CARTER.

Calorimetric researches. IX. Heat of combustion of d- and meso-tartaric acids, racemic acid, and some derivatives. J. Coops and P. E. VERKADE (Rec. trav. chim., 1925, 44, 983-1011; cf. A., 1925, ii, 490).—Ammonium hydrogen meso-tartrate (m. p. 167°) is prepared by adding solid mesotartaric acid to concentrated ammonia until neutral to methyl-orange; the same amount of the acid is then added and the solution evaporated and crystallised. meso-Tartramide (m. p. 187-187-5° with slight decomposition) separates slowly from a solution of ethyl meso-tartrate saturated with ammonia at 0°. The molecular heats of combustion at constant pressure (mol. wt. in mg.; 15° cal.) are: d-tartaric acid, 275-1; racemic acid, 273-0; meso-tartaric acid, 275-7; ammonium hydrogen d-tartrate, hydrogen racemate, and hydrogen meso-tartrate, 341-7, 339-5, and 341-2, respectively; methylammonium hydrogen d-tartrate and hydrogen racemate, 508-0 and 506-0, respectively; ethylammonium hydrogen d-tartrate and hydrogen racemate, 665-4 and 663-1, respectively; aniline hydrogen d-tartrate and hydrogen racemate, 1079-3 and 1077-3, respectively; benzylamine hydrogen racemate and hydrogen meso-tartrate, 1229-9 and 1231-5, respectively; d-tartramide, 427-0; mesotartramide, 426-4; d-tartaric diethylamide, 1064-1; racemic acid diethylamide, 1064-3; meso-tartaric diethylamide, 1065-3. The heat of racemisation of solid d-tartaric acid is thus 21+01 Cal. Both in the crystalline state and in dilute aqueous solution, the symmetrical intramolecular inactive acid has a greater free energy content than the asymmetrical optically active isomeride. meso-Tartaric acid has a larger heat of combustion, but a smaller dissociation constant than d-tartaric acid, in contradiction to the rule of Stohmann (J. pr. Chem., 1889, 40, 357). The heats of combustion of succinic, l-malic, and the tartaric acids show that replacement of a hydrogen atom by a hydroxyl group does not produce a constant difference, and hence the heat of combustion is not an additive quantity. Both racemic acid and hydrogen racemates are racemic compounds. W. HUME-ROTHERY.

Calorific value and constitution. M. F. BARKER (J. Physical Chem., 1925, 25, 1345–1363).-Previous empirical expressions for calculating the calorific values of carbon compounds take little or no account of constitutive effects. The contributions of like atoms and of the CH2-group have been taken as constant, which, except in the case of hydrogen, is not permissible. The molecular calorific value of an organic compound depends on its constitution; that of carbon varies and becomes less as the disposition of valency bonds approaches that in the symmetrical tetrahedral positions. The thermal effects accompanying oxidation of the carbonyl group and the combination of a hydrogen atom with a hydroxyl group, deduced from the calorific values of diphenyl, benzil, and benzoin, are 60-7 Cal. and 12.9 Cal., respectively. The combustion of diatomic hydrogen is similar to that of hydrogen in a hydrocarbon, and since the calorific value of hydrogen is the same in all the organic compounds studied, it is adopted as the basis of calculation.

The heat of combustion deduced for the carbon atom in methane (normal case) is 75.5 Cal.; for each of those of ethane, 83-1 Cal. This increase is ascribed to a decrease in the angle between the valency bonds accompanying the change from methane to ethane. In ascending an homologous series, this increase becomes less. Replacement of the four hydrogen atoms in methane by the same group (e.g., in tetramethylmethane) restores symmetry to the molecule and the central carbon atom attains the original value. The value of the change -C:C→ CO2 is 97.7 Cal., of the same order of magnitude as the combustion of elementary carbon: similarly with "benzenoid" carbon. The linking -CC- has a value 120-5 Cal. in acetylene to 124-5 Cal. in dipropargyl.

COOH

The contribution of the carboxyl group is not equivalent to one carbonyl plus one hydroxyl group. Results point to mobility of a hydrogen atom giving rise to an additional potential H ÕH hydroxyl, thus X< X.C<OH' and favouring association. Discrepancies between calculated and observed values disappear when calorific values of the vapours of the acids are taken. The aldehydic group is equivalent to one carbonyl and one hydrogen group. Among the simpler aromatic hydroxy-compounds, only the dihydroxybenzenes are anomalous, possibly because of tautomeric effects. The results of various observers indicate that the benzene molecule is best represented by Ladenburg's prism formula. L. S. THEOBALD.

