23 III.-Researcl~es on Vanadium. Part 111. By HENRY E. ROSCOE,B.A. Ph.D. F.R.S. (Abstracted from the Philosophical Transactions 1870 p. 317.) I. &TALL10 VANADIUM. INthe second part of my 6‘ Researches on Vanadium,”* I stated that metallic vanadium absorbs hydrogen. This conclusion has been fully borne out by subsequent experiment ; and it appears that the amount of absorbed or combined hydrogen taken up by the metal varies according to the state of division &st of the chloride (VC1,) from which the metal is prepared and secondly and especially of the metallic powder itself. When the metal containing absorbed hydrogen is exposed to dry air it slowly takes up oxygen water being formed and the vanadium undergoing oxidation to a substance which resembles the metal in its appearance but possesses a darker grey colour and has a less brilliant metallic lustre than vanadium itself.At this point the oxidation stops although in most air it proceeds still hrther. Thus a portion of pure dichloride was reduced in hydrogen; of the reduced substance free from chlorine 0.2666 grm. yielded on complete oxidation 0.4441 of V,O, corresponding to a percentage of 93.6 of pure metal. On expo-sure to the air for some days this substance absorbed oxygen losing its brilliant metallic lustre; and when it was burnt in a current of dry oxygen water was given off thus :-(1) 0-4232 grm. gave 030502 grm. of water and 0.6615 gin. ~2% (2) 0.2695 3 9 0.0315 0.414 grm. 3 99 V,% or (1) gives 87.8 p.c.vanadium; 1.3 p.c. hydrogen; 10.9 p.c oxygen. (2) gives 86.7 p.c. vanadium ; 1.3 p.c. hydrogen ; 12.0 p.c. oxygen. f Phil. Trans. 1869 p. 691; Chem. SOC. J. [2] vie 344. ROSCOE'S RESEARCHES ON VANADIUM. The difficulty of obtaining metallic vanadium perfectly free from admixture of oxide mas again rendered evident. Pure tetrachloride was prepared in quantity and from this the dichloride was made. On heating this dichloride to whiteness for forty-eight hours a substance was obtained which gained on oxidation 70.7 per cent. and therefore still contained a slight admixture of oxide. Tho reducing action of sodium on the solid chlorides was next examined ; in this case the reduction takes place rapidly but quietly in an atmosphere of hydrogen at a red heat and may be best conducted in strong iron tubes proved air-tight under hydraulic pressure of 200 lbs.on the square inch. Explosions occur when the tetrachloride is heated with sodium. The substance thus obtained by the action of sodium was found on lixiviation to be free from chlorine and on washing it mas found to separate into two parts-(1) a light and finely divided black powder (trioxide) soluble in hydrochloric acid which remains in suspension ; and (2) a heavier grey powder (insoluble in hydrochloric acid) which is deposited and which by repeated washing may be entirely fi-eed from the lighter trioxide. This bright grey powder con- sists of metallic vanadium mixed with more or less oxide. If the finely divided metallic powder after drying in wacuo be reduced at a low red heat in a current of pure hydrogen it takes fire spontaneously when cold on exposure to air or oxygen a distinct flame being seen playing on the surface whilst water is formed.In one experiment a product thus pre- pared contained 91.1 per cent. of metallic vanadium (0.354 substance gave 0.574 grm. V,O,). This substance exposed for some weeks to dry air slowly absorbed oxygen and on roast-ing gave a percentage increase of 53-75 (0.453 grm. yielded 0.6965 V,O,) whilst 0.034 grm. or 7.5 per cent. of water was at the same time formed. This shows that the point of oxida-tion at which the metal containing hydrogen becomes stable in dq air corresponds nearly to the oxide V,O. A similar slow change in the appearance of the metal has been noticed in some portions of the metallic powder placed on microscopic slides.11. VANADIUX AKD BRONIKE. 1. Vanadium Oxyiribromide or IGmaclyl T&!mmiclP VOBr3 molec. wt. = 307*3.