abundance. These attempts to effect the conversion into active lactic acid having failed, the rest of the ethylic chloropropionate, 8-7 grams, was dissolved in aqueous alcohol, and hydrolysed by adding to it slowly the calculated quantity of barium hydrate dissolved in 100 c.c. of warm water. The solution was neutral, and was found to have a slight lævorotation, from which it appears that the activity of barium chloropropionate is in the opposite sense to that of the acid. The chloropropionic acid was liberated by adding the calculated quantity of sulphuric acid. To the filtered solution of the acid, which was dextrorotatory, rather more than the quantity of silver oxide required to form the silver salt was added; silver chloride was formed in abundauce on warming the solution on the water bath, and the solution became very acid. The completion of the reaction was ascertained by boiling some of the filtered solution with excess of potassium hydroxide and finding that no more chloride was produced. The filtered solution was next saturated with sulphuretted hydrogen, again filtered, and then concentrated by evaporation. The solution of lactic acid thus obtained was lævorotatory, and on neutralising it with zinc carbonate the rotation was reversed. The deposit of crys *talline zinc salt obtained from the concentrated solution, after being recrystallised and dried at 130°, was found to contain 26.90 per cent. of zinc; the calculated percentage for anhydrous zinc lactate is 26-87. The water of crystallisation in the salt was 14:01 per cent., the calculated percentage for 2H2O being 12.88. The salt, therefore, contained a considerable amount of inactive zinc lactate, a result which was to be expected from the observed activity of the ethylic chloropropionate from which it was recovered. The presence of the inactive salt was also borne out by the specific rotation, a determination of which gave the following result: t = 8°, l = 4°, c = 7·48°, a = +1·72°, hence [a]p = +5.75°. The specific rotation of the pure active salt, under similar conditions, is +-6-29°. The chief part of the racemisation probably occurred during the production of the ethylic chloropropionate from ethylic lactate, for the zinc lactate from the latter contained, as has already been shown, much less racemised salt. Lactic acid was set free from the inverted zinc salt by sulphuretted hydrogen, and its solution was found to be lavorotatory. The remarkable inversion of optical activity, discovered by Walden, of which we have furnished another illustration, seems at first sight inexplicable by the current theory of the asymmetric carbon atom. Armstrong (Proc. Chem. Soc., 1896, 45) has, however, offered a simple and ingenious explanation of the apparent anomaly which quite accords with the ordinarily accepted views. He supposes that by the action of phosphorus pentachloride on a hydroxy-acid, an inter mediate compound containing the complex &<, may be formed, , which, by the action of hydrochloric acid is resolved into a chloro-acid and phosphorus oxychloride, and points out that if in this reaction. "the attack became directed by the phosphorus, so that the chlorine took up the position of the phosphorus, complete inversion would be effected." Walker found, in preparing the active ethereal halogen-propionates from lactic acid, that a small yield of the ethereal salt was obtained, and that in each case a large quantity of a high boiling compound was produced, which contained phosphorus, and was strongly active in the sense opposite to that of the propionate. We had a similar experience in preparing ethylic chloropropionate from lævorotatory etbylic lactate. The fraction of oil boiling above 80° gave, as already stated, a rotation of -3.67° in the 100-mm. tube. The oil went rapidly into solution on boiling with water, and on neutralising this solution by adding zinc carbonate, much zinc phosphate was precipitated, which was separated by filtration. The zinc salt of the acid produced in the reaction was much more soluble in cold water than in hot, and was obtained in the crystalline state by warming the solution and filtering. Its solution in water was lævorotatory. The acid, obtained by decomposing the salt with sulphuretted hydrogen, was a syrup, and was strongly lævorotatory in aqueous solution. It contained phosphorus, as phosphoric acid was abundantly formed on boiling it with nitric acid. The action of phosphorus pentachloride on the active ethereal lactates will be further examined in order to ascertain if the bye-product referred to is of the nature suggested by Armstrong. United College of St. Salvator and St. Leonard, LII.