ii. 48 ABSTRACTS OF CHEMICAL PAPERS. Inorganic Chemistry. Formation of Ammonia and Hydrogen Cyanide in the Distillation of Goal. OSKAK SIMMERSBACH (b'hhl uncl &ise?z 19 14 34 1153-1159 1209-1213).-The coal to be tested is enclosed in a porcelain tube which is provided with a tlie8rmocouple and is heated in an inclined position in an electric furnace. Before beginning an experiment the air is displaced by carbon dioxide previously purified and dried. The upper end of the tube is con- nected with a manometer containing liquid paraffin with a gas- sampling vessel and with the absorption apparatus. The latter consists of two U-tubes containing glass beads and 1 1 sulphuric acid a tube packed with cotton wool and a long 10-bulb tube containing suspended ferrous hydroxide. Beyond this is the gasometer.A t the end of the experiment the carbon dioxide is washed out by means of purified nitrogen. Tar is recovered by washing out the porcelain tube sampling vessel absorbing vessels and connecting tubes with chloroform and washing this solution with dilute sulphuric acid to remove ammonia. The acid solutions are united and used for the ammonia estimation. Hydrogen cyanide is estimated in the contents of the ferrous hydroxidei tube. A Kjeldahl estimation of nitrogen in the tar and coke is also made. For coke i t is necessary to boil vigorously with concentrated sulphuric acid then add potassium sulphate and after thorough boiling complete the solution by the cautious addition of small quantities of potassium chlorate. Various analyses are given and it is shown from experimentsINORGANIC CHEMISTRY.ii. 49 with a cod from Upper Silesia that the percentage of nitrogen in the coke steadily falls with increasing temperature of distillation whilst that in the gas passes through a minimum a t 850° and then rises rapidly. The percentage of ammonia is a maximum a t 850-900° and o,nly diminishes slightly up to 1200O; 3.5 to 476 of the total nitrogen is found in the tar. The proportion of hydrogen cyanide increases slowly with the temperature. The presence of water vapour increases the proportion of cyanide and diminishes that of ammonia. C . H. D. The Vapour-Pressure Lines of the System Phosphorus. A. SMITS and S. C. BOKHORST (Proc. K . Akad. Wetensch. Amsterdam 19 14 17 678-680. Compare A. 1914 ii 533).-Furtlier investigation of the vapour pressure of red and of violet phosphorus indicates that these cannot be considered as enantiotropic modifications.%'hen equilibrium between solid and vapour is attained the data yield a continuous vapour pressure-temperature curve and the authors draw the conclusion that the red and the violet forms of phosphorus differ only in respect of the size of the particles the violet represent- ing the more coarsely crystalline form. The behaviour of red or violet phosphorus on volatilisation is considered to afford strong evidence in support of the theory of allotropy put forward by Smits (A. 1910 ii 195 400). Accord- ing t o this theory red or violet phosphorus consists of a mixture of two or more kinds of molecules in equilibrium and these pseudo- components are supposed to exhibit very considerable differences in volatility. The observed fact that abnormally high vapour pressures are recorded when red o r violet phosphorus is rapidly heated to the temperature of the measurement is readily accounted for by the absence of internal equilibrium and the excessive volatnilisation of the more volatile pseudo-component.Above 450° the rate a t which the internal equilibrium is reached becomes con- siderable and the abnormally high vapour pressures are no longer observed. H. AT. D. A Spongy Modification of Sodium Amalgam. ENRIQUE V. ZAPPI (Anal. SOC. Quim. Argentina 1914 2 210-211).-A specimen of 20% sodium amalgam contained in an imperfectly closed vessel for more than a year gradually lost its metallic appearance and developed a lead-grey colour and foliated structure.Water de- composed it with formatio'n of mercurous oxide and a spongy sub- stance similar in appearance to ammonium amalgam. A. J. W. Influence of the Gaseous Impurities of Silver on the Values of the Atomic Weights Determined by Classical Methode Atomic Weights of Chlorine and Phosphorus. PH. A. GUYE and F. E. E. CERMANN (Compt. rend. 1914 159 992-996).