showed a regular increase with decreasing concentration of bromine. E. E. WALKER.

Velocity of the reaction between hydrogen peroxide and iodine ions. J. A. CHRISTIANSEN (Z. physikal. Chem., 1925, 117, 433-447). The influence of neutral salts on the velocity of this reaction was investigated, potassium iodide being used as a source of iodine ions. Potassium chloride, bromide, and nitrate cause a slight acceleration, potassium sulphate, oxalate, and chromate a slight retardation, and the alkaline-earth chlorides a marked acceleration. The temperature coefficient, between 0° and 25°, of the acceleration of the reaction by hydrogen ions (supplied by hydrochloric acid) was also determined. L. F. GILBERT.

Mechanism of the reaction between iodine, iodine ions, and hydrogen peroxide. J. A. CHRISTIANSEN (Z. physikal. Chem., 1925, 117, 448456).-Theoretical. A mechanism is suggested which is in harmony with Abel's data (A., 1921, ii, 180). The observations of Bray and Livingstone (A., 1923, ii, 473) on the course of the reaction between hydrobromic acid, bromine, and hydrogen peroxide are similarly explained. L. F. GILBERT.

Dependence of the rate of alkaline hydrolysis on the constitution of the alcohol. II. L. SMITH and H. OLSSON (Z. physikal. Chem., 1925, 118, 99-106; cf. A., 1922, ii, 701).-The kinetics of the bimolecular reaction in aqueous solution between sodium hydroxide and n- (3.93), iso- (3.54), sec.(0-816), and tert.-butyl acetates (0-081), n- (44-7) and sec.-propyl glycollates (13-8), have been investigated at 20°. The numbers in parentheses represent the values of the velocity coefficients at this temperature. The rate of hydrolysis varies with the nature of the alkyl radical and of the acid. When the alkyl acetates investigated in this and previous researches are arranged in the order of decreasing reactivities, the series obtained is identical with the corresponding series for the rate of ester formation from alcohols and acetic anhydride (Menschutkin, A., 1888, 901) and for the rate of formation of the corresponding ethyl alkyl ethers (Sagrebin, Z. physikal. Chem., 1900, 34, 149). J. S. CARTER.

Dependence of the rate of alkaline hydrolysis on the constitution of the alcohol. III. Temperature coefficients. H. OLSSON (Z. physikal. Chem., 1925, 118, 107-113).-The velocity coefficients for the reaction between sodium hydroxide and the two propyl acetates, the four isomeric butyl acetates, isoamyl acetate, and the two propyl glycollates in aqueous solution at 0.2°, 10°, 20°, 30°, and 40° have been determined. The temperature coefficients (10° intervals) decrease with increasing temperature according to Arrhenius' formula, there being no maximum value in the region 10-20° as recorded by Trautz and Volkmann (A., 1908, ii, 824).

J. S. CARTER.

Effect of differential aëration on corrosion. Electrode potential measurements. A. L. MCAULAY and F. P. BOWDEN (J.C.S., 1925, 127, 2605-2610). The four zones distinguished by Evans

Ꭰ

(A., 1925, ii, 688) in a sheet of metal partly immersed in water have been investigated by means of electrode

potential measurement using 0-1N-sodium chloride as electrolyte. Iron and zinc surfaces tend to exist in one of two normal states, one a more electronegative state characteristic of pure metal and state characteristic of aërated regions. For zinc, the corroded regions, and the other a less electro-negative difference in single electrode potential between these states is about 75 millivolts, for iron it is about 200 millivolts. These differences may be increased if the oxidation and corrosion respectively are very drastic. Experiments with drops of electrolyte on metal plates have shown that surfaces having a clean, bright appearance may be in either of these two states and then change rapidly from one state to the other with changing conditions. E. E. WALKER.

Mechanism of reduction. V. H. J. PRINS

(Rec. trav. chim., 1925, 44, 1051-1055; cf. A., 1925, ii, 1169).-Measurements of the rate of dissolution of lead in acetic acid in presence of nitromethane show that the reaction is much slower than in presence of nitrobenzene and that it is a true reaction velocity which is observed, not a velocity of diffusion. The critical concentration is calculated to be a 1-125M solution of nitromethane. The reactivity of the nitro-group in nitrobenzene is about 65 times that of the nitro-group in nitromethane, the difference being attributed to the positively charged carbon atom of the benzene nucleus. The mechanism of the reduction is discussed in detail. G. M. BENNETT.

