EXCHANGE Hbe TIlnix>ersitE of (Ibicago Mercuri-Organic Derivatives L The Mercurization of Aromatic Amines and Its Relation to the Theory of Substitution. IL A Study of the Nitranilines. Determination of the Position of the Mercury in the Mercurized Nitranilines. A DISSERTATION SUBMITTED TO THE FACULTY OF THE OGDEN GRADUATE SCHOOL OF SCIENCE IN CANDIDACY FOR THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF CHEMISTRY BY ISADORE MEYER JACOBSOHN Private Edition Distributed by THE UNIVERSITY OF CHICAGO LIBRARIES CHICAGO, ILLINOIS 1922 Reprinted from the Journal of the American Chemical Society, Vol. XUII. No. - August, 1921 and Vol. XUV, No. 4, April, 1922. Zlbe lUntYJersits of Mercuri-Organic Derivatives L The Mercurization of Aromatic Amines and Its Relation to the Theory of Substitution. II* A Study of the Nitranilines* Determination of the Position of the Mercury in the Mercurized Nitranilines. A DISSERTATION SUBMITTED TO THE FACULTY OF THE OGDEN GRADUATE SCHOOL OF SCIENCE IN CANDIDACY FOR- THE DEGREE OF DOCTOR OF PHILOSOPHY DEPARTMENT OF CHEMISTRY BY ISADORB MEYER JACOBSOHN Private Edition Distributed by THE UNIVERSITY OF CHICAGO LIBRARIES CHICAGO, ILLINOIS 1922 Reprinted from the Journal of the American Chemical Society, Vol. XLIII, No. 8, August, 1921 and Vol. XLIV, No. 4, April, 1922. MERCURI-ORGANIC DERIVATIVES. THE MERCURIZATION OF AROMATIC AMINES AND ITS RELATION TO THE THEORY OF SUBSTITUTION. 1 The principles underlying the mercurization of aromatic compounds have occupied the attention of one of us (Kh.) for the past two years. The important fact that, in a mono-substituted benzene derivative, the mercury enters para or ortho to the group already present in the molecule, was pointed out early by Dimroth, 2 and has been substantiated by the vast amount of experimental work done by others. This fact of ortho and para substitution is particularly striking when one attempts to apply the theories of substitution to mercurization since mercury never enters meta to the group already present in the molecule. Even where the usual rules of substitution require that a rneta derivative be formed, e. g., with benzoic acid, benzophenone, and nitrobenzene, only ortho sub- stituted mercury compounds have been isolated. The phenomenon of mercurization was considered by Dimroth to be similar to that of nitration, sulfonation, etc., in the sense that the po- sition taken by the entering mercury in compounds containing groups now considered electro-negative, e. g., amino or hydroxyl groups, was quite in agreement with the rule of Crum Brown and Gibson. He considered the exceptional behavior of nitrobenzene, benzoic acid, and benzophenone as anomalous reactions which needed to be accounted for. However, until recently, no effort has been made to explain the orientation of mer- cury in the benzene molecule. 3 The rule of Crum Brown and Gibson, being purely empirical, offers no suggestion as to the mechanism of the reaction, nor does it help to predict whether a compound could be mercur- ized or not. Therefore, it appeared desirable to apply some of the mod- ern theories of substitution to the orientation of mercury in the benzene molecule. In the present paper, the authors will confine themselves to the mechan- ism of mercurization of aromatic amines and also attempt to show that the mechanism suggested for mercurization may be extended to explain sat- 1 This work was carried out in cooperation with Dr. M. S. Kharasch and was published in J. Am. Chem. Soc., 43, 1894 (1921), in an article by M. S. Kharasch and I. M. Jacobsohn. 2 Dimroth Ber., 35, 2853 (1902). 3 vSchoeller, Schrauth, and Liese (Ber., 52B, 1777(1919) apply the theory of Arm- strong for substitution in aromatic amines to mercurization, the limitations of which theory will be pointed out later. 