-Taking into account the amounts of carbon monoxide and water vapour recently found by them t o be prelsent in highly purified silver after fusion in hydrogen (compare A. 1914 ii 727 740) the authors VOL. CVIII. ii. 4ii. 50 ABSTRACTS OF CHEMTCAL PAPERS. consider that the values obtained by the use of such silver in determinations of the atomic weight of this metal may be 0.0045 too high.They further discuss the effect of such an error on the atomic-weight determination of a number of other elements based on the use of silver as a standard and find the necessary correc- tions t o be of the order 0.005 t o 0.020 and exceptionally as much as 0.040. Thus for chlorine the atomic weight should be 35.460 instead of 35-455 and for phosphorus 31,007 instead of 31.028. W. G. The Hydration of Portland Cement. A. A. KLEIN and A. J. PHILLIPS (J. Washington Acad. Sci. 1914 4,573-576).-1n connexion with the investigation of the hydration of Portland cement ex- periments have been made on the hydration of mono-calcium aluminate 5 3-calcium aluminate tricalcium aluminate niono- calcium silicate P-orthosilicate y-orthosilicate and tricalcium silicate.Observations were also made with lime burned at different temperatures and ground t o various degrees of fineness and with commercial cements cliaracterised by a high silica a low silica a high iron and a high magnesia content. The only crystalline product observed in the hydration of the aluminates was hydrated tricalcium aluminate (3CaO,Al,O,,xH,O) and this is only formed with a large excess of wat.er. The 5 3- aluminate and th2 mono-calcium aluminate eliminate amorphous hydrated alumina and form the crystalline hydrated tricalcium aluminate. I n contact with a solution of calcium sulphate the three aluminates also yield a compound of the composition 3CaO,A1,0,,3CaSO,,xH,O (" sulpho-aluminateI "). It crystallises in long prismatic needles.Monocalcium silicate and the y-orthosilicate do not undergo hydration and the P-orthosilicate oiily takes up small quantities of water after long periods. Tricalcium silicate hydrates readily and quickly yields crystallised hydrated lime and amorphous hydrated orthosilicatle. I n the hydration of c'ement the first constituent t o react is the aluminate which forms amorphous hydrated tricalcium aluminate and sometimes amorphous hydrated alumina. The sulpho- aluminate crystals are also formed and hydration of the low burned or finely ground lime also takes place. These changes occur within a few hours after the gauging of the cement. Within twenty-four hours the tricalcium silicate begins t o hydrate and this change is generally completed within seven days.Between seven and twentyeight days the amorphous aluminate commences to crystallise and the 6-orthosilicate which is the least reactive compound also begins to react. Reference is also made t o the' influence of these changes on the mechanical properties and t o the action of small quantities of electrolytes on the initial setting of the cement. The Ternary System Ca0-A120,-Si0,. G. A. RANKIN [with Optical Study by FRED. E. WRIGHT] (Amer. J. Sci. 1915 [iv] 39 1-79).-The equilibrium relations exhibited by the binary com- H. M. D.IN ORGANIC CHEMISTRY. ii. 51 binations of these three substances have been described in previous papers (A,) 1906 ii 770; 1909 ii l015) and a partial account of the ternary system has also been given (A. 1911 ii 725) in which certain quintuple points and boundary curves were provisionally located.As a result of further experimental work the equil- ibrium relations have been more exactly defined and the present paper gives a detailed record of the results. The plane projection of the concentration-temperature model shows the existence of fourteen different regions each of which corresponds with a different solid phase. The characteristic points in this diagram are recorded in a series of tables which give respectively (1) the composition of the compounds their melting points and