I.

Catalytic combustion. Union of carbon monoxide and oxygen in contact with a gold surface. W. A. BONE and G. W. ANDREW (Proc. Roy. Soc., 1925, A, 109, 459-476; cf. A., 1906, ii, 434). -The rate of combination of a moist, theoretical mixture of carbon monoxide and oxygen, in contact with a gold surface at about 300° in a "normal" state of activity (i.e., the maximum reached after a period of use), is always directly proportional to the pressure of the mixture. The catalysing power of the surface may be much reduced by cooling to the ordinary temperature over a period of days, or by prolonged evacuation. The reattainment of "normal" activity takes many hours. The catalysing power of the surface, when in "normal" activity, can be highly stimulated by previous exposure to either carbon monoxide or oxygen at the experimental temperature. When either of the two reacting gases was present in excess, the rate of combination was proportional to the partial pressure of the carbon monoxide, which was thus the controlling factor.

The rate of reaction depends on the presence of combining gases are activated by the surface, and moisture in the gases. It is concluded that both that the activation is not confined to a unimolecular adsorbed layer, but extends to more deeply occluded gas. Structural changes in the metal probably play a part in the phenomena, although no change in the surface could be detected microscopically. The gold was introduced in the form of a gauze.

S. BARRATT.

Sulphuric acid from a quadruple mixture. R. SAXON (Chem. News, 1925, 131, 372—373).—Electrolysis in series of solutions of ferrous ammonium sulphate and of equimolecular mixtures of copper and ferrous ammonium sulphate yields a 9% solution of sulphuric acid in the first case and a 12% solution in the second case in 3 hrs. After 7 hrs.' electrolysis, a ferrous ammonium sulphate solution will contain 14% of free sulphuric acid, whilst an equimolecular solution of nickel, copper, ammonium, and ferrous sulphates electrolysed in series with the ferrous ammonium sulphate solution will contain 19% of free acid. A. R. POWELL.

Surface catalysis in photochemical processes. H. S. HIRST and E. K. RIDEAL (Nature, 1925, 116, 899-900). It is improbable that the diverse results obtained by various investigators for the rate of combination of gases irradiated in silica tubes by a mercury-vapour lamp can in general be attributed either to an inhibiting factor or to the propagation of chains. A mercury surface illuminated at the ordinary temperature with radiation, including strong emission of the resonance line, 2537 A., promotes combination in various mixtures of gases, but no catalytic effect has yet been observed with other metals. In the absence of hydrogen, ozone can be identified by the tailing of the mercury, but in presence of hydrogen, mercuric oxide is formed on the mercury surface and the walls of the vessel. Condensation of formaldehyde and its polymerides from carbon monoxide and hydrogen is accompanied by auto-retardation of the rate of reaction. This catalytic effect is believed to account for the high rates of combination frequently recorded.

A. A. ELDRIDGE.

Decomposition of ozone in red light. G. KISTIAKOVSKI (Z. physikal. Chem., 1925, 117, 337— 360). An apparatus is described which has been used for the investigation of the decomposition of ozone at high concentrations in red light. Empirical equations are derived with which the observed course of decomposition is in agreement. The apparently contradictory results obtained by previous workers (von Bahr, A., 1910, ii, 949; Warburg, A., 1913, ii, 652; Weigert, A., 1915, ii, 813; Griffith and others, J.C.S., 1923, 123, 2752, 2767) are due to erroneous interpretation of the experimental data, which agree approximately with those of the author. Helium, argon, nitrogen, carbon monoxide, carbon dioxide, and oxygen retard the reaction, the retarding influence increasing from left to right of the series.