520K- isfactorily orientation of other groups in the case of aromatic amines. It will be shown further, in a condensed critical survey of the theories of others, that none of these is completely satisfactory. For example, most of them fail to account for the fact that dimethylaniline is mercur- ized readily. The theory which the present authors have been led to adopt as a result of their extended experimental work, briefly stated, is as follows. The mercurization of an aromatic amine proceeds in two stages : first, the for- mation of an ammonium salt with the mercuric acetate or chloride to give a compound of the general type I. Ri Ri N R 2 XHg N l\ \ X HgX II R 2 This compound is then rearranged to the more stable configuration hav- ing the mercury radical attached to carbon, (Type II), the para position being usually favored. 4 If this position is already occupied, the mer- cury goes to the oriho position. Furthermore, it is only the ammonium salt that undergoes rearrangement, 5 the instability of the salt being due to the presence of an electropositive group or atom. on the amino nitrogen. 6 This, it will be noted, is the application to mercurization of the theories of substitution of Armstrong, 7 Bamberger, 8 Chattaway 9 and others, and of the general theory of catalysis advanced by Stieglitz, 10 Acree 11 and others 4 The fact of rearrangement may be due to a smaller energy content on the part of the rearranged compound. This relationship was shown to hold by Rother and Stoermer (Ber., 46, 260 (1913)) for a number of compounds, e. g., cinnamic acid and the isomeric allocinnamic acid. 5 Acree and Johnson, Am. Chem. J., 38, 265 (1905), have demonstrated that, in the rearrangement of N-chloro-acetanilide, it is the salt and not the free base that rearranges. They fail, however, to attach any importance to the positive charge on the rearranging group, a fact which has already been pointed out by Stieglitz and Stagner, /. Am. Chem. Soc., 38, 2057 (1916). 6 Acree and Johnson, Am. Chem. J., 38, 265 (1905), have shown that, when N-chloro- acetanilide is treated with hydrogen bromide, the speed of the rearrangement is a thousand times greater than when hydrogen chloride is used, and that the end prod- uct of the rearrangement is -bromo-acetanilide, instead of -chloro-acetanilide. 7 Armstrong, /. Chem. Soc., 51, 258, 583 (1887). 8 Bamberger, Ber., 27, 586 (1894). 9 Chattaway, J. Chem. Soc., 75, 1048 (1899). 10 Stieglitz, Proc. Congress of Arts and Sci. (St.Louis), IV, 278(1904); Am. Chem. J., 39, 29 (1908). Acree, ibid., 37, 410 (1907). that acid catalysts act through salt formation. 12 However, the theory ad- vanced by the present authors diverges from that of Armstrong, Bamberger, and Chattaway, and also from recent theory of mercurization of Schoeller, Schrauth and Liese 3 in a few very important aspects. The theories advanced by the above mentioned authors assume that the entrance of a group into the nucleus of an aromatic amine is always preceded by sub- stitution on the amino nitrogen, thus, o H -NH, + HONO 2 > < >-N-NO 2 . However, as an explanation, this is not adequate to account for the mech- anism of substitution of tertiary amines, of the type of dimethyl aniline which contains no replaceable hydrogen. Furthermore, they do not em- phasize the nature and importance of the catalysts used in substitution, i. e., strong acids. That the r61e played by acid catalysts is that of salt formation has been pointed out very clearly by Stieglitz 10 in his work on the catalysis of imido-esters, as well as by Acree. 11 We find thus, that the previous hypotheses have to be modified, if they are to explain consistently all the facts of substitution without fur- ther assumptions. The hypothesis advanced by the writers to explain the mercurization of aromatic amines lends itself admirably to this pur- pose. We shall now proceed to illustrate as briefly as possible the basis for this hypothesis and then show that it is quite compatible with the oriho and para orientation of other groups in the case of aromatic amines. That mercury has a great affinity for amino nitrogen and for carbon is best illustrated by the readiness with which the mercury-ammonium complexes are formed, and by the great stability of mercury-carbon com- pounds. 