-The Hydriodides of Hydroxylamine. By WYNDHAM R. DUNSTAN, F.R.S., and ERNEST GOULDING. In a previous paper (Proc., 1894, 138), the authors showed that methylic iodide reacts at the ordinary temperature with excess of a solution of hydroxylamine in methylic alcohol, much heat being evolved. The solution of hydroxylamine is prepared by the interaction of molecular proportions of hydroxylamine hydrochloride and sodium methoxide dissolved in methylic alcohol. On evaporation, a mixture of hydriodide of trimethylhydroxylamine and hydriodide of hydroxylamine crystallises out; these compounds are separated by dissolving in methylic alcohol and fractionally precipitating with ether. The hydriodide of hydroxylamine thus obtained contains three molecular proportions of hydroxylamine combined with one molecular proportion of hydrogen iodide (NH2OH)3,HI. By further evaporation of the residual solution, we have isolated a second, and apparently definite, hydriodide containing two molecular proportions of hydroxylamine combined with one molecular proportion of hydrogen iodide (NH2OH),HI. The normal salt, NH2OH,HI, could not be found. The same hydroxylamine hydriodides are produced by the interaction of hydroxylamine and ethylic iodide. Both the salts are well crystallised and comparatively stable, especially the trihydroxylamine compound. They have since been obtained by the direct combination of hydroxylamine (dissolved in methylic alcohol) and hydrogen iodide (dissolved in a little water); after evaporation, one or other salt or a mixture of the two, depending on the proportion of hydroxylamine used, crystallises out. We have not succeeded in crystallising any other definite hydriodide of hydroxylamine, and it is remarkable, especially in view of the stability of the normal hydrochloride, that no method we have tried, either direct or indirect, has furnished a crystalline normal hydriodide, NH2OH,HI. When a solution of hydroxylamine in methylic alcohol is mixed with the calculated quantity of aqueous hydrogen iodide to form this salt, the solution quickly darkens in colour from the liberation of iodine, and when evaporated, either by heating or by exposing it in a desiccator over sulphuric acid, further decomposition occurs. Decomposition, under various conditions, of hydroxylamine hydrosulphate, (NH2OH)2,H2SO,, with calcium iodide, barium iodide, or potassium iodide did not lead to the isolation of the normal compound. Trihydroxylamine hydriodide, (NH2OH),,HI, forms thin, crystalline plates which do not undergo any change in dry air, but in moist air gradually deliquesce, losing hydroxylamine. It is soluble in water and in methylic and ethylic alcohols, but not in ether; if this salt is recrystallised, either from hot methylic alcohol or especially from water, it gradually loses hydroxylamine, passing by degrees into the dihydroxylamine compound, which accounts for the higher percentages of iodine found in the analyses of certain specimens recorded below; but this loss does not occur if a little free hydroxylamine is present in the liquid. The salt is acid to litmus paper, and behaves as a powerful reducing agent. When gently heated in the dry, state, it suddenly decomposes in the neighbourhood of 100°, nitrogen and water being evolved and ammonium iodide left as a residue. We find that the amounts of these products are formed in accordance with the equation (NH, OH)3,HI= NHI + N2 + 3H2O. Various specimens of the salt have been analysed, the iodine being estimated in the usual manner as silver iodide, and the hydroxylamine by titrating the aqueous solution with standard iodine in the presence of magnesia. I. NH,O= 41.03 per cent.; I II. NH2O = 41.1 per cent. ; I = 58.5 per cent. 584 per cent. III. NH2O = 43.7 per cent.; I = 55.7 per cent. IV. NH2O = 43 59 per cent. V. NH,O= 43.8 per cent. Calculated for (NH2O),,HI; NH2O 436, I= 55.9 per cent. Specimen III was recrystallised from water in the presence of an excess of hydroxylamine; specimens