Determination of arsenic. I. BANG (Biochem. Z., 1925, 161, 195-209).-The material is incinerated with sulphuric acid and nitric acid is dropped in from a special container. After driving off excess of the latter, reduction is effected by ferrous ammonium sulphate in the presence of potassium bromide and chloride, followed by distillation into sodium hydroxide. To the distillate are added sodium hydrogen carbonate and a crystal of potassium iodide and the solution is titrated with 0.005N-iodine.

C. RIMINGTON.

II.

Azido-dithiocarbonic acid. Determination of the azido-dithiocarbonate radical.

A. W. BROWNE and G. B. L. SMITH (J. Amer. Chem. Soc., 1925, 47, 2698-2702).-The azido-dithiocarbonate radical resembles the halogenoid radicals. and may be determined by titration of the free acid with alkali, by precipitating and weighing the silver salt, by converting this into silver chloride and weighing, by titrating with silver nitrate solution by Gay-Lussac's or by Volhard's method, or by titration with alcoholic iodine. These all give slightly low results, the best being Volhard's method.

A. GEAKE. Micro-determination of sodium. E. TSCHOPP (Helv. Chim. Acta, 1925, 8, 893-900; cf. A., 1921, ii, 655; 1924, ii, 123, 413, 500).-Volumetric, colorimetric, and electrolytic methods for micro- and semimicro-analysis of biochemical products are described in detail. The complex sodium cæsium bismuth nitrite, 6NaNO2,9CsÑO2,5Bi(NO2)3, may be precipitated under suitable conditions. A solution of the complex on electrolysis deposits the bismuth quantitatively at the cathode. R. A. MORTON.

Electrometric study of the separation of silver iodide, bromide, and chloride. H. T. S. BRITTON (Analyst, 1925, 50, 601-604).-Since the solubility products of the iodide, bromide, and chloride of silver are of the order 10-16, 10-13, and 10-10, electrometric titrations of ammoniacal solutions of the three halides with silver nitrate solutions, in which the concentrations of the ammonia and silver nitrate are suitably varied, enable the continuous change in silver-ion concentration to be made sufficiently gradual to permit of the complete precipitation of one halide before the separation of the next begins. A silver electrode placed in the beaker containing the ammoniacal halide solution was connected through a salt bridge of saturated potassium nitrate solution with a normal calomel electrode. The E.M.F. of this combination were measured after the addition of each amount of silver nitrate. Using 3-37N-ammonia solution and 0.5N-silver nitrate solution, the bromide and chloride were separated quantitatively from the bromide, but not the iodide. With 0.1N-silver nitrate solution and (i) 18N-, (ii) 8N-, and (iii) 4Nammonia solution it was possible with (i) and (ii) completely to separate the silver iodide (by filtration through a Gooch crucible), but not with (iii). The experiments are illustrated by means of curves. The following mean values at 18° were found, [Ag•][I'] = 7 x 10-17; [Ag'][Br'] = 2.8 x 10-13; [Ag.][CI']=1.8 × 10-10. D. G. HEWER.