transition points; (2) the temperature and composition of the binary eutectics; (3) similar data for quadruple points other than eutectics ; (4) the temperature and composition of ternary eutectics ; (5) similar data f o r quintuple1 points other than eutectics. A further table is given in which the optical-crystallo- graphical properties of the various unitary binary and ternary compounds are recorded.The temperature-concentration diagram permits of the prediction of the course of crystallisation of any fused ternary mixture. Applied to the case of Portland cement clinker the results indicate that this is essentially a mixture of the compounds 3CaO,SiO 2Ca0,Si02 and 3CaO,A1,0 together with some 5Ca0,3A1,03 and possibly a small amount of free lime. This conclusion is in agree- ment with the results obtained in recent work on commercial Port- land cement clinkers. H. M. D. The Supposed Allotropy of Zinc. CARL BENEDICKS and RAQNAR ARPI (Zeitsch. anorg. Chem.1914 ‘88 237-254).-Several authors have described a transformation point of zipc between 300° and 350O. A further point atl about 170° was observed in commercial “zinc puriss.” by Benedicks in 1910 but was not confirmed by Werner (A. 1913 ii 1057). Kahlbaum’s distilled zinc containing only 0.008% of impurities has been examined in the cast and also in the drawn condition. The drawn metal after annealing a t 350° is almost as flexible as lead wire. The electrical resistance-temperature curves are not well suited to showing small irregularities so that a different graphical method is employed. The abscisg being t the ordinates are W - k t where k is a constant. The highly purified zinc gives a smooth curve without any transformation point. Between 15O and 300° the resistance is closely represented by the formula W=5.45(1+ 0.0039t + 0~0000017t~). The additio’n of 0.5% of cadmium or lead gives rise to1 breaks in t3hei curve those for the cadmium alloy being at about 170° and 350° respectively. These points are attributed to the formation of limited solid solutdons of cadmium in zinc the upper temperature’ being that a t which the solidus is intersected and the lowcr that a t which the curve of solubility in the solid state is crossed.Lead has a similar but smaller influence and is usually present in smaller quantities. 4-2ii. 52 ABSTRACTS OF CHEMICAL PAPERS. Tlie experiments of Colien and Heldernian (A 1914 ii 127) are criticised on the ground that the1 ziiic used becomes partly con- verted into a basic salt by contact with the zinc sulphate solut$ion and that the washing with dilute acid may have caused a loosening of texture which would account for the diminution of density. c'.H. D. Allotropy of Zinc 111. ERNST COHEN and W. D. HELDERMAN (t'roc. K. Akud. Wetensch. Amslerdam 1914 17 641-644. Compare A. 1914 ii 127 65Z).-A criticism of the conclusions drawn by Benedicks and Arpi (preceding abstract) from their observations on the influence of temperature on the electrical conductivity of zinc. The change in the physical propertiea of zinc at higher temperatures cannot be attributed to the presence of impurities. That the change is due t o allotropic transformation has been shown by the authors in experiments with pure zinc. 13. M. D. The Allotropy of Lead. I. EENST COHEN and W.P HELDERMAN (PFOC. A'. Akad. Wetensch. Amsterdam 1914 17 822-828. Compare A. 1914 ii 652).-Experiment# are described which are said t o indicate that ordinary lead consists of a metastable mixture of several allotropic modifications of the metal. The lead used in these experiments was very pure containing only 0.001?6 of copper and 0.0006% of iron. The metal was melted chilled in water and cut into small blocks which were immersed in a solution prepared from 400 grams of lead acetate 1000 C.C. of water and 100 C.C. of nitric acid (D 1.16). After remaining for some days a t the ordinary temperature it was found that the lead shows marlred signs of structural change and after three weeks this becomes so pro- nounced as to suggest complete disintegration of the metal.The change is illustrated by a series of photographs of the metallic surface. The