319 XLVI1.-On Nitro-Compounds (Part 11). With Remarks on! Isomerism. By EDMUND J. MILLS D.Sc. F.C.S. THEexperiments recorded in the preceding part* resulted in a classification of nitro-componnds according to their products of reduction and in a contribution to the solution of the problem of the isomerism of the two nitranilines. The present part is chiefly devoted to the consideration of the two nitrobenzoic acids and some general remarks are appended on the subject of isomerism itself. (1) Nitrobenzoic Acid. By the name a@ha-nitrobenzoic acid I designate the solid pro- duct of the action of absolute nitric on ordinary benzoic acid. This is at present the only trustworthy process for obtaining it. The preparation is somewhat tedious thorough nitration being effected only after eighteen hours boiling of the mixture and nitric acid being repeatedly added to replace that which is lost by evaporation.After cooling and addition of water nitroben- zoic acid separates out ; it may be purified by continued washing with the precipitant. The term beta-nitrobenzoic acid may be applied to that variety which was obtained by treating crude toluol with nitric acid by Fischert (who named it “paranitxobenzoic acid ”) and Wilbrand and Beilsteint (who distinguished it as nitrodracylic acid ”) and in all probability much earlier by Abel.5 P-Nitrobenzoic acid has been examined with considerable detail by several chemists who have formed from it illarge number of derivatives similar to but nut identical with those derived from the sister compound.My object was to ascertain if possible the source of this difference. In preparing the second acid Wilbrand and Beilstein’s method was followed precisely with the excep- tion of the final crystallisation from alcohol which was found unnecessary. The fact that dinitrotoluol accompanies this body in large quantity is a very remarkable circumstance. The mem- 8 Journ. Chem. SOC. [21 ii. 153. f Ann. Ch. Pham. cxxvii 137. 2 Ann. Ch. Pharm. cxxviii 257. 5 Ann. Ch.Pharm. lxiii 313. 2a2 3fILLS ON NITRO-COMPOUNDS. bers of the aromatic marsh-gas series in fact cannot be advancer1 to higher stages of nitration than the first without inducing a complicated reaction. P-Nitrobenzoic acid differs in a most marked manner from the a-modification.The latter is said to dissolve in 400 parts of water* at 10' C. ; the former according to two determinations of my own in 2,630 parts at 14'. The powder of the less soluble acid is highly electric. The formation of this substance may be approximately expressed by the following equation- 6C;HS + 17HNO3 = 5CjHG(N0,)2 + /3CjHj(NOJO; + 15H20+ 3N204 which was arrived at by weighing the first two of the specified prorlticts correction being made for the solubility of the latter in the necessary washings. Found. Theory. 7-\ Dini tro toluol 23.8821 grm. 84.84 5 mol. 84.49 9 P-Nitrobenzoic acid 49670 , 15.15 1 mol. 15.51 28.1491 100-00 1oo*oo Combustions and determinations of fusion-point were made both of the two nitrobenzoic acids and the dinitrotolnol the results yielding very closely the well established nrimbers cf recent observers.The action of hydriodic acid on the two nitranilines having borne fruit in some addition to our knowledge of the more inti- mate nature of those bases very naturally suggested its extension to the two iiitrobenzoic acids which are immediately related to them. Experiment showed that boiling hydriodic acid of tnode-rate strength produces the corresponding aim i dohenzoic hydrio- dates with liberation of iodine. These salts may be obtained in tolerable purity by evaporating the liquid containing the products of tlir reaction to dryness heating the residue to 1OO"as long as iodine continnes to escape dissolving in water filtering and crystallising.A small quantity of a brown substance removed in this process is probably clue to the oxidisiiig influence of free iodine. * I find the acid to dissolve in 352.5 parts of water at 13' C. Higher numbers given by other authors probably refer lo a subBtance obtained with the aid of sul-phuric acid and having a higher fusion-point than normal nitrohenzoic acid (128"). MILLS ON NlTltO-COMl’OUh’DS. The a-hydriodate dissolves easily in cold cthcr alcoliul or hot water crystallisiug readily from these solvents. It is a com- paratively stable body. The P-hydriodate dissolves less readily and is soon coloured by exposure to air especially if the temperature be raised. Neither of these bodies produces a.coloured reaction with argentic nitrate. The following iodine determinations were made :-(a)0.5740 grm. substance dried at looo gave 0.5145 grin. argentic iodide. (p) 0.4323 grin. substame dried at looo gave 0.4218 grm. argentic iodide a. Found. P. Theory. C‘iH (HSN)Oz.HI. Iodine 48-44 47.27 417.93 Hence the nitrobenzoic acids belong to the class of “anlido-gens.” In order to institute a comparison between the temperatures of attack of these substances by hydriodic acid the same method was employed to which I have already alluded