L. F. GILBERT.

Mechanism of the photochemical reaction between hydrogen and chlorine. II. A. L. MARSHALL (J. Physical Chem., 1925, 29, 1453—1461). -The amount of photochemical reaction between hydrogen and chlorine increases with the total préssure (A., 1925, ii, 883). A preliminary study of the effect of pressure over the range 0.001-6-0 cm. has now been made by a method in which a mixture of the two gases is drawn through a quartz vessel illuminated by a quartz mercury arc. After exposure, the chlorine and hydrogen chloride are removed

by liquid air and the pressure of the hydrogen is measured. The results are accurate only when this pressure greatly exceeds that of the chlorine.

The number of molecules of hydrogen chloride formed per quantum of light absorbed increases rapidly with pressure; over the above range the quantum yield increases from 20 to 25,000 mols. (approx.). At a constant total pressure of 5.9 cm., the quantum yield is unchanged when the light intensity is increased twenty times, a result not in agreement with Baly and Barker (J.C.S., 1921, 119, 653; cf. also Chapman, A., 1924, ii, 668). L. S. THEOBALD.

Microbalance. II. Photochemical decom

position of silver chloride. E. J. HARTUNG (J.C.S., 1925, 127, 2691-2698).-The action of light on silver chloride has been investigated gravimetrically by the same method as that employed for Films of silver silver bromide (A., 1925, ii, 57). weighing less than 0.5 mg. were deposited on vitreous silica, ignited, chlorinated, and then weighed on a Steele-Grant microbalance. The film was chlorinated by exposure to chlorine diluted with air, concentrated chlorine being found to act more slowly. The silver chloride film was sealed up and insolated as described for the bromide (loc. cit.) in the presence of air, nitrogen, or hydrogen at low pressure (0-001 or 10 mm.), copper or potassium hydroxide being present to absorb the chlorine evolved. The maximum percentage loss of chlorine on insolation was 91.1% in air, 89.9% in nitrogen, and 94.8% in hydrogen. It is shown that silver and chlorine are the only products of decomposition, there being no evidence

of the formation of subchloride. In the chlorination of silver deposits, optimum concentrations of chlorine were found to exist both for fresh and for previously chlorinated deposits. E. E. WALKER.

Nature of the photohalides and related substances. R. FEICK and K. SCHAUM (Z. wiss. Phot., 1925, 23, 389-412).-See B., 1925, 1013.

Behaviour of silver iodide in the photo

voltaic cell. A. GARRISON (J. Physical Chem., 1925, 29, 1406-1407).-A reply to the criticism of Price (A., 1925, ii, 680).

L. S. THEOBALD.

Selective action of polarised light on starch grains. E. C. C. BALY and E. S. SEMMENS (Nature, 1925, 116, 817).-Polemical against Jones (Ann. Bot., 1925, 39, 651). When starch grains in water are placed in a Petri dish or a flask and illuminated with a strong beam of polarised light they are hydrolysed. A. A. ELDRIDGE.

[Group of volatile hydrides.] F. PANETH and E. RABINOVITSCH (Ber., 1925, 58, [B], 2446—2448 ; cf. A., 1925, ii, 760).-A reply to Hantzsch and Carlsohn (A., 1925, ii, 1043) and Carlsohn (ibid., 1044). H. WREN.

Hydrates and hydrogels. VII. Isomeric hydrogels of aluminium hydroxide. R. WILLSTÄTTER, H. KRAUT, and O. ERBACHER (Ber., 1925, 58, [B], 2448—2458; cf. A., 1924, ii, 767, and previous abstracts).-Aluminium hydroxide-x is prepared as a white, somewhat plastic gel when a solution of ammonium alum is precipitated with an excess of

ammonia and the precipitate rapidly washed with water containing ammonia and subsequently dried with acetone; the success of the method depends entirely on rapidity of manipulation and exact observance of conditions completely specified in the original memoir. The compound, which has the composition Al(OH)3, readily passes within a few hours or a day into aluminium hydroxide-ß, which is converted by 10% ammonia at 100° into a gel poorer in water, whereas the x-compound is only transformed into a coarsely disperse condition. The second modification is likewise unstable, but several months are required for its conversion into aluminium hydroxide-y. The composition of the latter compound, which behaves towards ammonia in the same manner as the x-compound, is sometimes exactly, sometimes approximately, that of an orthoanhydride. The hydrogels-a and - behave as distinct chemical compounds, the second of which is much less basic than the first. Hydrogel-a dissolves in cold 0.1% hydrochloric acid within 20 min., immediately in cold 2% or hot 0.5% acid, whereas hydrogel-ẞ is not noticeably soluble in cold 5% acid. Both basic and acidic properties are lost in hydrogel-y, which does not dissolve in cold dilute or moderately concentrated hydrochloric acid or in 0.1N- or N-sodium hydroxide. These properties are not in any way connected with the degree of dispersion of the colloid, since a typical y-preparation absorbs invertase more freely than good specimens of the x- and B-gels. The differences between the x- and y-compounds are therefore due either to isomerism or to polymorphism.