transformation is notl due to chemical action but by means of pyknometric measurements and observations with a dilatometer it is shown that the change is accompanied by changes in density. These variations in density are not' always in the same direction and obviously depend on the previous history of the metal. The authors draw the conclusion that there are more than twos allotropic forms of the metal in the samples under examination. Closely connected with the above observatioiis is the fact that when pure metallic lead is immersed in the solution mentioned above or in a concentrated (30%) solution of lead nitrate a lead tree is formed in a few days both a t the ordinary temperature and a t 50°. H.M. D. Tensile Properties of Copper at High Temperatures. GUY D. BENaoUGH and D. HANSON (J. Inst. Metals 1914 12 56-88. Compare A. 1912 ii 763).-Tensil.e tests of copper in a vertical machine the specimen being heated by gas show that the proper- ties of tlhe metal are greatly influenced by the composition of theINORGANIC CHEMISTRY. ii. 53 surrounding atmosphsre. The ductility a t 800° and upwards is increased considerably by the presence of oxygen. Rolled copper which has noti been annealed is stronger but less ductile than the same copper which has not been annealed after rolling a t all temperatures up to 1000° when an atmosphere of carbon dioxide is used. With annealed copper the fracture passes through the crystals a t temperatures below 700° and between the crystals at temperatures above 750O.A t intermediate temperatures i t may take either course. Oxygen or arsenic alloyed with copper de- creases the strength and increases the ductility a t high tempera- tures. An atmosphere of hydrogen has a weakening effect above 720° probably owing t o reduction of oxide. The results are best explained by assuming the existence of an amorphous intercrystalline cement which is stronger than the crystals a t low and weaker a t high temperatures (Zoc. c i t . and aIso Rosenhain and Ewen A. 1913 ii 119; Humfrey ibid. 778). This cement is probably not identical with Beilby's amorphous matrerial formed by mechanical deformation. c f . H. D. Absorption of Gases by Refined Copper. W. STAHL (Metall unnd Erx 1914 11 470-482 ; from J .SOC. Chem. Ind. 1914 33 1158).-The absorption of gases by molten copper increases in general with the temperature up to a certain point with increasing purity of the metal and with the partial pressure of the gas. The presence of nickel or platinum however has a favourable influence on the absorption. Copper in the solid o r liquid state appears not to' absorb water vapour carbon dioxide nitrogen argon or helium. At 1420° 61 grams of copper absorb 0.15 C.C. of carbon monoxide; this small quantity markedly affects the physical properties of the refined metal. Hydrocarbons are not occluded by copper but are decomposed into ca,rbon and hydrogen the latter being absorbed and the former oxidised. The absorption of hydrogen in and its diffusion through copper has been detected even a t 650O.Up t o 1 500° the absorption of hydrogen increases almost linearly with the! temperature exceptl a t the melting point of the metal where a sudden increase occurs. The solubility per 100 grams of copper is 0.1 mg. a t 650° and 1.4 mg. a t 1500O; with both solid and liquid copper the solubility of hydrogen increases as the square root of the pressure. The conductivity of copper is not affected by absorbed hydrogen. By heating copper containing copper oxide in an atmosphere of hydrogen the gas penetrates the metal and reduces the oxide and the steam produced escapes by disintegrating the metal. The dissociation pressure of cuprous oxide is equal t o the partial pressure of atmospheric oxygen 0.21 atmos.a t 1662O. When the oxide is in solution in the copper its dissociation pressure is lowered so that a t 1600O i t is certain that thermal dissociation of the dissolved oxide does not occur and the absorption of oxygen a t this temperature is due to chemical combination not t o physical solution. The preceding results indicate that during the refining of moltenii. 