Two methods were finally employed. First, the gallium was melted in a small air-jacketed test-tube in an easily adjustable thermostat, after the fashion of the Beckmann freezing point apparatus, or that so often used by one of us for taking transition temperatures (see for example, Richards and Wells, Proc. Am. Acad., 1902, xxxviii., 431; Z. physik. Chem., 1903, xliii., 465). In this apparatus about 10 g. of the best gallium was used. Its melting point remained constant until only a very few crystals remained.

Another method which gave precisely the same result consisted in determining the temperature at which a glass-stoppered dilatometer, filled with a mixture of crystallised and liquid gallium under water, showed neither decrease nor increase in the volume of its contents with time. About 8.5 g. of crystallised gallium was used. The water was pure and was freshly boiled and quickly cooled; so prepared it does not oxidise gallium seriously in a few hours. The scale on the capillary of the dilatometer showed change of about 0'4 mm. during the melting of o'o g. of gallium. The dilatometer was immersed to the beginning of the index tube in a thermostat at 29°, and the temperature was slowly raised until melting was indicated by the sinking of the water level in the capillary. After perhaps a quarter of the metal had melted, the temperature was slowly lowered until the column remained stationary.

The small thermometer used with each method was very carefully standardised with reference to the Parisian hydrogen scale through one of the Harvard Baudin thermometers, both being read with a cathetometer to within o'002°.

The direct determination of the melting-point in the small Beckmann apparatus gave the value 29.752° for the melting point. With the dilatometer method, at 29.755°, the column slowly descended, whereas at 29745 it slowly rose, indicating the value 29.75° as the true melting point. Since the two methods agreed, this value (29.75°) may be taken as correct for the purest gallium which we have thus far prepared.

IV. Density of Solid and Liquid Gallium. The striking expansion of gallium upon solidifying is of peculiar interest, since this property is possessed by so few substances. Some previous investigators have ascribed it to the presence of impurities (Rudorf, Abegg's "Handbuch," loc. cit.). The work which follows shows, however, that undoubtedly the purest gallium behaves in the same way as the somewhat impure metal first studied.

The form of pycnometer employed was that used recently in the determination of the density of lead isotopes (Richards and Wadsworth, Journ. Am. Chem. Soc., 1916, xxxviii., 222). All the usual precautions were taken to ensure constancy and definiteness of temperature, absence of air bubbles (the pycnometer was filled in a vacuum) and protection from evaporation. Both the solid and liquid gallium were weighed under freshly boiled and quickly cooled pure water. Without precautions, gallium, on solidifying in a pycnometer invariably breaks it by the expansion which then occurs; but if the pycnometer is constantly agitated during slow cooling the crystalline mass does not attach itself to the vessel, nor subject the walls to outward pressure. If solidification

against the walls begins to take place, it is easily overcome by partial melting and resolidification. When the density of the liquid was determined after that of the solid, the pycnometer was placed in an evacuated vessel in order to remove any trace of hydrogen which might have been formed by the long standing under water. Promptness is advisable since even with the purest water a slight action takes place after a number of hours, and, of course, both bubbles of gas and the film of oxide cause a decrease in the observed density of the metal.

The densities of a number of samples were determined. The first sample, A, contained several per cent of indium and a trace of zinc. It was obtained from the hydroxide dissolved in sodium hydroxide, which carries with it indium hydroxide. From Sample B most of the indium had been removed by fractional electrolysis. Sample C had been refractionated electrolytically from a solution 2 N in free sulphuric acid and the metal had been heated in a vacuum for three hours to expel zinc. Sample C, was another preparation of the same kind. Neither was entirely free from impurities. Therefore Sample D was prepared with much greater care by fractional electrolysis, but it began to melt on the surface 0.2° below the true melting point of gallium. Sample E, the purest, has already been described above. (To be continued.)

THE CHEMICAL INDUSTRY IN THE ARGENTINE REPUBLIC.

THREE Argentine companies manufacture sulphuric acid (53°- 56° Baumé) by the chamber process with imported sulphur from Italy, Chile, and the United States. There is a great demand for Sicilian sulphur, since, being of greater purity, and burning without ash. Transport of sulphuric acid by sea is very difficult, owing to long voyages in the tropics, the home industry being thus favoured. Nevertheless, local works find it diffi- . cult to compete with imports.

