tion of the substance as the sulphur or chlorine, in a manner similar in other respects to that described for simple hydrocarbons.'

In pursuing this research some difficulty was experienced, as was anticipated, in finding a substance which would absorb and retain the whole of the chlorine, under conditions that would at the same time insure that every trace of the carbonic acid and water should pass through unabsorbed.

The search for this substance was confined to the oxyds of the heavy metals, as these alone, from their strong affinity for chlorine, and weak affinity for carbonic acid, seemed to give encouragement of success.

The difficulty, however, in finding such a substance was chiefly due to the circumstance that most of the chlorids of these metals are either too volatile, or begin to suffer decomposition at too low a temperature; it being requisite that the absorbing substance, and the newly formed chlorid of the same, should bear to be heated sufficiently to prevent both condensation of water and absorption of carbonic acid, and at the same time avoid a temperature high enough to occasion any appreciable decompo sition of the chlorid.

This question of temperature became, therefore, a prominent one in the investigation, as evidently the success of the process must depend, in a great degree, on the proper management of the temperature of the absorbing substance, within such limits as might be found to give satisfactory results. Hence, my first step was to devise means to secure the necessary control of the temperature of that part of the combustion tube which should contain this substance.

For this purpose was constructed a sheet-iron air-bath or chamber, A, fig. 1, provided with two holes-one on each side-to

AM. JOUR. SCI.-SECOND SERIES, VOL. XLII, No. 125.-SEPT., 1866.

receive the combustion tube, and a tubulure in the top for a ther mometer. One end of the air-bath is made to rest on the com. bustion furnace, and the other, which projects a few inches from the front of the furnace to make room for a lamp, is supported by a leg resting upon the table. The bulb of the thermometer is placed in a central position, in the interior of the bath, close by the side of the combustion tube.

The temperature of the air-bath, and consequently of the substance contained in the combustion tube within, is easily regulated by means of a Bunsen's burner placed under the front end of the bath, as shown in fig. 1. With the exception of the airbath, the apparatus employed is the same as that used in the analysis of substances containing sulphur, a full description of which is given in the papers above referred to.

The substance that I have found best adapted to absorb the chlorine, for substances easily combustible, is brown oxyd of copper, prepared by precipitation with potassa and ignition over a gas flame.

Difficultly combustible substances, like chloroform, are not completely burnt in oxygen in contact with asbestos alone, but require the presence of a body having affinity for chlorine; otherwise there is formed a liquid body, difficultly volatile,probably a chlorid of carbon,-which condenses in the vacant part of the tube, from b to c, fig. 2, and which cannot be entirely burnt off and save the analysis. In such cases the absorbing substance is mixed with the asbestos occupying the back part of the tube, where the combustion takes place. It is evident that oxyd of copper would not answer for this purpose, as at so high a temperature dichlorid of copper would be formed, which, being insoluble in dilute acids, would interfere with the determination of the chlorine. Oxyd of zinc has been found to give good results with such substances.

The preparation of the combustion tube, and the arrangement of the mixture of asbestos and the absorbing substance, is the same-except in the case last mentioned-as in the analysis of substances containing sulphur, as shown in fig. 2, viz., the space between a and b, about

10 inches in length, is packed with pure asbestos; between b and c,a space of about two inches,-being left vacant, a plug of asbestos

2.

is placed at c; the space between c and d, 4 to 5 inches in length, is filled with an intimate mixture of asbestos and brown oxyd of copper; and, finally, a plug of asbestos is placed at d.

After the combustion, the chlorid, together with the excess of oxyd, is extracted from the asbestos by means of dilute nitric acid.

To facilitate the removal of what may adhere to the sides of the tube, the apparatus shown in fig. 3 will be found serviceable as in the analysis of sulphur compounds.

I. Experiments with Oxyd of Lead and with Oxyd of Copper, placed in the anterior end of the combustion tube, as absorbents of Chlorine in the analysis of substances difficultly combustible.

The substance selected for analysis, as a test of the process for that class of bodies which are difficultly combustible, containing but a small percentage of hydrogen, was commercial chloroform. The preparation employed was first subjected to redistillation.

Its boiling point was found to agree essentially with that assigned to pure chloroform in Gerhardt's Traité de Chimie. When the usual tests were applied, no impurity could be detected.

3.