Heat of solution of gypsum at the maximum solubility. E. LANGE and F. DÜRR (Z. physikal. Chem., 1925, 118, 129-139).-The heat of solution in water of gypsum has been determined at 22.5°, 27-8°, 33-4°, and 37-6°, the values for the molecular heats of solution being -590, -300, 0, and 230 cal., respectively. Contrary to the statement of Colson (A., 1925, ii, 37), the zero value for the heat of solution occurs at a temperature which, within the limits of experimental error, is that at which the solubility is a maximum.

J. S. CARTER.

Calculation of some characteristic constants of free ammonium. A. BALANDIN (Z. physikal. Chem., 1925, 118, 114-118).-According to formula derived by the author (A., 1924, ii, 719; 1925, ii, 637) the heat of formation of ammonium is -17,800 cal., the molecular volume, 47-9 cm.3, and the density in the solid state, 0.356. J. S. CARTER.

Analogies and differences in behaviour of the various forms of energy in reversible and irreversible transformations. E. DENINA (Gazzetta, 1925, 55, 638-645).-A mathematical paper, in which the various formulæ derived from the second law of thermodynamics are reduced to the same type independent of the form of energy involved. T. H. POPE.

Vapour-pressure lowering as a function of the degree of saturation. I. I. BENCOWITZ (J. Physical Chem., 1925, 29, 1432-1452).-A relation

expressing the vapour pressure of aqueous solutions of non-volatile solutes as a function of temperature and solubility has been deduced, using the degree of saturation as a fundamental method of expressing concentration. The equation has the form log AP =K[1/T-a(1-log S/b)], where AP is the lowering of the vapour pressure, T the absolute temperature, and S the degree of saturation, i.e., the ratio of the number of g. or mols. of solute in solution to the number present at saturation, in a given weight of solvent. K, a, and b are constants. The three postulates used in deducing this relationship are as follow: (1) the coefficient [(OP log AP)/(1/T)}]s is constant (equals K), and (2) is independent of S, and (3) the coefficient [(@K log S)/(1/T)}]AP-1 mm. is

constant.

Existing data are used to verify these. The relation is shown to hold for 31 salts, which include the commoner salts of the alkali metals and ammonia, the chlorides and bromides of the alkaline earths, and the sulphates of beryllium, nickel, copper, and zinc. The values of the constants, K, a, and b are also given for these salts. The value of a is the reciprocal of the absolute temperature at which a saturated solution has a vapour pressure lowering of 1 mm.

From the above equation, there follow two generalisations, (1) the ratio of the lowering of the vapour pressure, at any degree of saturation, to that of a saturated solution, at a given temperature, is independent of temperature, and (2) the value of this ratio is a simple function of the degree of saturation, i.e., AP/APS=SKalb L. S. THEOBALD.

Free energy of ions measured by capillary electrode. P. B. TAYLOR (Physical Rev., 1924, [ii], 23, 556).-Using a cell with one mercury electrode contained in a capillary and completely polarisable, the other being reversible to the anion of the electrolyte, it is shown that T-To+Cf(V-Vo), where V is the fall of potential across the cell, Vo the E.M.F. of the reversible electrode, T the surface tension of the polarised electrode, To a constant, whilst C' depends only on the concentration of the electrolyte. Curves may be constructed yielding values of the difference in free energy of the anion at two concentrations. The ratio of the change in free energy of the anion to that of the electrolyte is 0-50 for potassium and sodium chlorides, but 0-85 for potassium hydroxide above 0.1N. A. A. ELDRIDGE.

Electrical conductivity in benzene solutions. S. JAKUBSOHN (Z. physikal. Chem., 1925, 118, 3136). The electrical conductivity of solutions of aluminium bromide monothiohydrate (AlBr3,H2S) over the concentration range 20-54% by weight increases with increasing concentration of solute; the values of the specific conductivities at 25° of the 20% and 54% solutions are 0-81 × 10-6 and 3.54X 104, respectively. The mol. wt. as determined by the cryoscopic method over the concentration range 1-22.5% increases with increasing concentration, being 271 in the 1% solution and 359 in the 22.5% solution. Electrolysis results in the cathodic liberation of hydrogen and the anodic deposition of bromine which reacts with the solvent to form bromine

derivatives. The decomposition voltage at 25° for a 45% solution is 0.74 volt. J. S. CARTER.