The transformation of aluminium hydroxide-a into the B-compound is invariably accompanied by a diminution in the proportion of combined water in the product (after treatment with acetone). The product reabsorbs moisture in the course of a few days, and its composition thereby gradually approximates to that of aluminium orthohydroxide. It appears most probable that the B-compound is a polyaluminium hydroxide, 4Al(OH),-H2O, which is converted by ammonia into the hydroxide, 4Al(OH)3—3H2O. Aluminium hydroxide-C, prepared according to the method of Willstätter and Kraut (A., 1923, ii, 493), is frequently used for the absorption of enzymes, but, as now shown, the process may lead to the formation of either the a-, 3-, or y-modifications; the nature of the product is most readily established by its behaviour

towards ammonia.

H. WREN.

Hydrates and hydrogels. VIII. Aluminium hydroxide gel of the formula, AlO.OH. R. WILLSTÄTTER, H. KRAUT, and O. ERBACHER (Ber., 1925, 58, [B], 2458-2462).-Aluminium hydroxidegradually loses water when heated in a current of dry air, but the composition of the residue remains constant between 212° and 239°, 16-6% H2O being present. The mineralised hydroxide-a loses water without break at 150-250°. Aluminium hydroxidesẞ and have constant compositions in the temperature intervals 214-239° and 242-272°, the water content being 19.1-19.5% and 14.5-15.4%, respectively. In these cases, the product appears to consist mainly of aluminium metahydroxide, AlO-OH, mixed with aluminium oxide or polyaluminium

hydroxide. Homogeneous aluminium metahydroxide is obtained when any of the aluminium hydroxide gels is heated rapidly in a sealed tube with 10% ammonia to 250°, maintained at this temperature for 8-9 hrs., and subsequently treated with acetone. The new product retains the gel structure and is without basic or acidic properties. Its chemical individuality is established by its stability over a range of 200°. The selective adsorptive capacity of aluminium metahydroxide for enzymes is remarkable. Thus from yeast autolysates the invertase is adsorbed one twenty-fifth, the maltase one-fifth, as much as by the customary aluminium hydroxide gels, so that from such mixtures it is possible to adsorb two-thirds of the maltase without the passage of invertase into the solid phase; the homogeneous enzyme is obtained by elutriation with diammonium hydrogen phosphate. Since the most markedly selective action is exhibited by the aluminium hydroxide gel which reacts least readily with acids or alkali hydroxide, it is no longer possible to consider adsorption as an effect of the opposed electrochemical nature of enzyme and adsorbent. The result cannot be ascribed to surface action, but must be attributed to affinity relationships which cannot yet be defined exactly. H. WREN.

Calcium

8

aluminate]. A. F. O. GERMANN and C. R. TIMPANY "phosgeno-aluminate" [chloro(J. Physical Chem., 1925, 29, 1423-1431).—The existence of the compound COAL,Clg, which behaves as a weak acid and is designated "phosgeno-aluminic acid," has been previously postulated (Science, 1925, 61, 71). The present investigation is to test the [chloroaluminate], CaAl,Cl, prepared as described theory put forward. Calcium" phosgeno-aluminate " by the authors (A., 1925, ii, 1085), is the salt used. ductivities at 25° and 0° for solutions in carbonyl The vapour pressures, densities, and electrical conchloride varying in concentration up to supersaturation at 35% (approx.) have been measured by methods previously used (cf. A., 1925, ii, 196, 288, 1066). The complete vapour-pressure curve for the position of the crystals which commence to separate system CaAl, Cl,-COCI, at 25 is given. The comwhen the solution contains 33.5% of the compound CaAlCl corresponds pound CaAlCl corresponds with the formula CaAlCl3,2COC1. The mol. wt. of the solute, using values from the curve in the usual vapour-pressure formula, is 766 (Ca,AlCl16-756).