Experiment 1.-A mixture of oxyd of lead and asbestos was placed in the anterior end of the combustion tube, between c and d, fig. 2, as previously described. As chlorid of lead was supposed to bear a pretty high temperature, without volatilization or decomposition, the use of the air-bath was omitted in this experiment, and the oxyd gently heated with a small flame from the combustion furnace. The combustion had not proceeded far when it became apparent, from deposition of minute drops of liquid on the sides of the vacant part of the tube,-from b to c, fig. 2, that the combustion of the chloloform was incomplete, although no doubt could exist as to the presence of an excess of oxygen. This deposit of liquid, which, as already stated, was supposed to be a chlorid of carbon, was found to be difficultly volatile, suffering partial decomposition, and leaving on the tube a brown deposit, which was not entirely removed by ignition in a stream of oxygen. The high temperature employed to burn off this deposit occasioned excessive heating of the posterior end of the mixture of lead oxyd and asbestos; and this may have been the cause, to some extent, of the excess in the determinations of carbon and hydrogen, although subsequent analyses indi cate that the sample of chloroform under examination contained a larger percentage of these elements-particularly of the latter -than belongs to pure chloroform. This experiment gave 11.47 per cent of carbon, and 1:87 per cent of hydrogen. Theory

gives 10-07 per cent of carbon, and 0.85 per cent of hydrogen. The mixture of asbestos and oxyd and chlorid of lead was removed from the tube, and treated in the usual manner with a solution of bicarbonate of soda to obtain a soluble chlorid. This operation was found extremely tedious. Even after treatment for more than two weeks, with occasional fresh portions of the bicarbonate and frequent agitation, the decomposition of the lead chlorid was still found to be incomplete, and the operation was abandoned. As this is given in the text books as a good process for the separation of chlorine from chlorid of lead, I am led to presume that in this case the excess of heat employed gave rise to the formation of an oxychlorid, which is, doubtless, more slowly acted upon by the bicarbonate. This single experiment does not, therefore, prove that oxyd of lead may not be employed in this process with good results, when used for easily combustible substances, and excessive heat is avoided. But it will, unquestionably, be found preferable to use a substance which will give directly a soluble chlorid.

Experiment 2.-This experiment was conducted as the last, with only this difference, viz. that oxyd of copper was substituted for the oxyd of lead. No better results, however, were obtained. The reappearance of the difficultly volatile liquid in the vacant part of the tube, while there was assurance of there being no deficiency in the supply of oxygen, served to confirm the impression gained by the preceding experiment,-that chloroform could not be completely burnt in oxygen alone, but that a substance having affinity for chlorine would have to be mixed with the asbestos, at the point where the combustion takes place. II. Experiments with Oxyd of Zinc, mixed with the asbestos in the posterior part of the combustion tube, as absorbent of Chlorine in the analysis of substances difficultly combustible.

As already indicated, the chief object of this set of experiments was to determine whether the presence, at the point where combustion takes place, of an oxyd capable of combining with the chlorine would have the effect to prevent the formation of the difficultly volatile liquid above mentioned, and thus remedy that defect in the process.

Experiment 1.-In this experiment, three grams of oxyd of zinc were intimately mixed in a mortar with the quantity of asbestos necessary to fill the space between a and b, fig. 2, and that part of the tube then packed with this mixture in the usual manner. A similar mixture composed of asbestos and only one gram of oxyd of zinc was placed between c and d. The space between b and c was left still vacant, in order to be able to observe the effect. On account of the volatility of the chlo

H. Rose, Chimie Analytique, new French edition, p. 801.

rid of zinc, it was deemed advisable to retain the use of the airbath to control the temperature of the anterior portion of the combustion tube, which, in this experiment, was not allowed to exceed 160° C. The result was, as anticipated, that no such condensation of liquid between b and c occurred. In order to gain from this experiment some idea of the degree of volatility of chlorid of zinc under such circumstances, the two columns of asbestos were treated for chlorine, separately. The solution obtained from the anterior column was found to contain but a trace of chlorine, giving only a milkiness with nitrate of silver; showing that the chlorid of zinc does not travel far through a column of asbestos from the point where the flame plays directly on the tube.

Results of the Analysis.-0·2067 gram of chloroform gave 0-0798 of carbonic acid, 0·0276 of water, and 0·7372 of chlorid of silver.

Experiment 2.-In this experiment, the whole length of the combustion tube from a to d was packed with a mixture of asbestos and four grams of oxyd of zinc. The temperature of the anterior end of the combustion tube was regulated, as in the previous experiment, by means of the air-bath.

Results of the Analysis.-01339 gram of chloroform gave 0.0506 of carbonic acid, 0·0156 of water, and 0.4768 of chlorid of silver.

These two analyses, agreeing as they do so closely, indicate that the chloroform analyzed contained larger percentages of carbon and hydrogen,-especially of the latter, and a correspondingly smaller percentage of chlorine than the theoretical quantities; occasioned, probably, by the presence of some impurity. This view is supported by calculations made on the assumption that the excess might have arisen from volatilization of chlorid of zinc, or from incomplete absorption of the chlorine; which would make the chloroform contain from two to six per cent more than the theoretical quantity of chlorine. These results are regarded, therefore, as satisfactorily establishing the utility of this process in the analysis of chloroform. But the