Pseudo-acids. G. B. SEMERIA and A. PICHETTO

(Atti R. Accad. Sci. Torino, 1925, 60, 241-250). Measurements have been made of the equivalent conductivities of aqueous solutions of camphoroxime, of its sodium salt, and of the sodium salt of nitrocamphor, at a series of temperatures between 25° and 50°. The conductivity of the sodium salt of camphoroxime increases between 25° and 35° with dilution more rapidly than that of the other sodium salt, but between 40° and 50° the relation is reversed, For the former temperature range, the degree of ionisation of the nitrocamphor salt is greater, and for the latter range, lower, than that of the camphoroxime salt. The pseudo-acid character of the two compounds is confirmed by the gradual neutralisation which their salts exhibit with hydrochloric acid, which has been followed by conductivity measurements. Equilibrium between the two forms is reached in the shorter time by nitrocamphor. Refractometric measurements with the two salts lead to nearly equal values for the molecular refractivities, the small differences being attributed to the difference in structure. Camphoroxime [R 60-5499, sodium salt of camphoroxime [R] 62-2605, sodium salt of nitrocamphor [R] 62-9818. The difference between the molecular refractivities of the two sodium salts is 1.7, in conformity with the rule of Muller and Bauer (A., 1903, ii, 705). F. G. TRYHORN.

The

Conductivity and electrolysis of iodine trichloride in acetic acid. B. P. BRUNS (Z. physikal. Chem., 1925, 118, 89-98).-Iodine trichloride dissolves in anhydrous acetic acid to give a clear solution. In the moist solvent some iodic acid is precipitated as a result of the reaction: 31C1, + 6H20=9HC1+ HI+2HIO3. Conductivity measurements at 18° show that the value of the molecular conductivity passes through a minimum at a dilution, v=10 litres. irregularities in the temperature-conductivity curves in the region 20-30° are regarded as furnishing evidence of the existence of complexes in solution. Electrolysis results in the liberation of chlorine in double the amount anticipated on the basis of Faraday's law. A decomposition potential either does not exist or is extremely small. It is suggested that iodine trichloride forms complexes with the solvent and that the ionic dissociation is: ICl,,n(CH ̧·CO2H) — ICl3′′+(CH2•CO2H)n ̈. J. S. CARTER.

Concentration E.M.F. in solutions containing acid. A. R. GORDON and C. WEBER (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 26-27).-The P.D. across a liquid surface separating two air-free solutions of copper sulphate in maximum conducting sulphuric acid, when current is flowing, is given by the equation P.D.=CR+Blog, C1/C2, where c and Ca

are the copper concentrations in the two solutions and B is the Nernst factor for bivalent ions (RT/2F). Experiments with cuprous chloride in 3N-hydrochloric acid and with silver sulphate in N-sulphuric acid give B=RT|F. The transport number of silver in these solutions is 0.023 for N/40 solutions and 0.011 for N/80-solutions. J. S. CARTER.

Thermodynamic potential difference at the boundary of two liquid phases. II. S. WosNESSENSKY (Z. physikal. Chem., 1925, 117, 457–460). -Cells similar to those previously studied (A., 1925, ii, 673) have been examined, amyl alcohol being used sisting of mixtures of potassium hydroxide and as non-aqueous solvent. With electrolytes conphosphoric acid, or of potassium hydroxide and citric acid, singular points were obtained in the E.M.F.composition curves corresponding with the formation of the primary, secondary, and tertiary salts of the respective acids. When N-potassium chloride was used as the electrolyte in a cell of zero E.M.F. and butyric acid was added to the half-cell containing observed until the concentration of the acid was about the aqueous solution, no change of E.M.F. was 0.5N. Valeric acid produced no change in the concentration range investigated (up to about 0.1N). These effects seem to be connected with the surface activity of the acids. L. F. GILBERT.

