a black deposit of the dichloride or a violet deposit of the trichloride is slowly formed on the surface of the amalgam.

The dichloride dissolves in water and in alcohol with a brown coloration, but it is insoluble in ether. On exposure to the air, the aqueous solution deposits titanic acid; the alcoholic solution remains clear, but contains organic chlorine products. With potassium thiocyanate and ether, titanium dichloride exhibits a delicate and characteristic reaction; the ether acquires a dark-brown colour. Robinson and Hutchings (Amer. Chem. J., 6, 74) describe a method of preparing metallic titanium by the action of sodium on titanic chloride in sealed tubes at 130°. The author finds that this reaction yields a considerable quantity of titanium dichloride, but little if any metallic titanium.

Titanium trichloride mixed with dichloride is precipitated when sodium amalgam is added to a solution of the tetrachloride in hydrochloric acid. The powder dissolves in water, forming a violet solution. When sodium amalgam acts on a solution of potassium titanofluoride and hydrochloric acid, a green solution is produced which does not contain any dichloride. On the further addition of sodium amalgam, the solution becomes colourless, and finally deposits a black precipitate of Ti(OH)2. If the solution contains only a small quantity of acid, the sesquioxide is precipitated.

The three chlorides exhibit the following reactions:-
:-

TiCl4.

Soluble in

TiClą.

TiCl3.

Water or alcohol... readily (brown) readily (violet or

readily (colourless)

Titanium sesquioxide is not formed as Ebelmann states (Ann. Chim. Phys. [3], 20, 394) by the action of hydrogen on titanium dioxide at a red heat. The product has the composition of Ti;O12. W. C. W.

Vanadates of the Alkaline Earths. By O. MANASSE (Chem. Centr., 1886, 773-774).—The normal barium salt, BaV20, + H2O, is obtained by adding barium chloride to a solution of the normal potassium salt. Precipitated from the acid potassium salt, it is often obtained as a yellowish-white powder mixed with orange-red crystals. An acid barium salt, Ba3V10028 + 19H2O, is formed by precipitating the acid potassium salt with a soluble barium salt. The precipitate first formed slowly changes into red-yellow crystals, mixed with the yellowish-white powder of the preceding salt, the solution gradually becomes quite colourless. It is soluble in water to the extent of 1 in 5000. The salt is obtained in beautiful, red, prismatic crystals, when equivalent amounts of potassium divanadate and barium chloride are mixed with a large excess of acetic acid. Barium orthovanadate, BaVO, cannot be obtained, owing to its great instability; it is decomposed into barium pyrovanadate and barium hydroxide at the time of its formation.

Strontium Salts.-The normal salt, SrV2O. + 4H2O, is described, and also acid salts, having the composition Sr3V8023 + 14H2O, Sr V14039 +30H2O, and SrV8О2 + 11H2O. When potassium divanadate and strontium chloride are allowed to crystallise together, the following double salts are obtained:-a and B, K,VO39 + 20H2O+ 3(Sr.V14O39 +20H2O); 3(Sr,V1109 + 30H,0) + KV11039 + 30H2O; and a y-salt, K.V14O39 + 18H2O + Sг ̧V1Óз9 + 18H2O, and also KV10027 +12H2O.

Calcium Salts.-The normal salt, CaV2O, + 4H2O, is obtained by allowing a mixture of normal potassium vanadate and calcium chloride to evaporate. The acid salt, Ca,VO23 + 15H2O, is formed when equivalent amounts of potassium divanadate and calcium chloride are mixed with excess of acetic acid. The tetravanadate, CaV,O+6H2O, is obtained in red crystals by evaporating potassium divanadate and calcium chloride. Other salts prepared were:-Ca,V14028 + 7H2O, CaзV16043 + 26H2O, and CaVO + 4K2VO + 22H2O, potassium calcium tetravanadate.

Magnesium Salts.-The normal salt, MgV,O. + 6H2O, is prepared by boiling finely powdered vanadic acid with excess of magnesium carbonate. Acid salts, 2(MgaVe017) + 19H2O and Mg3V16028 + 28H2O, were also prepared.