in Part I of these researches. I may be permitted to add here a brief description of the apparatus which the accompanying figure will help to explain. A beaker A more than two-thirds filled I\ ith oil of vitriol is adjusted on a ring-stand in the ordinary manner ;its edge supports a thick cork through the centre of which a hole is bored to admit the tube u in which the reaction is to be performed.This tube contailis a smaller om (the ‘‘weighing-tube,” in B which the given substance is weighed) not shown in the figure which is kept vertical by small beaks or outward pro-cesses from its edge. An 322 MILLS ON NITRO-COMPOUNDS. inclined position would interfere with the result. The tube a is furnished at its orifice with a cork having two apertures-one central to admit a thermometer and one lateral to permit the escape of air and vapours. The thermometer is always brought to the same position previous to an experiment by reference to the stem of the ring-stand and any other convenient line preferably vertical in the vicinity.The tube R is placed at a suitable dis- tance for observing the gradual obscuration due to the free iodine of the band of light occupying the proximate part of the cylindrical bulb of the thermometer It is merely furnished at b with an upright glass thread to aid the eye in returning to the Same position and on its upper part with a spirit level. It admits of motion both vertically and horizontally. Both it and the ring-stand should he left unmoved throughout a series of observations. The operator's back is towards the light and the experiments must be perfoi*med in the shadow of a partially opaque screen to avoid the direct illumination of the beaker A.The bath of oil of vitriol in A has its surface on a level with that of the liquid in a. It should be heated uniformly and with tolerable rapidity. The "attack-point " is the temperature at which the cylindrical por- tion of the bulb of the thermometer ceases to be visible. The results thus procured show a remarkable difference between the nitrobenzoic acids-a difference greater even than that which obtains in like manner between the two nitranilines. The num-bers are given in a table accompanying the present paper,* and curves are drawn to facilitate an immediate comparison. The necessary details of the apparatus for which alone these results are here taken to hold good are as follow Oil of vitriol in the bath ... . 250 C.C. Internal diameter of a .. .. 13.62 m.m. Diameter of thermometer .. . . 3.80 , Thickness of weighing tube .. .. 0.37 , Hence width of obscured annulus .. 4-54 , Volume of hydriodic acid employed .. 5-00 C.C. Weight of substance taken . . . . 0.20grm. Rate of heating .. .. .. 1' in 11 seconds. (2) Action of Nitric on Benzoic Acid. The experiments of Wilbrand and Beilstein already alluded to have shown that 6-nitrobenzoic acid may be successively * See opposite page. MILLS ON NITRO-COMPOUNDS. 323 transformed by known reactions into 6-amidobenzoic p-azo-amidobenzoic and benzoic acid. It was left an open question by those chemists whether the substance last produced was absolutely identical with ordinary benzoic acid or different from it.The resemblance is indeed most remarkable. It has been already pointed out that in fusion-point solubility in water and form of the calcium-salt* there is no practical distinction between these bodies. I may add that their ethylic ethers are precisely alike in odour a point on which frequent stress is laid in distinguishing isomeric substances. In order to decide this point if possible I submitted both ordinary benzoic acid and that derived from &nitro-benzoic to the action of nitric acid in the manner already de- scribed at the commencement of this paper. The liquid finally obtained was not precipitated with water but evaporated to dryness at 100'. When the free nitric acid had completely vola- tilised and the product had been allowed to cool a gummy mass presented itself in which crystals speedily made their appearance.The whole was transferred to a quantity of filter-paper and heated to 100' for a few hours. White crystals mere then found on the upper surface of the paper while a very easily fusible substance had soaked down to its interior. The crystals are readily purified by means of alcohol in which as well as in ether they rapidly dissolve in the cold; in boiling water they dissolve but sparingly; CURVESILLUSTRATING THE ATTACK OF ALPHA-NITROBENZOIC AND BETA-NITROBENZOIU ACID. 1.691 1.674 0-," 1.655 4 ?1 .rl 8 + 1.574 3 1.547 x" "0 1.512 $ 1,491 z$ 1.465 1-452 1 *439 JC Wil brand and Beilstein loc. cit. MILL$ ON NITRO-GO31POUNDS in water at 15' scarcely at all.With caustic potash there is first a violet and then a red coloration ; but not solution unless the liquid has been weak. These remarks with the exception of the reaction just mentioned (which belongs to the derivative from the source S),apply equally to both bodies. The fusion-point of the nitro-compound (from ? P-Senzoic acid) was 201"*9(cor.) ;. its composition that of dinitrobenzoic acid. 0.2655 grm. substance dried at 100° gave 0.3802 grm. carbonic dioxide and 0.0569 grm. water. Found. Theory. Carbon 39-05 39.62 Hydrogen 2.38 1 *89 The readily fusible body occurring with this is extracted from the paper by ether the solution filtered evaporated to dryness and the residue heated to 100" till constant.Thus prepared it probably contains a small quantity of the preceding compound with which it agrees closely in its reactions; it is however more soluble in alcohol or ether. It melts at 88' (cor.) and scarcely solidifies on cooling. 