The molecular conductivity of calcium chloroaluminate in carbonyl chloride is greater than that of aluminic acid ") at all concentrations. The concepaluminium chloride in carbonyl chloride (" phosgenotion, based on chemical grounds, of the latter as a weak acid is thus supported, although, it is pointed The data for the physical properties mentioned above out, additional evidence for the theory is still required.

are tabulated.

L. S. THEOBALD.

"Phosgeno-aluminates" [chloroaluminates] of sodium, strontium, and barium. A. F. O. GERMANN and D. M. BIROSEL (J. Physical Chem., 1925, 29, 1469-1476; cf. Germann and Timpany, A., 1925, ii, 1085, and preceding abstract; Kendall, Crittenden, and Miller, A., 1923, ii, 387).-Some new

"phosgeno-aluminates "[chloroaluminates] have been prepared by allowing carbonyl chloride and aluminium chloride to react in a sealed tube with metallic chlorides. The solution was decanted and fractionally crystallised from carbonyl chloride. The vapourpressure curves at 25° have been determined and are discussed in the systems formed by the chloroaluminates of barium, strontium, and sodium, respectively, with carbonyl chloride. The data also are tabulated. The m. p. of the salts BaAlCl3, CaAl,CI, SrAlCl ̧, and NaAICI are 295°, 280° (with loss of aluminium chloride), 325°, and 155-5°, respectively; the solubilities in carbonyl chloride (omitting the calcium salt), 52.5%, 52.3%, 36.5% at 25°. Compounds of the following composition have been identified : 3BaAl C1,8COCI2; BaAl2CI, COCl2; 5SrAl2Cl2,9COCl2; and SrAl,CI,,COCI,. The sodium salt forms no such compound at 25°. "Calculations of mol. wt. from the lowering of the vapour pressure indicate 15-20 atoms of metal in the strontium and sodium compounds. L. S. THEOBALD.

Solid hydrides of arsenic, antimony, and bismuth. E. J. WEEKS and J. G. F. DRUCE (Rec. trav. chim., 1925, 44, 970-974; cf. A., 1925, ii, 700).—Amorphous arsenic dihydride, As, H2, is formed as a brown powder when a solution of alkali hydroxide is electrolysed with an arsenic cathode and a platinum anode, the electrodes being separated by a porous pot. It is also produced when a solution of arsenic trichloride in dilute hydrochloric acid is added to an ethereal solution of stannous chloride; 2AsCl2+4SnCl2+2HCl=As2H2+4SnCl4. Solid antimony dihydride is similarly prepared by electrolysis, and also by the reduction of antimony salts by zinc in acid solution. Solid bismuth dihydride, BiH, is prepared by adding a solution of bismuth trichloride in hydrochloric acid to a mixture of zinc and acid while the latter is evolving hydrogen. It is a grey powder which decomposes when heated in a vacuum, and reacts violently with fused potassium nitrate. W. HUME-ROTHERY.

Hydrates and hydrogels. IX. Silicic acid. R. WILLSTÄTTER, H. KRAUB, and K. LOBINGER (Ber., 1925, 58, [B], 2462—2466; cf. Mylius and Groschuff, A., 1906, ii, 160).-Solutions of monosilicic acid are not suitably prepared by the interaction of sodium silicate and hydrochloric acid, since the subsequent dialysis is too slow and is accompanied by great loss. It is preferable to add silicon tetrachloride slowly to a well-stirred mixture of ice and water or to pass the vapours from the boiling chloride by a rapid current of air into water at 0° to -3°. The bulk of the hydrochloric acid is immediately removed by precipitation with silver oxide at 0°, whereby about 10% of the silicic acid passes to the precipitate. Very cautious addition of the final quantities of silver oxide permits the removal of 99% of the hydrochloric acid, but the remainder cannot easily be eliminated in this manner owing to the solvent action of silicic acid on silver oxide; it is removed by dialysis. Solutions of z-silicic acid, free from chloride, can be rapidly concentrated at 15° in the high-vacuum distillation