54 ABSTRACTS OF CHEMICAL PAPERS copper the absorption of carbon monoxide predominates during the end of the “hard poling ” period and a t the beginning of the “ green poling ” period whilst the absorption of hydrogen pre- dominates during the latter period when the oxygen has been removed. The disintegration of copper which occurs during solidification has been traced to occluded sulphur dioxide formed by oxidation of the sulphur present; the gas is evolved with (‘ spitting” during solidification but the cold metal still retains a considerable quantity of the gas.Sulphur dioxide does not diffuse through solid copper below 1000°. The reaction CuzS -t 2Cu,O = 6Cu + SO is reversible ; the equilibrium of the four substances at all temperatures has been Behaviour of Copper-Zinc Alloys when Heated in a Vacuum. W. E THORNEYCROFT and THOMAS TURNER (J. Inst. Metals 1914 12 214-229. Compare T. 1912 181 585; 1913 103 1534).- When a-alloys of copper and zinc are heated in a vacuum the rate of volatilisation is nearly proportional to the temperature but this proportionality is not observed in the alloys richer in zinc. With more than 40% of copper in the alloys the separation is quantita- tive but with alloys containing less copper a part of the copper volatilises with the zinc.The volatility of zinc is increased by alloying with small quantities of copper. proved in a number of investigations. c. s. C?. H. D. The Critical Point at 460’ in Zinc-Copper Alloys. 0. F. HUDSON (J. Inst Metals 1914 12 89-110).-The critical point observed a t 460° in brasses containing from 37 to 60% of zinc and regarded by Carpenter as a eutectoid point (A 1912 ii 764; 1913 ii 138 139) is considered to be due to a polymorphic transforma- tion of the j3-constituent. Structural resolution of /3 is not observed on annealing a t 440° even in presence of a slight excess of a o r y. Further when a heterogeneous alloy is prepared by casting a y-alloy on to copper so that a zone of /3 is obtained bordered above by p + y and below by a+@ the zone of p increases on annealing at 435O proving that a phase P’ is stable below the transformation point.The j3’ zone is also formed by diffusion of zinc into solid copper or a-brass a t 450O. It appears that j3 is more readily supersaturated with a than with y. When bronze containing 10% of tin is immersed in molten tin a t 300° diffusion takes place and a zone of the &constituent is formed in immediate contact with the a-bronze without an inter- mediate zone of a + 6 . This condition is also observed in the above experiments with zinc. [The Influence of Carbon on the Corrosion of Brass.] ARNOLD ’PHILIP ( J . Inst. Metals 1914 12 133-167).-1n a series of experiments to det,ermine the influence of contact with carbon on the corrosion of brass sections of condenser tubing containing 29% zinc and 1% tin are placed in close contact with broken coke C.H. D.INORGANIC CHEMISTRY. ii. 55 in a stream of sea-water. A series of similar specimens in contact with pebbles instead of coke is used for control. It is found that the corrosion is in all cases greatly accelerated by contact with carbon owing to local electrolytic action. This conclusion is opposed to that arrived a t by Bengough and Jones (A 1914 ii 366) and the causw of the difference are discussed a t length. C. M. D. Binary Copper-Zinc and Ternary Copper-Zinc-Lead Alloys. N. PARRAVANO (Gaxzatta 19 14 44 ii 475-502)-[ With C. MAZZETTI and R. MoRETTI.]-T~~ melting-point diagram for copper-zinc alloys differs from those of Shepherd (A.1904 ii 662) and of Carpenter and Edwards (A. 1912 ii 1057) only as regards the equilibrium horizontals at 905O 838O TOOo 595O 545O and 425O for which the corresponding durations have been determined. The temperature of the eutectoid transformation 6-y + E which the author finds to be 545O was given by Shepherd (Zoc. cit.) as 450° and by Tafel (A. 1908 ii 846) as 550O. Of the copper-zinc-lead alloys which exhibit a liquid miscibility gap seven series were examined the percentages of lead constant in the members of any one series being 2 4 10 15 20 30 and 50 respectively. I n the first series the lead remains uniformly