FOOD AND DRUGS ANALYSIS 49 ORGANIC ANALYSIS. Analyses, Purification, and Chemical Properties of Agar-Agar. C. R. Fellers. (J. Ind. andXng. Chem., 1916,8,1128-1133.)-A number of old analyses of agar are collect,ed, and sixteen new analyses of agar from various sources submitted. These sixteen analyses s’how a remarkable uniformity in composition. High per- centages of aah (5 per cent .) or silica (1 per cent .) content tire indicative of an inferior product.Considerable amounta of nitrogenous substances (1 to 3 per cent.) were found in all the samples.50 ABSTRACTS OF CHEMICAL PAPERS A method of preparing purified agar is described. It consists essentially in washing the agar shreds in dilute acetic acid, washing out the acid, and precipit,ating, whilst hot, a 5 per cent.filtered solution of the agar by pouring into 4 or 5 volumes of 95 per cent. alcohol or acetone. This treatment removes much of the nitrogenous matter . G. C. J. Estimation of Arsenic in Organic Compounds. A. J. Ewins. ( J . Chem. Soc., 1916, 109, 1355-1358).-The substance (0.1 to 0.2 grm.) is mixed with about 0.25 grm. starch, 10 grms. of potassium sulphate, and 20 G.C.sulphuric acid, in a Kjeldahl flask, digested, and subsequently boiled until colourless. The oooled solu- tion is neutralised with sodium hydroxide, faintly acidified with sulphuric acid, and made alkaline again with sodium bicarbonat,e, of which an excess (5 to 10 C.C. of a saturated solution) is added. The solnt.ion of alkali arsenite is then titrated with 2G iodine, using starch as indicator, until the appearance of the characteristic deep blue colour.The method has been applied to a large number of organic com- pounds of arsenic, and has never been found to fail, whereas Lehmann's method (Apoth. Zeit., 1912, 27, 545), excellent for salvarsan and neo-salvarsan, fails with benzarsenic acid and with dimethyl- and diethylbenzarsenic acids. G . C. J.Determination of Light Oils in Coal Gas, and Description of Still for. separating the Light Oils from the Absorbing Oil. D. H. Duvall. ( B e t . and Chern. Eng., 1916, 15, 626-627; through J. SOC. Chem. In&., 1917, 36, 25.)-After removal of hydrogen sulphide and naphthalene by successive treatment with sodium hydroxide and picric acid, measured quantities of gas are passed through a train of five or six bottles containing a suitalcile absorbent, preferably oleic acid.The light oils are then separated from this solution by distillation. The still and water-cooled worm condenser are made of copper. After raieing the temperature of the " benzolised " oleic acid to about 100" C., steam is admitted from a separate generator, and a temperature of 180" C. is finally attained.The distillate is sepamted from water and subjected to distillation. The new distillate is saponified with sodium hydroxide to removelthe last traces of oleic acid, then washed with water, drjed with calcium chloride, and fractionally distilled in an Engler flask. Mineral Constituents of Cotton Lint. J. H. Barnes. ( J . ~ O C . CAem. Ind., 1916, 35, 1191-1195.)-Consignments of Indian cotton which had given irregular results in dyeing with aniline colours were found to contain from 0-073 to 0-014 per cent.of magnesium chloride. As this might have been artificially added by the cotton- growers to increase the hygroscopic capacity of the cotton, 18 genuine specimens of cotton lint of American, Egyptian, and Indian origin were examined with the following results : Moisture, 2.23 to 5.45 ; and ash (on dried lint), 1.27 to 3-99 per cent'.The ash contained 4.44 to 31.72 per cent. of silica and silicates, and 2.23 to 6.55 per cent. of chlorine. The results showed that the samples of Chiniot (Punjab) cotton contained an unusually high amount of magnesium chloride, but that t,here was no ground for assuming that this had been intentionally added.The total amount of ash appears t.0 have been under-estimated in the past, high results having beenORGANIC ANALYSIS 51 attributed to foreign mineral matter preslent as dirt in the bale. The ash of genuine cotton may be upwards of I per cent., and the composition of the ash may vary within wide limits. It ia highly probable that this will affect the reaction of the fibres towards dyestuffs, and i t may also affect the tensile strength and keeping qualities of the cotton.C. A. M. Use of Methylene Blue as a Reagent in Chemical Analysis. Application of the Process to the Detection and Estimation of Perchlorates in Chile Saltpetre. A. Monnier. (Arch. Sci. p h p . mi?., 1916, [iv.], 42, 210-216; through J. Chern. Xoc., 1916, [ii,], 110, 639-640.)-The following acids in dilute solutions,.preferably in the form of their alkali salts, give coloured, crystalline precipitates with a solution of methylene blue. Hydriodic acid gives a deep blue precipitate, showing bronze-green by reflection ; perchloric and hydroferricyanic acids, violet precipitates, bronze-green by reflection; persulphuric, dicbromic, and permanganic acids, rose-violet precipitihi?s, bronze-green by reflection; metavanadic, molybdic, and tungstic acids, deep blue precipitates.The precipitate with a persulphate is readily diatinguished from. that with a perchlorate in that the former when calcined leaves a slight residue showing the reactions of a sulphate, and the latter when heated decomposes violently with deflagration.The amount of potassium perchlorate can be readily estimated colorimetrically in a sample of Chile saltpetre by this method. To 20 C.C. of a 5 per cent. solution of the crude nitrate 1 C.C. of a 0.3 per cent. solution of methylene blue in water is added, and the mixture left overnight. A crystalline precipitate forms, and the supernatant liquid is tinted blue. This colour is compared with that of standard tubes containing varying amounts of pure potassium perchlorate. If the perchlorate is present t o the extent of lees than 0.2 per cent., a 10 per cent. solution of crude nitrate, and if more than 0.5 per cent., a 2.5 per cent. solution of crude nitrate, is used. If the crude nitrate contains any iodide, this can be first removed by shaking the solution with moist silver oxide.