When a few drops of nitric or sulphuric acid is added to these salts, a hydrate of vanadic acid, HV2O6, is formed, which is much more soluble than the crystallised vanadic acid.

The author determined the vanadic acid in the above salts by reducing the acid and titrating with permanganate. He found that the reduction of vanadic acid is best effected by hydrogen sulphide, or by sulphurous anhydride with a simultaneous passage of carbonic anhydride.

A compound, VO17, was obtained as a dark blue or black powder with metallic lustre, by dissolving vanadic acid containing oxide in potassium hydroxide. It can be regarded as a mixture of V2Os+ 3V20, or of V10, + 2V204.

G. H. M.

Valency of Bismuth. By A. MICHAELIS (Ber., 20, 52—54).— See this vol., p. 368.

Action of Hydrogen Peroxide on Bismuth Salts. By K. HASEBROEK (Ber., 20, 213-218).-When commercial (3 per cent.) hydrogen peroxide, rendered alkaline with ammonia (or potash), is added to a bismuth salt and the mixture gently warmed, evolution of oxygen occurs and a characteristic orange-yellow precipitate of bismuth pentoxide is formed; the reaction is of such delicacy that 1 part of bismuth in 100,000 can be detected by its means. To obtain the pentoxide in the pure state, it is necessary to work in an atmosphere free from carbonic anhydride, otherwise absorption of the gas and formation of bismuthyl carbonate takes place. The air-dried bismuth pentoxide contains less water than corresponds with 1 mol. HO; it is a bright orange-yellow, amorphous powder, insoluble in water and alkalis, but soluble in dilute acids, solutions of varying shades of red being formed when dilute nitric and sulphuric acids are used as solvents. When heated with alkalis or alone, it becomes brown, and by further heating melts to form crystalline bismuth trioxide. Treated with hydrochloric and sulphuric acids, it evolves chlorine and oxygen respectively, whilst hydrogen peroxide is readily decomposed by it.

W. P. W.

Atomic Weight of Gold. By G. KRÜSS (Ber., 20, 205-210).— The collective results obtained in a determination of the atomic weight of gold are given. Two salts were employed, the neutral auric chloride, AuCl, (in solution), and potassium gold bromide, KAuBr. The gold chloride was obtained by passing chlorine over finely divided gold; to remove excess of chlorine, the product was exposed for a day to air dried by passage over sulphuric acid and phosphorus pentoxide, then allowed to remain for several weeks over frequently renewed soda-lime, and finally decomposed by water into a solution of neutral auric chloride. The potassium gold bromide was prepared by adding finely divided gold to potassium bromide and bromine in aqueous solution; the salt is anhydrous and only very slightly hygroscopic.

The values found for the atomic weight (O = 15.96) are given in the table (p. 340).

The author adopts the value 196-64 as that of the atomic weight of gold. W. P. W

Sublimed Auric Chloride. By G. KRÜSS (Ber., 20, 211-213). -When a continuous current of dry chlorine is passed over powdered gold, and the temperature gradually raised, it is found that at 140° a small quantity of reddish-brown vapour is formed which condenses to a yellowish-red sublimate; if the temperature is kept constant the sublimation soon ceases, and the gold is entirely converted into aurous auric chloride. When the temperature is raised to 180-190° this compound decomposes, chlorine is evolved, and green aurous chloride and a small quantity of volatile gold chloride are formed. At 220-230° a small quantity of gold chloride again volatilises, whilst the aurous chloride decomposes into gold and chlorine, and at 300° the gold, contrary to Debray's statement (Compt. rend., 69, 985), remains unaltered in the current of chlorine. On allowing the temperature to fall these changes take place in the reverse order. The amount of volatile gold chloride, which analysis shows to be auric chloride, AuCl,, and not a higher chloride as stated by Prat (Compt. rend., 70, 840), is very small; thus, 100 repetitions of the process with a layer of powdered gold 30-35 cm. long, yielded less than 0-12 gram of auric chloride in the form of a sublimate consisting of lustrous, reddish-brown needles. W. P. W.