dis- tributed throughout the whole mass of the metal. I n the series containing 4% Pb the latter remains moderately evenly distributed with the alloys rich in copper; but as the proportion of copper increases beyond 40% pronounced liquation begins and this pheno- menon becomes accentuat'ed as the percentage of lead present is .augmented. Lead is virtually undissolved in either the a- or the y-solution the lead in lead brasses being merely mixed with the crystals of the brass; this conclusion drawn from the durations of the arrests is confirmed by the results of Carpenter's investigation of the influence of addition of lead on the critical point at 470" (A.1913 ii 139). These results are confirmed by photomicro- graphs of three of the alloys containing 20L Pb. T. H. P. The Quaternary System KCl-CuCl,-BaCl,-H,O. F. A. H. SCHREINEMAKERS and (Miss) W.C. DE BAAT (Proc. K. Akad. IVetelzsch. Amsterdam 1914 17 781--783).-The equilibrium relations in this quaternary system have been described in a previous paper (A. 1913 ii 63). The solubility data at 40° and 60O are now com- municakd in detail. €1. M. D. The Basic Sulphates of Copper. H. FONZES-DIACON (Bull. Soc. chim. 1914 [iv] 15 723-727. Compare Smith Phil. Mag. 1843 [iii] 23 500).-A study of the conditions governing the formation of the basic sulphates of copper CuSO,,3Cu0,4~€,0 and CuS0,,3Cu0,5H20 and of the composition of Bordeaux mixture prepared under different conditions. The yield of basic sulphate for the same strength solutions is greater when the sodium carbonate is poured into the copper sulphate than in the reverse process. The precipitate from the same weight of copper sulphate,ii. 56 ABSTRACTS OF CHEMICAL PAPERS.is richer in basic sulphate when a very dilute solution of sodium carbonahe is poured intlo a moderately concentrated solution of copper sulphate. On slowly pouring a solution of sodium carbonate into1 a solution of copper sulphate continuously st’irred i f the Soh- tion is rich in copper sulphate tho basic sulphate C”uSO4,3CuO,4H,O is formed but as tlhe concentration diminishes the basic sulphate CuS04,4Cu0,6H20 is produced. I f an excess of sodium carbonate is used these basic sulphates are converted into’ the basic carbonate CuC‘03,Cu0,2H,0. I f the supernatant liquid is allowed t o remain a precipitate of one of the basic sulphates is formed. W. G.’ Supplement to the Phase Diagram of Aluminium-Magnesium Alloys.H. SCHIRMEISTER (Metall und Erz. 1914 11 522-523; from J. SOC. Chern. Ind. 1914 33 1159).-The alloys (200 grams in each case) were fused in a nickel wire electric furnace and the progress of the cooling was registered by a thermo-element and re- cording galvanometer. The first stationary points were found a t temperatures which gradually fell from 657O to 550° as the mag- nesium increased from 0 t o 2076. Second stationary points denoting a eutectic point were found at 580-590° with alloys containing 2-15% of magnesium and a second similar eutectic point was found a t 455O with alloys containing 12-150/ of mag- Electrolytic Iron. L. GUILLET (J. Iron Steel Inst. 1915 90 66-81).-Elect~rolyt~ic iron deposited on a large scale with a revolving cathode and high current densit’ies using a solution which is neutralised and depolarised by the addition of iron oxide is hard and brittle and contains as much as seven volumes of gas consisting of hydrogen and carbon monoxide.It has a fine struc- ture resembling that of martensite which becomes less distinct on annealing but only finally disappears a t 800-900° giving place to the normal structure of pure iron. Annealing below 700° has velry little effect on the properties. Using a solution prepared from a pig-iron containing 1% of phosphorus the iron has been found t o contain C 0.004 Si 0.007 S 0.006 and P 0.008%. The posi- tion of the critical points is unaffected by the presence of hydrogen but an additional absorption of heat a t 530-690° is observed in the heating curves of tho crude iron which do not recur on cool- ing.The electlrical properties have been examined by Broniewski (A. 1913 ii 288). Solubility of Hydrogen and of Nitrogen in Iron. E. JURISCH (Stahl und Eisen 1914 34 252).