hygroscopic nature, to conduct the experiments in a closed space, and an apparatus has been devised and is described by means of which this is easily rendered possible. A thermometer divided to one-tenth of a degree was used, and this was found to give sufficiently accurate results. Experiments are detailed with naphthalene and picric acid, acetanilide and benzil, and the isomeric diacetyl-compounds of 3-diphenylglyoxime.

In conclusion, the author recommends the use of Raoult's method in those cases where vapour-density determinations are not possible. H. C. Isonitroso-compounds. By E. BECKMANN (Ber., 21, 766-769). -The application to the ketoximes of Raoult's method for the determination of the molecular weight by the lowering of the freezing point, has led to the important result that the molecular weights are twice as large as has hitherto been assumed. The compounds experimented with were acetoxime and camphoroxime; the solvent being benzene.

With benzaldoxime and anisaldoxime, results were obtained pointing to molecular weights about one and a half times as great as those previously adopted; these results are, however, doubtless due to partial dissociation of the double molecule. A. J. G.

Desiccation of Gases. By J. D. VAN DER PLAATS (Rec. Trav. Chim., 6, 45-59).-An extended historical summary of the various methods which have been proposed for the desiccation of gases, together with an account of experiments made by the author. Anhydrous calcium chloride is more efficient than calcium chloride containing 2 mols. H2O, but the difference diminishes at low temperatures. Calcium oxide absorbs moisture more slowly than anhydrous calcium chloride, and leaves twice as much water in the air as calcium chloride containing 2 mols. H2O. Fused potash acts rapidly and much more completely than anhydrous calcium chloride, but less rapidly and less completely than concentrated sulphuric acid. Potash solution of sp. gr. 125 loses a considerable quantity of water if a current of air is passed through it for some time.

Ammonia is almost completely dried by fused potash.

Calcium chloride is a convenient but not very efficient desiccating agent. If the solution has been evaporated at 180°, it always contains calcium oxide, which cannot be neutralised even by the prolonged action of carbonic anhydride or hydrogen chloride so long as the salt remains dry.

Phosphoric anhydride is the best of all desiccating agents at present known, but it is inconveniently light and bulky. Not unfrequently it contains phosphorous anhydride, but this can be removed by Stas's method of distilling it in dry air.

Concentrated sulphuric acid is cheap and convenient, allows the speed of the gaseous current to be watched, and dries gases almost perfectly. It gives off no vapours, and when cooled its efficiency is increased. It must not contain sulphuric anhydride, and the presence of 6 to 8 per cent. of water is desirable in order to ensure absence of the anhydride in the gas which is passed through the acid. Some

times the acid contains sulphurous anhydride, but this is readily removed by a current of dry air or by boiling the acid. Sulphuric acid will dissolve carbonic anhydride, but this gas is removed in a few minutes by passing a current of air into the liquid. The distribution of sulphuric acid with a view to expose a large surface is best effected by means of broken glass. Pumice should be boiled with sulphuric acid containing a little nitric acid in order to remove chlorides and fluorides and metallic oxides which might absorb oxygen. Sulphuric acid, especially if cooled, is practically as efficient as phosphoric anhydride, and the errors due to its employment are certainly not greater than those arising from the use of india-rubber connections or from neglect to remove, by heating, the layer of moist gas condensed on the surfaces of the glass vessels.

India-rubber cannot be completely dried on a sand-bath, but should be kept over sulphuric acid in the dark for a long time. Tubing with walls 5 mm. thick allows air and aqueous vapour to pass extremely slowly, and its power of absorbing carbonic anhydride is at a minimum. Drying tubes, &c., which have to be weighed should be kept under a bell-jar containing vessels filled with the same desiccating agent as the tubes themselves contain. C. H. B.

Delicate Thermometer for Lecture Purposes. By S. YOUNG (Chem. News, 56, 261).-For this purpose the author recommends an air thermometer of the form commonly employed to show the expansion of air with rise of temperature, but using a volatile liquid such as ether in place of mercury or sulphuric acid; to make the column of liquid visible at a distance, a little absolute alcohol coloured with aniline red is added. The apparatus employed by the author had the following dimensions: the cylinder at the bottom containing the air confined by ether was 17 mm. in diameter and 130 mm. long, the narrow tube extended 700 mm. above this cylinder and had a diameter of 2.8 mm., above this there was a reservoir large enough to hold the ether when the tempeature got beyond the range of the thermometer. The total rise of column for 10° in the thermometer described was 510 mm.; for 1° between 0° and 1° the rise = 40 mm., whilst between 9° and 10° it = 82 mm. D. A. L.