Micro-determination of calcium and magnesium in organic liquids. L. CONDORELLI (Rend. Accad. Sci. Fis. Mat. Napoli, 1925, [iii], 31, 73-83). -When precipitated from an albuminous solution, calcium oxalate is included in a gelatinous envelope which resists the action of water and should be washed with dilute ammonia solution. The micro-methods developed by the author for determining calcium and magnesium in organic liquids are as follows: 1 c.c. of the blood or other liquid is evaporated to dryness and ashed in a platinum crucible. The ash is dissolved in 0-5 c.c. of N-hydrochloric acid and the solution transferred, by means of a thin pipette bent at right angles, to a 12 c.c. centrifuge tube, the crucible being washed successively with 0-5 c.c. of the acid, and 0.5 c.c. of water, and twice with 0.5 c.c. of 3% oxalic centrifuge tube. After addition of a drop of methylacid solution, and the wash liquors also placed in the red solution, the tube is immersed in a boiling waterbath and the solution neutralised by slow addition of N-ammonia solution from a micro-burette. Next 9% ammonium oxalate solution are added and the 1 c.c. of ammonium chloride solution and 1 c.c. of liquid is stirred well with a small glass rod, which is then washed into the tube with sufficient water to bring the total volume to 6 c.c. the total volume to 6 c.c. After 24 hrs., the liquid is centrifuged for 20 min., the clear solution being pipetted into a dry test-tube and replaced by 5 c.c. of water, which is removed after a brief centrifuging; this operation is repeated with two further quantities of water. Five c.c. of N-sulphuric acid are then added and, with the tube immersed in a boiling waterbath, the solution is titrated with 0.01N- or 0.005Npermanganate. With a liquid such as blood which contains iron, care must be taken that this is not precipitated with the calcium. To this end, the acid with which the ash is treated is introduced into the bottom of the centrifuge tube and centrifuged rapidly for 1-2 min., the clear liquid being then pipetted off into the precipitation tube; each of the washing liquids is treated similarly, the subsequent operations being as described above.

In an aliquot part of the liquid decanted from the calcium precipitate, the magnesium is precipitated as magnesium ammonium phosphate by addition of 1 c.c. of 2.5% ammonium phosphate solution and 2 c.c. of 10% ammonia solution. After 24 hrs., the crystalline precipitate is washed with three separate quantities of 5 c.c. of 10% ammonia, which are pipetted off after centrifugation. About 3 c.c. of the last wash water are neutralised with dilute sulphuric acid and treated with 1 c.c. of molybdic acid solution (50 g. of pure ammonium molybdate dissolved in 1000 c.c. of cold N-sulphuric acid) and 2 c.c. of quinol solution (20 g. dissolved in 1000 c.c. of water and 1 c.c. of concentrated sulphuric acid added); if no blue or green colour develops in 5 min., the washing is complete. The precipitate is then dissolved in 1 c.c. of N-sulphuric acid, which, together with five or six quantities of 1 c.c. of wash water, is transferred to a 25 c.c. flask, the whole being treated with 1 c.c. of the molybdic acid solution and 2 c.c. of the quinol solution, and, after 5 min., with 10 c.c. of a solution prepared by filtering a mixture of 2000 c.c. of 20% sodium carbonate solution with 500 c.c. of 15% sodium

sulphite solution. The solution is made up to 25 c.c. with water and mixed, the colour being compared in a Duboscq colorimeter with that obtained similarly from 10 c.c. of N/5000-potassium dihydrogen phosphate solution, which corresponds with 0.04864 mg. of magnesium. These methods yield highly accurate results. T. H. POPE.

Equilibrium between metal chloride, hydrogen sulphide, metal sulphide, and hydrogen chloride and its use in separating metals. Cadmium, bismuth, and lead. W. MANCHOT [with G. GRASSL and A. SCHREEBERGER] (Z. anal. Chem., 1925, 67, 177-195).—The maximum quantity of hydrogen chloride that may be present in a solution for complete precipitation of cadmium by hydrogen sulphide at 20° is about 5 g. per 100 c.c., and of bismuth, 14.5 g. per 100 c.c. No precipitation at all occurs when the hydrogen chloride exceeds 9.5 g. per 100 c.c. in the case of cadmium, 16.5 g. per 100 c.c. in the case of bismuth, or 7.2 g. per 100 c.c. in the case of lead. With hot solutions, these values are considerably reduced, so that by regulation of the temperature and acidity it is possible to separate cadmium from antimony and mercury, and bismuth and lead from mercury by treatment of the solution with hydrogen sulphide. As cadmium sulphide has a tendency to be carried down by mercury sulphide, it is advisable to use an acidity of 19 g. of hydrogen chloride per 100 c.c. Antimony is separated from cadmium by treating a hot solution containing 8 g. of hydrogen chloride per 100 c.c. with the gas so as to obtain a black, crystalline antimony sulphide, adding an equal volume of hot water, filtering hot through a Gooch crucible, and washing the precipitate with hot water containing 4.5 g. of hydrogen chloride per 100 c.c. A. R. POWELL.

