THE PO1,TSULPHIDES OF THE ALKALI METALS. PART 1. 177XXLI.L.-The Polysulphides of the Alkali Metals. Part I.The Pohysulphides of Sodium.By ALEXANDER RULE and JOHN SMEATH THOMAS.THE preparation and constitution of the polysulphides of the alkalimetals have attracted the attention of many workers, but even atthe present day the chemical individuality of many of the productsdescribed as polysulphides remains in doubt.Bloxam (Thesis, London, 1898) has given a detailed r6sumk andcriticism of work done on this subject up t o the year 1898, andsince then important contributions to our knowledge of the poly-sulphides have been made by Bloxam (T,, 1900, 77, 753), Kusterand Heberlein (Zeitsch. anorg. Chem., 1905, 43, 53; 44, 431),and Biltz and Dorfurt (Ber., 1905, 38, 123; Zeitsch.nnorg. Chem.,1906, 48, 29'7; 50, 67).The method almost universally employed for the preparation o lthe polysulphides is the action of varying amounts of sulphur 011aqueous solutions of the monosulpliides or hydrosulphides of themetals, but this method has been shown by Bloxam to be unsatis-factory in the case of sodium and potassium. Sodium nionosulphideundergoes considerable hydrolysis in aqueous solution, and Bloxamfound that the solid products obtained after the action of sulphur011 such solutions always contained thiosulphate. When the hydro-VOL. cv. 178 RULE AND THOMAS : THE POLYSULPRIDESsulphides were used, the products were not, as a general rule, thosecorresponding with the proportions of sulphur used in the reaction.On the other hand, Biltz and Dijrfurt (Zoc.cit.), using the formermethod, were able to obtain definite polysulphides of rubidium andcaxium.The only definite sodium compound obtained by Bloxam wastetrasodium nonasulphide, Na4S,,14H,O.Kuster (Zoc. cit.) determined the solubility of sulphur in aqueoussolutions of sodium monosulphide, but he did not isolate any solidproducts. His conclusions are of interest, for he shows that mix-tures of polysulphides are formed, and that complex equilibria existbetween the different substances in solution. He points out thatthe tetrasulphide is characterised by particular stability, anobservation which is confirmed by the work of the present authors.Bottger (Annulen, 1884, 223, 335) investigated the action ofsulphur on alcoholic solutions of the hydrated monosulphide, andobtained the hydrated di-, tri-, tetra-, and penta-sulphides, butBloxam, on repeating these experiments, was unable to confirmBottger’s results.Any difficulties which are to be attributed t othe disturbing factor of hydrolysis might obvioixsly be overcome bythe use of alcohol a,s a solvent, granted that it were possible inthe first place to obtain the mono- or hydro-sulphide in the pureanhydrous condition ; until recently, however, no satisfactorymethod had been descrjbed for the preparation of the anhydrouscompounds.I n a previous paper (T., 1911, 99, 558) one of us showed that,by the action of hydrogen sulphide on alcoholic solutions of sodiumethoxide and subsequent precipitation with ether, practicallyquantitative yields of pure sodium hydrosulphide were obtained,and in a later paper (T., 1913, 103, 871) the authors mentionedthe fact that sulphur readily reacted with the hydrosulphide inalcoholic solution forming polysulphides, hydrogen sulphide beingevolved a t the same time.A systematic investigation of thisreaction has been carried out in order to determine if definite poly-sulphides could be obtained, and the results are described in thepresent paper.A series of experiments was performed in which varying quantitiesof sulphur were added to alcoholic solutions of sodium hydro-sulphide of the same concentration throughout. An approximatelysaturated solution of the hydrosulphide was prepared by adding2 grams of metallic sodium to 40 C.C.of absolute ethyl alcohol ina flask fitted with a short reflux condenser and a gas delivery tube.The solution of sodium ethoxide thus obtained was saturated withdry hydrogen sulphide, and the excess of the gas was afterwardOF THE ALKALI METALS. PART I. 179removed by heating the solution to boiling on the water-bath andpassing through it a vigorous stream of dry hydrogen. The solutionwas then heated t o boiling on the water-bath until all the sulphurhad dissolved, the current of hydrogen being maintained through-out;.I n some cases the solution was allowed t o cool and the productprecipitated with dry ether. In other cases the water wassrunout of the condenser and the alcohol was evaporated off until aconsiderable amount of solid had separated on the walls of theflask and in the solution.Amounts of sulphur corresponding with the di-, tri-, tetra-, andpenta-sulphides, as