13 It is the combination of this property of mercury to form stable nitrogen and carbon compounds, and the relatively greater stability of the latter, which enables us to attach mercury to the carbon of the benzene ring in the case of aromatic amines. Within the scope of a series of investigations being carried out by the one of us, work was undertaken upon the mercurization of the N-alkylated nitro-anilines. It was found that mercurization does not take place when ^-nitro-dimethylaniline is treated with mercuric acetate in alco- holic solution. Though various experimental conditions were tried, the 12 In this connection, the work of Reilly and Hickinbotham (/. Chem. Soc. t 117, 103 (1919)) on the rearrangement of w-butylaniline is quite significant. They found that only those salts which are capable of combining with the amino group effected the re- arrangement. 13 The stability of the mercury-carbon union is a relative term. It depends to a large extent upon the position of the mercury in the molecule and upon the nature of the neighboring groups. This will be discussed in a later paper. See also Stieglitz, Kharasch, and Hanke, /. Am. Chem. Soc., 43, 1186 (1921). 6 efforts were futile. It seemed, however, in view of the results of an in- vestigation carried out with other nitro-anilines, 14 that mercurization should take place unless there was some property of the mercurizable nitro-anilines which was not shared by the ^-nitro-dimethylaniline. Reference to the literature at once revealed the fact that p-nitro-dimethyl- aniline is an extraordinarily weak base. It is precipitated from even cone, hydrochloric acid solution as the free- base instead of its hydrochloride, and it does not combine with chloroplatinic acid. 15 Further reference to the literature showed that its oriho and meta isomers form salts more read- ily. An investigation of the reactions of these isomers with mercuric acetate showed not only that the reactions are completed within a very short time but at a comparatively low temperature. An attempt was then made to mercurize ^-nitro-diethylaniline. This met with the same fail- ure as that of the corresponding dimethyl derivative. That this failure to mercurize is not due simply to the fact that the para position is occu- pied was then demonstrated by the fact that ^-nitro-monomethyl- and ^-nitro-mono-ethylanilines are mercurized without difficulty, the mercury being oriented to the position ortho to the amino group. These facts, taken in conjunction with the intermediates already iso- lated 16 and those obtained by Schoeller, Schrauth and Liese, 3 the complex salts of aromatic amines isolated by Klein 17 and the known general fact of the great affinity of mercury for an amino nitrogen, led us to adopt the hypothesis that mercurization proceeds first by the addition of the mercury salt to the amino group, i. e., to form an ammonium salt. This ammonium salt then rearranges to a compound in which the mercuric radical is attached to carbon. 4 From this theory of the mechanism of mercurization, one of the most logical deductions to be made is that, of the aromatic amines, only those which are capable of forming salts will mercurize. Among other tests, one to which this deduction has been put was an attempt to mercurize tri- methyl-phenyl-ammonium acetate, a compound in which the formation of a double salt with the mercuric acetate used is impossible. This com- pound was not mercurized although a water suspension of dimethyl- aniline mercurizes within one or two minutes, with great evolution of heat. 14 Unpublished work of Kharasch, Lommen, and Jacobsohn. 