Iodometric micro-determination of copper. N. SCHOORL and H. BEGEMANN (Rec. trav. chim., 1925, 44, 1077-1086).-A micro-determination of copper by the ordinary iodometric method is described, the titration being carried out with a 0.001 N-thiosulphate solution, and the end-point observed in a dark room with special white illumination and a comparison solution. Liberation of iodine by atmospheric oxidation is prevented by the use of acetic acid for acidification. The titration is satisfactory in presence of iron salts if sodium phosphate or fluoride is added. The separation of the copper by electrodeposition on the microanalytical scale is also described, the metal being then dissolved in nitric acid, the solution evaporated to dryness, and the copper determined by titration. Amounts of 0.05–0.5 mg. of copper are determined with an accuracy of 1%. G. M. BENNETT.

the passage of a current of 13-1.5 milliamp. at 1.5-2.0 volts for a few minutes is sufficient to produce a silver-white amalgam on the cathode. This is examined microscopically, and 0.5 mg. of mercury per litre can be detected. By evaporating 25-30 c.c. of the test solution to a few drops in a vacuum desiccator over sulphuric acid, the sensitivity of the test is increased a thousandfold. A. GEAKE.

Volumetric determination of iron in hydrochloric acid solution. W. MANCHOT and F. OBERHAUSER (Z. anal. Chem., 1925, 67, 196—197).— The ferric chloride solution is reduced with sulphur dioxide or with metallic cadmium and after removal of the excess of reducing agent, 1 g. of potassium fluoride and 20 c.c. of 15% hydrochloric acid are added, followed by an excess of 0-1-bromine in N-potassium bromide solution. The excess of bromine is determined by titration with arsenious acid. Phosphoric acid may be used as a decolorising agent in place of potassium fluoride. A. R. POWELL.

Determination of ferrous oxide in insoluble silicates. O. HACKL (Z. anal. Chem., 1925, 67,

197-204). Decomposition of silicates by means of sulphuric and hydrofluoric acids in a neutral atmosphere and titration of the ferrous sulphate solution obtained gives low results for ferrous oxide. It is suggested that more accurate values would be obtained by the addition of a known weight of potassium dichromate during the decomposition, followed by a gravimetric, volumetric, or colorimetric deterthat fluorides do not affect the precipitation of ferric Preliminary tests indicate hydroxide by ammonia or potassium carbonate, the

mination of the excess.

reaction between chromate and silver nitrate in acetic acid solution, or the colorimetric determination of chromium. Ammonia precipitation of the iron, however, results in a slight loss of chromate, but potassium carbonate yields a good separation of ferric hydroxide and chromate. A. R. POWELL.

[Detection of nickel.] L. A. TEST and D. L. SCOLES (Proc. Indiana Acad. Sci., 1924, 34, 163164). The test for nickel dependent on the retardation of the change to pink cobaltous hydroxide of the blue basic cobalt compound obtained on addition of excess of sodium hydroxide to a solution of a cobalt salt is not trustworthy unless the sodium hydroxide solution is freshly prepared or free from silicate.

CHEMICAL ABSTRACTS.

Determination of uranium. G. E. F. LUNDELL and H. B. KNOWLES (J. Amer. Chem. Soc., 1925, 47, 2637-2644).-See B., 1925, 1016.