Nitric and hydrochloric acids, the first in commercial degrees (38° and 42°), are manufactured from Chile saltpetre and native salt. Chemically pure sulphuric and hydrochloric acids are, however, manufactured. The main part of the acetic acid imported is glacial, at 99°. Two works near Cordoba produce pyroligneous products by dry distillation of algarrobo wood, chiefly acetates of lead and lime acetone, acetic acid, which are more expensive than imports. However, the demand exceeds the supply. Local production of tartaric acid is limited to one or two vineyards, wine lees being obtained in large quantity at Mendoza. In 1916 the United States began to export this acid to the Argentine. During the last eight years, imports of citric acid from France, England, Germany, and Italy, represented an average of 132 metric tons per year. The United States commenced exports in 1914. The acids sell better in crystals than powder. Imports of phenic acid are very irregular; England is the chief manufacturer. Tar oil acids are preferred to wash cattle

The oils should contain at least 10 per cent phenol, and some manufacturers say 25 per cent. Boric, carbonic, hydrofluoric, and tannic acids are also imported from Germany, France, England, and, since 1916, the United States has supplied a certain amount of carbonic acid. Gaseous carbonic acid is manufactured locally in breweries. Imports of alum, which were more than 6300 metric tons from 1913 to 1916, fell to 400 tons during the last two years. The United States imports now represent 50 per cent. Hydrated alumina is employed to a limited degree for lakes and colours, textiles and glass. Large quantities of home-made sulphate of aluminium are employed to purify drinking water in Buenos Ayres, &c. Local manufacture is only one-quarter of the imports. Ammonia is chiefly imported. The average imports during the last seven years were 540 metric tons per year, 90 per cent being from America. Liquid ammonia is chiefly English. Arsenic and arsenical compounds are employed for glass, insecticides, and colours. The demand is for white, yellow, and red arsenic (99 per cent), in barrels of 200 kilogrms., and arseniates of lead and soda in boxes of 5 kilogrms. The United States now heads the list of imports.-Chimie et Industrie, February, 1921.

PROCEEDINGS OF SOCIETIES.

MINERALOGICAL SOCIETY.

March 22, 1921.

DR. A. E. H. TUTTON, F.R.S., in the Chair.

THE following papers were read :--

"The Vibrations of a Crystalline Medium." By Prof. H. HILTON.

The paper attempts to give an indication of the kind of vibrations which the molecules of a crystal may be expected to make about their positions of equilibrium. The case of an orthorhombic crystal in the form of a rectangular parallelepiped is considered in detail, and the normal modes of the molecular motion completely determined.

"Augite from Nishigatake, Japan." By Prof. R. QHASHI.

attracted material in order to determine how far they supported the idea that in meteoric stones the ratio of MgO to FeO in the magnesium silicates varied directly with the ratio of Fe to Ni in the nickel-iron. The results showed that the ratios of MgO to FeO in the olivine and pyroxene respectively were about 4 or 5 for Adare, and 2 and 3 for Ensisheim.

Mr. W. BARLOW exhibited models to represent the atomic structure of calcite and arragonite.

NOTES.

ROYAL SOCIETY.-At the Meeting of June 2, the Bakerian Lecture will be delivered by Dr. T. M Lowry, F.R.S., and Dr. P. C. Austin, on "Optical Rotatory Dispersion."

ROYAL INSTITUTION.-A General Meeting of the Members was held on April 4, Sir James Reid, Bart., Vice-President, in the Chair. Mr. S. D. Bles, Mr. S. T. Covington, Mrs. L. W. Dent, Mr. B. G. Donne, Prof. Godfrey, Com. A. C. Goolden, Mr. A. Mallalieu, Mr. W. Marshall, Mr. E. W. T. L. Brewer Williams, and Sir John Wormald were elected Members. The Chairman reported the death of Lord Moulton, a Member of the Institution, and a resolution of condolence with the relatives was passed.

ROYAL INSTITUTION.-On Tuesday next, April 19, at 3 o'clock, Prof. Arthur Keith resumes his Lectures at the Royal Institution on "Darwin's Theory of Man's Origin (In the Light of PresentDay Evidence)"; on Thursday, April 21, Mr. H. S. Foxwell begins a course of two lectures on "Nationalisation and Bureaucracy"; and on April 23, Mr. H. Y. Oldham delivers the first of two lectures on the "Great Epoch of Exploration : (1) Portugal; (2) Spain." The Friday evening Discourse on April 22 will be delivered by Sir James Walker on "Electro-Synthesis in Organic Chemistry," and on April 29 by Sir Frank Dyson on "Advances in Astronomy."

MARKET FOR ANT EXTERMINATORS IN COLOMBIA. --The Commercial Secretary to H. M. Legation at Bogota (Lt.-Col. G. W. Rhys Jenkins) reports that there is an urgent demand in the Santa Marta district for a good ant exterminator for use on the local coffee plantations. It appears that practically all the estates are infested with quantities of large brown ants, which devour everything in sight, including the young coffee shrubs. The Commercial Secretary has received an enquiry as to United Kingdom sources of supply of a suitable exterminator, and particulars will be furnished to any United Kingdom firms interested, on application to the Department of Overseas Trade

"The Chemical Composition of the Adare and Ensisheim Meteorites." By Dr. G. T. PRIOR.