16 We found, however, that a hydrochloride of -nitro-dimethylaniline could be prepared by passing dry hydrogen chloride into a benzene solution of the base. 16 Kharasch, Lommen, and Jacobsohn (unpublished work) have isolated the com- TT pound,' by treating an alcoholic solution of w-nitroan iline with 2 C 2 H 3 O 2 N an aqueous solution of mercuric acetate. 17 Klein, Ber., 11, 743(1778). The theory here advanced as to the mechanism of mercurization of aro- matic amines adapts itself as already stated, quite readily to the expla- nation of the orientation of other groups upon substitution in the nucleus of aromatic amines. 18 Not only does it explain satisfactorily the cases of ortho and para substitution but, as will be shown later, it explains also the fact that a mixture of all three isomers is obtained when an aromatic amine is nitrated in cone, sulfuric acid solution, as well as the more recent ob- servations of Vorlander 19 that, upon bromination and nitration, the quater- nary ammonium salts yield only the meta derivative. To explain these cases of meta substitution, Vorlander advances a the- ory of the benzene nucleus in which he assumes that the hydrogen atoms in the benzene molecule have positive tensions, 20 and that it is the tension within the nucleus which determines the position taken by the entering substituent. He assumes, further, that the nature of the tension on the group already substituted in the molecule exerts an influence on the arrange- ment of the tensions within the nucleus. Thus for nitrobenzene he as- sumes Structure I. I while, for aniline, Structure II is assumed. Further, substitution takes place by absorption with the subsequent loss of a molecule of acid or water 21 thus : Vorlander makes the additional postulate, in order to explain the cases of meta substitution in aromatic amines, that, upon salt formation, there is a change in the tension throughout the molecule, e. g., upon treating 18 It is self-evident that this theory is applicable also to such rearrangements as those of hydrazobenzene, diazo-aminobenzene, the Fischer and Hepp, etc. It is to be noted that, in all cases, where the rearranging compound is a weak base, the reaction must be carried out in the anhydrous media to insure salt formation. 19 Vorlander, Ber., 52B, 263 (1919). 20 In order to avoid committing himself to any theory as to the nature of valence, Vorlander uses the term "tension" and denotes the kind of tension by -f- and . While he does not assume that they are electrical in nature, the properties assigned for them are such that, for the purpose of discussing substitution, there is little distinction to be made. 21 It is rather peculiar that Vorlander does not take into account the fact that bro- mine is Br + Br~, unless by an error in print, since the existence of positive bromine has been proved conclusively. (W. A. Noyes, J. Am. Chem. Soc., 23, 460 (1901); Stieglitz, ibid., 23, 796 (1901) ; Walden, Z. physik. Chem., 43, 385 (1903). 8 aniline with sulfuric acid, the following change is assumed to take place. Such an assumption is entirely unwarranted. Not only are there no ex- perimental facts upon which to base such an hypothesis, but the assump- tion made involves an upheaval taking place within the molecule, as well as a change of tension between the carbon and groups attached to it. Such a supposition as a change of tension would also demand that, in a compound such as aniline nitrate, analogous to aniline sulfate, the ten- sion on the nitrogen bound to the phenyl group must be positive. ^ON0 2 This compound should then, according to the views of Vorlander, give a meta derivative when dropped into cone, sulfuric acid. On the contrary, however, it gives only o- and p-nitro derivatives. 22 Similarly, the ni- tration of aromatic quaternary ammonium compounds demands, accord- ing to Vorlander, that the tension between the ammonium nitrogen and the carbon of the benzene nucleus be positive (II). This is hardly com- patible with the fact that a water solution of the hydroxide of the com- pound (II) when evaporated gives a small quantity of trimethylamine and w-nitrophenol. 