CVI.-The Action of Sulphuric Acid on Fenchone.

By JAMES E. MARSH.

Ir fenchone is warmed with strong sulphuric acid, sulphur dioxide is given off and acetoxylene [Me: Me: Ac=1:2:4] is formed. This compound has been obtained synthetically by Claus by the action of acetyl chloride on orthoxylene in presence of aluminium chloride (J. pr. Chem., 1890, [ii], 41, 396). Armstrong and Kipping (Trans., 1893, 63, 75) obtained the same compound by the action of sulphuric acid on camphor; they found that camphor is not readily acted on by sulphuric acid at a temperature below 100°, that it was most advantageous to heat the mixture to 110°, that acetoxylene is not the only product of the reaction, and that it was not practicable to separate the compound in a pure state by fractional distillation; they effected its separation by preparing the crystalline phenylhydrazone from it and decomposing this by hydrolysis.

When fenchone was warmed with five times its volume of strong sulphuric acid, the action began at a temperature below 50° with evolution of sulphur dioxide. The evolution of gas was rapid at 80°, and at 100° was complete in a few minutes. I used small quantities of not more than 10 c.c. of fenchone in one operation. It appeared preferable to keep the temperature at about 80°, and there was little or no charring. After cooling, the solution was poured into water and distilled in steam. The distillate was shaken out with ether and the ethereal solution distilled under reduced pressure. After the ether and water had come off without condensing, the residue distilled almost entirely at 131° under 20 mm. pressure; it amounted to about 70 per cent. of the fenchone taken. If we take into consideration the loss arising from the number of operations which the fenchone had undergone, and that only 10 grams were used, the yield of acetoxylene probably approximates to that required by theory. It was redistilled and the fraction boiling at 131° under 20 mm. pressure was analysed. It gave

C=806; H=8.3. C10H120 requires C-810; H=8·1 per cent.

The acetoxylene so obtained was a nearly colourless oil smelling somewhat of cinnamon. Its specific gravity was found to be 0-9968 at 20°/4°. When treated with hydroxylamine hydrochloride and caustic potash in alcoholic solution, it gave a good yield of the oxime which crystallised from methylic alcohol, forming colourless crystals melting at 86-87. Claus gives the melting point at 88-890, Armstrong and Kipping at 84.5-85.5°.

To further establish its identity and constitution, a small quantity

was oxidised by warming gently with a slight excess of bromine dissolved in 4 per cent. caustic soda solution, until bromine was scarcely liberated by addition of acid. On filtering from the crystals of carbon tetrabromide, and on addition of dilute sulphuric acid, a thick precipitate was thrown down which, after crystallising from dilute alcohol, melted at 163°. On analysis the product gave

C=716; H=6·7. CH10O2 requires C=720; H= 6.6 per cent. The silver salt was prepared and, on analysis, gave

Ag=418. CH,Me, COOAg requires Ag=420 per cent.

The acid thus obtained by oxidation of acetoxylene is paraxylic acid [CH: CH2: COOH = 1: 2: 4].

It seems reasonable to suppose that the almost quantitative conversion of fenchone into acetoxylene ought to have an important bearing on the constitution of fenchone. I have attempted in a paper on "The Constitution of Camphor " (Trans. Oxford Univ. Jun. Sci. Club N.S., 1898, 110) to show how the formation of acetoxylene from camphor, established by Armstrong and Kipping, may be explained; I here only wish to point out that fenchone is probably still more nearly related in constitution to acetoxylene than is camphor, judging from the ease with which this compound is obtained from fenchone compared with the difficulty of its production from camphor. The following formula for fenchone was put forward last year by Wallach (Annalen, 1898, 300, 320) and by Gardner and Cockburn (Trans., 1898, 73, 708) independently:

The coincidence of two separate lines of research leading to the same conclusion might be regarded as affording at least a presumption in favour of the formula. But the coincidence is less remarkable seeing that the formula is constructed on the model of Bredt's camphor formula. Further, neither Wallach nor Gardner and Cockburn express themselves as satisfied with the formula, and bring forward objections to it which seem to me not less convincing than their arguments in its favour. To these objections I think ought to be added the formation from fenchone of acetoxylene which it seems impossible to reconcile with the formula proposed.

I have from time to time ventured to criticise some of the formulæ proposed for members of the terpene group, partly on the ground that they do not account for the derivatives of benzene which are obtained from them by comparatively simple reactions. Nor does it seem to

me obvious why it should be regarded a sounder principle to base the constitution of closed chain compounds, such as the terpenes, on the products obtained by breaking down their cyclic structure, than on those products in which a ring remains. At present there seems to be an inclination either to disregard the latter entirely, or in certain cases to make a selection of those benzene derivatives which may be accounted for. It is not unreasonable to require that any constitution assigned to a member of the terpene group must account for both the aromatic and the fatty derivatives obtained from it.