0.37635 grm. substance dried over oil of vitriol gave 0.6044 grrn. carbonic dioxide and 0.1035 grm. water. Theory. Found. C7H4(N02)aOP CTH (NO;)02 Carbon 43.79 39.62 50.30 Hydrogen 3.06 1.89 2-99 This product therefore is intermediate in composition between mono-nitrobenzoic and dinitrobenzoic acid but decidedly approaches the latter. The white crystals obtained from common benzoic acid as mentioned above melt at 137'05 (COI.) ; after re-crystallisation from dilute nitric acid washing and drying the fusion-point is 136'03 (tor.)-a very slight alteration.They dissolve in 422.2 parts of water at 13"; and constitute I believe a definite com-pound. 0.3391 grm. substance dried at IOO" gave 0,6225.grm. car-bonic dioxide and 0.1006 grm. water. Theory. Found CrHj(NO2)02 Carbon 50.07 50.30 HJ-hg:.r% 3-29 2-99 MILLS ON NITRO-COMPOUNDS. Hence these crystals have the composition of mononitrobenzoic acid but are not physically identical with it. I have not yet ex- amined the more fusible body which accompanies them and which appears to have hitherto escaped the notice of' chemists. The following experiments have a very important bearing on those just described :-(3) Simultaiieous Action of Sulpphuric uizd iYitiic Acid on Ordiiiary Henzoic Acid.When benzoic acid is heated to 100"in a closed vessel with sulpburic and nitric acids and the liquid cooled and mixed with water a white crystalline precipitate falls the yellow mother- liquid retaining in solution at least one other product. The nature of the result varies according to the time of digestion and the quantity and strength of the mixed acids. (1) The precipitcLte.-IIt is immaterial whether benzoic or a-nitrobenxoic acid bas been used as the starting point. In either case the product continues for some time to exhibit the composi- tion of mononitrobenxoic acid with an increased fusion-point and (it cannot be doubted) a decreased solubility. This fusion-pint is always higher than I28",provided the first stage of nitration has been reached.One precipitate of this kind coincided remarkably with the white crystals last described in the preceding section. The sub- stance fused at 136O.4 (cor.). 0.3225 grm. substance dried at loo" gave 0.5956 grm. carbonic dioxide and 0.0933 grm. water. Theory. Found. C7H,(N02)02 Carbon 50.37 5w30 Hydrogen 3.21 2.99 If the action be continued beyond six hours especially if abso- lute nitric acid be used the nitration advances a stage further and dinitrobenzoic acid is obtained. This appears to be a deFtnite product and has already been studied by Voit ,* Neither his results nor my own however agree with those of Cahour s ; the time of preparation is not short nor is the fusion-point of the * Ann.Ch. l'harm. xcix 100. MILLS ON NITRO-COJIPOUNDS. acid a low one. On the contrary it melts at 206'07 (cor.) I may add also the following numbers :-0.4196 gym. substance dried at looo gave 0.6190 carbonic dioxide and 04821 grm. water. Found. Theory. Carbon 40.23 39.62 Hydrogen 2.18 1.89 (2) The mother-liquid.-This is filtered from the precipitate and the sulphuric acid it contains exactly removed by means of barytic hydrate nitrate or carbonate. The residual solution is evaporated to dryness ultimately at 100'. The cooled crystalline mass which is invariably coloured yields it perfectly colourless product on extraction with carbonic disulphide. This body may be either mononitrobenzoic acid or a mixture of that compound with small quantities of dinitrobenzoic acid.The nitrobenzoic acid obtained in this way hsed first at 127O.5 (cor.) ; after cooling to solidification it melted at 122O.4 (cop.). Another preparation melted at 122"*3(cor.). It gives a red colouration with caustic potash. 0.3508 grm. substance dried at loo" gave 0.6393 grm. car-bonic dioxide and 0*1080grm. water. Found. Theory. Carbon 49-70 50.30 Hydrogen 3-42 2.99 As an instance of the mixed result I may adduce the following numbers :-0.2812grm. substance dried at loo",gave 0.4921grm. carbonic dioxide and 0.525 grm. water. Found. Carbon 47*73 Hydrogen 3-26 (4) Simultaneous Action of Sulphuric and Nitric Acid on &Nitro-benzoic Acid. When the above named acids are heated together to a tempera- ture of about 115O for several hours cooled and cautiously stirred with a moderate quantity of water a precipitate falls after some MILLS ON NITRO-COMPOUNDS.time. The yellow mother-liquid yields a second product on ap-propriate treatment. If the temperature of digestion be only 1004 only @-nitrobenzoic acid (with unchanged fusion-poia t) separates on the addition of water. (1) The precipitate.