* The amount of acetone formed in this experiment was found to be 0·0013 gram, or 0.01 per cent.

The Gases from Copper Acetate heated in a Vacuum.

Although many chemists have observed the evolution of gas which occurs when cupric acetate is heated, no two of them agree as to the composition of the gaseous mixture. Thus, Gehlen finds the mixture to be 6 vols. of carbon dioxide and 5 vols. of carbon monoxide; Vogel, 4 vols. of carbon dioxide and 1 vol. of carburetted hydrogen; Roux, a mixture of carbon dioxide and a combustible gas; Chenevix, 2 vols. of carbon dioxide and 3 of hydrocarbon.

In the present experiments, the gases collected after washing with water consisted only of carbon dioxide and carbon monoxide, approximately 4 volumes of the former to 1 of the latter. In some of the earlier experiments, it was found that the gases contained some acetic acid vapour, which had to be washed out before they were analysed. When the last form of washing bulbs (shown in Fig. 3) was used, the acetic acid was completely stopped. As none of the chemists above-mentioned washed the gases with water, it is probable that some acetic acid vapour was mixed with the permanent gases and constituted the "hydrocarbon" which several mention.

The method of analysis was as follows. The gas was collected over mercury at the exit of the Sprengel pump in tubes holding about 40 c.c. The amount of carbon dioxide was estimated by absorption with potash and the residue exploded with an excess of oxygen. A second absorption by potash and, in some cases, an estimation of the residual oxygen by absorption with pyrogallol or explosion with hydrogen, completed the analysis. We found, as others must, that when the

gas in which oxygen is to be estimated by pyrogallol is unmixed oxygen, the potash used must be very strong.

In some cases, one portion of gas was divided into two parts and each half analysed separately. Portion XI, Expt. 32, and Portion II, Expt. 34, were treated in this way, and the numbers obtained were:

A table is given on p. 1396 which shows the results of the analyses of the gases from different experiments, from which it will be seen that the mean ratio CO2: CO is approximately 4:1, that in the first portions of gas collected the proportion of carbon dioxide is invariably greater than this, and that it then diminishes as the decomposition of the salt proceeds.

The temperature of the salt was kept within a degree or two of 230° and the pressure of gases upon it at 2 or 3 cm. The time required for the collection of each portion of gas was about three-quarters of an hour. The evolution was rather more rapid at the beginning and showed a sudden decrease at the end.

In most cases, the gas was collected under as low a pressure as possible. It was found inexpedient to reduce the pressure below about a couple of centimetres of mercury, because water then evaporated quickly from the washing-tube into the sulphuric acid drying-tube and there was a danger of sulphuric acid spirting back into the bulbs. In Expt. 35, an attempt was made, by collecting the gases at different pressures, 1/3, 2/3, &c., of an atmosphere, to see whether the pressure influenced the ratio of carbon dioxide to carbon monoxide. It was found that at the higher pressures the proportion of carbon dioxide was slightly increased.

As the gas formed and collected consecutively in separate tubes had been found to consist wholly of carbon dioxide and carbon monoxide, in the later experiments only a few complete analyses were made, and in most cases the carbon dioxide was absorbed and the residue taken to be carbon monoxide. Where, in the tables, the numbers would add up to 100 exactly and the spaces have been left blank, the carbon monoxide was determined by difference.

In Expt. 30, the analysis was made, during the repair of the gas analysis apparatus, by absorbing and weighing the carbon dioxide present, passing the rest through a tube over red hot copper oxide, and absorbing and weighing the carbon dioxide produced.

It will be seen that the differences in the proportion of the two gases, in portions collected successively, far exceed the differences shown by two analyses of the same portion. There are, therefore, at least two changes, of which the gases are products, whose relative rates vary slightly as the decomposition of the salt proceeds, but are not greatly affected by changes of temperature or pressure.

The Liquid Products.

The liquid products of the decomposition of normal copper acetate at a gradually increasing temperature are water, acetic acid, and a trace of acetone. The amount of acetone found in these experiments is extremely small, for example, in Expt. 32, it was 0.01 per cent. With some other acetates, for example, lead and calcium, acetone is the principal product, and some previous observers, working on a large scale, have obtained acetone from copper acetate. The brothers Derosnes, using more than 20 kilograms of the salt, got several ounces of acetone. The cause of the difference would seem to be that in distillations on a large scale substances produced at a lower temperature come into contact subsequently with the sides of the retort and with other substances at a much higher temperature, and thus secondary changes occur. This complication is avoided when we work with small quantities and raise the temperature very gradually. In the present experiments, it has been noticed that if the experiment is prolonged the top of the sublimate becomes discoloured, and if greatly prolonged, some of it is destroyed and a coppery deposit is formed on the sides of the sublimate chamber. In a good experiment, the amount of this decomposition is very small, and is probably comparable in amount with the trace of acetone formed. If the acetone were a primary product of the decomposition of the cupric acetate, it would not be formed in such very small quantities. If, however, it is a product of the decomposition of the sublimed cuprous acetate, which is itself only a small product of the reaction and is only decomposed

to a small extent, the minute amount of acetone formed is accounted for. A possible equation is :

Estimation of Water, Acid, and Acetone.

After the washing-bulbs had been weighed, they were emptied and washed, the liquid was made up to a known volume, and a portion taken for titration with standard alkali.

The acetone was estimated in another portion by Messinger's method, which depends on the conversion of the acetone into iodoform by a standard solution of iodine in presence of potash solution, and estimation of the residual iodine by standard sodium thiosulphate and starch. Three molecules of iodine are required to convert one molecule of acetone into iodoform. The method was first tested with known quantities of acetone and found to be satisfactory. The water was estimated by difference.

The Sublimate.

The white substance, here called the sublimate, has been known for a long time as a product of this decomposition. Originally described as "flowers of copper" it was afterwards found to be cuprous acetate. The small amount of it formed (1/1280 of the amount of the salt decomposed is given as the quantity by Lassone) has, however, hitherto prevented any analysis of it being made. The amount obtained varies in different experiments; the largest amount obtained in these experiments was 4.21 per cent. (Expt. 34). Also it is very easily decomposed. As ordinarily prepared, by heating cupric acetate, it is apt to contain traces of acetic acid, and this causes it to turn green when exposed to the air; if it is free from acetic acid, it is fairly permanent. Water turns it yellow, forming cuprous oxide or hydroxide. When it is obtained by heating the salt in a tube through which hydrogen is passed, it forms beautiful, transparent, leafy crystals, and the first portion that appears when the salt is heated in a vacuum is always crystalline. It can be resublimed in a current of gas (hydrogen, carbon monoxide, &c.), but decomposes on strong heating, leaving a coppery deposit.

The analysis of the sublimate was made as follows. The sublimate chamber, after being sealed off, was cleansed from all traces of oil. A file mark was then made near one end and a crack led nearly round it, so as to allow air to enter slowly. After the air had entered, it was weighed. The tube was then parted at the crack and the contents washed out into a porcelain dish. The sublimate resisted the action of the water for a short time (it was not very easily wetted by water)