apparatus until they contain 5-7% SiO2; they are completely transparent and mobile. Rapid evaporation without gelatinisation has not been effected with solutions containing more than 7-10% SiO2. Precipitation of gels from these solutions is best effected by ammonia and much ammonium chloride. The water content of the gels which have been treated with acetone never exceeds 22% (calculated on SiO2) and generally lies between the values required for H2SiO and H.Si,Og. Dialysis experiments establish the existence of at least two forms of a-silicic acid, one of which readily, the other slowly, passes through a diaphragm. They are regarded as monosilicic acid and oligosilicic acids formed from a few molecules of the former; in contrast with them are the polysilicic acids of the ẞ series formed by further condensation. The latter acids readily precipitate egg-albumin, whereas the former do not. It is remarkable that a-silicic acid is somewhat volatile with steam. H. WREN.

Bismuth nitrates. M. PICON (Bull. Soc. chim., 1925, [iv], 37, 1365-1375).-The water content of basic bismuth nitrates cannot be accurately determined from the loss of weight at 110°, but only by decomposition by heat in the presence of copper, and absorption of the water formed. The nitrate content is not accurately determined by boiling with sodium hydroxide or by the oxalic acid method, but results correct to 1% are obtained by boiling with a ferrous salt solution in an atmosphere of carbon dioxide, followed by titration with permanganate. The effect of conditions of formation on the composition of the medicinal compounds is determined. When the salts concerned are kept in a vacuum at 15° in the presence of phosphoric oxide, the following changes take place (1) Bi(NO3)3,5H2O →→ Big(NO3)6,3H2O; (2) Bi(OH)2(NO3)(NO,BiO)12,5H2O; (3) (NO3)18Bi20021,7H2O → (NO3)18Bi2002,6H2O; (4) (NO3)10B112013,8H2O → (NO3)10B11201,4H20. The action of heat on these salts is complex and pure bismuth trioxide is not obtained below 425°. W. HUME-ROTHERY.

Additive compounds of hydrogen chloride and sulphates of the heavy metals. F. EPHRAIM (Ber., 1925, 58, [B], 2262-2267).-Attempts to prepare anhydrous "chloro-acids" by the action of dry hydrogen chloride on lead, silver, cadmium, nickel, cuprous, cupric, mercurous, or mercuric chloride over a wide temperature interval were unsuccessful. On the other hand, hydrogen chloride readily combines with certain metallic sulphates, yielding compounds SO of the type [XH X Addition of H2 or Cl2 hydrogen chloride occurs readily with the sulphates of metals of which the chlorides do not readily evolve hydrogen chloride when treated with concentrated sulphuric acid, and the temperature at which evolution of the gas commences is very approximately the dissociation temperature of the complex. The evolution of hydrogen chloride from sulphuric acid and a chloride appears, therefore, not to be an ionic reaction. The additive products generally lose about 1.5HCl at the temperature of dissociation, whereas the

4

remainder is evolved only at a gradually increasing temperature; the production of additive compounds containing 0.5HCl is not definitely established. The following compounds are incidentally described and the temperatures of dissociation for the mean pressure of 713 mm. are recorded: PbSO,,2HCI (32°); CdSO,,2HCl (128-135°); CuSO4,2HCl (83°); HgSO4,2HC1; Ag2SO4,2HCl, which does not exhibit a definite temperature of dissociation, but yields a series of solid solutions when heated until it ultimately becomes converted into a mixture of silver chloride and silver hydrogen sulphate. Zinc sulphate and hydrogen chloride appear to yield only a series of solid solutions; the temperature of dissociation of the adduct lies below that of the customary freezing mixture. H. WREN.