23 This theory must, therefore, be further modified if it is to be made to fit all the facts of substitution in aromatic amines. Similarly, it has been necessary to make additional postulates to all pre- vious theories in an attempt to build a workable theory of substitution. Yet, with all these postulates, none of the older theories can be made to fit all the facts of substitution. However, no new postulates need be made to the theory advanced by the present authors. Once it is assumed that all ortho and para substitution in the hydroquinoid nucleus of aromatic amines takes place through intermediate salt formation, it follows that, when such salt formation is impossible, as in the quaternary ammonium salts, substitution must take place in the meta position or, as in such spe- cial cases as that of mercurization, it is not accomplished at all. Not only does the theory advanced explain the case of meta substitu- tion in the quaternary ammonium salts, but it explains also the fact that a mixture, of all three isomers is obtained upon nitration of an aromatic amine in cone, sulfuric acid solution. It explains, further, the fact that 22 Bamberger, Ber., 28, 400 (1895). 23 Staedel and Bauer, ibid., 19, 1939 (1885). 9 the yield of the meta derivative increases with the amount of sulfuric acid used. Here it will be seen that an equilibrium mixture of both the amine and the ammonium salt exists in such a solution. Therefore, in such a mixture both types of substitution, that dependent upon intermediate compound formation, giving oriho and para derivatives, and the sub- stitution of a salt yielding a meta derivative, are found. Again, as the volume of sulfuric acid is increased, the equilibrium will be shifted to form more of the sulfate, thus increasing the yield of the meta derivative. Furthermore, the theory here advanced also offers an explanation of certain peculiar facts of rearrangement. For example, the rearrange- ment of /3-phenyl-hydroxylamine 24 in alcoholic solution with hydrogen chloride as catalyst, gives not only -amino phenetole but also -chloro- aniline, the more strongly electro positive chlorine atom rearranging. The theory advanced predicts that N -substituted aromatic amines, when treated with a catalyst capable of being oxidized by the group al- ready substituted in the molecule will yield a derivative having the more strongly electro positive group in the para position. It will be seen, therefore, that the present theory explains from a sin- gle standpoint the rearrangement and substitution in the case of aromatic amines, without the necessity of any additional postulates. It is also far more adequate for the explanation of these facts and furnishes a more workable hypothesis than any of the other theories of substitution. Experimental Part. Method used for Mercurization. The method used for the preparation of the compounds herein described is, briefly, as follows. One mol of the compound to be mercurized was dissolved in a small volume of alcohol and boiled with an aqueous solution containing . 9 mol of mercuric acetate, until a test portion gave no black precipitate of mercuric sulfide when treated with ammonium sulfide. This reaction mixture was filtered hot. Upon cooling; the acetate of the mercurized compound crystallized from the solution. The filtrate from the above was treated with an aqueous solution of sodium chloride, whereupon the chloride of the mercury compound soon precipitated. The total yield of mercurized product calculated upon the basis of the mercuric acetate used, was almost quantitative. An Attempt to Mercurize -Nitro-dimethylaniline and -Nitro-diethylaniline. As stated in the theoretical part, several attempts were made to mercurize p-nitro- dimethylaniline, all of which met with failure. The method above outlined was used, as well as several other methods, none of which yielded a mercurized product. Similarly, attempts to mercurize />-nitro-diethylaniline were without success. o-Nitro-/)-Acetoxymercuri-dimethylaniline, (C6H 3 (l)N(CH 3 ) 2 (2)NO 2 (4)HgO 2 C2H,.) This compound was prepared from o-nitro-dimethylaniline and mercuric acetate. 