C. J.

Electromotive behaviour of aluminium. DE GRUYTER (Rec. trav. chim., 1925, 44, 937-969). -The freezing-point diagram of the system aluminium-mercury has been determined. No compounds or transition points exist, the liquidus falling from the m. p. of aluminium, at first gradually and later very rapidly, to an eutectic of practically pure mercury. A solid solution of mercury in aluminium, containing approximately 8 atoms % of mercury, exists at the ordinary temperature. The P.D. of aluminium-mercury alloys in contact with dry alcoholic solutions of aluminium chloride and with solutions containing both mercury and aluminium salts has been measured. The P.D. for pure aluminium in aqueous salt solutions has been determined in the absence of oxygen. For amalgamated aluminium in normal salt solutions it is 1.590 volts, relative to the calomel electrode. The influence of different salts and of hydrogen-ion concentration on the potential of aluminium has been examined, and the complex results interpreted in reference to Smit's theory of allotropy. Mercury is a positive and aluminium hydroxide or oxygen a negative catalyst for the internal reaction Alsolid Alt+30 solid, and the passivity of aluminium is due to the slowness of the (-) reaction.

Pseudohalogens. II.

residue.

solid

W. HUME-ROTHERY.

(I) The fulminic iodine, (II) Equilibrium between selenocyanogen, and the corresponding silver salts. (III) Polypseudohalides. L. BIRCKENBACH and K. KELLERMANN (Ber., 1925, 58, [B], 2377 -2386; cf. A., 1925, ii, 568).-I. Comparison of the decomposition potentials of mercury fulminate and mercury cyanate shows the fulminic to be slightly more strongly electronegative than the cyanic residue. could not be measured, since the product of the action The decomposition potential of potassium fulminate of potassium amalgam on mercury fulminate could not be purified completely from potassium hydroxide. II. The equilibrium between silver iodide, silver selenocyanide, selenocyanogen, and iodine in the presence of ether (cf. loc. cit.) is established for mixtures in stoicheiometric proportions when 86%

of the iodine is present as silver iodide and 14% remains in the ethereal solution.

III. The tendency of the alkali halides to yield polyhalogen salts can be measured by potentiometric titration. Standard ethereal solutions of halogens are added to standard alcoholic solutions of halides and the changes of potential against an unattackable electrode are measured in the usual manner. Potassium iodide and potassium selenocyanide are thus titrated with iodine and selenocyanogen solution and sudden alterations of potential are observed corresponding with the formation of the salts KI,, K(SeCN)I2, K(SeCN),I, K(SeCN). The tendency to form such compounds attains its maximum in the case of cæsium; caesium triselenocyanide has been isolated.

The simple halogen atoms appear univalent, since the periphery contains seven electrons and by the addition of a further electron yield the complete octet. The idea is extended to the pseudohalogens, and it is pointed out that the azide residue has fifteen electrons, eight of which surround and hold together the complex and are externally inactive chemically, whereas the remaining seven condition the halogen-like behaviour of the group. Similarly, the sum of the electrons in the cyano-, thiocyano-, selenocyano-, and telluracyano-residues is fifteen and a similar arrangement may be assumed. H. WREN.

Decomposition potentials and polarisation of certain heavy metallic chlorides dissolved in anhydrous pyridine. R. B. MASON and J. H. R. B. MASON and J. H. MATHEWS (J. Physical Chem., 1925, 29, 1379-1393; cf. Müller, A., 1923, ii, 287; 1925, ii, 133, 134). -The decomposition potentials and polarisation curves of certain metallic chlorides in anhydrous pyridine have been studied in order to test the behaviour of the reference electrodes, which consisted of an amalgam covered with a paste of the salt of the metal contained in the amalgam. The reference electrodes Hg-Cd,CdCl2 and Hg-Zn,ZnCl2 are the most satisfactory. The decomposition potentials of saturated solutions of zinc, cadmium, cuprous, and mercuric chlorides at 30° are, respectively, 1-75, 1, 0.5, and 0.65 volts. The curves for lead chloride and for mercuric sulphate in pyridine show no changes in direction. Solutions of cupric chloride were difficult to work with owing to low solubility and low conductance, and the results obtained depend on the method of preparation of the anhydrous salt. The cuprous chloride solution conducts well, but the formation of a compound at the anode makes polarisation difficult to follow. Where possible, both total polarisation and polarisation at either electrode have been followed. The solutions of lead, mercuric, and cuprous chlorides behave as if a depolariser for small amounts of chlorine is present. Cadmium and zinc chlorides in pyridine give results which are similar to those obtained with water as solvent.