well as large excess of sulphur, were used.I nthe case of each product an estimation of the polysulphide sulphurwas made by the method described by Kuster and Heberlein(Zoc. cit.), in addition t o t2le determinations of sodium and totalsulphur.Using the proportions for the disulphide, no solid separated fromthe solution until some of the alcohol had been driven off. A paleyellow product was obtained, which, after drying in a vacuum, wasclearly not homogeneous. On treating a portion of the solutionwith more sulphur, hydrogen sulphide was evolved, indicating thepresence of unchanged hydrosulphide.In another experiment the solution was allowed, to cool, andwas then treated with exce~s of dry ether.A very pale yellowprecipitate was obtained which became almost white on drying ina vacuum, but gave a yellow solution in water. It contained aconsiderable amount of alcohol, which was evolved on heating :0.3484 gave- 0.3477 Na,SO,.0.7829 ,, 002349 S. *(S)=30.00.0.4619 ,, 1.7773 BaSO,. S=52.84.Na2S2=2 : 1 : 2.Na = 32.38.Na : (S) : S =2 : 1-33 : 2.34.Using the proportions for the trisulphide and treating the solutionwith ether, a viscid, yellow precipitate was obtained, which wasdifficult to filter. It was freed as well as possible from the solution,and allowed to remain in a vacuum desiccator over phosphoricoxide.The dry product was orange-yellow, but was not homo-geneous. It was very hygroscopic, and a solution in alcohol, ontreatment with sulphur, evolved hydrogen sulphide.It was obvious from their appearance and behaviour that thesesubstances were mixtures, and since they contained unchangedhydrosulphide, the sulphur had apparently reacted with only aportion of the latter. This fact may be explained by assuming that* (S) indicates " polysulphicie " sulphur.N 180 RULE AND THOMAS : THE POLYSULPHIDESthe reaction results in the formation of a polysulphide higher thanthe di- or tri-sulphide.Sodium Te trusuZphide.-Using the proportions of sodium andsulphur for the tetrasulphide, a very deep red solution was obtainedon boiling, and the sulphur dissolved completely.After boilingfor about an hour in a current of hydrogen, a small quantity ofglistening, yellow crystals had separated out. The solution wasconcentrated to about 5 C.C. by allowing the alcohol vapour toescape through the condenser, and the solid product, which formedcrystalline crusts on the walls of the flask, was filtered off froiiithe hot solution, washed with a little alcohol, and dried in avacuum desiccator over phosphoric oxide.The product was quite homogeneous, and was dark yellow, witha curious olive-green tinge.0.7071 gave 0.5758 Na,SO,.0.6362 ,, 0.3481 S. (S)=54.71.0.2048 ,, 1.0936 BaS04. S=73*34.Na2S, requires Na = 26.44 ; (S) = 55-17 ; S = 73.56 per cent.On heating in a capillary tube the substance became orangereda t 115-120°, began t o sinter a t 258O, and melted to a dark-redliquid a t about 267O.Sodium tetrasulphide is extremely hygroscopic, and readily dis-solves in water to form a clear, deep orange solution, which becomesdark red on heating.The solution soon begins to deposit sulphurwhen allowed to remain in the air. Alcoholic solutions behavesimilarly, but appear to be even more sensitive to the action of air.A microscopic examination of the crystals suspended in a xylenesolution of Canada balsam showed them to be quite homogeneous,and to consist of perfectly defined cubes.In the case of the tetrasulphide, therefore, it is possible to obtaina pure product by simply using the necessary proportions of sulphurand the hydrosulphide.The tetrasulphide may also be obtained by treating the con-centrated alcoholic solution with ether in excess.A red oilseparates, and it may be converted into a viscid, yellow, crystallinesolid by continued stirring. The product is difficult to filter, andit contains alcohol, which can be almost entirely removed by allow-ing the substance to remain in a vacuum over phosphoric oxide;it gradually shrinks and assumes the characteristic brownish-yellowcolour of the anhydrous tetrasulphide. The substance appears tocling very tenaciously to the last traces of alcohol, and attemptsto remove them by heating in a current of hydrogen always led t oslight loss of sulphur. The same difficulty has always been encoun-tered in the attempts to prepare anhydrous polysulphides from theNa = 26.37OF THE ALKALI METATS.PART I. 181hydrated products (compare Biltz and Dorfurt, Zoc. cit.), partlyowing to hydrolysis, and, in the case of the higher polysulphides,partly owing to direct dissociation.The product of the precipitation by ether is probably a definitealcoholate, but it is hoped that further