The calculated density for CHO, is 14.94.

The still higher-boiling condensation-products would be very difficult to obtain pure, and were not analysed. There appeared, however, to be a definite body boiling at about 250°, under a pressure of 100 mm. A table is appended of the bodies obtained and their boiling points.

XI.-The Alkaloids of Nux Vomica. No. II. On Brucine.

By W. A. SHENSTONE, Professor of Chemistry at Clifton College, Bristol.

In a former paper "On the Nux Vomica Alkaloids" (Chem. Soc. J., 1881), I showed that it is highly improbable that igasurine has any existence, and stated my intention of examining the two other alkaloidal constituents by limited oxidation.

I have made a number of experiments with brucine by various methods, employing large quantities of material. I find that the action proceeds very rapidly to the production of oxalic and carbonic acids, especially with alkaline reagents. When such complete oxidation is prevented, the products are amorphous unworkable substances.

Strychnine behaves in much the same manner, but the action is not so rapid, and I may possibly return to this part of the subject in a

complete examination of strychnine, which the results given in this paper have induced me to undertake.

Strecker, Liebig, Gerhardt and Laurent, and others have reported experiments which have led them to suppose that brucine, when treated with dilute nitric acid, yields among other things either methyl or ethyl as nitrite or nitrate. I have not met with any corresponding statements concerning strychnine, and as the formula of brucine suggested that it might be a dimethoxyl-derivative of strychnine, I was led to make the experiments which I now report.

When brucine is heated in sealed tubes to 100° C. with several times its weight of strong solution of hydrochloric acid, it is not much affected. When, however, from 7 to 15 times its weight of the acid is employed, and the temperature is raised to 130-140°, on opening the tubes there is a great rush of a gas which is colourless, rather soluble in water, more so in alcohol, and not attacked in the cold by solid potassium hydrate.

This gas burns with a luminous green-edged flame giving clouds of hydrochloric acid, and when carefully freed from hydrochloric acid and passed through a red-hot tube, yields hydrochloric acid, which has been recognised by the ordinary reactions of that substance.

Although there was little doubt that the gas was methyl chloride, I thought it worth while to determine the amount of carbon dioxide produced by exploding a measured quantity of it with oxygen.

The gaseous contents of a tube in which brucine had been treated with large excess of acid were collected over mercury, and qualitative examination having shown that they consisted of a mixture of the gas in question with air, water-vapour, hydrochloric acid, and a little carbonic acid, the last three were removed, and a measured quantity of the mixture of air and chlorinated gas was exploded as usual with -proper proportions of oxygen and air. The carbonic acid formed and the residual nitrogen were determined as usual. The following are the numbers (corr.) obtained. From them I calculated the volume of the chlorinated gas in the mixture examined, and so was able to know the relation between the gas burnt and the carbon dioxide which it produced:

Vol. of chlorinated gas and air taken = 5.77 c.c.

These numbers showed that 412 c.c. of the gas given off by brucine yielded 4.5 c.c. of carbon dioxide.

As 4:12 c.c. of ethyl chloride would have given 8-24 c.c. of carbon dioxide, this result leaves no doubt as to the nature of the gas in question. But as the excess of carbon dioxide is rather large, I may state that it is due to the following circumstance, which I think is not generally known:

I find that on exploding a mixture of methyl chloride, oxygen, and air, the inside of the eudiometer becomes coated with a white solid, which is shown by the action of potassium hydrate to be mercurous chloride. This makes the tube rather opaque, and therefore, though I had no difficulty in observing where the liquid stood in the tube, I found it impossible to read the level of its surface with perfect exactness, and I was careful to read rather too high than too low.

When tubes containing strychnine and solution of hydrochloric acid in similar proportions to those employed in the experiments with brucine are heated, there is good evidence that the alkaloïd is much attacked, but little or no evolution of gas occurs. In an experiment in which 1 gram of strychnine was heated with 15 grams of acid for many hours at temperatures gradually rising to 170°, 7 c.c. only of gas were given off on opening the tube, and the air in the tube was found to contain carbon dioxide.

It is probable therefore that the difference in the formulæ of brucine and strychnine may be correctly accounted for by writing that of the former C2H20(CHO),N2O2, i.e., strychnine, in which 2 atoms of hydrogen are replaced by 2 methoxyl-groups.*

-

In order to learn if the quantity of methyl chloride produced agreed with this hypothesis, I made the following experiments:(a.) 1 gram of dry brucine with 20 grams of acid heated to 160° for 7 hours, then to 180° for 7 hours, gave 646 c.c. at 0° and 760 mm.

(b.) 1 gram of dry brucine with 8 grams of acid heated to 140° for 6 hours, then to 160°, 180°, and 190° for 4 hours at each temperature, gave 707 c.c. of methyl chloride at 0° and 760 mm.