-When the vapour of perfectly dry and ROSCOE'S RESEARCHES ON VANADIUM. pure bromine is passed over vanadium trioxide (V,O,) heated to redness dense yellowish-white fumes of the oxytribromide are formed in the heated portion of the tube and these con- dense together with the excess of bromiue to form a dark red transparent liquid. In order to free the oxytribromide from excess of bromine the mixed liquids must be rectified in cucuo as the temperature of decomposition of the oxybromide lies (under the ordinary atmospheric pressure) below its boiling-point.By distilling under a pressure of 106 millims. of mercury in a current of perfectly dry air the whole of the bromine was got rid of before the thermometer rose to 45' C. The trans- parent liquid remaining in the retort had a dark red colour gave off white fumes on exposure to moist air and when thrown into water produced a light yellow solution of a vanadic salt. It is possible to distil the oxybromide under diminished pressure with but slight decomposition occurring although when heated under the atmospheric pressure it suddenly solidi- fies at 180' C. splitting up into the oxydibromide and bromine. Under a pressure of 100 millims. of mercury the oxytribromide volatilizes without decomposition between 130' to 136O C.The following analytical results were obtained:-Analysis No. 1 was made from a portion of oxytribromide which had not been dis-tilled ; No. 2 from a portion of the same substance after further treatment with dry air at 63'; No. 3 from another preparation which had been distilled in vucuo and in which the bromine determination is too high owing to traces of free bromine. The composition of the oxytribromide is as follows :-Calculated. Found. Mean. 7-I I. 11. 111. Y V = 51.3 16.6'3 16.52 16.87 16.62 16.67 Br = 240.0 78-10 79.62 79-10 80.48 79.36 0 = 16.0 5.21 7 --307-3 100*00 The colour of the oxytribromideis somewhat redder than that of bromine and it is more transparent in thin hlins and much more translucent than bromine.The oxytribromide slowly decomposes at the ordinary atmo- spheric temperature into bromine and oxydibromide; it is very deliquescent and hygroscopic and cannot be formed in presence ROSCOE'S RESEARCHES ON VANADIUM. of moisture. The specific gravity of the oxytribromide at 0' is 2.9673 and at 14'05 it is 2.9325. 2. Vanadium Oxydibromideor VunadylDibromide VOBr, molec. wt. = 227*3.-This substance forms suddenly when the oxytri- bromide is heated to temperatures above MOO and it is slowly produced by the same decomposition at lower temperatures. The oxydibromide is a yellowish-brown solid body in appear- ance resembling ochre; it is very deliquescent and on heating in the air it loses all its bromine and is converted into the pentoxide.Thrown into water it dissolves furnishing a blue solution of hypo-vanadic (V,O,) salt. Its composition is as follows :-- Calculated. Found. Mean. V .... 51-3 22.57 23*40 21-50 22-45 Br,. ... 0 .... 160.0 16.0 70.39 7.04 71.75 - 70.11 - 70.93 - - -7- 227.3 100.00 3. Vunadium T?ibrornide VBr, molec. wt. = 291.3.-This body condenses as a greyish black opikque amorphous subli- mate when dry bromine vapour is passed in excess over vanadium nitride heated to redness. Brown vapours are given off which soon condense in the cooler portions of the tube. The tribromide is a very unstable compound losing bromine even at the ordinary temperature in dry air and being con- verted into V,O when gently heated. It deliquesces rapidly on exposure to moist air giving rise to a brown liquid in this respect resembling the tricliloride; but on addition of a few drops of hydrochloric acid the brown liquid changes to the green colour characteristic of a solution of vsnadous salt (V,O,).No free bromine is evolved when the tribromide dis-solves in water. In order to prepare the tribromide pure nitride of vanadium contained in a porcelain boat was intro- duced into a combustion tube and after all the air had been displaced by dry carbonic acid the part of the tube containing the nitride was heated to redness the other part of the tube being kept at such a temperature as to volatilize any excess of bromine which might pass over. After all the nitride had burnt away the bulb containing the bromine was sealed off and a current of dry carbonic acid passed over the solid bromide ROSCOE’S RESEARCHES ON VANADIUM.to displace all traces of free bromine. A second method of pre-paring the tribromide is to pass bromine vapour over a mixture of vanadium trioxide and charcoal; in this reaction the oxytri- bromide is first formed then the oxydibromide and lastly the tribromide VBr,; but this plan is not to be recommended as the tube soon becomes stopped up by the formation of these solid compounds. The bromide thus prepared mas not analysed but it presented exactly the same appearance and properties as the tribromide obtained by the first method. No higher compound of bromine and vanadium than the tri- bromide could be obtained.The volatile liquid passing into the bulb in the first preparation was carefully rectified and it was all found to distil over at the boiling point of bromine leaving only a small quantity of the