information on this pointwill be provided by an investigation of the system sodium tetra-sulphide-ethyl alcohol, which is now being carried out by the authors.Using the proportions of sodium and sulphur corresponding withthe pentasulphide, an interesting result was obtained. The wholeof the sulphur dissolved in the boiling solution, and, on concen-trating to a small bulk, a crystalline product separated on thewalls of the flask; it wa8 collected, washed with alcohol, and driedover phosphoric oxide.The product was yellow, and not homo-geneous. On treating it' with water, a residue of sulphur remainedundissolved, and the substance was therefore thoroughly extractedand washed with carbon disulphide, then with ether, and dried in avacuum over phosphoric oxide. After collecting the washings andallowing them to evaporate to dryness, a considerable residue ofsulphur remained.The final product possessed all the characteristics of the tetra-sulphide. It was soluble in water, forming a clear solution, andmelted a t about 267O. A portion of the product, which still con-tained a little ether, was analysed:0-4986 gave 0.3978 Na,SO,.0*4050 ,, 0.2161 S.(S)=53.36.0*2947 ,, 1.5471 BaSO,. S=72*09.Na = 25.84.Na: (S): S=2: 2'97: 4.01.Na,S4 requires Na = 26.44 ; (S) = 55*17 ; S = 73-56 per cent.Using the proportions for a possible hexasulphide, some indicationwas obtained of the separation of a polysulphide higher than thetetrasulphide. The solid product was not homogeneous, but con-tained particles of a red substance. On treatment with water, acopious precipitate of sulphur was obtained. A portion of thesolid was extracted and washed with carbon disulphide, but aftercontinued treabment in this way the substance was only partlysoluble in water and still left a residue of sulphur. The finalproduct was not homogeneous, and contained carbon disulphide,which was evolved on heating. Its behaviour resembles that ofrubidium pentasulphide described by Bijtz and Dorfurt (Zoc.cit .),and it is possible that the substance contained a certain amount ofsodium pentasulphide :0*4880 gave 0.3287 Na,SO,.0'2571 ,, 1,2734 BaS04. S = 68-02,Na = 21.82.0,4634 ,, 0.2443 S. (S) =52*71.Na: (a) : S =4 : 6.92 : 8-96182 RULE AKD THOMAS : THE POLYSULPHIDESThe ratio corresponds very closely with that required by thenonasulphide, NaaS9, a hydrate of which is described by Bloxam,but the substance is almost certainly a mixture, and, moreover,contains carbon disulphide.As will be seen from the figures given below, the limit ofsolubility of sulphur is reached when the proportions of sodiumand sulphur approach Na2: S,.The results of these experiments, which were carried out underprecisely similar conditions, may be briefly summarised by statingthat, only when the proportions for the tetrasulphide are used is itpossible to obtain a pure product.Below these proportions theproducts are probably mixtures of the t e t r h l p h i d e and unchangedhydrosulphide. A t the pentasiilphide stage the solid product is amixture of the tetrasulphide and sulphur, whilst with largeramounts of sulphur there is some indication of the presence ofhigher polysulphides in the solid which is always a mixture.Kuster and Heberlein (Zoc. cit.) found that aqueous solutionsof the monosulphide, under certain conditions of concentration andtemperature, were able to dissolve sulphur up to a point repre-sented by the formula Na,S,.2i.No systematic investigation of the solubility of .sulphur inalcoholic solutions of the hydrosulphide has yet been caried out bythe authors, but one determination made just below the boilingpoint of the solution shows that under these conditions a stillgreater proportion of sulphur is dissolved, even when the solubilityof sulphur in alcohol is taken into consideration.It is probable,therefore, that even more complex polysulphides are present inalcoholic solutions than those recorded by Kiister and Heberleinas existing in aqueous solutions.I n the apparatus described above a 2N-solution of sodium hydro-sulphide in absolute alcohol was treated with an excess of sulphur,and, after the hydrogen sulphide had been removed by boiling, thetemperature was maintained a t 8l0 for about four hours, thecontents of the flask being kept thoroughly agitated by means ofa current of dry hydrogen. After allowing to settle in an atmo-sphere of hydrogen, a portion of the solution was removed by meansof a specially constructed vacuum pipette, and introduced intobromine water.When oxidation was complete the solution wasmade up t o one litre, and sodivm and sulphur were determined inaliquot portions :250 C.C. gave 0.2512 Na,SO,.250 C.C. ,, 2.8447 BaSO,. S =0*3907.Other