Determination of the amount and composition of the nickeliferous iron in these meteorites had shown that Adare contained 18 per cent, in which Fe: Ni=11, and Ensisheim 3 in which Fe: Ni=34. Analyses were made of the un

Street, London, S. W.1.

THE AUTOMATIC FURNACES, LTD., have issued a small pamphlet presenting some carefully arranged data on the effects of overheating high carbon steel. The Company have for years specialised in precision heat treatment of steel, and now are undertaking to harden samples of manufacturers' own tools for a comparative test against similar articles hardened by the maker. The pamphlet in question and full information can be obtained at the Company's offices, 281-283, Grays Inn Road, W.C.1.

CHEMICAL NEWS April 22, 1921

THE CHEMICAL NEWS.

VOL. CXXII., No. 3184.

181

FURTHER STUDIES CONCERNING

GALLIUM.

ITS ELECTROLYTIC BEHAVIOUR, PURIFICATION, MELTING POINT, DENSITY, COEFFICIENT OF EXPANSION, COMPRESSIBILITY, SURFACE TENSION, AND LATENT HEAT OF FUSION.*

By THEODORE W. RICHARDS and SYLVES ER BOYER. (Continued from p. 177.)

THE results for the density of the less pure material need not be given in full, but will be summarised.

Sample A (containing indium) was found, as an average of a number of determinations, to have the density of 6162 as liquid at 29° and 5'975 as solid at 20°. No great difference was observed between the metal thus solidified in the pycnometer or solidified in air a little below the melting point on a block of paraffin. On the other hand, this impure gallium when solidified by dropping into cold water from a pipet, was found to give a lower result, about 5'90, probably due to included water, but possibly to be referred to the sudden solidification of a more bulky unstable alloy, which may not appear when the cooling is slow. Correcting the values for the solid to 29° by means of the coefficient of expansion, mentioned later, the value at 29° is found to be 5'974. The change in specific gravity of this impure sample on melting is therefore found to be o·188; of Sample B, considerably purer, the change at 29°, corrected in the same fashion, was from 5.893 to 6084, or o'191; for sample C (several preparations) the average for four determinations of the solid was found to be 5.889, whereas four determinations of the liquid gave the value 6'079, a difference of o'190.

Sample D, which was very nearly, but not perfectly, pure, gave, with three different preparations, the following values (two apiece) for the liquid at 29.8°, 6·090, 6·094, 6'093, 6'097, 6·098, 6099; an average of 6095, whereas for the solid at 296°, the following five values were found: 5.907, 5'908, 5'907, 5'905, 5'903, an average of 5'906, and a difference of o189.

*From the Journal of the American Chemical Society, February, 1921

given in full Table I. In these cases, the gallium was introduced into the pycnometer in the liquid state and solidified there under pure boiled water. The density of the solid was determined at 29'65°, and that of the liquid at 29.8° on the hydrogen scale. The respective densities of pure liquid and solid at the melting point are, therefore, 6'0947 and 5'9037, the difference being 0191; and the respective specific volumes of the two states of the metal are o 16408 and 016939 cc., the difference being 0.00531 CC. Allowing for its coefficient of expansion, the density of the solid at 20° may be taken as 5.907 and its atomic volume as 11.85, if the atomic weight is 701.

Evidently, since the further purification between D and E produced no essential change in the density of either solid or liquid, and since the substance last measured had been heated for a long time in vacuum and purified by crystallisation, the densities recorded above for the gallium in the two states cannot be far from the truth. Table II., reiterating the change of density on crystallisation in the several specimens of increasing purity, shows conclusively that the impurities have nothing to do with the change of volume, since the variations are not greater than the possible variation due to experimental error. TABLE II.

Supposing that no great contraction or expansion takes place when indium is dissolved in gallium (which seems reasonable because of the similarity of the two metals), the percentage of indium in Sample A (its specific gravity being 7:31), is easily calculated approximately from the liquid to be 5.6 per cent, and from the solid to be 4.8 per cent. Evidently, Sample A must have contained about 5 per cent of indium. If contraction occurred on the solution of one metal in the other, the percentage thus calculated would be higher than the true value, and vice versa.

9'0452 9'0452

Wt. in

vac.

5.8956

Corrected vol. H2O. 5'9200

5.8950 5.9211

10. 1864 10.1864

5.7103

5'7347

7'4056

16709

6'097