Delicate colour reactions for zirconium, hafnium, and fluorine by means of hydroxyanthraquinones. J. H. DE BOER (Rec. trav. chim., 1925, 44, 1071-1076).—The colour test for zirconium (or hafnium) with alizarinsulphonic acid (A., 1924, ii, 705) depends on the formation of a complex of 1 atom of zirconium to 1 mol. of the sulphonic acid. Addition of an excess of hydrochloric acid does not discharge the colour, but the resulting solution is a sensitive reagent for fluorides, detecting as little as 10 g. of fluoride ions in 1 c.c. of water, the reddish

violet colour being turned yellow by removal of the zirconium as a complex fluoride. A method is described, based on these facts, for the volumetric determination of zirconium by titration with a fluoride solution, using alizarinsulphonic acid as indicator. Fairly accurate results are obtained by an empirical comparative method, the titration being carried out side by side with a solution of a known amount of zirconium having a similar concentration of hydrochloric acid. The results are independent of the presence of titanium, cerium, erbium, yttrium, manganese, or aluminium. Ferric salts must be reduced before titration by addition of an excess of stannous

chloride.

The reactions of a number of other hydroxyanthraquinone derivatives with zirconium are described; colorations are produced wherever the substance contains two hydroxyl groups in the 12-positions. Hafnium gives similar reactions, but a slight difference is detected in the reaction with rufigallic acid (1:2:35:6: 7-hexahydroxyanthraquinone), the colour developed being somewhat more easily discharged by hydrochloric acid with hafnium than with zirconium. G. M. BENNETT.

Crystalline digallic acid as a precipitant. M. NIERENSTEIN (Analyst, 1925, 50, 604-605). Since all recent work goes to show that gallotannin is a mixture, and even when purified by the usual methods contains free gallic acid, its replacement in analysis is suggested by such a trustworthy crystalline reagent as digallic acid, which possesses all the properties of gallotannin and is readily prepared from gallotannin or gallic acid.

D. G. HEWER.

Precision X-ray spectrometer for chemical Precision X-ray spectrometer for chemical investigations. G. L. CLARK, H. C. WEBER, and R. L. HERSHEY (Ind. Eng. Chem., 1925, 17, 11471150). A description is given of the source of power, tube housing, spectrometer, make-and-break switch, auxiliary shielding, ionisation chamber, and potentio:

metric control of the installation. Current delivered from a transformer at 240,000 volts is rectified by a rotating full-wave synchronous rectifier. The tube used is a 140,000-volt Coolidge tube with a tungsten target; interchangeable with this is a water-cooled molybdenum target tube operated at 35,000 volts. Angles can be read to 0-2 sec. The quadrant electrometer needle is charged to 45 volts, and the ionisation chamber to 90-200 volts.

Precise automatic pressure regulator. L. E. DAWSON (J. Physical Chem., 1925, 29, 1408-1414).-The regulator differs slightly in form from that previously described by the author (A., 1924, ii, 246), but the same principle is involved. The regulator may now be used for pressures greater, as well as less, than atmospheric. Improvements in adjustment give a high degree of accuracy in regulation. No adjustment of the apparatus is needed to compensate differences in atmospheric pressure when the pressure is measured with a manometer which has one end open to the air and the other connected with the system to be maintained under constant pressure. This only holds, however, when the latter is mechanically produced; in distillation processes, for example, changes in atmospheric pressure must be counterbalanced by changes in the setting of the regulator. Pressures approximately 100-250 cm. of water up to 50 cm. of mercury above atmospheric can be maintained with no fluctuation in the manometer levels. A system for the regulation of pressure when the source of pressure fluctuates is also described. L. S. THEOBALD.

Viscosity measurements by means of Mohr's balance. C. STICH (Pharm. Zentr., 1925, 66, 781783).-A simple method of measuring relative viscosities is described, in which the rate at which a horizontal plate, suspended in the liquid from one arm of the balance, is displaced by a suitable weight on the opposite pan, is determined.

N. H. HARTSHORNE.

Analytical crucible. J. D. M. SMITH (J.S.C.I., 1925, 44, 539-540T).-Practically all the disadvantages of ordinary crucible methods of analysis result directly from the exterior-fitting lids. These disadvantages are all simultaneously overcome by the use of a lid having the vertical flange serrated instead of entire, the lid fitting wholly inside the crucible. Used with squat-shape crucibles of silica or platinum, this completely obviates the usual losses due to decrepitation of solids, spurting of liquids and pasty masses, and the "creeping of heated volatile liquids over the outside of the crucible. Details are given of the application of such crucibles to the determination of nearly all the known metals as anhydrous sulphates, oxides, or free elements, whether in the form of pure compounds or as undried precipitates obtained by wet processes of separation.