considerations also render it probable that a t and beyondNa=0*0812.Atomic ratio, Na = 1 : S = 3'45, corresponding with Na, : S6.9OF THE ALKALI METALS. PART I. 183the tetrasulphide stage polysulphides higher than the tetrasulphideare present in the solution.A very noticeable feature of thereaction is the increase in the depth of colour of the solution withincrease in the amount of sulphur added. The tetrasulphide is ayellow substance, but the solution from which it separates in thepure state is dark red. With larger amounts of sulphur the solutionbecomes quite opaque.Valuable information as to the nature of the substances presentin the solution after the action of sulphur on the hydrosulphide,as well as to the course of the reaction, may be gained by deter-mining the amount of hydrogen sulphide evolved during thereaction.According to the equation 2NaHS -t- SS =NaZSx+l +H2S, it isobvious that the amount of hydrogen sulphide formed is strictIyproportional t o the amount of hydrosulphide involved in thereaction.There are three powibilities, namely: (1) that a series of poly-sulphides is formed according to the amount of sulphur added, asrepresented by the above equation where x=1, 2, 3, 4, 5.. . . Inthis case, if the reaction proceeded t o completion, the amount ofhydrogen sulphide evolved would always be the same, and wouldcorrespond with the complete decomposition of the hydrosulphide.It has already been shown that the reaction does not proceed inthis manner.(2) Partial decomposition of the hydrosulphide, resulting in theformation of an equilibrium mixture of different polysulphidesand the hydrosulphide.(3) The formation of only one polysulphide, whatever the pro-portion of sulphur present, involving the decomposition of anamount of hydrosulphide suEcient to produce the polysulphide inquestion.In this case the amount of hydrogen sulphide evolvedwill vary directly with the amount of sulphur present up to acertain maximum corresponding with complete decomposition of thehydrosulphide, and will then become constant. The minimumamount of sulphur required t o bring about complete decompositionwill indicate the composition of the polysulphide.A series of determinations of the amount of hydrogen sulphideevolved was carried out in order to ascertain which of the threepossibilities mentioned applies t o the reaction,The apparatus used is shown in Fig. 1. It consists of a round-bottomed Jena-glass flask of about 400 C.C.capacity, fitting bymeans of a ground gas-tight joint to a reflux condenser. A gasdelivery tube and the tube of a dropping funnel are sealed into thelower portion of the condenser, and dip nearly to the bottom o184 RULE AND THOMAS : THE YOLYSULPHIDESthe flask. A long tube sealed to the top of the inner tube of thecondenser is bent twice a t right-angles, and connected with theabsorption apparatus by means of a securely wired rubber joint.I n the apparatus figured it was found that perfect absorption wasattained with the use of a minimum quantity of absorbent.Before each experiment t k whole apparatus was carefully dried,and the absorption tubes were then connected. A weighed quantityof pure : sodium hydrosulphide WL.S introduced into the flask, aFro.I .hour, a conetantweighed quantity ofsulphur which hadbeen recrystallisedfrom carbon disul-phide was added, andthe flask was at oncefitted to the condenser.A current of hydrogendried by sulphuricacid and freed fromtraces of acid by pass-ing over solid potass-ium hydroxide wasthen passed tlwoughthe apparatus untilthe air was displaced.The amount of abso-lute ethyl a l c o h 01necessary to form anapproximately 1-5N-solution of the hydro-sulphide used was runin from the funnel,and the solution wasgradually heated toboiling on the water-bath. The boiling wascontinued for about anstream of hydrogen being maintained.In one series qf experiments ammoniacal hydrogen peroxide wasused as an absorbent, but later it was found more convenient toemploy concentrated bromine water.I n each case the sulphur wasweighed as barium sulphate.The results obtained are expressed graphically in the diagram(Fig. 2).I n the diagram (Fig. 2) the values for hydrogen sulphide andsulphur arp, calculzted for one gram-molecule of hydrosulphideOF THE ALKALI NETALS. PART I. 185The curve indicates that there is, a t first, a gradual increase inthe amount of hydrogen sulphide obtained with increase in theamount of sulphur added. The first part of the curve lies veryclose to the theoretical straight line representing the values forhydrogen sulphide, if the tetrasulphide were the only polysulphideformed, but there is a gradual divergence from this line as theamount of sulphur is increased.All points on the curve lie belowthis line, and therefore it is practically certain that little, if any,di- or tri-sulphide is formed, otherwise larger values for hydrogensulphide would be expected. This conclusion agrees with the resultsof the experiments described above.At the tetrasulphide stage, considerably less hydrogen sulphideFIG. 