624 XLVII.-The Action of Heat on the Halts of Tetramethylammoniurn. By A. T. LAWSON and NORMAN COLLIE PhD. F.R.S.E. THE following experiments on the action of heat on the tetramethyl-ammonium salts were undertaken in order t o ascertain in what respects the compound ammonium salts when subjected to the action of heat resembled and also in what points they differed from the corresponding phosphorus and sulphur salts. The action of heat on a considerable number of the trimethyl-sulphine salts (chiefly the salts of sulphur acids) has been investi-gated by Crum-Brown and Blaikie (J. pr. Chern. [a] 23 395); whilst the action of heat on the tetramethylphosphonium salts is the subject of a separate paper by one of us (p. 636) but the decomposition that the tetramethylammonium salts suffer when heated has only been studied in a very few cases.Hofmann (Ber. 14 494) has noticed that the hydroxide is easily decomposed by heat-(CH,)aN*OH = (CH,),N + CH,*OH, whilst Thompson (Bey. 16,2339) has found that the cyanide volatilises unchanged. The salts that tetramethylammonium hydroxides form with the more common acids have not received much attention since Hofmann discovered these interesting compounds nearly 40 years ago. Duvillier and Buisine (Ann. Clzi9n. Phys. [ S ] 23 331) have shown that methyl nitrate and also methyl bromide when heated with metbylamine or with dime thylamine yield te trame t h ylammonium Compounds but they did not prepare many of the salts Hofmann mentions that crystalline salts can be prepared by neutralising the base with sulphuric oxalic or nitric acids and he also prepared the chloride but he gives no details as t o the solubility crystalline form, or other properties of these compounds.The bromide and cyanide have been prepared by other workers and also a number of double salts ; but the chief work on the tetramethylammonium compounds Seems to have been directed towards the preparation of a number of addition products obtained by the action of chlorine bromine or iodine on the halogen salts (Waltzien Annahz 99 1 ; Stahlschmidt, Jahresber. 1863 403 ; Dobbin J. Chem. Soc. Trans. 1886 846). The starting point from which we prepared all our saltis was the iodide of tetramethylammonium. It was made according t o the method suggested by Hofmann. Methyl alcohol saturated with ammonia gas was digested with methjl iodide in sealed tubes at a temperature of 1.00-120".It was found better to put the methy ACTION OF HEAT ON SALTS OF TETRAMETHYLAMXONIUM. 625 iodide in a separate tube (open at one end) and not to mix it with the ammoniacal methyl alcohol till after the tube which contained both had been sealed. The yield varied considerably in different tubes several tubes after heating for only a few hours being filled with crystals while the yield in other tubes even on prolonged heat-ing a t a temperature of 140" did not seem to be materially increased. We believe this was due possibly to the presence of small quantities of water in the methyl alcohol. The iodide of tetramethylammonium can be easily separated from the iodide of ammonium and from the hydriodides of mono- di- and tri-methylamine by crystallisation from water in which it is not very soluble.The mean of several analyses of the salt gave 63.1 per cent. of iodine the theoretical amount required by (CH,),NI being 63-2 per cent. iodine. When this salt is heated i t decomposes a t a temperature not much fihort of a low red heat without melting giving trimethylamine free iodine and other products. It probably first splits up into trimethyl-amine and iodide of methyl-and at the high temperature both the trimethylamine and the iodide of methyl are partially decomposed (the latter almost completely), (CH,),NI = (CH3)SN + CHJ, Action of Heat on Bromide of Tetramethylanzmonium. This salt was prepared by neutralising hydroxide of tetra-methylammonium with hydrobromic acid and evaporating the solution over the water-bath.On allowing the salt to remain over sulphuric acid in a vacuum it soon crystallised. An analysis of the salt gave tha following results 0.430 