L. S. THEOBALD.

Effect of variation of current and concentration on polarisation in a lead cell. J. T. BURT-GERRANS and H. R. HUGILL (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 26).-Two co-axial lead cylinders were used as electrodes, the arrangement of the apparatus affording a uniform electric field between

the cylinders. With acid more dilute than 0.05N as electrolyte the fall in E.M.F. during the discharge is due to depletion of acid in the electrolyte, but with maximum conducting sulphuric acid the fall is due to exhaustion of active material on the electrodes. Polarisation, which occurs at the cathode only with currents down to 0.03 amp., is manifested by a sudden increase in the voltage. The time from the beginning of the charge to the rise in polarisation voltage varies inversely as the current, and changes in a regular manner with the previous history of the cell (preliminary charging and discharging in preparation for test charge). Polarisation effects can only be observed separately from other changes in potential difference with acids which are more dilute than 0.5N. No regular variation with acid concentration observed. J. S. CARTER.

Photomicrographic study of the evolution and disappearance of gas during the passage of electricity through glass. J. B. FERGUSON and O. W. ELLIS (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 34).-Photographs which illustrate the points stressed by Rebbeck and Ferguson (A., 1924, ii, 840) have been obtained. With the reappearance of gas the presence of a brown colour in the glass was noted. The colour may be made to come and go by proper regulation of the direction of the electrolysing current. Photographs show that the colour originates in minute circular spots with dark centres which later grow and spread over the glass. The colour is not attacked by concentrated nitric acid. Microscopical examination shows the coloured spots to be in the same plane and very thin. The discontinuous nature of the coloration and the way in which gas bubbles form in different tubes indicate that glass does not behave as a homogeneous medium when electrolysed.

J. S. CARTER.

Electrolysis of soda-lime glass. M. J. MULLIGAN (Trans. Roy. Soc. Canada, 1925, [iii], 19, III, 35-36). In an extension of the work of Schulze (Ann. Physik, 1913, 40, 335) the electrical migration of silver into glass from anodes of metallic silver and of aqueous solutions of silver nitrate at 100° has been observed. Silver was also found to diffuse into glass at 100° from an aqueous solution of silver nitrate without the use of the current. J. S. CARTER.

and electroultrafiltration. A comparison. E. Dialysis and ultrafiltration, electrodialysis, HEYMANN (Z. physikal. Chem., 1925, 118, 65–78).— The kinetics of the removal of crystalloids from sols by dialysis, ultrafiltration, electrodialysis (Freundlich and Loeb, A., 1925, i, 96) and electroultrafiltration (Bechhold and Rosenberg, A., 1925, ii, 668) under ideal conditions are derived. Strong electroare considered and expressions for the rate of removal lytes are removed from sols by electrodialysis or electroultrafiltration about 10 times more rapidly than by ultrafiltration and about 100 times more rapidly than by dialysis. Where weak electrolytes are concerned, as in the separation of gelatin from its degradation products (amino-acids etc.) (Knaggs, Manning, and Schryver, A., 1923, i, 1144), the advantages of the electrical method are nullified and ultrafiltration is

most suitable. Removal of non-electrolytic crystalloids by ultrafiltration is considerably influenced by the mol. wt. of the crystalloid and by the nature of the dispersed phase. With hydrophobic colloids (rapidly removed by ultrafiltration) and crystalloids of high mol. wt. (100-400) complete separation is attained 20-50 times as rapidly by ultrafiltration than by dialysis. With hydrophilic colloids (slowly removed by ultrafiltration) separation is attained about 5-15 times as rapidly. If it is necessary to free a protein solution from electrolytes and also from nonionised degradation products of high mol. wt., recourse must be had to electroultrafiltration.