-This has tlie composition of dinitrobenzoic acid. It dissolves very readily in cold alcohol or ether and to a considerable extent in hot water. The solution is acid to litmus. With caustic potash it gives a beautiful red colouration. The fusion-point of one preparation was 161O.9 (cor.) ; after keeping several months during which it was once heated on the water- bath for a few hours the fusion-point was found to be 184O.3 (cor.) Another preparation gave the number 166O.5 (tor.)$ 0-2870grm.substance dried at loo" gave 0.4207grm. carbonic dioxide and 0.0618 grm. water. Pound. Theory. Carbori 39.98 39.62 Hydrogen 2.39 1.89 (2) The mother-liquid.-This was treated precisely as that men- tioned in the preceding section. The product however having been found to be impure from calcium-salt was mixed with a little water and primary potassic oxalate dried and extracted with ether. The latter solvent on evaporation left the body in a pure state and almost free from colour. The substance melts at 153'-1 (cor.) to a clear liquid; dissolves with remarkable ease in cold alcohol or ether but only slightly in water and gives with potash or ammonia a beautiful deep red colouration.Rapidly heated alone it explodes. After fusion it solidifies with extreme reluc- tance. 0.2920grm. substance dried over oil of vitriol gave 0.4256 grm. carbonic dioxide and 0.0697 water. Theory. Pound. C7H,(NOJ201 Carbon 39.75 3962 Hydrogen 2.65 1.89 This appears to be the only product remaining in the mother- liquid. Under tlie circumstances in which I examined it it probably contained a small amount of the "precipitate." MILLS 05 NITKO-COMPOUNDS. (5). Geizernl Observations. The remarkable difference between the temperatiires at which a-nitrobenzoic and B-nitrobenzoic acid are attacked by hydriodic acid points clearly to a difference between the nitryl radicals they contain.This it will be remembered was the conclusion arrived at in the case of the two nitranilines and it may doubtless be extended to many other isomeric nitro-compounds. Whether the nitryl of a-nitrobenzoic acid is or is not identical with that of a-nitraniline is still an open question. The experiments described in sections (2) (3) and (4) throw considerable light on the point they were intended to decide. It appears from (2)-though from equal weights of the bodies not having been taken the results are not absolutely precise-that benzoic acid prepared from gum benzoin is uitrated only with difficulty whereas that obtained from P-nitrobenzoic acid is nitrated with remarkable ease. Under the circumstances meu- tioned a mono-nitrocomponnd is obtained from the former a di-nitrocompound from the latter.Now if the resistance to nitra-tion which ordinary benzoic acid manifests could be sharply overcome it is clear that we ought to obtain P-nitrobenzoic acids from it. This condition is satisfied by the employrnent of a mis-ture of sulphuric and absolute nitric acid; a substance is pro- duced having the composition of di-nitrobenzoic acid and melting at 206’-7. But benzoic acid derived from 6-nitrobenzoic acid yields to nitric acid alone di-nitrobenzoic acid melting at 20io-9. This is as close an approximation as can be fairly expected. A further proof of the strong resistance which common benzoic acid presents to nitration is found in the contents of the mother- liquids obtained after digesting with sulpliuric and nitric acid and precipitating with water.That from a-nitrobenzoic acid contains altered a-nitrobenzoic acid ;ihat from ,@-nitrobenzoic acid con-tains di-nitrobenzoic acid. On the whole therefore I consider that there are two species of benzoic acid. Gregory* long since observed the occurrence of two potassic benzoates. I intend to examine these interesting bodies still further persuaded that they are the true sources of all the phenylic isomers. * Ann. Ch I’ha.rm. lxxxvii 125. 31ILLS ON PJ1‘1’RO-COMPOli N 1)s. (6). Kemurks on Tsornc?&m. The great problem of isomerism still pressing for solution has hitherto been treated almost solely in accordance with atomic views.The statical theorists who have adopted these and who snppose all bodies constituted of atoms fixed in space hold that isomeric substances are produced by a variation of the position of some radical or radicals in the same molecule The attempts to express this in formulze and the invention-of a corresponding phraseology are matters almost too familiar to require special notice. OD the other hand the principle of atomic motion-so largely accepted in physics so little in chemistry-does not seem to have ever been considered available for speculations in this direction. There have been however partial indications of the attainment of a theory having no reference to corpuscular notions. Thus Graham* long since enunciated a peculiar doctrine of chemical action and bestowed the names “zincous ” and “chlo-rous” on the two opposite functions of one and the same element at the instant of disturbance.Brodie,? also claimed for sub- stances at the moment of chemical change a two-fold aspect or “polar” antithesis. Kekul61 and others pointed out that in certain bodies containing replaceable hydrogen one-half of this exhibits a different property from that shown by the other half; ‘I and they employed the terms alcoholic ” and “saline ” to ex-press the difference. Now these distinctions in function con-sidered as belonging to certain relative weights of known radicals have no necessary connection with the doctrine of atoms though frequently associated with it in ordinary scientific language. Let us consider a special case of the phenomenon we are dis- cussing.a-Nitrobenzoic acid is produced by the action of nitric upon benzoic acid from gum benzoin ;P-nitrobenzoic acid results from the more complicated reaction between nitric acid and toluol where oxidation as well as nitration takes place. Omitting for the present the oxidation observed in preparing the latter isomer it is obvious that the difference in character whereby we distin-guish two kinds of nitryl is conferred during the formation of the Substances containiug them. The nitryl of the nitric acid is of * Elemmts (1842) p. 204. 9 Phil. Trans. 1850 p. 762. Lehrhuch. vol. i. p. 180. MILLS ON NITRO-COMPOUNDB. one nature; and this nature becomes altered in one or both cases of its transference.If therefore we would account for the alteration we must do so by exploring further the reaction itself in other words it is a question of force. That the functional change in nitryl does not depend wholly by any means on the composition of the residues with which it is associated is clear from what we know of the nitranilines and nitrobenzoic acids. That it does not arise from what is termed the ‘‘satureability ” of these residues is equally obvious. That-nitryl influences and is influenced by the nature of what is com- bined with it is no doubt true to a certain extent; but if the properties of the nitranilines be compared in detail with those of the nitrobeneoic acids it will become evident that the change alluded to has a logical claim to be considered a distinct pheno- menon.The numerous and remarkable oppositions furnished by isomer- ism whether exhibited with respect to solubility polarisation fusion-point basicity or the like appear to present us ouly with the following principle of co-ordination,-that we may induce a permanent change in the properties of radicals by virtue of certain reactions. The formation of isomeric bodies indeed probably depends on precisely the same laws as that of bodies having differences of composition. The substitution of a radical with certain functions for one with certain other functions but the same composition differs only in this last respect from the exchange of a radical for one completely distinct ;and without reference to isomerism it is easy to see that mere composition has a very subordinate bearing on the properties of a substance.When we reflect that of the radicals R r (having the same composition) we may have n R and m r simiiltaneously in a com- pound where n + m is a constant quantity but where n and m may vary the most complicated cases have a basis of classifica-tion. Thus if the somewhat formidable array of the turpentines owe its number to the opposite natures which hydrogen is un-doubtedly capable of assuming,* we should have here n i-rn = 16; and we might of course anticipate more isomers in this group than we are at present acquainted with. Again with regard to * For an instance of this see especially Hofmann on Ilydra obenzol Proc. Roy. SOC.xii 578.SIMPBON ON THE SYNTHESIS OF TEIBASIC ACIDS. 331 the hydric tartrates of which there are three well-established species. If we conceive their isomerism to be also due to the functional difference of hydrogen ‘md that here we at once account for their number and can deduce from the small percentage value of the hydrogen concerned that whatever difference there is must be very small. These however are merely hypotheses on two special cases; but I may remark with respect to the latter that there is no evidence whatever to justify the common statement that ‘‘ inactive ” hydric tartrate is a com- pound of the two fc active ” species if it be meant that the last two have respectively half the molecular weight of the first. The general formula of an isomeric group may be written as follows :--A a &c.being the two opposite forms respectively of the ulti-. mate component radicals ; the italics represenbing their number and being constant in SUM for each of them. The real problem of isomerism appears to be as follows:-Under what conditions of action does chemical force induce a per-manent alteration in one or more functions of radicals? This I conceive to be the problem of every-day chemistry itself which is occupied for the most part with the production of permanently different bodies.