Circulation apparatus for gases. N. G. CHATTERJI and G. I. FINCH (J.C.S., 1925, 127, 2464-2466).-An improvement of the apparatus described by Bone and Wheeler (ibid., 1903, 83, 1074). It may also be used for the circulation of liquids. R. CUTHILL.

Mechanism of chemical gas reactions. M. KRÖGER (Z. physikal. Chem., 1925, 117, 387-431).Apparatus is described for volatilising solids and for obtaining part of the vapour in "unidimensional " streams, i.e., in which no collisions of the molecules take place, either between themselves or between the molecules and the walls of the containing vessels. Such a stream of iodine, passed through a platinum

cylinder heated at 1100°, was found, within the limits of experimental error, to have suffered no decomposition into atoms. It would appear generally that a unimolecular reaction cannot take place if collisions cannot occur. Arrangements are described by which streams of two different vapours may be made to impinge on one another and the products examined. Preliminary experiments have been made with the vapours of cadmium and iodine, cadmium and sulphur, and cadmium and selenium.

L. F. GILBERT.

Apparatus for demonstrating thermal transformations of steels and anomalies in special alloys. P. CHEVENARD (J. Phys. Radium, 1925, [vi], 6, 264-272).-Apparatus has been designed primarily for illustrating lecture courses in metallurgy. The experiments include a neat method for demonstrating recalescence, using a series of rings of ferronickel (25% Ni) mounted alternately with steel rings on a spindle; the whole is first heated to redness and on cooling there is a very decided difference in brilliancy between the alternate rings. A wire dilatometer enables an audience to follow the change in volume of an alloy. The polymorphic transformations of a specimen are contrasted with the

expansion of baros (10% chromium, 90% nickel). This alloy has no thermal anomaly, and its expansion is optically compared with that of the specimen and anomalies indicated on a screen. The apparatus is suitable for quantitative investigations on the mechanism of the tempering process. Another instrument, the thermomagnetic tracer, records the magnetisation-temperature curve of a ferromagnetic substance. The thermo-elastic oscillometer and a galvanoscope for wires under torsion exhibit in a striking manner the positive thermo-elastic anomaly of certain ferronickel alloys. In certain iron-nickel alloys such as "invar" the form which is stable in the heated state is not only more dense but also more rigid than the form stable in the cold. Whilst in "invar the anomalous contraction due to the magnetic transformation scarcely compensates for the normal expansion, the anomalous increase in rigidity is much greater than the normal decrease. Thus the elastic modulus may increase with temperature. This phenomenon is ingeniously demonstrated. In elinvar," the anomaly has been modified by suitably changing the alloy so that no indication of its existence is found by using this new alloy with the apparatus devised. None of the devices is unduly complicated. R. A. MORTON.

Meteorite (eucrite) fallen in the Haute-Volta on June 27th, 1924. A. LACROIX (Compt. rend., 1925, 181, 745-749).—The meteorite, belonging to the eucrite class, was practically unbroken by the fall, and weighed 18 kg. It contained 3.48% of free silica; 34.93% of total felspars (orthose 1.11%, alkite 1.57%, anorthite 32-25%); 57-94% of total pyroxenes (CaSiO3, 9-40%, MgSiO3, 16-20%, FeSiO3, 32-34%); 0.46% of magnetite; 1-22% of ilmenite; 0.57% of pyrrhotite; 1-25% of chromite; 0.34% of apatite. The general composition agrees closely with that of the eucrite type of meteors which have occasionally fallen elsewhere. W. HUME-ROTHERY.