Lecture Experiment for Demonstrating the Valency of Metals. By B. LEPSIUS (Ber., 21, 556–561).—The method which Nilson and Petterson have recently described for determining the atomic weights of the rare earths can be employed for demonstrating the valency of metals. A quantity of the pure metal, proportional to its atomic weight, is heated in gaseous hydrogen chloride, the volume of liberated hydrogen being proportional to the valency of the metal. For various reasons many of the metals offer great difficulties in carrying out this experiment, and the author recommends thallium, zinc, and aluminium as the most suitable examples. The hydrogen chloride, which must be perfectly dry, is obtained by acting on solid sublimed ammonium chloride with concentrated sulphuric acid in a Norblad's apparatus; it is passed through a combustion tube which contains weighed portions of the above-named metals (twice the

atomic weight) placed at intervals of 10 cm. from one another. The liberated hydrogen is collected in a graduated three-limbed glass apparatus filled with a 5 per cent. solution of potash and provided with a mercury seal. By heating the metals successively, the hydrogen evolved from each can be collected separately and the volumes compared directly. F. S. K.

Inorganic Chemistry.

The Composition of Water by Volume. By A. SCOTT (Proc. Roy. Soc., 42, 396-400).—The ratio by volume in which oxygen and hydrogen combine at 0° and 760 mm. is redetermined. The apparatus is so arranged that both gases are measured in the same vessel, a separate vessel being used for exploding. After explosion, the residue was analysed by exploding with oxygen or hydrogen and the impurity (nitrogen and carbonic anhydride) determined. The oxygen was obtained from potassic chlorate and from mercuric oxide prepared from the nitrate, the hydrogen was obtained by electrolysis. The ratio obtained is 1994: 1, which with 15.9627 for the density of oxygen, gives 16.01 as the atomic weight of oxygen. H. K. T.

The Weldon-Pechiney Process for the Manufacture of Chlorine from Magnesium Chloride. By J. DEWAR (J. Soc. Chem. Ind., 6, 775-790). This process, which was patented in June, 1884 (Eng. Pat. 9306), has recently been worked at Salindres on an experimental plant designed for the daily production of 1 ton of chlorine, and may be briefly described as follows:-The raw material employed is hydrochloric acid, the operations being (1) dissolving magnesium oxide in hydrochloric acid; (2) preparing magnesium oxychloride; (3) crushing, breaking, and sifting the oxychloride; (4) drying the oxychloride; and (5) decomposing the oxychloride. The magnesium oxide to be dissolved in hydrochloric acid is a portion of that which results from the fifth operation. The hydrochloric acid also results in part from the fifth operation and the remainder from the decomposition of salt. The solution obtained is evaporated down to the point at which it will contain not more than about six equivalents of water, and is then converted into oxychloride by mixing one equivalent of magnesium chloride with one and a third equivalents of magnesium oxide. This operation lasts only about 20 minutes, but during this time the whole mass becomes very hard and during solidification disengages much heat. The product is then in the form of solid pieces of different sizes along with a small quantity of powder. It is necessary to reduce this material to fragments, of which the largest shall not exceed the size of a walnut, and further to clear them of dust which might, when in the decomposing furnace, prevent free passage of air through the mass; this is effected by crushing it between cylinders bristling with diamond points, and

then passing it through a rotating sieve. The oxychloride is then dried, this operation being necessary, as in the decomposition a larger quantity of free chlorine and a smaller amount of hydrogen chloride are produced if the material contains less water, and the decomposition is performed at a higher temperature, both conditions being realised by drying previous to decomposition. The dried oxychloride is then decomposed in a special apparatus consisting of a series of furnaces provided with decomposing chambers heated by a movable regenerative burner, the oxychloride being decomposed in a current of air. The author's theory of the decomposition which goes on in the furnace is that in the first stage there is a rapid evolution of steam, the steam decomposing a portion of the magnesium chloride with formation of hydrogen chloride which passes off with the vapour of water. Anhydrous magnesium chloride and magnesium oxide remain, and this mixture then undergoes decomposition into magnesium oxide and chlorine by the action on it of atmospheric oxygen. The products are drawn off from the furnace by means of a diminished pressure steadily maintained, an aspirator being employed which acts through an ordinary hydrochloric acid condensing tower, a number of sandstone bonbonnes and a glass tube refrigerator, the latter being in immediate connection with the decomposing furnace. The hydrochloric acid is condensed in these apparatus, whilst the mixture of air and chlorine passes on through the aspirator and is employed for the manufacture of chlorate: for this purpose, it is driven into special apparatus, wherein the chlorine acts on calcium hydroxide and gradually transforms it into a mixture of calcium chlorate and calcium chloride. D. B.