2.0 20 40 ti0 80 100 120 141R’a2S2 No& Na& Ka& = Ka2S6Grums of mslphur added per 56 gram of NaHS.cq Bmnine trscd as absorbent.0 J1mmoniacnl H,02 used as absorbent.is evolved than would be the case if the tetrasulphide were theonly polysulphide in solution. This is in accordance with the factspreviously mentioned, which point to the probability that a t thetetrasulphide stage higher polysulphides are present in the solution.The latter must therefore contain an equilibrium mixture of tetra-sulphide, higher polysulphides, and unchanged hydrosulphide.Beyond the pentasulphide stage the values for hydrogen sulphideare nearly constant. I f we consider the highest value obtained,which lies slightly off the curve and represents the amount ofhydrogen sulphide evolved when excess of sulphur is used, it willbe seen that this value is not very much lower than the maximumobtainable186 RULE AND THOMAS : THE POLYSULPHIDESConsidering the course of the curve as a whole there is ano very great divergence from the course of the straight lines repre-senting the theoretical values for hydrogen sulphide assuming thatthe tetrasulphide is the only polysulphide formed in the reaction,so that it is fair t o draw the coiiclusion that the tetrasulphide isthe predominating compound present.Small quantities of higherpolysulphides are also present, and the concentration of these higherpolysulphides probably increases with amount of sulphur added.It bas already been shown that a t the tetrasulphide stage theonly solid product which separates is the tetrasulphide itself, evenwhen the solution is evaporated completely to dryness. The higherpolysulphides- must theref ore decompose into tetrasulphide andsulphur as the evaporation proceeds, thus :Na,S, --+ Na,S, + S,or else react with the hydrosulphide still present as the concen-tration of the latter increases.I n either case the result will bethe formation of more tetrasulphide, and these processw will con-tinue until the whole of the alcohol is removed. Since the amountof sulphur present was that required to form the tetrasulphide,none will separate out in the free state with the solid product.At the pentasulphide stage there is again no separation of poly-sulphides higher than the tetrasulphide under the conditions ofexperiment. In this case also it must be assumed that the higherpolysulphides decompose into tetrasulphide and free sulphur, andas the latter is in excess the solid product must necessarily be amixture of the tetrasulphide and sulphur.Kuster and Heberlein mention that in aqueous solutions in whichthe proportions for the pentasulphide are used, that substance isonly obtained after a' large quantity of the tetrasulphide hasseparated out, whilst in the case of other bases the only productis a mixture of the tetrasulphide and free sulphur. Kuster andHeberlein point out that it does not necessarily follow that thetetrasulphide is the polysulphide present in predominating amount,but that its solubility product may be exceeded sooner than inthe case of the higher polysulphides. At the hexasulphide stagein the authors' experiments, there is certainly some indication ofthe separation of a higher polysulphide, and it therefore seemsprobable that a t this stage the concentration of the higher poly-sulphides is considerably increased.On evaporation of the solution,the compound decomposes to some extent into tetrasulphide andsulphur, as before, but as the evaporation is moderately rapid, acertain proportion separates in the free state.The extent t o which ionic dissociation takes place in alcoholicsolutions of such high concentration is no doubt very small, anOF THE ALKALI METALS. PART I. 187in the absence of any data it is unsafe t o introduce any reasoningwith regard to the solubility products.Considering the results generally, there is little doubt that underthe conditions described the tetrasulphide is always the chiefproduct of the action of sulphur on alcoholic solutions of the hydro-sulphide.Action of Metallic Copper on Alcoholic Solutions of thePol ysulphides.When solutions of the polysulphides are treated with excess offinely divided copper and boiled for some time, complete reductiontakes place, and a colourless solution is obtained.On removingthe black copper sulphide by filtration and