tribromide in the bulb. Some difficulty was experienced in obtaining satisfactory ana- lytical results with the tribromide owing to the fact already observed by Stas,* that bromide of sillrer when boiled with excess of nitrate of silver carries down with it some of this latter salt inclosed in the bromide and that this impurity cannot be got rid of by washing. Owing to this admixture of nitrate of silver the bromine determinations usually come out about two per cent. too high whilst the vanadium determinations gave constant numbers agreeing much more nearly with the calculated results.Thus in four analyses of the tribromide prepared on different occasitjns the mean percentage of bro- mine was found to be 84-15 the calculated percentage being $2.4 ;whilst the vanadium determinatioiis of the same portions gave 18.57 per cent. iustead of 17.6 per cent. In order to lessen as much as possible this error the precipitated bromide of silver was reduced in hydrogen until no further diminution of weight occuri*ed and the percentage of bromine ca,lculated from this loss. Calculated. E’ound. Mean. -I/ V = 51.3 17.6 18.46 18.80 18-53 18.59 Br = 240.0 S2-4 81-21 80.58 80.78 80.86 291.3 100-0 98-67 99.38 99-30 99-45 Experiments made with the bromine employed which had been rectified over potassium bromide and was carefully tested Y Xtas Recherches sur les Lois des Proportions chimiques p.156. ROSCOE'S RESEARCHES ON VANADIUM. for chlorine and iodine and shown to be pure proved that a similar excess of weight occurred on precipitation with nitrate of silver. In one experiment the percentage of bromine thus found was 100.96 and in a second experiment 101.41. The bromine determinations of the oxybromides are similarly all too high from the same came. 111. VANADIUM AND IODINE. When vapour of iodine is passed over the red-hot nitride of vanadium contained in a tube no action whatever takes place the nitride after the operation remaining perfectly un- changed. Vanadium trioxide is likewise unacted upon by iodine at all temperatures.IV. METALLIC VANADATES. In the first part of these researches" I pointed out (1) that the vanadates analysed by Berzelius prepared by boiling vanadic acid with the alkaline hydroxides and by double de- composition must be considered as meta- or monobasic vana- dates (2) that the so-called bi-vanadates aiialysed by Von Hauer,t and prepared by acting on the metavanadates with acids are anhydro-salts and (3) that the naturally occurring vanadates are tribasic salts and tht sodium ortho-vanadate is formed when one molecule of vanadium pentoxide is fused with three molecules of carbonate of soda three molecules of carbon dioxide being expelled. I have now to describe the preparation and properties of some characteristic members of these three classes of vanadates as well as those of a fourth new class viz.the tetrabasic or pyro-vanadates. Determination of ?7amdium in the Xoldle Vanadates.-The separation of vanadic acid from the metals of the alkalies by means of chloride of ammonium as proposed by Von Hauer is apt to give too low results both as regards the vanadium and the alkali. It is almost impossible to prevent traces of amnio-nium metavanadate from dissolving and on ignition even with the greatest care some portions of the finely-divided vanadium pentovide are invariably carried off when the aminonia escapes. * Philosophical Transactions 1868 (Baherian Lecture) ; C'hcm. SOC.J.[a] vi 322. +-Jonm. prac. Chem. Bd. lxix p. 355 1536. ROSCOE'S RESEARCHES ON VANADIUM.On the other hand the volatilization of the comparatively large quantities of sal-ammoniac which must be employed in order to ensure the complete precipitation of the vanadium almost always entails a considerable loss of the fixed alkaline chlo- rides. A far more accurate plan for the separation of vanadium is the precipitation of the soluble vanadate by acetate of lead when basic lead vanadate is precipitated which is so insoluble that a portion when finely powdered and boiled in water did not dissolve in sufficient quantity to enable the lead to be detected in th'e filtrate with sulphuretted hydrogen. This salt is also insoluble in acetic but dissolves readily in nitric acid liberating vanadic acid which separates out but dissolves completely when the liquid is warmed.In the analysis of a soluble vanadate this iiisoluble lead salt is collected on a filter dried at 100' C. and weighed; a given quantity of the dried salt is then dissolved in nitric acid the lead precipitated by pure sulphuric acid and the lead sulphate determined with the usual precautions of evaporation with addition of alcohol &c. The lead sulphate thus obtained is. (contrary to B er zelius's statement) quite free from vanadium whilst the vanadic acid in the filtrate is obtained perfectly pure and well crystallised on evaporation and ignition. The filtrate from the lead vana-date freed from excem of lead by nieans of sulphuric acid and evaporated yields the alkaline sulphate not containing a trace of vanadium.Sodium Vanadates. 