24 See, however, Stieglitz, Am. Chem. /., 46, 327 (1911 10 The reaction was completed after half an hour's boiling. The reaction mixture was filtered hot in order to remove the small amount of mercurous acetate formed in the reaction. It was then cooled and the precipitate collected on a filter. The precipitate was then washed with alcohol, and dried to constant weight in vacuo over sulfuric acid. Analyses. Subs., 0.2759: dry N 2 , 16.20 cc. (22 and 741 mm.). Subs., 0.2604: HgS, 0.1427. Calc. for Ci Hi 2 O 4 N 2 Hg; N, 6.61; Hg, 47.23. Found: N, 6.63; Hg, 47.26. The compound thus obtained is bright yellow; m. p. 160. It is crystalline, and is soluble in the common organic solvents. o-Nitro-/>-Chloromercuri-dimethylaniline. (CH(l)N(CH)(2)NOs(4)HgCI.HjO). ' This compound was washed well with water to remove sodium chloride and finally with alcohol, and dried in vacuo over sulfuric acid to constant weight. Analyses. Subs., 0.2635: dry N 2 , 15.60 cc. (21.5 and 741 mm.). Subs., 0.3233: AgCl, 0.1096. Calc. for C 8 HnO 3 N2HgCl : N, 6.69; Cl, 8.46. Found. N, 6.71; Cl, 8.38. The substance is red and amorphous. It is soluble in acetone and in boiling alco- hol. It melts at 185, with decomposition. w-Nitro - p - Acetoxymercuri - dimethylaniline, (C 6 H 8 (l)N(CH 3 )2(3)NO 2 (4)HgO 2 - C 2 H 3 ). This preparation was obtained upon treating w-nitro-dimethylaniline with mercuric acetate. A half hour's boiling was sufficient to carry the reaction to com- pletion. For analysis, the compound was dried to constant weight over sulfuric acid. Analyses. Subs., 0.2924: dry N 2 , 17.05 cc. (20 and 734 mm.). Subs., 0.2940: HgS, 0.1610. Calc. for Ci Hi 2 04N 2 Hg:N, 6.61;. Hg, 47.23. Found: N, 6.56; Hg, 47.23. The compound is obtained in the form of brilliant orange needles. It is soluble in the common organic solvents; m. p. 140. w-Nitro - p - Chloromercuri-dimethylaniline, (C 6 H 3 (l)N(CH 3 ) 2 (3)NO 2 (4)HgCl) . This compound was obtained from the filtrate of the preceding preparation. Analysis. Subs., 0.3438: AgCl, 0.1237. Calc. for C 8 H 9 O 2 N 2 HgCl: Cl, 8.84. Found: 8.90. The compound is red and amorphous. It is soluble in acetone and in boiling alco- hol. It melts at 220, with decomposition. ^-Nitro-o-Acetoxymercuri-monomethylaniline, (C 6 H 3 (l)NH.CH 3 (4)NO 2 (2)HgO 2 - C 2 H 3 ). This compound was prepared from -nitro-methylaniline and mercuric acetate. The product obtained was extracted with ether in order to remove unchanged ^-nitro- methylaniline, and the residue was recrystallized from alcohol. A dark red solid remained which was insoluble in alcohol, but the quantity was too small to be analyzed. Analyses. Subs., 0.2585: dry N 2 , 15.6cc. (21 and 736 mm.). Subs., 0.2981: HgS, 0.1673. Calc. for C 9 H 10 O4N2Hg: N, 6.84; Hg, 48.82. Found: N, 6.79; Hg, 48.38. The compound forms small yellow crystals melting at 197 with decomposition. It is soluble in acetone and in hot alcohol to which a few drops of glacial acetic acid have been added. When treated with a cone, potassium hydroxide solution, a brick- red compound is formed. Upon dilution, the compound again becomes yellow. p-Ni tro-o-Chloromercuri - monomethylaniline, (C 6 H 3 ( 1 ) N . H . CH 3 (4) NO 2 (2) HgCl- H 2 O). This compound is obtained from the filtrate of the preceding compound. Analysis. Subs., 0.2972: AgCl, 0.1036. Calc. for C 7 H 9 O 3 N 2 HgCl: Cl, 8.75. Found: 8.62. The compound is yellow and crystalline. It is soluble in acetone and in boiling alcohol. It melts at 215, with decomposition. 11 The position of the mercury in this compound was determined by the method of Dimroth. 