concentrating thesolution in a current of hydrogen a white solid separates out whichappears to be a mixture of the hydrosulphide and monosulphids ofsodium (compare T., 1913, 103, 871).Action of Metallic Sodium om Alcoholic Solutions of theTe trasuJphi.de.When metallic sodium is added to a solution of the tetrasulphidethere is an immediate separation of a yellow, crystalline solid.This product, on examination, exhibited all the characteristicsof a polysulphide, and analysis showed i t to be the disulphide.Aseries of experiments with varying amounts of sodium alwaysresulted in the same product being obtained. The action is one ofreduction of the higher polysulphides present in the aolution bythe nascent hydrogen produced in the action of sodium on alcohol,but in this case the reduction only proceeds as far as the disulphide.No hydrogen sulphide is evolved during the reaction owing to thepresence of sodium ethoxide with which it reacts forming thehydrosulphide. The presence of the 1at;ter in the solution afterfiltration was proved by adding sulphur, when hydrogen sulphidewas a t once evolved.Addition of a solution of sodium ethoxide t o a solution of thetetrasulphide simply salts out the latter compound, but the productis always contaminated wizh sodium ethoxide, and the washingnecessary to remove it results in a considerable loss of the tetra-sulphide.Analysis of such a product resulted as follows :0.3515 gave 0.2907 Na,SO,.0.3521 ,, 1.8456 BaS04.S=71*98.0'5941 ,, 0.3186 S. (S)=53*62.Na = 26.79.Na,S, requires Na = 26.44 ; S = 73.56 ; (S) = 55-11 per cent188 THE POLYSULPHIDES OF THE ALKALI METATS. PART 1.The figures indicate that a small amount of sodium ethoxide wasstill present.Sodium JJiszilphide.The preparation was carried out by dissolving 2 grams of sodiuiiiin 50 C.C. of absolute ethyl alcohol and saturating the solution withhydrogen sulphide. 4.17 Grams of sulphur were added, and thesolution was boiled on the water-bath for about one hour, a rapidcurrent of hydrogen being passed through it.In one experimentexcess of sodium (4 grams) was added t o the hot solution, andafter boiling for a short time the bright yellow precipitate wascollected, washed with absolute ethyl alcohol, and dried in r2vacuum over phosphoric oxide :0.3166 gave 0.4066 Na,SO,.0.2697 ,, 1,1362 BaSO,. S=57*96.Na = 41.60.0.3902 ,, 0.1102 S. (S) =28.24.NazSz requires Na = 41-82 ; S = 58.18 ; (S) = 29.09 per cent.Usiiig lower proportions of sodium for effecting the reduction,a similar precipitate was obtained, but analysis showed that theratios of sodium to polysulphide sulphur and to total sulphurwere, as a rule, slightly low. It is probable that small quantitiesof the tetrasulphide separate out with the disulphide owing t o thesalting out action of the sodium ethoxide mentioned above. Thedisulphide is only sparingly soluble in alcohol, and it ought there-fore to be possible to wash out the more soluble tetrasulphide, butit was noticed that continued washing, especially with hot alcohol,appeared to bring about slight decomposition of the product, andwhite patches appeared on its surface.Sodium disulphide is a bright yellow, micro-crystalline powder.I t dissolves readily in water, forming a deep yellow solution, which,unlike that of the tetrasulphide, becomes only slightly da,rker onboiling.It is only sparingly soluble in cold alcohol, but on heatingwith alcohol an intensely green solution is produced, which becomesyellow and cloudy on further heating. This occurs to some extentin the case of the tetrasulphide, and appears t o be due t o slightdecomposition and the probable liberation of sulphur. Thedisulphide is apparently more sensitive in this respect, as a minuteamount heated with excess of alcohol gives a bright blue solution,a phenomenon which has been noticed in the case of other sulphurcompounds. The green coloration is evidently the superimposedeffect of the blue colour due to decomposition and the yellow colourof the actual solution of the polysulphide.The disulphide behaves very like the tetrasulphide on heating,as it becomes red, and finally fuses t o a dark red liquidRESEARCHES ON RESI1>UAL AFFINITY AND CO-ORDINATION. 189Since the di- and tetra-sulphides have been obtained in the pureanhydrous state, it is now possible to settle the question of theexistence of polysulphides higher than the pentasulphide, and alsoof the various intermediate polysulphides which have been described.With this object in view, we are a t present making use of thedi- and tetra-sulphides for carrying out an investigation of thefreezing-point curves of the system sodium-sulphur.INORUANIC LABORATORIES,UNIVERSITY OF LIVERPOOL