the several fractions were not completed. In stating the results, it has been deemed desirable to give not only the percentage composition of the several fractions, but also the absolute volumes of each gas yielded, calculated upon 100 grammes of coal. As previous experience has shown the combustible constituents to be almost entirely members of the paraffin series of hydrocarbons; and as carbon monoxide and olefines have been proved to be absent, or present only in minimum quantities, these gases have been omitted in the statement given below.

TABLE. I.-WEIGHTS OF COAL TAKEN, AND TOTAL VOLUMES OF GAS EVOLVED IN VACUO AT ORDINARY TEMPERATURE AND AT 100° CENT.

* Volumes of gas given are in all cases expressed in cubic centimetres at 0° Cent., and a
pressure of 760 millimetres of mercury.

Since the results of the estimation of the paraffin hydrocarbons do not suffice to allow of a definite statement of the nature of the constituents of the mixture, instead of returning these in terms of a value for n in the general formula C, Han+2 it has been thought better to interpret the results in terms of a possible mixture of marsh-gas (CH) and ethane (C,H) or ethane and propane (C,H,) according as the analytical data appeared to warrant. All that is attempted to indicate by this form of statement is simply that the combustible gases are not pure marshgas, but that other gaseous members of this series of hydrocarbons are present in the gases. In the majority of cases these numbers are based upon the mean of two or more concordant sets of determinations.

The proportion of such hydrocarbon constituents is determined by taking a measured volume of gas from which oxygen, etc., have been absorbed, mixing it with a measured volume of oxygen and exploding, then by a series of operations determining the amount of the contraction and of carbon dioxide resulting, also the amount

* After withdrawal of sample b, ar accident to the pump resulted in an admission of air.

The pump was repaired, and after re-exhausting, the experiment continued.

of oxygen which has been used in the combustion. From these data, the information desired can be obtained, but owing to the difficulties surrounding the exact determination of the oxygen, the writer has, in computing the results, neglected the values found for the oxygen used. This mode of interpreting the results appears to be better than that used in the previous paper, since the amount of contraction and the proportion of carbon dioxide are data capable of more exact estimation, than is the determination of the amount of oxygen used; so the influence of experimental errors in making this last estimate is avoided. If this mode of calculation be applied to the analytical data obtained in the previous investigation, the proportion of combustible gas would then appear somewhat higher.

Table II. contains a statement of the results of the examination of the gases obtained in the several experiments, and from these a series of averages have been calculated, based upon the results of experiments Nos. III., IV. and V., and are contained in Table III. In the first column of figures are given the volumes in cubic centimetres of the several gases obtained from 100 grammes of coal; in the second column the composition of the gas is expressed in volumes per cent.; while in the third column of figures the analyses have been calculated on the assumption that the Oxygen existed as air.

These numbers call for one or two remarks. In the first place, the total volume of gas per 100 grammes of coal obtained at 100° Cent, is not very different from that given in the paper of the late Mr. McConnell and the writer. In experiments Nos. III. and IV. (Table II.), it will be noted that the gas expelled by steamheating was free from oxygen, and in one instance consisted entirely of combustible gas. In experiment No. IV., whereas the last extract at 100° Cent. was free from oxygen, the next extract at a higher temperature contained oxygen. The improbability of this being derived from the outside air (which would necessitate the assumption of a leakage from the tap) is rendered all the greater in the light of the results in experiment No. V., in which similar results were observed; and, as has been already pointed out, in this experiment a pump was employed without taps.

The composition of the gas obtained in the first extraction of experiment No. V. is worthy of special attention, as, in addition

TABLE III.-AVERAGE RESULTS OF EXPERIMENTS NOS. III., IV. AND V. A. GASES EXTRACTED AT ORDINARY TEMPERATURES AND AT 100 CENT.

B.- GASES EXTRACTED AT 100 CENT., AND INCLUDED IN A.

to carbon dioxide, it contains a considerable proportion of oxygen, and an amount in relation to the nitrogen in excess of that in which these gases occur in air. The coal used in this case was