1. Sodium Ortl~ovanadate,Na,VO +lGH,O.-When a mixture of three molecules of sodium carbonate and one molecule of vanadium pe'ntoxide is fiised until no further evolution of carbon dioxide is observed three molecules of CO have been expelled and a tribasic vanadate remains as a white crystalline mass. In one experiment in which a slight excess of sodium car- bonate was taken 0.5785 grm. V,O liberated on fusion 0.4185 grm. COT According to the equation- the weiglt of CO liberated by this quantity of vanadium pentoxide is 0.4182 grm. The mixture is easily fiisible at fist but becomes less so as ROSCOE’S RESEARCHES ON VANADIUM. the reaction proceeds; whilst to begin with the heat of a Bunsen’s burner is sufficient to melt the mass it is necessary to apply the heat of a blowpipe-flame to keep up the fusion when the decomposition becomes more nearly complete.On cooling the solidified mass acquires first a dark green colour and then passes through yellow until when cold it ‘becomes perfectly white and is found to possess a crystalline appear-ance. It dissolves easily in cold water but is insoluble in alcohol. Hot water must not be employed for dissolving the fused mass and as little cold water as possible. The cold strong aqueous solution must be instantly mixed with excess of strong alcohol ; two layers of liquid are then formed the upper one consisting of dilute alcohol the lower one of the saline solution. After standing for a fern hours the lower layer of liquid solidifies forming an aggregate of colourless needle-shaped crystals.These crystals which possess a strong alkaline reaction are washed with small quantities of alcohol then placed on a porous plate over sulphuric acid in vacuo and after remaining for a short time they may be taken out for analysis. The analysis yielded the following results :-Calculated. Found. Na,. ... 69.0 14-61 13-80per cent. V.... .. 51-3 10.86 10.86 ) 0 .. . . 64-0 13.56 -16H20.. 288.0 60.97 60.44 , 472.3 100*00 Sodium orthovanadate is an extremely unstable compound. Its aqueous solution slowly undergoes decomposition on stand-ing at the ordinary temperature of the air out of contact with atmospheric carbonic acid whilst at higher temperatures the same change takes place quickly.This decomposition consists in the formation of a new salt sodium tetravanadate the liquid becoming strongly alkaline whilst caustic soda is liberated according to the equation- 2(Na,VO,) + H,O = Na,V,O + 2NaHO. This remarkable reaction was carefully investigated as is seen in the sequel. I have not been successful in several attempts to prepare a ROSCOE’S RESEARCRES ON VANADIUM. tribasic sodium vanadate containing basic hydrogen. All the reactions which with the corresponding phosphates yield hydrogen-sodium salts give with the vanadate the tetrabasic compound above mentioned. The orthovanadates of most of the metals are insoluble compounds obtained by precipitating neutral solutions of the soluble metallic salts with a solution of orthovanadate of sodium.The reactions of the more important metals are as follows :-Reaction of the Orthovanndates. 1. Ferric salt .,. . . . Gelatinous precipitate of a light brownish -yellow coloui; soluble in hydrochloric insoluble in acetic acid. 2. Ferrous salt ..... Dark grey precipitate. 3. Manganous salt . . Brownish-yellow crystalline precipitate. 4. Zinc salt .. .., ,. White gelatinous precipitate. 5. Cobalt salt . . . ... Brown-grey gelatinous precipitate. 6. Nickel salt . . . . . . Canary-yellow crystalline precipitate. 7. Aluminium salt . . Bright yellow gelatinous precipitate soluble in excess of both reagents ; on boiling a white precipitate is again thrown down.8. Copper salt ...,.. Apple-green precipitate. 9. Mercuric salt .. . . Orange-yellow precipitate. The reaction which Serves best to distinguish the ortho- from the inetavanadates is that of the corresponding copper salts. Orthovanadate of copper possesses a bright apple-green colour whilst the metavanadate falls down a light yellow crystalline powder. 2. Tetrasodium Vanadate or Pyrovanadate Na,V,O +18H20.