108 ABSTRACTS OF CHEMICAL PAPERS. In organic C h e mi s t r y. Combination of Hydrogen Fluoride with Water. By R. METZR'ER (Compt. rend,, 1894, 119, 682-684) .-Neither anh ydrons hydrogen fluoride nor a mixture containing 30 per cent. of water solidifies at - 70", but a t about -45" the " pare " acid of commerce, which contains about 43.4 per cent. of hydrogen fluoride, gradually deposits sniall, hard, opaline masses. If some of this solid mass is placed in acid containing 70 per cent. of the fluoride and cooled to -70", it sinks and rapidly dissolves. When an acid containing 55 per cent. is slowly cooled to about -45", it is converted into small, truncated prisms which hare the composition HF,H20, and melt a t -35", the temperature remaining constant until the change is complete.The crystals fume in the air, dissolve readily in the cold concentrated acid, and are heavier than it and than the commercial acid, which has a sp. gr. of 1.15. Sitice the sp. gr. of the anhydrous acid is 0,988, i t follows that mixtures of the acid and water show a point of maximum contraction, which probably corresponds with the mon- hydrate. The monhydrate readily forms supersaturated solutions. There is no evidence of the existence of a hydrate HF,2H20, although the acid that, distils without change approximates to it in composi- tion. This acid solidifies when strongly cooled, but the crystals have no definite melting point, and seem to be a mixture of the mon- hydrate and ice. 0. EL. B.INOROANIU CHEMISTRY. 109 Formation of Ozone by the Action of Electric Oscillations.By E. WIFDE3IIBNN and G. c. SCHM~DT (Bnu. 1’hy.s. C‘hem., 1894, [2], 53, 924---827).-The authors show that if, in the preparation of ozone Iqr the influence of the silent discharge, meaiis are adopted to obtain oscillations of a regular periodic nature, the yield of ozone is in- ureased. They, however, only worked at low pressures where the yield of ozone was small, namely, 0.12 per cent., under the inost favour- able conditions. L. 31. J. Gases from Wood Charcoal at a High Temperature. By Domiom (C~mpt. ?-end., lS94,119, 733--735).-W hen wood charcoal is heated to redness out of contact with air, i t gives off from 170 to 250 litres of gas per kilo. The gas has the mean composition 100.00, and its antiseptic properties are superior to those of carbonic oxide.The greater part of the gaR is probably formed by the pyro- genic decomposition of solid or liquid carbon compounds condeused in the charcoal. It would seem that in cases of poisoniug by so- called charcoal fumes, the carbonic oxide is not simply ;t result of incomplete combustion. C. X. B. Influence of certain Metals on the Stability of Ammonium Amalgam. By G. MICHAUD (Anzey. Chenz. J., 1894, 16, 438-490). -Having examined the behaviour of a saturated solution of ammo- nium chloride with mixtures of sodium amalgam and anialgams of silver, aluminium, gold, bismuth, cadmium, cobalt, copper, iron, magnesium, manganese, lead, platinum, antimony, tin, and zinc respectively, the author arrives at the following conclusions :-(i.) That when the radicle ammonium is set free in presence of an alloy of mercury with one of the aforesaid metals, its stability is inferior to what it would be when associated with mercury alone.(ii.) That a, v e q small percentage of platinum absolutely prevents the forma- tion of the ammonium amalgam. (iii.) That the power of decreasing the stability of the radicle ammonium is greater i n the electro-nega- five metals, but is not proportional to their polarity. (iv.) That in spite of the contrary conclusion of several chemists (coinpare Lan- dolt, Zeit. f. Chein., 1869, 12, 429), the amalgam of ammonium does contain ammonium, for whenever a simple mixtuye of hydrogen and ammonia is evolved during the experiment, i t has not the power to dilate the mercury and to give it the characteristic appearamce of the amalgam of ammonium.Ammonium Orthophosphate. By I(. kit^ UT (Zeit. anoy. CLem., 1894, 7, 392).