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 chloride 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 gramme of chloroform gave 0.0798 of carbonic acid, 0.0276 of water, and 0.7372 of chloride 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 grammes of oxide 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. 0.1339 gramme of chloroform gave 0.0506 of carbonic acid, 0.0156 of water, and 0.4768 of chloride 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 chloride 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 analysis of this

body, containing as it does eighty-nine per cent of chlorine, and only eighty-five hundredths of one per cent of hydrogen, must be considered as an extreme case, and does not prove the process a good one for other classes of substances.

The next step, therefore, was to determine whether the process would be equally efficient in the analysis of substances rich in hydrogen, the combustion of which would give rise to the formation of a large quantity of hydrochloric acid. The substance selected for analysis, to settle this question, was chloride of amyl.

III. Experiments with Oxide of Zinc, as an absorbent of Chlorine in the analysis of substances rich in Hydrogen.

In these experiments, the oxide of zinc was employed in the same manner as above described for the analysis of chloroform. The chloride of amyl, which was the subject of analysis, was prepared in the usual manner. Its boiling-point was 102°, 8 corrected.

The following results of two analyses with oxide of zinc indicate that this oxide combined with and retained some of the carbonic acid. This result was not anticipated, as in the analysis of chloroform the determination of carbon was uniformly slightly in excess.

The Results of these two analyses are as follows:
:-

1.-0.1922 gramme of chloride of amyl gave 0.3513 of carbonic acid, 0.1854 of water, and 0.2528 of chloride of silver.

- not

* Since the above was written, I have observed, on reviewing my notes, only of experiments with oxide of zinc, but also with oxide of copper,— that in every analysis in which I made note of carbonization, or blackening of the asbestos in the combustion tube, which may sometimes occur from too rapid distillation of the substance, or, what amounts to the same thing, a deficiency in the supply of oxygen, there was a loss in the determination of the carbon, and generally, also, in that of the chlorine; while the hydrogen would agree pretty nearly with the theoretical quantity. I am, therefore, at the present writing, inclined to suspect that the carbonization may have had some connection with the deficiency in the carbon determinations in these instances, although the blackening would readily and completely disappear so soon as a sufficiency of oxygen was supplied. This momentary blackening of the asbestos occurred in both of the analyses of chloride of amyl with oxide of zinc, but, as already intimated, was not regarded at the time of serious consequence, as similar phenomena in the analysis of hydrocarbons by my process were generally attended with good results. It may, therefore, remain an open question, whether the oxide of zinc may not serve a good purpose in the analysis of substances of the class now under consideration.

2.-0.1657 gramme of chloride of amyl gave 0.3314 of carbonic acid and 0.1608 of water.

IV. Experiments with Oxide of Copper, as absorbent of Chlorine in the analysis of substances rich in Hydrogen.

In these experiments, for the reason previously stated, the oxide of copper could only be placed in the anterior end of the combustion tube, where it might be maintained at a tolerably low temperature. After two or three experiments, which were but partially successful, it became apparent that the range of temperature within which oxide of copper could be made serviceable to absorb the chlorine was probably rather limited.

It was observed, for example, that at 150° to 160° even brown oxide of copper, which had been but gently ignited, would fail to absorb nearly all of the chlorine, and consequently the determination of the carbon, and sometimes that of the hydrogen, would be in excess. In one experiment, in which the oxide of copper was kept at about 153° C., its appearance had suffered no change, and it was found to contain only 8.29 per cent of chlorine, or only about one quarter of the theoretical quantity. When a sufficiently high temperature is employed, on the contrary, the posterior end of the column of oxide of copper and asbestos has the appearance of being entirely changed into yellow chloride of copper, the rest of the column remaining, for the most part, of its original dark color.

In another experiment, with the oxide of copper kept at a temperature of about 160°, only about fourteen per cent of chlorine was obtained.

In both of these experiments the carbon determination was considerably in excess, and in one of them the hydrogen also. The oxide of copper employed had been strongly ignited.

Before proceeding further with these somewhat random experiments, it was deemed advisable to determine the temperature at which chloride of copper begins to give off chlorine, in order to know how far it would be safe to raise the temperature of the air-bath in conducting an analysis. By making use of the air-bath to regulate the temperature of the chloride of copper, this determination was easily made. During the heating of the chloride, a current of air from the air-gasometer was admitted through the tube in which it was contained.

Observations. At 243°, not a perceptible trace of chlorine was given off. After the lapse of fifteen minutes, at 250°, the nitrate of silver, into which the gas was conducted, was observed to be slightly milky; this may, therefore, be taken as about the temperature at which chloride of copper begins to suffer decomposition. At 267°, a solution of nitrate of silver was instantly precipitated.

Thinking that perhaps the small quantity of chlorine evolved under these circumstances might be taken up again and retained if oxide of copper were present, and possibly, also, that in that case a higher temperature might be safely employed, to make the conditions of the experiment conform in this particular to those which exist in an analysis, all but one inch of the chloride of copper was removed from the tube, and in its place was put a mixture of asbestos and oxide of copper, occupying a space of four inches in length, forward of the chloride. The experiment was then repeated. Prolonged heating in a current of air, and afterwards in oxygen, during which the thermometer rose to 350°, produced no reaction with nitrate of silver. From this it appears that the chlorine, which is given off below this temperature from chloride of copper, when this is mixed with oxide of copper, is absorbed and retained by the latter; hence, that so high a temperature as 350° may be safely employed for the air-bath in conducting an analysis by this process.

Analysis 1.- In this analysis the oxide of copper employed was prepared in the ordinary way and strongly ignited. The space in the tube occupied by the mixture of asbestos and oxide of copper was five inches in length, and contained just five grammes of the oxide. During the experiment, the temperature of the air-bath was maintained at about 350°. At the close of the combustion there was no appearance of chloride of copper, except in the first half-inch at the back end of the column of the mixture of oxide of copper and asbestos; showing that the temperature employed was favorable for rapid and complete absorption of the chlorine.

Results of the Analysis. — 0.1682 gramme of chloride of amyl gave

0.3486 of carbonic acid, 0.1633 of water, and 0.2233 of chloride of

Analysis 2.

The oxide of copper employed was of the same preparation as that used in Analysis 1. The space occupied by the mixture of asbestos and oxide of copper was only 3 inches in length, but contained the same quantity, viz. 5 grammes of the oxide of copper, as used in the previous analysis. The temperature of the air-bath ranged from 250° to 253°. At the close of the combustion, it was found that all but inch at the forward end of the column of mixed asbestos and oxide of copper had the appearance of containing chloride of copper. By comparison with the corresponding observation in Analysis 1, it will be seen that the appearance of the chloride extends over more than five times the space in this analysis as in the former, showing that with strongly ignited oxide of copper a temperature higher than 250°, even as high as 350°, is more favorable for the absorption of the chlorine. The following results of the analysis, however, are equally accurate with those of the preceding analysis.

0.1669 gramme of chloride of amyl gave 0.3457 of carbonic acid, 0.1612 of water, 0.2213 of chloride of silver.

Analysis 3. Under the impression that an oxide of copper which had been less strongly ignited might be effectual to absorb the chlorine at a lower temperature, I employed in this and the two following analyses a preparation of brown oxide of copper, obtained by precipitation with potash and ignition over an ordinary gas flame. In this analysis the temperature of the air-bath ranged from 150° to 158°. The space occupied by the asbestos mixture was four inches in length,