salt took 27.7 C.C. decinormal AgN03 solution = 51.5 per cent. bromine ; theory for (CH,),NBr = 51.9 per cent. bromine. It forms needle-shaped crystals which are deliquescent and their solubility in water is considerably greater than %hat of the iodide; 100 C.C. of water a t 15" dissolve 55.26 grams of the salt. The salt on heating to 300" in a vacuum gave off no gas b u t on raising the temperature above 360" the salt sublimed and condensed again on the walls of the tube (which was used as a condenser and was surrounded by a f'reezing mixture) in the form of a white powder.This solid was analysed and from a bromine determination proved to be pure bromide of tetramethylammonium. (It contained 51.6 per cent. bromine.) At the end of the experiment the whole of the salt had sublimed and scarcely a trace of permanent gas had been evolved. (CH3)JY"r = (CH,),N + CH,Br, Evidently the bromide on beating dissociates-and the two gases a t once recombine on cooling. VOL. LlII. 2 626 LAWSON AND COLLIE THE ACTION OF HEAT Action of Heat on Chloride of Tetramethylarnmonium. The chloride was prepared by treating the iodide with hydroxide of silver and neutralising the base thus obtained with hydrochloric acid. The solution was evaporated over the water-bath to the consistency of a syrup and allowed to stand in a vacuum over sulphuric acid when it soon crystaIlised.Attempts made to determine its solubility in water were unsuccessful on account of its great solubility ; it is also deliquescent. The salt was heated in a vacuum a t 150" till perfectly dry when a chlorine determination gave 32.4 per cent. of chlorine : (CH,),NCl contains 32.4 per cent. C1. When the salt is heated to above 360° it decomposes completely without the slightest charring yielding trimethylamine and a gas which burnt with a greenish flame and behaved in every way like chloride of methyl. I n one experiment 3.52 grams of salt yielded 750 C.C. of gas. The theoretical yield of methyl chloride would be about 800 C.C. The trimethylamine was converted into the chloropbtinate and gave the following numbers on analysis.Theory 0.144 gram salt gave 0.0527 gram Pt = 36.60 per cent. Pt. The gas was also analysed :-for (Me,HNC1),PtC14 = 36.93 per cent. Pt. Gas taken 8.0 C.C. Gas arid oxygen . 31.0 ,, After explosion and the addition of a few drops of water 19.8 ,, After addition of caustic soda. . 12.0 ,, This shows that one volume of the gas on explosion with excess of oxygen yields nearly its own volume of carbon dioxide which is the amount required by methyl chloride. The decomposition of the chloride is nearly quantitative and can be expressed by the following equation :-(CHS)4NC1 = (CH3)SN + CH3C1. Action of Heat on Fluoride of Tetramethy kammortium. After the extremely neat manner in which the chloride decom-posed when heated we had every reason to expect that the fluoride would decompose in an analogous manner yielding trimethylamine and methyl fluoride and the decomposition besides being of interest in illustrating the general method of decomposition of the tetra-methylammonium salts would also be useful in the preparation of organic fluorine compounds ON THE SALTS OF TETRA4METHYLAMMONIUM.627 Although fluorine itself has recently been isolated still on account of the great difficulty of uniting fluorine with carbon the number of organic compounds where fluorine is combined with hydrocarbon radicles of the paraffin series is small and their properties have not been much studied. The salt was prepared in a manner similar to that employed in the manufacture of the chloride The base was neutralised with hydro-fluoric acid and evaporated over the water-bath to a syrup.This solidified on cooling to a solid mass of radiating crystals. They were, however by no means dry for on heating at 100" in a vacuum a considerable amount of water was lost. Great difficulty was experienced in drying the-salt completely and it could only be accomplished by prolonged heating in a vacuum at 160". A deter-mination was made of the amount of water contained in the salt which had been dried over sulphuric acid. 0.900 gram salt heated at 160" in a vacuum lost 0.150 gram H20 = 16.6 per cent. H,O. Theory for Me4 N F H,O . Found. B,O . . . . 