-In reference to Schottlander’s remarks (this vol., ii, 64), the author points out that he had already shown in 1855 that the sparingly soluble crystals sometimes obtainzd on adding niiimonium phosphate solntion to ammoniacal liquids, have the corn- position POa(NH,),3H,0. E. C. B. Presence of Ammonia in Zinc Powder. By F. ROBINEAU and G. ROLLIN (Zeit. anal. Chem., 33, 594 ; from Jfo/dteur scientifique, [4], 7, 138).-Zinc powder contains ammonia, partly in a form COa, 9.14; 0, 0.26 ; CO, 18.08 ; H, 49.11 ; CHI, 16.04 ; N, 7.37 = G. T. M.110 ABSTRAOTS OF CHEMICAL PAPERS. soluble in hot water, partly also in combination, from which i t is set free by boiling with soda. It can be entirely freed from ammonia by boiling and washing with dilute sulphuric acid (1-loo), but, after drying in pwified air, ammonia can again be detected in it.Conversion of Black Mercury Sulphide into the Red Modi- fication and the Density and Specific Heat of the two Modifi- cations. Ey W. SrRiNCt (Zeit. anorq. Chern., 1894, 7, 371-383).- The ordinary black mercury sulphide cannot be converted into the red modification by pressure. Theoretically, it would require a pressure of 35,000 atmospheres. The author has obtained a new black modification of mercury sulphide by subliming ordinary mercury sulphide in the presence of a sufficient quantity of an inert gas. It is a micro-crystalline powder, is converted into the red modification by treatment with yellow ammonium sulphide, and has sp.gr. = 8.0395 at 17". When subjected to a pressure of 100 atmo- spheres, i t is easily converted into the red modification. The red sulphide, when heated a t 230-320", turns black, but ilegains its red colour on cooling, whether it be cooled slow!y 01- qnickly. After being heated at 410", it, however, remains black on cooling. The sp. gr. of the ordinary black modification varies from 7.6242 at 18.3" to 7-5610 a t 77.8". The sp. gr. of the red modification, obtained by treating the black modification with ammonium sul- phide, varies from 8.1289 at 21.6" t o 8.0902 at 77.7". The sp. gr. of the red modification, obtained by sublimation, varies from 8.1587 at 15.8" to 8.0978 at 77.7". If the numbers obtained between these temperatures are represented graphically, parallel curves are obtained, showing that the volume for all the modifications varies in the same manner.Similar parallel curves are obtained by representing graphically the specific heats of the three modifications at different temperatures. The specific heat of the black sulphide varies from 0.0835 at 25-30" to 0.1433 at 75-80"; that of the red sulphide from 0.0749 at 85-30" to 0.1270 at 75-80'. Temperatures of Transformation of Irons and Steels. By G. CEARPY (Compt. r e d , 1894,119, 735-737).-The metals were heated in a small furnace, the source of heat being a platinum wire, through which an electric current warns passed, and the temperatures were M. J. S. E. C. R. a,. -7 Steel containing ing.ing. e a t - Cool- - Carbon 0.07 per cent.. . Carbon 0.07 per cent., Carbon 0.08 per cent., Carbon 0.11 per cent., - ,, 0.09 ,, . . 722" 664" nickel 1.15 per cent. . 710 698 chromium0*75percent. - 675 tungsten 0.60 per cent. - 630 a2- r--7 Ifeat- Cool- ing. ing. 740" 730" 744 731 744 732 744 744 749 740 a,. -7 Heat- Cool- ing. ing. 865" 840" 903 860 835 - 860 789 923 877INORGANlC C HEMISTRT. 