26 The acetate was treated with two equivalents of a solution of potassium perbromide. The product was then extracted with ether and the extract purified from alcohol; m. p. 112-113. When mixed with 4-nitro-2,6-dibromo-monomethylaniline, prepared by the method of Blanksma, 26 the melting point was not lowered. The use of only one molecule of the bromine in solution yielded no definite compound. Since the position could not be thus in dispute the dibromo derivative was therefore isolated. /?-Nitro-0-Acetoxymercuri-mono-ethylaniline, (C 6 H 8 (l)N.H.C2H 5 (4)NO2(2)Hg.O2- C2H 3 ). This compound was prepared from -nitro-mono-ethylaniline, and was purified from the unchanged raw material by recrystallization from alcohol. Analyses. Subs., 0.2641: dry N 2 , 15.7 cc. (24.5 and 731 mm.). Subs., 0.3480; HgS, 0.1912. Calc., for CioHi 2 O 4 N 2 Hg: N, 6.61; Hg, 47.23. Found: N, 6.56; Hg, 47.17. It crystallizes from alcohol in small yellow crystals; m. p. 183. It is also soluble in acetone. When treated with cone, potassium hydroxide solution, a brick-red compound is obtained. Dilution of the solution restored the original yellow color. />-Nitro-0-Chloromercuri-mono-ethylaniline, (C 6 H 3 (l)N.H.C2H 8 (4)NO 2 (2)HgCl- H2Q). The compound was prepared in the usual manner. It is obtained in the form of an amorphous yellow solid; m.p. 218, with decomposition. For analysis it was dried to constant weight over sulfuric acid. Analysis. Subs., 0.4544: AgCl, 0.1573. Calc. for CgHnOaNaHgCl: Cl, 8.46; Found: 8.56. It is soluble in acetone and in hot alcohol. The position of the mercury in this compound was determined by the method used above for the corresponding methyl derivative. The melting point of the 4-nitro- 2,6-dibromo-ethylaniline thus obtained was 68-71. The melting point was not low- ered when a mixture of the compound with known 4-nitro-2,6-dibromo-ethylaniline was used. -Nitro-0-Bromo-mono-ethylamline, (C 6 H3(l)N.H.C2H 5 Br(4)NO2) . This com- pound was prepared by the addition of a solution of bromine in glacial acetic acid to a solution of an equivalent portion of -nitro-mono-ethylaniline in the same solvent. After standing for about half an hour, -the reaction mixture was diluted with water, and the precipitated compound collected on a filter. It was washed with water and a little alcohol. After crystallization from alcohol, the compound was dried to con- stant weight in vacua. Yield, 90%. Analysis. Subs., 0.2704: dry N 2 , 26.50 cc. (17 and 750 mm.). Calc. for C 8 H 9 2 N 2 Br: N, 11.44. Found: 11.39. The compound crystallized from alcohol in beautiful long yellow prisms which are highly refractive; m.p. 65 to 66. It is insoluble in water but soluble in all common organic solvents. When an alcoholic solution of the compound is treated with cone, potassium hy- droxide solution, the alcoholic layer assumes a red color. Upon evaporation of the alcohol, or upon dilution with water, the yellow compound is again obtained. -Nitro-0-Dibromo-mono-ethylaniline, (CHt(l)N.H.CtHs(2 > 6)Br(4)NOi). Thia compound was prepared in the same manner as the above monobromo derivative, in this case, with two mols of bromine. For analysis the compound was crystallized from alcohol. Yield, 90%. Analysis. Subs., 0.2761 : dry N 2 , 21.2 cc. (21 and 744 mm.). Calc. for C 8 H 8 O 2 N 2 - Br 2 : N, 8.68. Found: N, 8.74. 25 Dimroth, Ber., 35, 2033 (1902). 26 Blanksma, Rec. trav. chem., 31, 271 (1902). 12 This compound crystallized from alcohol in small yellow needles; m.p. 75-76. It is insoluble in water, but soluble in all common organic solvents. Upon treatment with a cone, potassium hydroxide solution, an alcoholic solution of the compound turns red. If the alcohol is evaporated, or the solution diluted with water, the color is restored to the original yellow. An Attempt to Mercurize Trimethyl-phenyl-ammonium Acetate. A solution of 0.5 mol of mercuric acetate was added to a solution of trimethyl-phenyl-ammonium acetate. Even after standing for two months at room temperature, no mercurized product could be isolated. The solution gave a heavy precipitate of mercuric sulfide when treated with hydrogen sulfide. Upon boiling another portion of the solution for 24 hours, no change was observed. A heavy precipitate of mercuric sulfide was again obtained upon treatment with hydro- gen sulfide. Summary. 