16.2 per cent. 16.6 per cent. The tube in which the salt was heated waq not etched showing Some of the dried salt that no hydrofluoric acid had been liberated.was analysed :-I. 0-500 gram salt gave 0.19'75 gra,m CaF2 = 19.2 per cent. 11. 0.2727 gram salt gave 0.1186 gram CaF = 21.1 per cent. fluorine. fluorine. Found. Calculated for rA-7 Me,NF. I. 11. 3'. . . . . . . . 20.4 per cent. 19.2 21.1 per cent. The great difference in the percentage of fluorine found is pos-sibly due to the great difficulty in obtaining the salt pure and dry at the aame time for if the salt be heated even as high as 160" in a vacuum it is not perfectly anhydrous and if the temperature be raised above that point it begins to slowly decompose. When the salt is heated to 180" in a vacuum it be,gins to decom-pose and yields trimethylamine and a gaseous substance. During the first experiments made on the action of heat on this salt the tri-methylamine was condensed by passing the products of the decompo-sition through a U-tube surrounded by a freezing mixture.Sub-sequently it was found better to absorb the trimethylarnine by p=rmice-2 u 628 LAWSON AND COLLIE THE ACTION OF HEAT stone moistened with sulphuric acid (which absorbed only the tri-methylamine). The trirnethylamine obtained was converted into the chloroplatinate and analysed (0.362 gram salt gave 0.132 gram P t = 36.6 per cent. ; theory for (Me3€€NC1),PtCI = 36.9 per cent. The amount of tetramethylammonium fluoride which was decom-posed was also noticed as well as the amount of gas evolved. 2.614 grams of the fluoride gave 400 C.C. of gas which proved to be fluoride of methyl.This gas was first prepared by Dumas and Peligot (Anmaleiz 15 59) by heating together potassium fluoride and potassium methyl sulphate. The gas prepared by heating the fluoride of tetramethylammonium was slightly soluble in water but more so in alcohol ; it burnt with a blue flame yielding hydrofluoric acid and it had a pleasant odour. On analysis it yielded its own volume of carbon dioxide :-Pt). Taken of gas After explosion. . 39.0 ,, After addition of caustic soda. 31.0 ,, 8.0 C.C. Gas and oxygen 51.0 ,, thus showing that 8.0 C.C. of gas yielded 8.0 C.C. of carbon dioxide. The fluoride of tetramethylammonium therefore decomposes in a manner similar t o the chloride :-Action of Heat on Nitrate of Tetramethylammonium. The salt was prepared by the action of silver nitrate on the iodide.Evaporated over the wate~bath to a small bulk the solution crystal-lises in long needles ; when pure it does not seem t o be perceptibly deliquescent. Several grams of the salt were heated. No decompo-sition occurred till the temperature had risen t o above 300". Slight blackening then took place but it was found that if the temperature was kept just at the melting point of the salt nearly the whole of it decomposed without much charring. A yellow liquid was found in the condensing apparatus and a small quantity of gas was produced, consisting chiefly of nitric oxide but on treatment with oxygen and water there remained a small amount, which from its properties seemed t o be methyl nitrate ; the quantity however was too small for identification.The yellow distillate was alkaline and smelt strongly of trimethylamine it was therefore warmed with water in order t o free it from the base and then treated with a small quantity of pure caustic soda j on warming a further quantity of trimethylamin ON THE SALTS OF TETRAMETHYLAMMONIURI. 