111 autoinatically recorded by a Le Chatelier pyrometer. The table on p. 110 gives the critical temperatures observed during heating and cooling, the critical points being distinguished by the letters a1, u2, a3 used by Osmond. With these very soft steels, the point a,, which is usiially more marked the higher the proportion of carbon, is only feebly indicated, and the point u3 varies greatly both with the nature of the metal and the rate of heating.The point u2, on the other hand, is very con- stant, and seems to indicate a distinctly reversible change, since it is produced at the same temperature during both heating and cooling. Arnold has also observed this constancy of the point a2 in soft steels. Mechanical tests made after tempering indicate that the recti- linear break in the curve of extension disappears when the metal has been heated to 750-880" and then tempered. When the same steels are heated, after having been drawn i n the cold, the curves, as a rule, are identical with those given by the annealed metal, and this is true also of the tempered metals, which indicates that the changes produced by tempering or hardening are not permanent in the cold, and disappear on annealing a t a temperature below the temperature of transformation.Byo rapid heating, however, curves were obtained in which the point 740 did not exist or was replaced by ,z convexity extending from 720" to 740". The facts observed show that the point a, (69&--'700') corresponds with the transformation of the carbon estimated by Eggertz's method, a transformation which markedly increases the hardness of the steel ; the point a2 (740") corresponds with the transformation of iron which is characterised by the rectilinear break in tlie curve of extension, a i d which slightly modifies the magnetic and mechanical properties ; the point u3 (860") corresponds with a second tr:msforniation of the iron which seems to affect more particuIarJy the magnetic properties.Action of Ferric Acetate on Potassium Iodide and Hydrogen Iodide. Ey K. SEUEX~T and R. ROHRER (Zeit. anoyg. Chem., 1894, 7, 393-405 ; see also this vol., ii, 17).--Ferric acetate and potassium iodide do not react with liberation of iodine, either in neutral solution or in the presence of a large excess of acetic acid. If, however, hydrochloric or sulphuric acid is added to the mixture, iodine is set free, and when the ratio of acid t o the ferric acetate is 3 equivalents t o 1 equivalent, the quantity of iodine liberated is equal to that liberated by an equivalent quantity of ferric chloride or sulphate. Ferric acetate liberates iodine from hgdrogen iodide, but when the proportions employed are 1 mol. hydrogen iodide and 1 equivalent of ferric acetate, the quantity of iodine liberated is considerably less than in the case of ferric chloride and ferric sulphate. When 10 mols, of hydrogen iodide are employed for 1 equivalent of ferric salt, 99 per cent.of the theoretical amouiit of iodine is liberated in all three cases. The addition of hydrochloric or sulphuric acid to the mixture Fe(C2H302)3 : HI, iucreases the amount of iodine liberated uutil it nearly approaches the amount liberated in t'he case of ferric C. H. B. chloride and sulphate. E. c. R.112 ABSTRACTS OF CHEMICAL PAPERS. Methods of Removing the Phosphoric acid as Alkali Phos- phate from Phosphates of Calcium and Iron. By E. A. SCIIXEIDEI~ (Zeit. aizorg. Cltem., 1894, 7, %86-391) .-The hydrogel obtained by dialysing a solution of ferric chloride saturated with normal calcium orthophosphate contains ferric oxide and phosphoric anhydride in the proportion 55.29 : 49-71 (Abstr., 1893, ii, 573).When the solutioll of t h i s hydrogel in ammonia is dialysed, nmnionium phosphate is removed ; after 13 days the ratio mas 83.37 Fe,O, : 17.63 P,O,. After another 15 days, the phosphoric anhydride amounted to 11.07 per cent,., and after another 19 d a p , tbe phosphoric anhydride amounted to 9-17‘ per cent. The slowness of the removal of phosphoric anhydridc after the first 13 days is probably due t o the reformation of ferric phosphate from the double compound of ferric phosphate and ammonium phosphate which is probably present in the solution. When the hydrogel is treated with potassium hydroxide.i t is almost quantitatively converted into potassium phosphate and ferric hydroxide. When treated with potassiuin carbonate, a double com- pound is formed, soluble in water, which, when allowed to remain in solution, deposits a precipitate. When this precipitate is sub- jected to dialysis, it contains phosphoric: anhydride and ferric oxide in the proportion 15.10 : 84-90, A solution of ferric phosphate in ferric sulphate, prepared by digesting normal calcium orthophosphate with n solution of ferric sulphate, and coutahiing 130.8 grams P205, 183.54 grams Fer03,. and 2:38*62 grains SO:3 per 1000 c.c., was diluted three successive times with water ; the precipitates SO produced contained phosphoric anhy- dride and ferric oxide in the proportion 1 : 1.36. 