Conditions underlying the attack of hydrogen chloride and ammonium halide on metals. K. A. HOFMANN and F. HARTMANN (Ber., 1925, 58, [B], 2466—2475; cf. A., 1925, ii, 685). The greater readiness with which copper is attacked by ammonium chloride vapour than by hydrogen chloride is explicable on thermochemical grounds, but the magnitude of the superiority of the former over the latter can be accounted for only if the much greater adsorption of ammonium chloride vapour is taken into account. In every case, the evolution of hydrogen from ammonium chloride is preceded by vaporisation of the latter and adsorption of the vapour. The previous view (loc. cit.) that copper reacts with undissociated ammonium chloride molecules is there

fore abandoned in favour of the hypothesis that the reactive system is an adsorbate, Cu... NH3,HCl. All those conditions which increase the adsorption of hydrogen chloride or of hydrogen chloride and ammonia within the temperature range 270-330° facilitate the reaction, which yields hydrogen in a remarkable manner. Thus the presence of carbon dioxide increases the yield of hydrogen from copper gauze of limited surface, whilst aniline hydrochloride and, in particular, dimethylaniline hydrochloride are more active than ammonium chloride, since the bases are less readily carried away by the gases than ammonia. Under like conditions, hydrogen chloride gives only a very small yield of hydrogen, which, somewhat unexpectedly, is lowered by the presence of water vapour. The reaction is not facilitated by mixing the hydrogen chloride with methyl alcohol, ethyl alcohol, ether, or diphenylamine. The activity of hydrogen chloride can be increased, not only by the presence of substances which facilitate adsorption, but also by increase of the adsorbing surface, which is most readily effected by substituting copper powder for the gauze. Under these conditions, the difference between the activity of ammonium chloride vapour and hydrogen chloride may be reduced to such an extent that the ratio of the amounts of hydrogen evolved is only 1.8:1 (instead of 140: 1); this order of magnitude is in agreement with that expected on thermochemical grounds. The superiority of ammonium chloride vapour over hydrogen chloride in chemical activity towards copper depends therefore mainly on its much greater adsorption, and is more pronounced as the surface of the copper is reduced. The decom

position of ammonium chloride vapour by tin is much less marked than by copper, since tin has but little affinity for ammonia.

If the conditions are sufficiently favourable for adsorption the chemical energy of the system is a measure of the extent of the change. Thus, ammonium bromide gives 1.7 times as much hydrogen with copper powder as does ammonium chloride, whilst ammonium iodide and copper gauze yield 2.6 times the amount of hydrogen obtained with ammonium chloride; these amounts agree with those expected from the heats of formation of cuprous bromide and cuprous iodide. H. WREN.

Different states of beryllium oxide. H. COPAUX and C. MATIGNON (Bull. Soc. chim., 1925, [iv], 37, 1359-1365).-See A., 1925, ii, 1192.

Occurrence of

atomic

number

dvi-manganese (atomic number 75) in manganese salts. V. DOLEJŠEK and J. HEYROVSKÝ (Nature, 1925, 116, 782—783).— An examination of the polarograph curves obtained in investigations of the electrolytic deposition potentials of manganous solutions, particularly when prepared from potassium permanganate, reveals nickel, cobalt, or iron, but may be due to that of ekahumps which cannot be due to the presence of zinc, manganese (atomic number 43) and dvi-manganese ("rhenium," 75). Manganese amalgam in contact with platinum foil was immersed for several days in a nearly saturated solution of manganous sulphate, the deposit removed from the platinum with concentrated hydrochloric acid, and, after dilution, neutralisation with sodium carbonate, and acidification with acetic acid, the solution was treated with hydrogen sulphide; the solution then contains manganese with about 2% of dvi-manganese. One of the humps is due to the element 75, but it could not be determined whether the other is caused by the element 43. Concentrated acidic solutions containing the element 75 are green, and the dry green chloride becomes black on keeping. Neutral solutions yield a yellowish-brown precipitate, probably by oxidation. Apparently higher valency compounds

of the element 75 are more stable than those of

manganese. Provisional spectroscopic results were obtained as follows: La1, 1430X., Lẞ1, 1235·3X., LB2, 1204-3X., Ly1, 1059 X., the copper Kß line being used as reference. The results of Noddack, Tacke, and Berg (Naturwiss., 1925, 26, 567) are with those of thallium than those of the element 75. inconclusive, their observed lines agreeing better

A. A. ELDRIDGE.