Chlorite in Transvaal marundite. J. ORCEL (Compt. rend., 1925, 181, 795-797).-Particulars are given of the analysis of optically active lamellar chlorite of greyish-green colour, found in marundite corundum-bearing rock, near Mays Kraal (East Transvaal). The specimen was of the grochauite type, of formula 4SiO2,2Al2O3,7MgO,6H2O, but with 9.42% of ferrous oxide in place of magnesia. The composition agreed closely with that of a chlorite from North Carolina, but with the ratio MgO: FeO lower in the Transvaal specimen, indicating the existence of a ferrous grochauite intermediate between true grochauite (MgO : Fe0=10 approximately) and the ripidolites or perromagnesium perchlorites (MgO: FeO<3).

W. HUME-ROTHERY.

Identity of fibrillar limonite with goethite. R. GAUBERT (Compt. rend., 1925, 181, 869—872).— Investigation of the effect of temperature on the

variation of the dispersion of the optical axes of goethite and of limonite and also of the action of hydrochloric and hydrofluoric acids on these minerals indicates that fibrillar limonite is impure goethite. The former phenomenon also serves to distinguish between goethite and lepidocrocite.

S. K. TWEEDY.

Variation with depth of certain salts utilised in plant growth in the sea. W. R. G. ATKINS and H. W. HARVEY (Nature, 1925, 116, 784-785).There is a marked decrease in på value (8.35—7.87) with increasing depth (0-3000 m.), due mainly to the removal of carbon dioxide by the phytoplankton, but partly to the high surface temperature altering the equilibrium between hydrogen carbonate and atmospheric carbon dioxide. Phosphate first appears at a depth of 50-75 m. and increases to 88 mg. per m.3 Nitrate nitrogen varies from 15 mg. per m.3 at the surface to 6 mg. at 75 m. and 265 mg. at 3000 m.; silica varies from 220 mg. at the surface to 1200 mg. at 3000 m. A. A. ELDRIDGE.

Sulphoselenium from Vulcano (Aeolian Islands). E. QUERCIGH (Rend. Accad. Sci. Fis. Mat. Napoli, 1925, [iii], 31, 65—69).—Amorphous vitreous specimens of sulphoselenium from the fumaroles of Vulcano have d below 4 and no between 2.544 and 2.675, corresponding with 83-95% Se (cf. Merwin and Larsen, Amer. J. Sci., 1912, 34, 43); they begin to soften below 100° and become gradually pasty and then liquid. Unlike the material described by Stromeyer (Pogg. Ann., 1824, ii, 410), they are free

from ammonium chloride. Crystalline sulphoseleniums, containing 98-71-99-06% S, 0-83-1-03% Se, and traces of tellurium and arsenic, are also found, and probably consist of mixed rhombic crystals (cf. Ringer, A., 1902, ii, 651; Pellini, A., 1909, ii, 805). T. H. POPE.

Relationships between radioactivity, density, helium content, and hafnium content in certain zircons. A. PIUTTI (Rend. Accad. Sci. Fis. Mat. Napoli, 1925, [iii], 31, 72-73).-No relationship is apparent between the hafnium contents and densities of zircons from different localities containing comparatively small proportions of hafnium, but the author's results confirm von Hevesy's observation that a connexion

exists between the percentages of hafnium present and the radioactivities of the minerals. Vesuvian zircon contains about 0.7% HfO2. T. H. POPE.

Structure of a- and ß-quartz. W. H. BRAGG and R. E. GIBBS.-See this vol., 13.

Barytes, cellestine, and anglesite. R. W. JAMES and W. A. WOOD.-See this vol., 13.

Magnetic rotatory power of paramagnetic vol., 14. minerals. H. K. ONNES and others.-See this

Gel of protein type (found) in the kieselguhr stratum of the Lüneburger Heide. F. V. VON HAHN.-See this vol., 23.