5 C.C.of the above solution, after dilution with 900 C.C. of water, contained 3.22 per cent. of the P,05 originally present, 8-72 per cent. of the Fe,03, and 92.62 per cent. of the SO,. ‘ h e author points out the application of the above results to thc manufacture of soluble phosphates from natural phosphates and Thomas slag. E. C. R. Complex Inorganic Acids : Phosphoduodeeimolybdic acid and Phospholuteomolybdic acid. By F. KEHRMANN and E. BOHN (Zeit. anovg. Chem., 1894, 7, 406-426) .-Phosphoduodecimolybdic acid is obtained by adding n cold solution of ammonium inolybdate to a cold solut.ion of sodium phosphate and nitric acid ; the precipitato thus obtained is dissolved in nitro hydrochloric acid and reprecipitated with water ; finally, it is crystallised from water containing a few clrops of nitric acid.It crystallises in beautifid oi’ange octahedra, and conbains by analysis phosphoric anhydride and molybdic anh.y- &ide in the proportion 1 : 83.89. The barium salt, 3Ba0,Pz05,24MOs, obtained by adding barium chloride to a concentrated solution of the acid, crptallises in bright yellow octahedra, resembling the bayium salt of the corresponding tungstic compound. Pho~~~ol,~teoncol2/bdic acid.-The amrnoitiicnz salt, 3(r\’H4),O,P,O,,lSM03 + 14Hz0, of tllis acid is obtained by precipitatiiig with aznmonium chloride t,heINORGANIC CHEMISTRY. 11s mother liquors of the barium salt of the duodeci-acid, after sepa- rating the bsrium therefrom with dilute sulphuric acid. It crystal- lises i n orange-red prisms: is stable i n aqneous solution at the ordi- nary temperature, but decomposes when heated with formation of ammonium phosphoduodecimoly-bdate.The potassium salt is obtained bey adding potassium chloride to the mother liquors of the ammonium salt. I t is more easily obtained by addiug potassium chloride to a, mixture of the duodeci-acid and phosphoric acid which has been allowed to remain for some time. It is more stable than the ammonium salt, and crystallises from boiling water without decom- position in short, orange-red prisms containing 14H20. The .free acid, 3H20,P20,,18M0, + Aq, is obtained by adding strong alcohol to a concentrated solution of the potassium salt acidified with sulphuric acid, then adding ether, and allowing the mixture to remain in a freezing mixture until the potassium sulphate has sepa- rated.It crystallises in long, transparent, orange-red prisms, which absorb water from the air, and effloresce when allowed to remain over sulphuric acid. When the aqueous solution is treated with potassium chloride, a quautitative yield of the potassium salt is obtained. Phosphoduodecimolybdic acid is obtained directIy by adding-yellow molybdic anhydride to tt boiling aqueous solution of phosphoric acid. The authors attempted to prepare the corresponding arsenic com- pound in the same way, but the compound obtained was arseno- luteomolybdic acid. Wheu a solution of ammonium molybdate containing nitric acid is mixed with arsenic acid in sufficient quan- tity for the formation of the 24-acid, a yellow precipitate of ammonium arsenoduodecimolybdate is formed.By boiliug this precipitate with nitrohydrochloric acid, the luteo-acid and molybdic acid are obtained. The amnzonizcm salt, 3( NHJ20,As,05,18M03 + 14H20, crystallises in beautiful, orange-red prisms, and is very soluble both in water and alcohol. The potassium salt, .3K20,As2O5,18MO3 + 14H20, is very similar to potassium phosplioluteomolybdate, and is insoluble i i r alcohol. A cold, concentrated, aqueous solution of the salt, 3K,O,P205,18M03 + 14H20, when treated with potassium hydrogen carbonate is converted into the salt, 5K20,P2O6,17MO3 (Abstr., 1894, ii, 384). The latter forms bright, lemon-yellow crystals, decomposes slowly at ordinary temperatures, and at once on boiling. was obtained in a similar manner i n bright-yellow nodules, but owing to its instability cannot be obtained pure.The authors give A lengthy table showing the reaction of the duodeci- and luteo-acids towards solutions of metallic salts and organic bases. E. C. R. Mixed Double Haloids of Antimony and Potassium. By C. H. HERTY (Amo.. Chem. J., 1894, 16, 490499).-The compounds VOL. LXVIII. ii. 9 The salt, 5 (NJL) 2 0 , p205, 17MO3,114 ABSTRACTS OF CHEMICAL PAPERS. SbC13Br3K,l+H20, 2SbCl3,Br3K3,2H20, and S bC13BrK,H,0, obtaincd by Atkinson (this Journal, 1883, 289), are i n all probability not tme salts, but isomorphous iiiixtures of the double chloride and double bromide. This view is in harmony with the observation that, after heating the supposed substance, SbC13,Br3K,,1$H,0, until a constant weight is obtained, the residue contains an appreciable quantity of antimony, and that therefore the determination of the halogens in i t by Atkinson (Zoc.cit.) affords no satisfactory evidence of the nature of the substance. The analytical results obtained by the author point to the formulse, 10SbCI3,23KCl and 10SbBr3,23KBr for the double chloride and bromide respectively, and negative the simpler formulse, SbCI3,3KC1 and SbBr3,3KBr, assigned to these compounds by Poggiale (Compt. rend., 1845, 20, 1180). G. T. M. Antimony Vermilion. By H. BAUBJGNY (Comnpt. rend., 1894, 119, 687-690) .-Antimony sulphide precipitated from solutions of the chloride by sodium tbiosnlphate a1 ways contains more or less antimony oxychloride, even when precipitation takes place in a strongly acid solution.The proportion of oxychloride is lower the higher the proportion of acid, and the higher the temperature at which precipita- tion takes place. The precipitate obtained by the action of the thiosulphate on solu- tions of tartar eme tic containing tartaric acid under the conditions described by Wagner, although cinnabar-red in colour, is practically pure antimony trisulphide, S b&, whether precipitation takes place at 80" or at the ordinary temperature. When the dried precipi- tate is treated with carbon bisulphide, very little sulphur dissolves, and the precipitate is practically insoluble i n tartaric acid. Atomic Weight of Bismuth. By R. SCHNEIDER (J. pr. Chem., 1894, [el, 50, $61--4?71) .-The author, in his earliest determinations (Anrt. Phys. Chem., 1851, 82, 303), found that the atomic weiFht of bismuth is 208 (0 = 16), and this number was confirmed by Marignac in 1883. More recently Classen (Abatr., 1890, 706) has found it to be 208.9, the same method being employed as was originally used by the author. I n view of this discrepancy, the experiments have been repeated with carefully purified material. Six experiments, in which an aggregate of more than 43 grams of metal were converted into the oxide, gave as a mean result 208.05 (0 = 16), the separate numbers varying from 207.84 to 208.15. The presence of lead in the metal used, to which Classen ascribed the difference between his results and the author's, would tend to raise the number obtained, as would also any loss of bismuth during the evaporation of the solution of the metal in nitric acid and the sub- sequent ignition of the nitrate. The author suggests the conversion of bismuth iodide into the oxide by ignition with mercuric oxide, or into the sulphide by treat- ment with mercuric sulphide, as independent methods by which further results might be obtained, but has not yet completed his experiments in these directions. A. H. C. H. B.