The sodium salt of formaldoxime thus generated will be dissociated into the tautomeric form of the oxime, which, condensing

HỌCẢNH,

with the nitromethane residue, will form the sodium salt of methazonic acid,

The different behaviour of the sodium salt of secondary nitropropane furnishing acetone, nitrite, and hydroxylamine when acted on by alkalis may be explained by its greater stability owing to the absence of the CH or CH, group which makes it less susceptible of intermolecular oxidation.

It is probable that the decomposition takes place by hydrolysis,

(CH3)2C:N OH + H2O
(CH3),C:Ŋ·OH+

=

(CH3)2CO+H, NOH

the "oxy-hydroxylamine" undergoing immediate decomposition with the production of hydroxylamine and nitrous acid,

The Reduction of Nitroparaffins and their Salts.

In the foregoing discussion, the reduction of the metallic derivatives of nitro-compounds to those of oximes is assumed to be possible. It has been previously stated (Dunstan and Dymond, Proc., 1894, 10, 139) that by the reduction of the nitroparaffins under certain conditions, the corresponding aldoximes and ketoximes are apparently produced, and it was suggested that these may possibly be intermediate products between the nitro-compounds and the substituted hydroxylamines, the formation of which has been observed by V. Meyer, Bamberger, and others.

In order to further elucidate this question, we have carried out a number of experiments on the reduction of the oximes, but in no case has the formation of any substituted hydroxylamine been observed. It therefore appears probable; that the substituted hydroxylamines obtained by the reduction of nitro-compounds are not derived from previously formed oximes, and, moreover, our results seem to indicate that the oximes and hydroxylamines are not produced together, but each separately and under different conditions. The nitro-compounds themselves (reduction in acid or neutral solution) yield the hydroxylamines, whilst their salts (reduction in alkaline solution) furnish the

oximes (or their reduction products). This may be readily demonstrated by warming a few drops of nitroethane with either zinc dust and dilute acid or with zinc dust and water. After a minute or two the solution when filtered is found to reduce Fehling's solution powerfully in the cold, owing to the action of B-ethylhydroxylamine. On the other hand, if nitroethane is dissolved in solution of caustic soda and reduced with zinc dust, the filtered liquid has no reducing property.

These results are in accordance with the view that the salts of the nitroparaffins possess a different configuration from that of the nitro-compounds themselves. The first stage of the reduction of the nitroparaffins and of their salts may therefore be expressed by the following equations:

Since the suggestion above mentioned was published, two cases have been recorded in which nitro-compounds have been directly reduced to the corresponding oximes. Hantzsch and Schultze (Ber., 1896, 29, 2252) have shown that phenylnitromethane, when reduced in alkaline solution with zinc dust or sodium amalgam, yields syn-benzaldoxime. Bamberger and Weiler (J. pr. Chem., 1898, ii, 58, 333) obtained 2-amino-3: 5-dimethylbenzaldoxime by reducing 2-nitro-3: 5dimethylphenylnitromethane (w-2-dinitromesitylene) by means sodium amalgam. It is noteworthy that in each case the reduction is effected in an alkaline solution.

SCIENTIFIC DEpartment,

IMPERIAL INSTITUTE, S. W.

of

CXXI.-Amount of Chlorine in Rain-Water collected at Cirencester.

By EDWARD KINCH, Royal Agricultural College, Cirencester. THE determination of chlorine in rain-water collected in the rain-gauge of the Meteorological Station at the Royal Agricultural College was commenced in 1870 and has been continued to the present time. Results of the determinations in the mixed rain-waters of each six months-October to March and April to September inclusive-up to September 30th, 1886, have been published (Trans., 1887, 51, 92). Results from October 1st, 1886, up to September 30th, 1900, are given below, together with some averages:

The results in the first years were more irregular than later on, partly because the method of determination of chlorine is improved, so in the final averages only the last 26 years have been taken.

It will be seen that the total deposit of chlorides is distinctly greater in the winter months than in the summer months, this being largely dependent on the prevalence of S.W. gales from the Bristol Channel. For example, the rainfall for the week ending December 31st 1898, during which S. W. gales prevailed, was 1.27 inches, and

contained chlorine 8.5 per million. Rain falling during a storm from S.W. by W. on January 16th, 1899, contained chlorine 44.79 parts per million, equivalent to 5·166 grains of salt per gallon. Other examples were given in the previous report.

