“The solution of the alkaloid in excess of sulphuric acid having been obtained, Mayer’s reagent was carefully added till no more precipitation was observed, the end of the reaction being ascertained when on filtering some of the nicotine solution into a watch-glass and adding a drop of the reagent, no precipitate was formed. With careful manipulation concordant results were obtained. “Other methods tried were as follows: “Volumetric method.—Ten or more grams of powdered tumbeki were distilled with a solution of sodic or potassic hydrate, the distillate being passed into a known volume of decinormal standard solution of sulphuric acid, and the amount of acid neutralized by the nicotine was determined by a standard decinormal solution of soda and the nicotine calculated. “By this method the results obtained were invariably too high owing to an appreciable quantity of ammonium salts contained in the leaves. Dr. Kissling[E] has also noticed the high percentages obtained by this method of estimating nicotine. “Kosutány treats the leaves with milk of lime till all the ammonia is driven off, and then extracts with water; shakes the aqueous solution with petroleum ether and proceeds as before. “This method was not found to give good results, for though the ammonium salts do not interfere with the reaction, yet the petroleum ether does not extract the whole of the alkaloid, and thus a low percentage is obtained. “Extraction by Ammoniacal Ether.—This consists in extracting the powdered leaves in an upright extractor, by an ethereal solution of ammonia, and either driving off the ether and weighing the residue as nicotine; or volumetrically estimating the residue by decinormal solution of sulphuric acid, or precipitating the alkaloid by platinum perchloride. In either case, whichever way the residue is estimated, the results are too high, owing to the difficulty of entirely getting rid of the ammonia. “The following are the percentages of nicotine in the tumbeki:— ‘_Ispahan._’—I. By Mayer’s Reagent. A. (midrib) 8·156 per cent. B. (leaf) 5·508 ” ” C. (leaf and midrib) 5·589 ” ” D. (leaf) 5·3865 ” ” —————— 5·4945 per cent. average. II. By Volumetric Method. By working on 10 grams = 7·2 per cent. By working on 50 grams = 7·228 ” ” ‘_Hidjaz._’—I. By Mayer’s Reagent. A. (leaf and midrib) 2·025 per cent. B. (leaf and midrib) 2·268 ” ” C. (leaf and midrib) 2·028 ” ” D. (leaf and midrib) 1·863 ” ” ————— 2·046 per cent. average. II. By Volumetric Process. A. 2·37 per cent.