1. A theory of the mercurization of aromatic amines has been advanced, postulating that mercurization is preceded by addition of a mercuric salt to the amino nitrogen, with the subsequent rearrangement of the mer- cury to the ortho or para position. This theory, without any additional postulates, has been shown to apply equally well to the introduction of other groups in the benzene nucleus in the case of aromatic amines. 2. This theory has been shown to apply also to rearrangements such as that of Fischer and Hepp, the nitro-amines, sulfamic acids, etc. 3. The limitations to other theories of substitution have been pointed out. 4. The preparation of mercury derivatives of >-nitro-methylaniline, ^-nitro-ethylaniline, o-nitro-dimethylaniline, and m-nitro-dimethylaniline, and the preparation of 0-bromo and 0,-bromo-acetanilide the melting point was not lowered. The mercury is shown thus to be in the position para with respect to the amino group. ^-Acetoxymercuri-acetyl-w-nitro-aniline, [C 6 H 8 (1)NH- (COCH 3 ) (3)NO 2 (4)HgO - OC 2 H 3 ]. This compound was prepared from -acetoxymercuri-w-nitro-aniline in the same manner as that described for the p-nitro-aniline derivative above. For analysis it was dried in vacuo over sulfuric acid to constant weight. Analysis. Subs., 0.2522: 14.24 cc. of dry N 2 (18 and 739.5 mm.). Calc. for Ci H 10 5 N 2 Hg: N, 6.40. Found: 6.45. This compound is white. It melts with decomposition at 230 . Determination of the Position of the Mercury in the Mercurized w-Nitro-aniline The method of procedure used in determining the position of the m-nitro-aniline derivative? was similar to that used for the para compound. The compound obtained 3 The compound thus obtained was insoluble in the common organic solvents and, owing to the presence of an appreciable quantity of metallic mercury formed in the course of the acetylation, was not analyzed. 15 when recrystallized from water melted at 135-139. When mixed with some known -bromo-w-nitro-acetanilide the melting point was not lowered. The position of the mercury is thus shown to be para to the amino group. To determine the position of the mercury in the dimercury-substituted w-nitro- aniline, the latter compound was suspended in ethyl acetate and boiled under a reflux condenser with acetic anhydride until the whole became white. 4 The icetyl derivative thus prepared was treated with 2 mols of potassium perbromide, in the manner previously described. The product obtained was then filtered, dried and extracted with ether. The ethereal extract was evaporated to dryness, leaving a residue of melting point 141-146 . Upon recrystallization from 50% alcohol it melted at 150 . When mixed with known 2,4-dibromo-3-nitro-acetanilide the melting point was not lowered. The com- pound was then de-acetylated with cone, sulfuric acid at 120. Upon dilution of this reaction mixture with water, a yellow product melting at 84-87 was obtained. Upon recrystallization from 50% alcohol, a product melting at 89 was obtained. When mixed with known 2,4-dibromo-3-nitro-aniline, the melting point was not lowered. This identifies the position of the mercury as being 2,4 with respect to the amino group (N0 2 =3). 4 In the course of the reaction, an appreciable amount of metallic mercury was formed. Since the compound thus obtained is insoluble in the common organic solvents, it could not be purified, and therefore was not analyzed. In conclusion, the writer takes pleasure in acknowledging gratefully his appreciation of the sympathetic encouragement and personal kindness which Dr. M. S. Kharasch has shown him in the guidance of his work. FINE OF 25 CENTS . BOOK ON THE SS OVERDUE. - 13 LD 21-100m-8,'34 Syracuse, N. Y. PAT. JAN 21, 1908 UNIVERSITY OF CALIFORNIA LIBRARY