629 was evolved which gave the characteristic chloroplatinate. The remaining caustic soda solution was treated with carbon dioxide and evaporated to dryness it was then extracted with absolute alcohol and the alcoholic solution of the sodium salt evaporated to dryness. Thiis treated a smd1 quantity of a soluble sodium salt was obtained. An attempt to prepare the silver salt wa,s unsuccessful owing to the reduction of the salt to metallic silver ; with mercurous salts a similar reduction took place and the sodium salt itself gave carbon mon-oxide when heated with strong sulphuric acid ; the salt in questtion was, therefore probably sodium formate.From this it will be seen that the decomposition of the nitrate is complex trimethylamine alone being formed in any quantity. I f methyl nitrate is also formed it is decomposed yielding as oxidation prodncts formic acid &c. and as reduction products methyl nitrite and nitric oxide. Action of Heat on Nitrite of Tetmmethylamrnonium. This salt was prepared by treating a solution of iodide of tetra-methylammonium with nitrite of silver. It is deliquescent and much more soluble in water than the nitrate When subjected to the action of heat it decomposed at a temperature above 300" with great rapidity ; trimethylamine was produced and at the same time a small quantity of an orange-coloured oxide of nitrogen mixed with a considerable amount of some other gas was given off.On treat-ment with caustic soda there was little diminution in volume but on adding oxygen the orange peroxide of nitrogen was produced which dissolved in the caustic soda solution. About half of the gas collected consisted of nitric oxide and the remainder was inflam-mable. This residual gas was mixed with an equal volume of oxygen (under the supposition that it was methyl nitrite) and exploded in a eudiometer. The result was unexpected for the two volumes oi mixed gases became nearly five volumes and the remaining gas was inflammable ; evidently the oxygen used was not nearly sufficient for its combustion and another experiment was therefore made with the following quantities :-Gas used .5.0 C.C. Gas and oxygen 37.0 ,, After explosion. . 24.0 , After treatment with caustic soda . 14.4 ,, Thus 5.0 C.C. of the gas yielded 9.6 C.C. of carbon dioxide or nearly twice the volume and the remaining gas in the eudiometer was nearly pure oxygen. This proved that the gas under examinatio 630 LARSON AND COLLIE THE ACTION OF HEAT did not contain nitrogen and also that two atoms of carbon were present in the molecule; in all probability the substance was methyl ether. For when mixed with its own volume of oxygen and exploded two volumes would give five volumes of mixed and inflam-mable gases :-CZHeO + 0 2 = co + co + 3H, Lpv-2 L-,___J whilst if it were exploded with excess of oxygen it should yield twice its volume of carbon dioxide :-2C2H60 + 602 = 4c02 + 6H20.Evidently then when the nitrite is heated it is decomposed into trimethylamine and nitrite of methyl the latter being further decom-posed into methyl alcohol iiitric oxide and oxygen :-2 vols. 5 vols. (CH,),N.NOZ = (CH3)3N + CHJTOZ. 4CH3N02 = 2(CH,),O + 4NO + 0 2 . The nitric oxide and oxygen on cooling combined forming some of the higher oxides of nitrogen which probably united with the tri-met h ylamine. Action of Heat on Acetate of Tetramethylamrnonii~m. The hydroxide of te tramethylammonium was neutralised with acet8ic acid in order to produce this salt ; the sclution was evaporated, and solidified to a mass of aeedle-shaped crystals when allowed to stand in a vacuum over sulphuric acid.The salt was highly deli-quescent. When heated it melted a t about 70° and a small quantity of water distilled ; a t 190-200" complete decomposition took place. The condenser which was surrounded by a freezing mixture con-tained a liquid which on the addition of water separated into two layers one of' which proved to be an aqueous solution of trimethyl-amine whilst the other was acetate of methyl. This was proved beyond doubt by the boiling point 57-58' (methyl acetate b. p. 56") and by its conversion into methyl alcohol and sodium acetate On treatment with caustic soda. The sodium acetate was converted illto the corresponding silver salt and then analysed.0.5395 gram salt gave 0.348 Ag = 64.5 per cent. ; theory f o r AgC2H3O2 = 64.6 per cent. Ag. Some of the liquid b. p. 57*8" was shaken with a con-centrated solution of acid sulphite of sodium but no crystalline double salt was formed. The decomposition of the acetate is there-fore quite simple :-(CH,),N*C2H302 = (CH3)sN + CH3.C2H302 ON THE SALTS OF TETRAMETHYLAMMONIUM. 