Playfair, H. E. Roscoe, A. W. Williamson, A. Crum-Brown, P. Griess, G. D. Liveing, J. E. Reynolds, E. Schunck, A. Voelcker. Secretaries.-H. E. Armstrong, J. Millar Thomson.

Foreign Secretary.-Hugo Müller.

Treasurer.-W. J. Russell.

Council.-E. Atkinson, Capt. Abney, H. T. Brown, W. R. E. Hodgkinson, D. Howard, F. R. Japp, H. McLeod, G. H. Makins, R. Meldola, E. J. Mills, C. O'Sullivan, C. Schorlemmer.

RESEARCH FUND.

During the past year grants from the Research Fund have been made as follows, on the recommendation of the Research Fund Committee:

£100 to Dr. Wright for the continuation of his research on the Determination of Chemical Affinity in Terms of Electromotive Force. £25 to Messrs. Cross and Bevan for the investigation of Certain Varieties of Cellulose.

£25 to Dr. Rennie for the further investigation of Australian Sarsaparilla.

£25 to Mr. Shenstone for the further investigation of the Nux Vomica Alkaloïds.

£15 to Mr. Watson Smith for the investigation of a Sublimate formed in the Manufacture of Aurin.

£25 to Professor Tilden for the investigation of the CH, Hydrocarbons and the Decomposition of Terpenes by Heat.

£10 to Mr. James for the investigation of Ethylene Chlorobromide and its Homologues.

266

£598 7 3

267

XXXII.—On the Estimation of Hydrogen Sulphide and Carbonic Anhydride in Coal-gas.

By LEWIS T. Wright.

THE estimation of hydrogen sulphide and carbonic anhydride in coal-gas is not without difficulties, for unpurified coal-gas is a complex mixture—a mixture whose components are far from being all known. In cases where the method of estimating the hydrogen sulphide and carbonic anhydride is absorptiometric, as in Bunsen's and other gas-analysis methods which depend on the principle of the measurements of absorbed volume, it is customary to term that portion of the gas absorbed by such reagents as manganic peroxide, ferric oxide, &c., hydrogen sulphide; and that portion of the gas which may be absorbable by potassium hydroxide and not by manganic peroxide, &c., is considered as carbonic anhydride. It must be admitted that the portion of the gas under analysis termed hydrogen sulphide (an amount perhaps closely and practically approximating to the true amount) is really hydrogen sulphide, plus other gases of more or less unknown composition; and further that the portion termed carbonic anhydride is really carbonic anhydride, plus other bodies.

With this reservation there can be no particular objection to such a method of stating analysis. But however valuable may be the absorptiometric methods for the estimation of hydrogen sulphide and carbonic anhydride in coal-gas, when the sample of gas has to be transported from its place of collection to the place of analysis, or where the supply may be small and large quantities not available, I think it possible that there are other methods more generally applicable, especially where the hydrogen sulphide and carbonic anhydride exist only as a very small percentage. In practice the united amount of these two impurities in unpurified coal-gas ranges from about 2 to 4 per cent., and in partially purified or finished coal-gas will range from that amount to nullity.

Being anxious to determine the degree of accuracy to which hydrogen sulphide and carbonic anhydride could be estimated in coalgas by absorptiometric gas-analysis, I made some experiments with an apparatus of the Frankland-Ward type, which are instructive and interesting with regard to the use of solid porous reagents.

Bunsen (Gasometrische Methoden, 1877) publishes some analyses from which, of course, it is not possible to draw safe conclusions in reference to the estimation of hydrogen sulphide in coal-gas, for the mixtures he analysed were simple and artificially prepared.

Operating with a manganic peroxide bullet moistened with syrupy phosphoric acid, according to the method fully described in his work, he obtained the following results :

The numbers in analysis 18 are very good indeed, but the difference in analysis 19, viz., 0.33 per cent. on the total volume of the mixture analysed, is serious.

The system adopted in my experiments can be briefly described. Test-tubes containing about 60 c.c. were filled over mercury with crude coal-gas (free from ammonia-vapours), and were kept sealed with mercury in small porcelain crucibles.

It is true that mercury is attacked by hydrogen sulphide, but the error due to this is infinitesimal when the gas is in contact with mercury for a few hours. In 20 to 48 hours I have noticed a slight loss of hydrogen sulphide from this cause. In the following experiments the gas was analysed as soon as collected :

One-half of the sample of coal-gas was passed into the laboratory tube which, together with the measuring tube, had been freshly washed out. It was then passed into the measuring tube, measured, and transferred to the laboratory tube; then treated with some solid reagent for hydrogen sulphide absorption for half an hour; then remeasured; again brought back into the laboratory tube, treated with a few drops of potassium hydroxide solution, and again measured.

The hydrogen sulphide and carbonic anhydride were thus separately determined. The laboratory and measuring tubes were again washed to bring them into the same condition as at the commencement of the first analysis. The remaining portion of the gas sample in the testtube was then passed into the laboratory tube, measured, and treated