X-Rays and the constitution of the hydrocarbons from paraffin wax. S. H. PIPER, D. BROWN, and S. DYMENT (J.C.S., 1925, 127, 21942200). The distances apart of the main cleavage. planes of the crystals of seven hydrocarbons, isolated from Scotch paraffin wax by Francis (ibid., 1922, 121, 1529), have been determined by X-ray examination. The reflexion spectra (from specimens melted on glass strips) yielded by iron K-rays from a Shearer tube were photographed on a Müller spectrograph. Comparative photographs were obtained from synthetic triacontane, dotriacontane, and tetratriacontane. With these as standards, values were obtained graphically for the number of carbon atoms in the molecule of each unknown hydrocarbon. These were found to be whole numbers, indicating that the paraffin wax hydrocarbons are of normal constitution, without sidechains. The respective numbers obtained were: 22, 24, 25, 26, 28, 29, and 31. The mol. wts. of the seven hydrocarbons, determined by the ebullioscope method of Menzies and Wright (A., 1921, ii, 622), indicated the following numbers of carbon atoms in the respective hydrocarbon molecules: 23, 24, 26, 26, 28, 29, 31. A. DAVIDSON.

Hydrogenation of ethylene in presence of metallic calcium. R. N. PEASE and L. STEWART (J. Amer. Chem. Soc., 1925, 47, 2763-2766).Reaction between ethylene and hydrogen was not observed when the mixed gases were bubbled through liquid sodium-potassium alloy at 300°. Hydrogen is adsorbed by metallic calcium at 200°, formation of the hydride apparently not taking place at this temperature. When a mixture of equal volumes of hydrogen and ethylene is passed, at 200°, over metallic calcium which has been previously saturated with hydrogen, rapid combination takes place and the resultant gas contains up to 92.7% of ethane. Ethane is also formed, but at a much slower rate, when ethylene alone is passed, at 200°, over the metal saturated with hydrogen, and a portion of the gases remains adsorbed on the catalyst. On treating the gas-saturated catalyst with hydrogen, a small proportion of a hydrocarbon, apparently methane, was obtained. The reaction appears to be of zero

order, and it is concluded that the intermediate formation of calcium hydride is improbable. F. G. WILLSON.

Stability of hexa-substituted ethanes. A. E. GRAY and C. S. MARVEL (J. Amer. Chem. Soc., 1925, 47, 2796-2802).-Dicyclohexylphenylchloromethane, m. p. 122°, obtained by the action of acetyl chloride in benzene on dicyclohexylphenylcarbinol (cf. Godchot, A., 1910, i, 104), yields s.-tetracyclohexyldiphenylethane, b. p. 170-171°/1.75 mm., n 1.6710, when treated with molecular silver in toluene. This hydrocarbon is not oxidised when heated in air at 330° for 30 min. Reaction between ethyl hexahydrobenzoate and excess of magnesium cyclohexyl bromide affords a product, b. p. 130-133°/2 mm., presumably impure dicyclohexylcarbinol. Tri-yy-dimethylAa-butinenyl chloride, m. p. 170-171°, obtained by the action of phosphorus trichloride on the corresponding carbinol in toluene, darkens on keeping, even in absence of light. When heated in toluene, it is converted into a yellow, crystalline solid, m. p. 210-220°, which contains chlorine. The chloride does not react with metals at the ordinary temperature, except with partial conversion into the above solid. When heated on the steam-bath with molecular silver in toluene for 15 hrs., the resulting mixture absorbs oxygen corresponding with the production of the free radical tri-yy-dimethyl-Aa-butinenylmethyl in 50% yield, the oxidised mixture containing the above solid, m. p. 210-220°, and a resinous, nonpurifiable material. The existence of this purely aliphatic, free radical, as well as the stability of the above tetracyclohexyldiphenylethane, are in accord with Lewis' explanation of the existence of free radicals. F. G. WILLSON.

Preparation of hexachloroethane by the chlorination of ethylene. G. E. MILLER (Ind. Eng. Chem., 1925, 17, 1182-1183).-See B., 1925, 1011.

Electropyrogenetic decomposition of ethyl alcohol. F. G. MÜLLER (Helv. Chim. Acta, 1925, 8, 821-825).-When an electric arc is maintained