631 Action of Heat on Benzoate of Tetramethylammonium. This salt was prepared in the same way as the acetate. It is a deliquescent salt but could be obtained in the form of long needles by allowing the salt to remain over sulphuric acid in a vacuum. When it was heated it melted a t 220-230" and at once decom-posed. There was no charring and the whole of the salt distilled by the time the temperature had risen to 250'.No gas was produced by the action of heat. The distillate was completely liquid and sepa-rated into two layers when water was added the one an aqueous solution of trimethylamine the other a liquid which when dried over calcium chloride and distilled boiled at 198" (methyl benzoate, b. p. 199"). It did not contain nitrogen and in order to be szre that it was methyl benzoate it was boiled with caustic soda. The distillate contained methyl alcohol and from the residue contailiing the sodium benzoate the silver salt was prepared. 0.230 gram salt gave 0.110 gram Ag = 4'7.1 per cent. Ag; theory for AgC,H50 = 47.1 per cent. Ag. An analysis was also made of the chloroplatinate of trimethylamine. 36.5 per cent. Pt was found while the theory for (Me3HNC1)2PtC14 = 36.7 per cent.Pt. The benzoate decomposes in exactly similar manner to the acetate :-Action of Heat on Xulphate of Tetrarnethylammoniu?n. This salt was obtained by neutralising the base with sulphuric acid. It is crystalline and very deliquescent and before the last traces of water could be removed it had to be heated t o about 160" in a vacuum. An analysis of some of the salt thus dried gave the following num-bers :-0.4342 grsm salt gave BaS04 0.417 gram = 39.57 per cent. SO,. Theory for (Me4N)2S04 . . . . . . . . . . . . . . = 39.34 ,) 7 7 The remainder of the salt was carefully heated (5.72 grams). It melted at 280" and at once began to decompose. The temperature was kept its near as possible to 290° till all effervescence had ceased and the loss that the 5.72 grams had suffered was 1.42 grams.The whole of this was found to be due to the trimethylamine no other sub-stance passing over into the condenser. The salt on cooling solidified t o a mass of deliquescent crystals which were broken up and analped. They were found to contain sulphur but their solution in water was not precipitated by a barium salt. The sulphur therefore had to be determined by it combustion of the substance (which Dr. Plimpto 632 LAWSON AND COLLIE THE ACTION OF HEAT kindly undertook using his new method f o r analysing such com-pounds). 0.196 gram salt gave 0.2517 gram BaS04 = 1'7.63 per cent. S. A determination of carbon and hydrogen was also made. 0.497 gram salt gave 0.590 GOz and 0.3765 HzO = 32.38 per cent.C and 8.41 per cent. H. Found. Calculated for -1 (CHd4N. (CHdSO4. I. 11. C . . 32.47 - 32.38 H . 8.11 - 8.41 S . . 17.29 17.63 N - '7.56 0 34-60 -- -The first decomposition which the sulphate of tetramethylam-monium suffers is the loss of trimethylamine whereby it is converted into the methyl sulphate of tetramethylammonium. (Me4N)2S04 = (Me4)N*MeSOd + Me,N. The 5-72 grams which were used should lose on heating 1.38 grams of trimethylamine and the amount found was 1.42 gram. The further action of heat on the methyl sulphate of tetramethylammonium was complex ; under these circumstances part of the salt passed over into the condenser unchanged while compounds possessing a very strong smell recalling that of sulphide of methyl were formed.There was also produced in small quantities a highly crystalline substance which resembled dimethylsulphone and much charred matter remained behind in the distilling flask. Action of Heat on Oxalate of Tetramethylramm,onizLm. Two methods were employed for the preparation of this salt-one by the action of oxalic acid on the base the other by the action of silver oxalate on the iodide. A deliquescent salt was obtained by both methods and was most difficult to obtain anhydrous. When allowed to remain in avacuum civer sulphuric acid a crust formed on the surface of the salt while the underlayer remained in a liquid state for weeks. The only method which gave at all a. dry product was heating the salt in a vacuum at 160". An analysis of some of the dry salt was made.0.590 gram salt gave 0.1355 CaO = 36.35 per cent. ($04. Theory for (Me4N),Cz04 = 37.28 9 2 The rest of the salt was subjected to the action of heat. Scarcely any decomposition occurred below 360" whilst above that temperatur ON THE SALTS OF TETRAMETHYLAMMONIUM. 