-The solubility of hydrogen in iron wire containing 0.0474 of carbon determined a t atmospheric pres- sum increases somewhat more rapidly than the temperature from 416O to 850O. The curve then rises steeply and suddenly until 930° is reached after which it is approximately parallel t o its original direction. The discontinuity is t o be attributed to the allotropio changer in the iron. The hydrogen is practically all nesium (compare Grube A 1905 ii 523). c. s. C. H. D.INORGANIC CHEMISTRY. ii. 57 liberated on slow cooling.Within all three allotropic ranges the solubility of hydrogen in iron is proportional to the square root of the pressure. The golubility of nitrogen in reduced iron powder increases very suddenly a t 930° and then slowly falls. The solubility is propor- tional to the square root of the pressure. Theory of Hardening and Constitution of Steel. EDWARD D. CAMPBELL (J. Iron XteeE Inst. 1914 90 l-l6).-The properties of steel depend largely on the nature of the carbides. It has been shown that the mean molecular weight of the hydrocarbons evolved when annealed steel is treated with an acid decreases as the per- centage of carbon increases. Hardened steel gives hydrocarbons of lower molecular weight than annealed steel. It is likely that carbides of lower molecular weight are more soluble in solid iron than those of higher molecular weight and have a greater harden- ing effect and that with increasing temperature dissociation of the more complex carbides occurs.Austenite may contain almost com- pletely dissociat'ed carbide. The fact that the hardness of hyper- eutectoid steel diminishes when the temperature of quenching is raised much above the p o h t a t which all the carbide dissolves corresponds with the theory as the hardness should be greatest when solution is complete and dissociation is a t a minimum. The facts of specific volume1 and electrical conductivity are also ex- plained qualitatively. c". H. D. The Transformations of Steels. H. DE NOLLY and L. VEYRET (J. Iron Steel Inst. 1914 90 165-180).-DiIatometric curves are given f o r a number of steels in suppotrt of the view that iron carbide is partly dissociated in solid solution (compare Campbell preceding abstract). C.€3. D. The Influence of Coalescence on the Properties of Steel and Alloys. A. PORTEVIN (J. Iron Steel Inst. 1914 90 204-212)- Bronze containing 16% of tin and consisting of the u- and &con- stituentfs is heated in a salbbath a t 525O for an hour and cooled from 5 2 5 O to 475O during five hours. After reheating rapidly to 525O the slow cooling is repeated. The eutectoid structure dis- appears the masses of 6 coalescing and the hardness measured by Brinell's test diminishes. The effect on a steel containing 25% of tin is less marked. Annealing a eutectoid steel a t TOOo for thirty hours causes coalescence of the pearlite and a great diminution of hardness.A steel containing 0.50% of carbon annealed at 800° for ten hours and very slowly cooled t o 700° shows segregation of the ferrit'e into parallel bands without coalescence of the pearlite whilst annealing for thirty hours a t 700° brings about coalescence. Methane is formed between 600° and 9 3 0 O . C. R. D. C. H. D. Behaviour of Slag Occluded in Steel. 11. Supposed Solubility of Inclusions of Slag in Liquid Steel. F. GIOLITTI and G. TAVANTI (Ann. Chim. Applicuta 1914 2 360-366).-Examinaz-ii. 58 ABSTRACTS OF CHEMICAL PAPERS. tion of the condition and quantity of the reaction slag occluded in a small mass of steel coded rapidly from the fused state and of that occluded in the same mass after re-heating followed by slow or rapid cooling renders untenable the theory of the solubility of this slag in the steel a t a high temperature.A marked difference exists between the' behaviour of the reaction slag and that of inclusions consisting largely of sulphides of iron and manganese. T. H. P. Molecular Magnitude of Metals in the Solid Phase. M. PADOA and F. BOVINI (Gaxxetta 1914 44 ii 528-534).