I and II were recrystallised from methylic alcohol, and specimens IV and V from the same solvent in the presence of a slight excess of hydroxylamine. Dihydroxylamine Hydriodide, (NH, OH),,HI, is formed when the proper proportion of hydrogen iodide (in water) is mixed with a very slight excess of hydroxylamine (in methylic alcohol), or it may be obtained by evaporating the mother liquor from the trihydroxylamine salt. The crystals, which resemble in appearance those of the trihydroxylamine compound, deliquesce more quickly in moist air. When recrystallised from water, iodine is liberated. The salt is acid to litmus paper, and acts as a powerful reducing agent. It is more soluble in water and in methylic and ethylic alcohols than the trihydroxylamine compound, and, like this substance, is insoluble in ether. It is best crystallised from its solution in hot methylic alcohol, although even in this case some of the salt is decomposed. Different specimens were analysed with the following results. I. NH,0 = 34.5 per cent. II. NHO 33.9 per cent.; I = 66.7 per cent. III. NHẠO = 334 per cent. Calculated for (NH,O)2,HI; NH2O = 34.02 per cent.; I = 65·45 p. c. From a solution of the di- and tri-hydroxylamine salts in methylic alcohol, the trihydroxylamine compound crystallises out first, and in this way a partial separation of the two salts may be effected. There seems to be no obvious explanation of the molecular associa tion which hydroxylamine exhibits in these salts. It is interesting to note that one of us has already shown that the simplest known oxime, formaldoxime, CH,NOH, also exhibits the peculiarity of forming salts of the same type, in which 3 mols. of the oxime are combined with 1 mol. of the halogen hydride, (CH2NOH), HCI, &c. (Dunstan and Bossi, Proc., 1894, 55). The instability of the normal hydroxylamine hydriodide as compared with hydroxylamine hydrochloride may be partly explained by the readiness with which hydrogen iodide reduces the single molecule of hydroxylamine to ammonia. In the case of hydrogen chloride, besides the stable, noriaal salt of the formula, NH, OH, HCI, Lossen (Annalen, 1871, 160, 242) has described two other compounds, one, (NH ̧ƆH),,HC, corresponding with one of the hydriodides now described, the other of the formula (NH,OH)3,HCI)2. The latter salt, however, requires re-examination, since the evidence for its existence offered by Lossen does not appear conclusive. We have not been able to isolate the corresponding hydriodide. Research Laboratory, Pharmaceutical Society, London. LIII.-The Determination of the Composition of a "White Sou" by a Method of Spectragraphic Analysis. By W. NOEL HARTLEY, F.R.S., Professor of Chemistry, Royal College of Science, Dublin. A FRENCH Coin, formerly known as a "white sou," was given to me by my colleague Professor J. P. O'Reilly, Foreign Secretary of the Royal Irish Academy, with the information that such coins were at one time commonly in circulation in France, but had now become obsolete. It was supposed that they had been made during the Revolution of 1798, being hurriedly struck from any dies which came to hand, and from an alloy into which such suitable metals had entered as were most easily obtainable and readily appropriated, such as copper roofing, bronze cannon, and church bells. It was suggested to me that it probably contained silver. Its colour was that of pale gold, that is to say, of the precious metal in the natural state when it contains about half its weight of silver. It was curious that it did not ring, but this might have been accounted for by the edges being slightly cracked, the more probable reason, however, being the peculiar composition of the coin which occasioned its cracking. As it was possessed of some historical interest, it became desirable that its composition should be ascertained without injuring it, and accordingly I resolved to make its analysis by the spectroscopic method, which may be more accurately described as spectrographic. The quantitative analysis by means of the spectrograph has been described by me in the Phil. Trans. (175, 49; also 325, 1884), and occasionally put into practice by executing analyses of the old brass coinage of James II and of silver coins of the reign of Edward I, Edward III, and Alexander of Scotland, which latter were found in a |