J. S. CARTER.

Rate of unimolecular reactions. D. ALEXIEV (Z. physikal. Chem., 1925, 118, 119-122).-A theoretical paper in which the importance of collision as a factor in unimolecular gas reactions is emphasised. A reaction mechanism is suggested which leads to an expression for the reaction velocity involving concentration to the first power only. J. S. CARTER. Derivation of the equation for the effect of temperature on reaction rate. R. C. TOLMAN (J. Amer. Chem. Soc., 1925, 47, 2652-2661).—The equation, -dC/dt=ke (E/RTC, connecting the rate of a first order unimolecular reaction with the temperature and the energy of activation (E) is derived without making any specific assumption as to either the rate or the mechanism of activation, and is thus shown to be at least approximately valid under a wide variety of conditions. The energy of activation is the excess per mole in the energy of the molecules that react over the average energy of the unactivated molecules. The corresponding equation, The corresponding equation, -dC/dt=kT2e-[(E+E')/RTICC', for a second order bimolecular reaction is derived both on the assumption that the reacting molecules have received their energies of activation preceding the collision that leads to reaction, and on the assumption that the processes of activation and reaction are merged in one collision. A. GEAKE.

Influence of inflammable and other gases on the explosibility limits of mixtures of gas and air. XI. Graphical representation. W. P. JORISSEN (Rev. trav. chim., 1925, 44, 1039-1047).See B., 1925, 980.

Thermal decomposition of nitrogen pentoxide at low pressures. H. S. HIRST and E. K. RIDEAL (Proc. Roy. Soc., 1925, A, 109, 526-540). The results of several investigators agree in showing that the thermal decomposition of nitrogen pentoxide satisfies the usual criteria of a unimolecular reaction. Observations have been made both on the gaseous system, and on solutions of the pentoxide in organic solvents. The surface of the glass containing vessel, the presence of inert gases and of the products of reaction, are factors which do not disturb the unimolecular law. The heat of activation, calculated in the normal manner, is about 24,700 cal. per g.-mol. The high velocity of the reaction precludes the possibility of this large activation energy being supplied either by inelastic thermal collisions, or by "blackbody" radiation, unless some "chain mechanism

66

is called into play. Such mechanisms have been suggested. One depends on the passage of a radiation quantum through a chain of successive molecules, each one re-emitting the quantum after reaction. The other supposes that an activated form of nitrogen dioxide is produced by the decomposition, and that such molecules, on collision, may cause further reaction. Both these theories would predict a diminution in the velocity of reaction at low pressures, owing to the greater probability of the rapid failure of a chain." This point has now been tested experimentally. The rate of decomposition of gaseous nitrogen pentoxide has been measured at pressures as low as 0.01 mm. of mercury. The procedure was to measure the pressure of oxygen resulting from the reaction by a platinum Pirani gauge, after removing the other gases into a cooled side-tube. No retardation of the decomposition was observed, but below a critical pressure of the total gaseous constituents in the reaction vessel-about 0.25 mm. of mercury-the reaction velocity began to increase, finally, at the lowest pressures, attaining five times its normal value and becoming approximately constant at this value. Possible disturbing factors such as the presence of impurities are carefully considered, but it is decided that the observed velocity increase is of real significance. All "chain" mechanisms are irreconcilable with the phenomenon. The results are interpreted by the assumption that a definite fraction of the activated molecules always undergoes decomposition irrespective of pressure, but that a larger fraction (about four-fifths) does not decompose if it collides within 10-6 sec. after activation, but is deactivated by collision. Calculations on this basis are in good agreement with experiment. A possible explanation of the two different types of activated molecules is offered, based on the fact that there are four NO linkings, and only one shared NO linking in the molecule of the pentoxide. Activation of one linking causes decomposition, activation of the shared linking invariably so; activation of the others, only after a time interval and if collisions do not intervene. S. BARRATT.

Rate of reaction of bromine with aqueous formic acid. D. L. HAMMICK, W. K. HUTCHISON, and F. R. SNELL (J.C.S., 1925, 127, 2715—2720).— The reaction between bromine and formic acid has been studied in dilute aqueous solution, using the Ostwald isolation method. The formic acid is completely oxidised to carbon dioxide. The reaction is of the second order, but the rate is retarded by the hydrobromic acid produced. From a study of the separate effects of the hydrogen and bromine ions, it is deduced that the reaction takes place between the formyl ions and free bromine molecules, i.e., those molecules of bromine which are not combined with bromine ions to give the complex ions Br,'. In order to obtain agreement with theory, it is necessary to assign to the constant for the equilibrium between bromine molecules, bromine ions, and tribromide ions a value considerably higher than that found by Jakowkin (A., 1896, ii, 514). This is justified by the fact that this author's experiments were carried out with higher concentrations and his constants