-This salt crystallises in beautiful six-sided tables. It is easily soluble in water insoluble in alcohol and is precipitated from aqueous solutions by the latter Iiquia in the form of white scales of a pearly lustre. The pyrovanadate can be readily obtained by fusing one molecule of vanadium pentoxide (V,O,) with two molecules of sodium carbonate (Na,CO,) dissolving and crystal- lising.It can also be obtained by the deconiposition of the orthovanadate in aqueous solution. As long as the tetrabasic salt contains free alkali from the decomposition of the ortho- ROSCOE’S RESEARCHES ON VANADIUM. vanadate the precipitate with alcohol forms oily drops which golidify after some time only whilst the pure salt is at once thrown down in the form of silky scales. If the fusion of vana-dium pentoxide with three molecules of carbonate of soda be not completed at a very high temperature the carbonate is not fully decomposed and the fused mass when dissolved in water crystallises at once in six-sided tables or if the solution be very concentrated in nodular groups of needle-shaped crystals.The tetrabasic salt is more easily fusible than the tribasic salt and on cooling from fusion it also forms a crystalline mass. - Calculated. Found. Mean. Na ,... 92.0 V2...... 102.6 0,....,. 112.0 14.58 16.27 17’77 /14-67 16.29 - 14.79 15.71 - 1457 15.60 - -16.19 - \ -16.06 - 14-68 16.06 - 18H20,. 324.0 51.38 50.69 63.M 51.44 - - 51.84 630.6 100.00 When a solution of tetrasodium vanadate is treated with carbonic acid the salt is decomposed into sodium carbonate which crystallizes out and sodium metavaaadate which being the more soluble salt remains in solution; thus :-Na4V207 + GO = 2NaV03 + Na2C03. The insoluble py-rovanadates precipitated in solutions of the various metals possess properties generally similar to those of the corresponding tribasic vanadates.Calcium Vanadates. If to a freshly prepared solution of trisodium vanadate a solution of chloride of calcium be added a white precipitate falls down whilst the liquid possesses a strongly alkaline re- action and absorbs carboiiic acid from the air. The precipitate is a mixture of calcium pyrovanadate and calcium hydroxide; the tribasic calcium salt therefore cannot thus be obtained as it at once decomposes as follows :-Ca3V,0 + H,O = Ca,V207 + CaH20,. Calcium Pyrovanadate Ca,V,07 +2iH20.-This compound is precipitated as a white amorphous powder when a solution of ROSCOE'S RESEARCHES ON VANADIUM. chloride of calcium is added to one of the tetrabasic sodium salt. The salt dried at loo0 C.exhibited the following corn-position :-Calculated. Found. / A -Ca ........ 80.0 23.56 23-23 V,. ......... 102.6 30.21 30.16 0,.......... 112.0 32.98 - 2,1H20.. .... 45-0 13.25 12.63 339.6 100*00 Ban'um Pyrovanadate Ba,V,O,.-The dibasic barium salt is anhydrous but otherwise it closely resembles the corresponding calcium compound. It is slightly soluble in water. For analysis it was dried at 100" C. Calculated. Found. -Ba ........ 274.0 56-08 54-69 V,. ......... 102.6 20.99 21-50 1124 22-93 -0,........... -- 488.6 100~00 Lead Thiadates. Three native lead vanadates are known. (a) Lead metavanadate Pb(VO,), occurs as Dechenite. (b) Lead pyrovanadate Pb,V,07 occurs as Descloizite. (c) Lead orthovanadate and lead chloride 3Pb,(VO,) + PbCI, occurs as vanadinite.1. Basic Pyrovanadate qf Lead 2Pb,V,07 + Ph0.-TVhen a solution of the tetra-sodium vanadate is mixed with a solution of lead acetate a pale yellow precipitate is thrown down and the liquid acquires an acid reaction. The properties of this salt have already been described. For analysis the salt was dissolved in nitric acid and (with the exception of No. 1)the lead precipitated by sulphuric acid. The sulphate of lead was found to be quite free from vanadium and the vauadic acid contained no lead provided care had been taken to remove all nitric acid by evaporation and if the liquid was mixed with alcohol before the lead sulphate was filtered. VOL. XXIV. D ROSCOE'S RESEARCHES ON VANADIUM.34 Calculated. Paund. Mean. A Pb ,... 1035.0 69-92 50.92 69.85 69.83 70.57 70.29 V,. . .... 205-2 1386 13.55 12.98 13.52 13-03 13*27 015. . .. 240.0 16-22 ---. 1480.2 1OO*OO 2 Lead Orthovanadate Pb,(VO,),.-The tribasic vanadate of lead falls as an insoluble nearly white powder when tribasic sodium salt is precipitated by lead acetate. 