633 the salt seemed to slowly sublime without melting ; decomposition occurred at the same time for trimethylamine and a gas were evolved. When the salt was more strongly heated however the evolution of gas became more rapid. The gas seemed to be composed of carbon dioxide mixed with nearly its own volume of carbon monoxide; a small quantity of an inflammable gas was also present. I n two experi-ments made with weighed quantities of the oxalate the following amounts of gas were obtained :-I.2.5 grams salt gave 500 C.C. of gas composed of 225 C.C. of carbon dioxide 225 C.C. of carbon monoxide and 50 C.C. of in-flanimable gas. 11. 2.4 grams salt gave 450 C.C. gas composed of 205 C.C. carbon dioxide and 205 C.C. of carbon monoxide and 40 C.C. of inflam-mable gas. As far as could be ascertained no methyl oxalate was formed ; this was obviously due t o the extremely high temperature a t which the oxalate decomposed. Trimethylamine was formed in considerable quantity and was converted in to the chloroplatinate aiid analysed. The decomposition of this salt is therefore not a simple one owing to the high temperature at which i t occurs. Probably methyl oxalate and trimethylamine are first formed and the former a t once decom-poses into carbon dioxide carbon monoxide and other substances, perhaps methyl ether.(Me4N),C,04 = 2Me3N + Me2C204, Me2C204 = Me,O + CO + GOz. The amount of carbon dioxide and carbon monoxide evolved by the 2.4 grams and 2.5 grams respectively agrees fairly well with the amount required by theory ; 2.4 grams should yield about 230 c.c., whilst 2.5 grams should yield about 245 C.C. of mixed gases. Action of Heat on the Acid Carbonate of Tetramethy lammonium. A solution of the base saturated with carbon dioxide was evapo-rated first over the water-bath and finally over sulphuric acid in a vacuum. The salt was crystalline but highly deliquescent and could only be obtained anhydrous by prolonged heating in a vacuum. When subjected to the action of heat it began to decompose a t 180" with effervescence but more rapidly a t a temperature of 210-2220".Tri-methylamine methyl alcohol and carbon dioxide were the only sub-stances produced. This decomposition, (CH,)JT*HCO = (CH3)3N + CH,*OH + COZ 634 LAWSON AND COLLIE THE ACTION OF HEAT seemed to be nearly quantitative 3 grams of the salt yielded 280 C.C. of carbon dioxide the theoretical yield according to the above equation being about 500 C.C. Attempts were made t o prepare the normal carbonate by dividing a solution of the base into two equal parts saturating one with carbon dioxide and then adding the other half to it. On evaporating and leaving the salt for two months in a, desiccator a white friable crystalline deliquescent salt was obtained. This salt when heated gave exactly the same resiilts as the acid carbonate and was probably a mixture of the acid carbonate with hydroxide of tetramethylammonium.Not a trace of carbonate of ethyl was produced. Possibly therefore the normal carbonate does not exist. A few other salts were examined the hydrosulphide the acid sulphite and the phosphate. The hydrosulphide is a very deliquescent salt and when heated to about 200" easily decomposes into trimethyl-amine and methyl mercaptan. The acid sulphite crystallises fairly easily from concentrated solutions and is not very deliquescent ; on heating this salt it melted at 180" and gave off some water of crys-tallisation leaving a white salt behind ; this did not melt or decom-pose till the temperature had risen above 300° it then split up in a complicated manner yielding trimethylamine methyl