In the valley of the Mississippi, in Ohio, Illinois, and Kentucky, are great deposits of petroleum, far removed from any mountain chain or volcanic vent, and the cases which have been cited of the limited production of hydrocarbons in the vicinity of, and probably in connection with, volcanic centers may be explained by supposing that in these cases the petroleum is distilled from sedimentary strata containing organic matter by the proximity of melted rock, or steam. Everything indicates that the distillation which has produced the greatest quantities of petroleum known was effected at a low temperature, and the constant escape of petroleum and carbureted hydrogen from the outcrops of bituminous shales, as well as the result of weathering on the shales, depriving them of all their carbon, shows that the distillation and complete elimination of the organic matter they contain may take place at the ordinary temperature. * * * * * ESTIMATION OF SULPHUR IN IRON AND STEEL. By GEORGE CRAIG. For wellnigh two years I have been estimating sulphur in iron and steel by a modification of the evolution process, which consists in passing the evolved gases through an ammoniacal solution of peroxide of hydrogen, which oxidizes the sulphureted hydrogen to sulphuric acid, which latter is estimated as usual. The _modus operandi_ is as follows: [Illustration] 100 grains of the iron or steel are placed in the 10 oz. flask, a, along with ½ oz. water; 1½ oz. hydrochloric acid are added from the stoppered funnel, b, in such quantities at a time as to produce a moderate evolution of gas through the nitrogen bulb, c, which contains 1/8 oz. (20 vols.) peroxide of hydrogen and ½ oz. ammonia. The tube, d, is to condense the bulk of the hydrochloric acid which distills over during the operation. When all the acid has been added and the evolution of gas becomes sluggish, heat is applied and the liquid boiled till all action ceases. Air is blown through the aparatus for a few minutes and the contents of c and d washed into a small beaker and acidified with hydrochloric acid, boiled, barium chloride added, and the barium sulphate filtered off after standing a short time. A blank experiment must be done with each new lot of peroxide of hydrogen obtained, which always gives under 0.1 barium sulphate with me. The whole operation is finished within two hours, the usual oxidation process occupying nearly two days; and the results obtained are invariably slightly higher than by the oxidation processes.