Taking all the chlorides as being in the form of sodium chloride, the yearly average deposit of common salt per acre for the past 26 years has been 36 lbs., and for the past 14 years it has been nearly 30 lbs.

G. Bellucci (Staz. Sper. agrar. Ital., 1888, 21, 255) made monthly determinations of the chlorides in the rain-water falling at Perugia, about 75 miles from the sea and 412 metres above sea-level, for 10 months in 1886, and 12 months in 1887. For 1887, the total deposit was equivalent to 37.8 lbs. of salt per acre. Generally, there was rather more in the winter than in the summer months, although the largest amount in 1887 was collected in May.

N. Passerini (Bol. Scuola Agr. Scandicci, 1893, 12) made determinations of the chlorides in the rain-water near Florence, about 47 miles from the sea on one side and 67 miles on the other, during 1890 and 1891. The chlorine per million varied from 0.168 to 24-177 on different occasions. In 1890, the average was 6.7 per million, and in 1891, 37 per million.

J. B. Harrison (Reports of the Agricultural Work in the Botanic Gardens, Demerara) has determined the chlorine in the rain-water collected at the Botanic Gardens, Georgetown, Demerara, for some years. The average rainfall for the six years 1890-1895 inclusive was 111 inches, the mean annual amount of chlorine 4.69 parts per million, and the average amount of common salt deposited per acre 186 lbs.

CXXII.-Hexachlorides of Benzonitrile, Benzamide, and Benzoic Acid.

By FRANCIS EDWARD MATTHEWS.

BENZONITRILE combines directly with chlorine to form a hexachloride, but, unlike benzene and monochlorobenzene hexachlorides the product appears to be formed in one modification only. The hexachloride obtained, however, is a very interesting substance, and taking it as a starting point for further work, some very unexpected results have been obtained.

The method of preparing the hexachloride is as follows. A considerable amount of benzonitrile (of late one pound has been taken at once) is placed in a large flask and covered with water. Chlorine is

then passed into the mixture to saturation and the flask is placed in a good light or even exposed to direct sunlight till all free chlorine has disappeared. The mixture loses its yellow colour in a few hours in summer and in from two to three days in winter, but if allowed to stand longer, it again gradually assumes a yellowish tint, apparently with the formation of a larger quantity of uncrystallisable substance. The amount of hydrogen chloride produced is inconsiderable. The saturation with chlorine is repeated four or five times, and after each saturation the colour of the chlorine takes noticeably longer to disappear.

The mixture is now subjected to steam distillation till no more benzonitrile comes over, a thick, heavy, yellowish oil remains behind which on cooling becomes semi-solid. This is transferred to a beaker, heated to about 60°, and treated with about twice its volume of cold glacial acetic acid; the oil at once goes into solution on stirring, but almost immediately the solution begins to deposit small crystals, and after standing for twelve hours a large crop is obtained.

The crystals are collected, well drained by means of the pump, and recrystallised from the smallest possible quantity of boiling glacial acetic acid, when large, transparent, colourless crystals are obtained, which are pure enough for most purposes, although they are still slightly sticky from the presence of a small quantity of some impurity, probably a chlorine substitution-product of benzonitrile, which is very difficult to remove by any other solvent than acetic acid. One or two further recrystallisations from almost any organic solvent will yield the substance in a state of purity.

Chlorine determinations were made with both the crude and pure substance.

Crude substance 0·1700 gave 04575 AgCl.
Pure substance 0.1572

Cl=66·58.
0-4282 AgCl. Cl=67-30.
CH,NCI requires Cl=67·40 per cent.

Benzonitrile hexachloride when crystallised from glacial acetic acid forms what appear to be rhombic prisms. It melts perfectly sharply and without decomposition to a colourless liquid at 157°, but on strongly heating it above its melting point, decomposes, evolving torrents of hydrogen chloride and leaving a residue of trichlorobenzonitrile. dissolves readily in all the ordinary organic solvents, from any of which it may be recrystallised. It is only very slightly volatile with

steam.

It shows great stability towards aqueous acids, and is not affected by strong hydrochloric acid at 200° or by boiling dilute sulphuric acid. Chromic acid in acetic acid solution is without action on it. Alcoholic caustic soda decomposes it in two stages; on gently warming the solution, the usual action of alkalis on hexachlorides takes place, with