Mineralogical Chemistry.

Amorphous Carbon (Graphitoïd) in the Saxon Erzgebirge. By A. SAUER (Zeit. Kryst. Min., 12, 527).-Amorphous carbon is widely distributed throughout the mica schists and phillites of the Erzgebirge. It is not identical with graphite or with anthracite. The author finds that it is identical with the amorphous carbon described by Inostranzeff (Abstr., 1881, 357), and suggests for the substance the name of graphitoïd. The analysis gave the following results :

Crystallisation of Native Copper. By E. S. DANA (Amer. J. Sci., 32, 413-429).-This elaborate memoir is based on a careful study of the beautiful collection of 60 specimens of native copper from Lake Superior, in the possession of Mr. C. S. Bement. The author also had the use of a large number of specimens belonging to the cabinets of Professor G. J. Brush, and of the Yale College

Museum. The paper is accompanied by illustrations of 54 varieties of crystalline forms of native copper. The planes observed by the author are coО∞, ∞s0, 0, ∞04, 0, ∞02, ∞0, 303, 202, 10, 110, 604. Of these, the trisoctahedron 202, and the hexoctahedra 110 and 604, are new to the species. B. H. B.

Crystallographic Notes. By W. G. BROWN (Amer. J. Sci., 32, 377-380).-1. Artificial Copper Crystals.-In a cell of a Calland gravity battery which had remained undisturbed for three months, there was found attached to the insulating wire coming from the copper plate, at the point where it passed the zinc plate, a pendant mass consisting almost entirely of copper crystals. The crystals were evidently accidentally formed by electrolysis. The upper part of the mass consists of spongy dendritic copper, whilst the lower and larger portion consists of definite copper crystals. By far the greater number of the crystals are twins, either according to the spinel law, twinning plane an octahedral face, or the twinning plane is that of a trapezohedron forming polysynthetic twins.

2. Artificial Crystallised Cuprite.-On the outside of the spongy portion of the copper described above, there is a purple layer of crystals of cuprous oxide. Individually, they are of a dark ruby-red colour; the forms present being combinations of the octahedron and cube. Another example of crystallised cuprous oxide was found on a German-silver camp spoon picked up on the site of a magazine constructed in 1863, on Morris Island, South Carolina. The short time in which the crystals have formed seems a noteworthy feature of this The spoon could not have been exposed more than 20

Occurrence.

years.

3. Crystallised Cerussite.-At Morris Island, among the debris of a battle-field, a dozen Minié balls were picked up. Upon various parts of the surface of these were patches of lead carbonate, sometimes amorphous and sometimes acicular. The balls had not been fired. It is possible that the formation of the carbonate was hastened by the rapid oxidation of the lead by the nitre of the gunpowder of the cartridge in the presence of moisture. B. H. B.

Mineralogical Notes. By A. GENTH (Zeit. Kryst. Min., 12, 487492). The author gives descriptions and analyses of the following minerals:-1. Native tin from the Aberfoil River, New South Wales; 2. Joseite and tetradymite from San José, Minas Geraes; 3. Galenobismuthite containing selenium from Falun, Sweden; 4. Silver bismuth glance from Lake City, Colorado; 5. Cosalite from the Alaska and Gladiator mines in Ouray Co., Colorada; 6. Beegerite from the Treasury Vault mine, Summit Co., Colorado; 7. Fahlerz from the Governor Pitkin mine, Lake City, Colorada; 8. Polybasite from the Terrible Lode mine, Clear Creek Co., Colorado; 9. Arsenical pyrites from North Alabama; 10. Brucite from Berks Co., Pennsylvania; 11. Ilmenite and oligoclase from the Carter mine, North Carolina; 12. Topaz from Stoneham, Maine; 13. Orthoclase from French Creek, Chester Co., Pennsylvania; 14. Muscovite pseudomorphs after nepheline from Wakefield in Canada; 15. Stilp