alcohol sulphur dioxide and small quantities of a volatile crystalline substance with a high boiling point but it was not obtained in amount sufficient for purification.The phosphate of tretramethylammonium is also a deliquescent salt. It was prepared by shaking a solution of the iodide with phosphate of silver ; the solution thns obtained was strongly alkaline. With am-monium nitromolybdate it gave a ligh t-yellow precipitate which con-tained tetramethylammonium phosphate and molybdic acid. When the phosphate was heated at a very high temperature it decomposed, giving trimethylamine and methyl alcohol which distilled while meta-phosphoric acid remained in the flask. Traces also of sulphide of methyl were apparent.Conclusion. It appears therefore that the action of heat on the salts of tetra-met hylammonium is usually of a simple nature and trimethplamine is always produced. If the salt heated be one which decomposes at a low temperature trimethylamine and a salt of methyl are the only substances produced. This is seen in the case of the fluoride, iicetate &c., (CH3)iNF = (CH3)2N + CH3F ON THE SALTS OF TETRAMETHYLAMMONIUK 635 but if the salt decomposes only at a high temperature it is often the case that the salt of methyl (which is no doubt at first produced) is decomposed ; for instance the nitrite oxalate &c. (CH3),"0 = (CH3)3N + (CHs)N02, 4(CH3)N02 = 2(CH3)@ + 4N0 + 0,. One salt aloiie decomposed in an unexpected manner namely the normal sulphate but Crum-Brown and Blaikie (Zoc.cit.) have noticed that when the hyposulphite and sulphite of trimethylsulphine are heated a similar splitting up of the molecule takes place-(Me3S),S03 = (Me,S)MeS03 + Me2S, (Me4N),S04 = (Me4N)MeSO4 + Me3N. It will be seen that the salts experimented with may be divided into two groups firstly those which are easily acted on by heat and which decompose at a temperature of about 200"; secondly those which are more stable and are only decomposed at temperatures above 300". Those belonging to the first class are the hydroxide, carbonate acetate benzoate fluoride and hydrosulphide ; whilst those which belong to the second class are the iodide bromide, chloride oxalate nitrate &c. Another point of interest is the solubility iu water of the salts of tetramethylammonium which seems to be in many cases exactly the reverse of the solubility of the ammonium salts for the iodide of tetramethylammonium is the least soluble of the tetramethyl-ammonium compounds with the halogens the fluoride and chloride the most soluble.Amongst the other salts the oxalate and sulphate are highly deliquescent whilst the nitrate is scarcely even hygro-scopic. When tlhe action of heat on the tetramethylammonium salts is compared with the action of that agent on the corresponding phosphorus and sulphur compounds it is seen that the decomposition which they undergo is similar. This is shown in the case of the acetate-(CH,)4N*CzH30 = (CH3)3N + C'H3*C2H302, (CH3)3S*CzH,O = (CH,),S + CH3.CzH302, (CH,),P*C2H302 = (CHs),PO + (CH,),CO.The apparent difference in decomposition which the phosphorus salt suffers is at once explained when we remember the intense affinity of that element for oxygen. Phosphorus sulphur and nitrogen are a group of elements which show a regular gradation in their affinity for oxygen and electronegative elements. When once phosphorus i 636 COLLIE ACTION OF BEAT combined with oxygen it is usuitlly no easy matter to separate these two elements consequently we find the method of decomposi-tion of the tetramethylphosphonium salts containing oxygen differ-ing in this respect from the corresponding ammonium compounds. The decomposition of the tetramethylphosphonium and the tetra-methylammonium chlorides by heat also illustrates this difference-2(CH,),PCl = 2(CH,),PHCl + CzH4, (CH,),NCl = (CH,),N + CH,Cl. As sulphur does not possess such a strong attraction f o r oxygen a8 phosphorus yet still possessing more than nitrogen for that element, we find that the sulphine salts resemble both the compound ammonium and phosphoniurn salts in their behaviour when heated