Nitrogen Chloride. By L. GATTERMANN (Ber., 21, 751–757).— The composition of nitrogen chloride has not yet been satisfactorily determined; weighed quantities of the chloride have never been analysed, but merely the ratio of nitrogen to chlorine determined in an unknown amount of the chloride, and, as is well known, the results obtained by different observers are by no means in accordance. In the present research, the experiments were carried on as far as possible in the glass case devised by V. Meyer (Ber., 21, 26).

The nitrogen chloride was prepared in the usual manner by the action of chlorine on a solution of ammonium chloride. No sign of the formation of the chloride was visible until more than half the chlorine was absorbed, when the formation of minute drops was noticeable in the layer of liquid drawn up by capillary attraction on the sides of the flask; this soon spread and formed a thin film on the surface of the liquid, and this film in turn gradually separated into larger drops, the heavier of which sank, but were again brought to the surface by the nitrogen bubbles formed in the slight decomposition which goes on. When no more drops are formed, the chloride by careful shaking was made to fall into a small leaden capsule provided with a handle. It was then poured, with the aid of a funnel, into a thinwalled separating funnel, the ammonium chloride solution removed by a pipette, and the nitrogen chloride repeatedly washed by shaking with water until the washings no longer gave a chlorine reaction. To

assist the removal of dissolved chlorine from the oil, air was repeatedly blown through it by means of a glass tube, so that the oil was disseminated through the liquid in small drops. The successful carrying out of these operations shows that nitrogen chloride is by no means so readily exploded as has usually been supposed. The purified oil was next transferred to a thin glass flask (an operation attended with much danger owing to the friction of the stopcock of the separating funuel), dried by agitation with a small, clean piece of fused calcium chloride, and poured into the weighing glass-a weighed cylindrical vessel of about 1 c.c. capacity, provided with a stopper not quite airtight. The weighing was effected in the usual manner, save that a glass screen was again employed. The weighed substance was then decomposed by aqueous ammonia, and the chlorine estimated as silver chloride. The results showed that the product of the action of chlorine on ammonium chloride is invariably a mixture of several chlorinated ammonias, and that the product was richer in chlorine the longer the oil was in contact with chlorine, but under no conditions was the trichloride obtained directly.

After washing out the ammonium chloride as described above, the separating funnel containing the crude product covered with a few drops of water was placed horizontally, and a moderately strong stream of chlorine passed through for half an hour. The oil so obtained, after washing and drying, was analysed, and proved to be pure nitrogen trichloride.

Nitrogen chloride explodes on exposure to a strong light-either bright sunshine or magnesium light, although not so readily in the latter case. As in the author's experiments, no explosions occurred for which a cause could not be assigned, it seems very probable that the so-called spontaneous explosions experienced by other observers may be attributed to the action of light.

About gram of nitrogen chloride was heated in a thin-walled tube immersed in a beaker filled with liquid vaselin. Up to 90° no change was observed, but about 95° a violent explosion suddenly occurred. As has been noticed before in the explosion of nitrogen chloride, the main force of the explosion was exerted in a downward direction.

A. J. G.

Pyrophosphates. By G. v. KNORRE and E. OPPELT (Ber., 21, 769-773). Pahl has stated (this Journal, 1875, 375, 774) that he obtained an acid calcium pyrophosphate by the action of oxalic acid on normal calcium pyrophosphate: although closely following his directions, the authors have failed to obtain any such salt, pyrophosphoric acid being formed, nor is any acid pyrophosphate formed by the action of pyrophosphoric acid on the normal salt.

By adding calcium chloride to a moderately concentrated solution of dihydrogen disodium pyrophosphate, a white, crystalline precipitate is obtained of the formula 2CаH2P2O,,Ca,P2O, + 6H2O, this can be recrystallised from hot water in which it is very sparingly soluble. When this is boiled for a long time with hot water, it is decomposed, and the residue, after thorough washing, has the composition.

Ca H2P2O, Ca2P2O, + 3H2O.