0.7245 of the sub-stance when decomposed by nitric acid and precipitated by sulphuric acid yielded 0.1515 of V,O, or contained 11.75 per cent. of vanadium9 the percentage required by the formula being 12.04 3. Lead Orthoomadate and Lead Chloride ;arttjicial Vanadinite 3(Pb,(V O,),)PbCl, or lead trivanado-chlorhy din 2 3v0"'1 Pb, If'oxide of lead vanadic acid and chloride of lead be fused together for a few hours in the proportions in which they are contained in the above formula the mass after slowly cooliiig is found to consist of a greyish-yellow crystalline substance in the interstices of which groups of needle-shaped crystals occur.The fused mass on boiling in water is soon reduced to a powder entirely consisting of fine crystals. This crystalline powder is boiled with water until no further trace of chlorine can be detected in the washings when it is dried ready for analysis. The crystals obtained were too small for measurement ; they were however seen to consist of hexagonal prisms ; the faces of the hexagonal pyramid could not be identified. The crystals have a yellow colour and possess the waxy lustre characteristic of the natural mineral.The following gives the composition of various specimen8 of natural vanadiniles compared with that of the artificial mineral :-Naturd ranadinites. Artificial Calculated. dimapan TViuddlka$l Wleklow. Beresesow2k vanadinites* 3(Pb,(T0,),)PbCl2. (Berzelius). (Rammelsberg). (Struve). il>r(zi Lead.. . . .. 73-08 70-40 71'20 68'72 73.76 71.96 71-57 Vanadium . 10.86 -9.77 13.15 9-54 11.11 -Phosphorus. ---1.34 -Chlorine . . 2-56 2.54 2.23 2.44 2.46 2.33 2.17 I Oxygen.. .. 13-55 --- ROSCOE'S RESEARCHES ON VANADIUM. The specific gravity of the artificial vanadinite at 12' C. ie 6.707 that of the natural mineral varies from 6.66 to 7.2*. Silver Vanadates I. Silver Orthovanadate or Tribasic Silver Vanadate Ag,VO, is precipitated as a deep orange-coloured powder when a freshly prepared solution of tribasic sodium salt is mixed with a perfectly neutral solution of silver nitrate.If the precaution of neutralizing the silver solution with carbonate of soda filtering and boiling be not adopted a salt ie precipitated which consists of a mixture of tribasic and tetrabasic silver salt. The colour of this mixed salt is lighter than that of the tribasic compound and it gives on analysis a percentage of silver and vanadium intermediate between the two salts. Silver orthovanadate is easily soluble in nitric acid and am- monia. For analysis it was dissolved in nitric acid the silver being precipitated as chloride and the vanadium estimated in the filtrate.Calculated. Found. Mean. Ag .. 324.0 73.75 74.12 73-54 73.83 V.. .. 51.3 11-67 11-59 11.94 11.86 0,.... 64.0 14.58 --439.3 looooo 2. Tetrabasic Silver Vanadate Ag,V,07.-This salt is pre-pared by precipitating a solution of pure tetrasodium salt with a neutral golution of silver nitrate. It is a dense yellow pre- cipitate settling very easily when the liquid is warmed and resembling in its appearance ordinary tribasic phosphate of ~ silver. Calculated. Found. Ag .......... 432.0 66-81 66.45 V ............ 102.6 15.87 15.99 O,.. .......... 112-0 17.32 - 646% 100*00 * Pyromorphite and apatite were prepared artificially for the &st time in 1852 by Manross (Ann. Ch. Pharm. lxxxii p. SeS) and afterwards by Deville and Caron and Debray.Mimetesite has also been recently artificially prepared by Lechartier (Comptes Rendus 1867 lxv p. 1'72). ROSCOE’S RESEARU~SON VANADIUM. From the foregoing experimentrs on the vanadats it appears (lj That the soluble tribasic salts are less stable at the ordinary temperature than the tetrtlbasic compound Na,VO, splitting up in solution into free caustic soda and the pyro- aalt. (2) That at a high temperature on the other hand the tri-basic form is the most stable V,O liberating three molemilea of CO when fused with carbonate of soda but forming a mono-basic (meta) salt when boiled wit,h a solution of alkaline carbonate. (3) That as the majority of the naturally occurring vanadates are tribasic compounds we may assume that these have been produced at a high temperature.(4) That in aqueous solutions the soluble pyrovanadates are easily decomposed by carbonic acid into an alkaline carbonate and a monobasic or metavanadate. Hence the order of stability of the different vanadatea a& the ordinary temperatures is as follows :-(1) Monobasic or metavanadates. (2) Tetrabasic or pyrovanadates. (3) Tribasic or orthovanadates. 111the phosphorus series the order of stability is (as is well known) exactly the reverse of this the tribasic phosphoric acid and soluble orthophosphates being most stable and being formed from the other two classes of acids and soluble salts either by ebullition alone or in presence of weak acids.