I wish to thank Mr. Gardner and Mr. Cockburn for kindly giving me the samples of fenchone with which these experiments were made.

UNIVERSITY MUSEUM,

OXFORD.

CVII. On a Method for Providing a Current of Gaseous Chloroform mixed with Air in any desired proportion, and on Methods for Estimating the Gaseous Chloroform in the Mixtures.

By A. VERNON HARCOURT, M. A., F.R.S., Lee's Reader in Chemistry at Christ Church, Oxford.

WHEN air is passed through a mixture of alcohol and chloroform, both liquids evaporate, and if the surface of contact of the air bubbles and the liquid is sufficiently extensive relatively to the total volume of air passed through, the maximum tension is reached of each vapour as given off from the actual mixture of the two at the actual temperature. From the mixed gases, the vapour of alcohol may be withdrawn by passing them through sulphuric acid, the gases which go forward being a mixture of air and chloroform in proportions dependent upon the proportion of alcohol to chloroform in the mixed liquid and upon the temperature of the liquid, which must not exceed that of the air around.

Since the density of alcohol is much less than that of chloroform, mixtures of the two liquids in different proportions differ much in density, and the composition of a mixture can be adjusted and kept constant by bringing it to a known density by additions of either alcohol or chloroform.

The method, resting upon these facts, which is here proposed consists in leading dry air through a bent tube to the bottom of a Wolfe's bottle, which is half filled with glass balls and charged with a mixture of alcohol and chloroform, and thence through two washing vessels containing respectively sulphuric acid and water. The air issuing from a number of small holes made in the lower branch of the

tube passes up by many channels among the glass balls into the liquid above. With a bottle of about 1200 c.c. capacity, this contact is sufficient to saturate air passing at the rate of at least 1 cubic foot per hour, as has been proved by analysing the gaseous mixture when produced at that rate, and with air passing at the rate of cubic foot per hour, and finding that the composition was the same.

The density of the liquid can be conveniently observed while the air is passing by means of two small bulbs, one of which floats and the other sinks when the liquid has the desired density. As the operation proceeds, the proportion of chloroform diminishes, and more is added until the lighter bulb again comes to the surface.

In order to ascertain what was the proportion of chloroform vapour mixed with air which had given satisfactory results as an anesthetic, two methods were tried.

I. The first method depends upon the action on chloroform of an alcoholic solution of caustic potash, which produces potassium formate and potassium chloride,

CHCl2+4KHO=3KCl + H·COOK + 2H2O.

This reaction takes place slowly, and is, perhaps, accompanied to a small extent by some other reaction, leaving other products. Even when an alcoholic solution of potash is heated with a small amount of chloroform to a temperature of 50-70° for some hours, the chloride formed does not correspond to the whole of the chloroform; and, consequently, quantitative estimations of chloroform depending upon this reaction yield results which are a few per cent. too low. Under constant conditions, however, the deficiency is nearly constant, and thus, by applying a correction, results are obtained which may be sufficiently accurate for the object in view.

A

For bringing about the requisite contact between the alcoholic potash and the mixture of chloroform and air, either the mixed gases may be led slowly through the alcoholic solution, or they may be collected in a bottle or flask and there treated with the reagent. short account may be given of the way in which each of these modes of testing was conducted. For the first, the air and chloroform were passed through a long and almost horizontal tube, the upper surface of which had been pressed in at intervals of a few centimetres, so that while the solution with which it was charged moved about freely, the bubbles of gas were arrested at successive stages. A lamp placed beneath the tube, near its inlet, heated every portion of the liquid in turn to a temperature at which the chloroform vapour which has been dissolved undergoes the above change, while the air about to escape at the opposite end gave up the last portions of chloroform to alcohol which had been freed from the chloroform already absorbed and was

of a lower temperature. When the first part of the experiment was complete, the liquid was neutralised with dilute nitric acid, and in a portion of it the mass of chlorine, and thus the mass of chloroform, was measured by adding to it in a porcelain dish two drops of a solution of potassium chromate and a standard solution of silver nitrate from a burette.

In experiments made with a view to test the method, air was driven by a water-pump through a small meter, and thence through the vessel in which it was mixed, under the conditions to be tested, with the vapour from a known weight of chloroform, and thence through the washing-tube in which the chloroform was arrested and decomposed. Some of the results obtained by this method were as follows:

In the course of these successive experiments, as the results fell too low, various further precautions were taken, such as avoiding contact between chloroform vapour and india-rubber, which has an absorbent action. But the deficiency could only be reduced from 6 or 7 to 3 or 4 per cent.

A plan, which led up to the method already described for making a mixture of air with chloroform suitable for anæsthetic purposes, was tested by this method. Mixtures of chloroform and alcohol were brought into a tube similar to that which held the alcoholic potash, and the measured volume of air was passed through the two, bubble by bubble. In the following table, the volume of chloroform vapour found has been increased by 4 per cent. :