-When a solid solution of concentration C separates from a dilute solution of concentration C and the ratio C, C is constant f o r solutions of different concentrations i t may be asserted that the solute has the same molecular magnitude in the solid as in the dissolved state. The authors have investigated three different alloys of tin cadmium and bismuth. The first two metals form a eutectic and limited solid solutions which are transformed at a temperature just above looo into the compound Sn,Cd; tin and bismuth and also cadmium and bismuth give eutectics but no solid solutions.The ternary alloys examined which solidify at about' ZOO" present no further complications. I n the three ca,ses the ratio of the con- centration of the cadmium with respect t o the tin in the crystals to that in the liquid has the values 0.986 0.985 and 0.957 re- spectively. The solid solutions separating from different liquid solutions of cadmium in tin have therefore approximately the same concentration the coefficient of distribution being sensibly constant. The conclusion is drawn that cadmium has the same molecular magnitude in solid solution in tin as in liquid solution that is i t is monatomic.These results tend t o confirm Stoffel's data concerning these ternary alloys (A. 1907 ii 357). T. H. P. The Allotropy of Antimony. ERNST COHEN and J. C. VAN DEN BOSCH (Proc. K. Akad. Wetensch. Amsterdam 1914,17,645-647). -Evidence has been obtained that so-called metallic antimony is a metastable system which consists of more than two allotropic modifications. From experiments with a dilatometer containing metallic antimony which had been melted rapidly cooled and then powdered a transition point was found at about 1 0 1 O . This agrees with the observation that the density of the substance was lowered on heating f o r four days at 102.5O. That the equilibrium relations are of a somewhat complicated character is shown by the fact that a dilatometer which had been heated for fifty minutes a t 150° before being submitted t o ex- amination a t 9 6 O exhibited for some time a diminution in volume a t this temperature but later the volume showed a steady increase.From observations of this kind it is infcrred that a t 96* metallic antimony consists of a mixture of a t least throe allotropic forms. H. M. D.MINERALOGICAL CHEMISTRY. ii. 59 Reduction of the Oxides of Antimony and Bismuth by their Sulphides. W. R. SCHOELLER (J. SOC. Chem. Ind. 1915 34 6-9).- Mixtures of antimony sulphide (Japanese stibnite) and antimony tri- or tetra-oxide (not antimonious oxide) are heated t o incipient; fusion or a t a higher temperature (965O) in a porcelain boat in a current of dry carbon dioxide.The Sb,O in the residue is ex- tracted by 5% tartaric acid and is estimated iodometrically. The sulphur dioxide evolved is oxidised by a solution of iodine and is estimated in the usual manner. Metallic antimony is produced the best yield being obtained by heating the sulpliide and trioxide (1 2 mol.) a t 965O but the process has the disadvantage that both antimony sulphide and trioxide volatilise in the current of carbon dioxide. Reduction to the metal does not occur when t'he mixtures are fused beneath sodium chloride in a crucible. Antimony is pro- duced when stibnite alone is heated gently in a current of air and subsequently at about 950° in a current of carbon dioxide. The observations recorded above are briefly discussed in con- nexion with the metallurgy of antimony. The oxide and the sulphide of bismuth react readily either in a current of carbon dioxide or beneath sodium chloride in a crucible. Sulphur dioxide is evolved a t a comparatively low temperature and globules of metallic bismuth are formed by heating more strongly. The yield is 90%; a small quantity of basic bismuth sulphate is also formed. c. s. The Effect of Hydrogen on the Annealing of Gold. JOHN PHELPS (J. Inst. Metals 1914 12 125-132).-The temperature at which hard-rolled gold is softened by annealing is raised by hydrogen. Thus the tempsrature of annealing when heating is continued for thirty minutes is raised from 150° t o about 300° by the presence of 0.02 t o 0-04 part of hydrogen per thousand. The effect is destroyed by melting the gold in air or oxygen. C. 11. D.