(c.) 1 gram of dry brucine with 15 grams of acid heated to 150°, 160°, and 170°, gave about 70 c.c. of methyl chloride.

(d.) I gram of dry brucine heated with 15 grams of acid to 145° * In adopting the formula C2H22NO2 for strychnine, I am aware that Claus and Glassner (Ber., 14, 773) have suggested that strychnine is a body of variable composition, containing sometimes 22 and sometimes 21 atoms of carbon in its molecule. After carefully reading their memoir, however, I have not felt that this hypothesis can be accepted until more exact results are brought forward, and some corresponding difference in property is established.

for 6 hours, then to 150° for 8 hours, and finally to 180° for 6 hours, gave 83 c.c. of methyl chloride at 0° and 760 mm. (e.) As it appeared that prolonged heating at low temperatures gave the greatest yield of methyl chloride, I employed in this final experiment the lowest temperature that seemed likely to answer, and continued the application of heat for longer periods than in the previous experiments. 1 gram of dry brucine heated with 15 grams of acid for 14 hours to 135°, then for 13 hours to 145°, then to 150° for 7 hours, and finally to 160° for 7 hours, gave 902 c.c. of methyl chloride at 0° and 760 mm.

At 0° and 760 mm. 1 gram of dimethoxystrychnine would theoretically yield 113 c.c. of methyl chloride. Each of my experiments gave more than half that volume, none of them gave more than that amount, and, under the most favourable conditions (Experiment e), I obtained, as will be seen, 79 per cent. of the theoretical yield.

The non-gaseous contents of the tubes after heating were black, sometimes rather tarry liquids; these, on slight dilution, gave a precipitate which was soluble in pure water. Alkalis precipitated a base from its solution which underwent instant change on exposure to the air; the base formed an insoluble platinum salt which also was too unstable for examination. As strychnine is also profoundly attacked by hydrochloric acid, I am not inclined to regard this base as likely to be the compound C21H18N2O2, but rather in all probability a product of further change; and this opinion is confirmed by the fact that it does not appear to be affected by nascent hydrogen, which might be expected to convert the compound C21H18N2O2 into strychnine or an isomeride of strychnine.

As there was no hope of obtaining results of value with this substance, I decided to try the effect of hydriodic acid on brucine at somewhat lower temperatures. My results are worth a brief record, but are not conclusive. When brucine or its hydriodide is heated gradually up to 100° with eight times its weight of solution of hydriodic acid (sp. gr. 1·9), and maintained at that temperature for two hours, a black solid separates, from which hot dilute hydrochloric acid extracts a new base. A further quantity of the same base remains in solution. When precipitated by alkalis it at once undergoes very rapid oxidation. On evaporating an acid solution of the hydrochloride in a vacuum over sulphuric acid it is also destroyed Addition of solution of iodine in potassium iodide threw down the periodide as a chocolate powder, with a slight odour of iodine when damp. This appeared to be permanent at first; but over sulphuric acid it continued to lose weight steadily, although very slowly after

the first day or two. When this substance was placed in a vacuum over sulphuric acid, the loss was much increased for the first few hours, and then continued pretty steadily for some days. I was proceeding to analyse it in the hope that the change of weight was due only to the loss of small quantities of iodine, but I found, on treating it with solution of sulphur dioxide, that the solution obtained had a distinctly brown tint at once, and as such a change of colour had been previously found to accompany the oxidation of the base, I unwillingly concluded that even in this combination it had not altogether escaped change, and that no value could be attached to quantitative results obtained with it.

The base may be easily recognised by a very fine colour-reaction. If ammonia is added to an acid solution of it, and this is at once extracted with chloroform, the solution in chloroform leaves on evaporation a greenish solid which dissolves in solution of hydrochloric acid with a fine coloration, blue when seen in thin layers, purple in thicker layers. The colour of the solution remains unchanged for many days.

I present these somewhat imperfect results to the Society now, because I have come to the conclusion that it will in future be most profitable to continue my endeavours to make out the nature and relation between the Nux vomica alkaloïds by an examination of strychnine. My present results show that there is reason to think that brucine is a dimethoxyl-derivative of strychnine. The removal of these methoxyl-groups gives rise to bodies of such unstable character, that for the present their further examination seems less likely to repay the labour they demand than experiments made upon strychnine, which may very probably yield more manageable products.

XII.-The Behaviour of the Nitrogen of Coal during Destructive Distillation; with some Observations on the Estimation of Nitrogen in Coal and Coke.

By WILLIAM FOSTER, M.A., Lecturer on Chemistry at the Middlesex Hospital.

WHEN Coal is submitted to destructive distillation in close vessels, as in coal-gas manufacture, ammonia-gas with numerous other substances is evolved. It is usual for writers on the subject to consider that the greater portion of the nitrogen of coal takes the form of ammonia-gas or its compounds. Professor Roscoe in his small manual has the following sentence :-" Coal contains about 2 per cent.