The original of this book is in the Cornell University Library. There are no known copyright restrictions in the United States on the use of the text. http://www.archive.org/details/cu31924004084665 y Cornell University Library /6D 441.B84 On the constitution of gallein and coer 3 1924 004 084 665 CORNELL UNIVERSITY. Department of Chemistry/ On thfe Constitution of Gallein and Coerulem. A DISSERTATION Submitted to the University Faculty for the Degree of Doctor of Fhilos6phy; CHARLES EDWARD BREWER. 1900. EASTON, TA. : THE CHEMICAL fUBHSHING COMPANY. • ., ' 1900. CORNELL UNIVERSITY. Department of Chemistry. On the Constitution of Gallein and Coerulein. A DISSERTATION Submitted to the University Faculty for the Degree of Doctor of Philosophy. CHARLES EDWARD BREWER. 1900. BASTON, FA.: THE CHEMICAL PUBLISHING COMPANY 1900. ? ACKNOWLEDGMENT. This work was undertaken at the suggestion of Dr. W. R. Orndorff and prosecuted with his constant cooperation. I desire to express my grateful appreciation not only of his unfailing and sympathetic interest in the work, but also of his many kindnesses shown me personally. I acknowledge gladly my indebtedness to Dr. A. C. Gill, of Cornell University, for accurate description and measurement of crystals. My thanks are tendered also Professors Caldwell and Dennis for their uniform kindness to me. May, igoo. HISTORICAL. I. GAUEIN. It was in 1871 that Adolph von Baeyer' discovered gallein. He made it by heating together phthalic anhydride and pyrb- gallol. The product was purified by dissolving in alcohol and precipitating with water. He described it as a solid, red by reflected light, blue by transmitted light. Later 2 he described it as a brown-red powder, or small metallic-green crystals. He found it to be soluble in alcohol with dark-red color, in potas- sium hydroxide with splendid blue color, which, however, it loses on continued exposure to air. In ammonium hydroxide it dissolves with purple color. He regarded the compound as derived from pyrogallol by loss of water according to the reaction : 3C.H.O, = C lt H 14 0, + 2H.O. Pyrogallol. Gallein. The results of his analysis of the new compound agreed with the above formula. Later 3 he tried the action of phthalic anhydride upon other hydroxyl derivatives of benzene, among them phenol and resor- cin. He found that these reacted in a manner similar to that of pyrogallol, phenol giving phenolphthalein and resorcin giving fluorescein. From the analysis of phenolphthalein he was led to consider it as made up of one residue of phthalic an- hydride and two residues of phenol, and represented the reac- tion thus : C 9 HA + 2C 6 H 6 = C 10 H 14 O 4 + H 2 0. Phthalic Phenol. Phenolphthalein. anhydride. Fluorescein, from its similarity to phenolphthalein and also from the results of analysis, he regarded as derived from 1 molecule of phthalic anhydride and 2 of resorcin, thus : 1 Ber. d. chem. Ges., 4, 457- "Ibid., 4.555- 3 Ibid., 4. 658. C e H 4 O s + 2C.H.O, = C 50 H,A + 2H a 0. Phthalic Resorcin. Fluorescein, anhydride, A study of these reactions led him to change his views of the reaction by which gallein was made. 1 He regarded it as made according to the following equation : C 8 H 4 0, + 2C,H,0 3 = C, H„O 7 + 2H a O. Phthalic Pyrogallol. Gallein. anhydride. Thus it would be the phthalein of pyrogallol. By reduction in alkaline solutions by zinc dust, Baeyer dis- covered that phthaleins lose color and are converted into closely related compounds which he called phthalins. The structure of gallein is so closely related to that of the phthaleins in general that a summary of work done in estab- lishing the structure of the latter is necessary for an intelligent study of the structure of the former. Fluorescein was the first of the phthaleins to be investigated and have assigned to it a constitutional formula. The work was done by Baeyer and his assistants. 2 Fluorescein was made by heating together resorcin and phthalic anhydride. The purified product proved to be a yellow crystalline powder which dissolved in alkalies with a beautiful green fluorescence. Results of the analysis pf the compound pointed to formula C a „H M 0, as representing its composition. On the other hand, E. Fischer 3 prepared pure fluorescein by first making its acetate, purifying it by recrystallization, and then saponifying it with alcoholic caustic potash. Fluorescein thus prepared was analyzed and the results pointed to C, H„O, as its formula. Baeyer' s next object was to determine the number of hydroxyl groups contained in fluorescein. This he did by making its acetyl and benzoyl derivatives in the usual way. He found that in each case two acid residues were introduced — that the compounds formed were the diacetate and the di- 1 Ber. d. chem. Ges., 4, 663. 2 Ann. Chem. (L,iebig), 183, 2ff. 3 Ber. d. chem. Ges., 7, 1211. benzoate of fluorescein. Further, he made what he termed the monoethyl and the diethyl ethers of fluorescein, and the dichloride. The fact that in all but one of the derivatives thus far mentioned two substituting atoms or groups replaced hy- droxyls or the hydrogen atoms of these hydroxyl groups, pointed to the presence of two hydroxyl groups in fluorescein. Quite a series of substitution-products in which two or more hydrogen atoms were replaced, gave information of other char- acteristics of fluorescein. Thus, dinitro and tetranitro fluores- cein ; mono-, di-, and tetrabrom fluorescein (eosin) ; salts of eosin ; erythrin (ethyl ester of eosin) ; diethyl ether of eosin — all followed in rapid succession. By fusing fluorescein with potassium hydroxide Baeyer found that decomposition follows in two stages. First, he obtained what he called monoresorcin phthalein, to which he gave the formula : /OH y COC 8 H s < C 6 H 4 < \OH. x COOH When heated further this compound yielded, besides resor- cin, benzoic acid and carbon dioxide. The last two com- pounds came from the phthalic acid residue. As the result of his investigation, he concluded that without doubt fluorescein was a diketone and should be given the for- mula : /OH /COC.H/ C.H 4 < >o • \COC.H 3 < X)H The formation of the anhydride between the hydroxyl groups in the resorcin residues was, he said, in harmony with an observation made by him in the preparation of phenol- . phthalein. Two compounds were in reality formed in the syn- thesis of phenolphthalein — one a phthalein proper, the other a phthalein anhydride. The relation of the two is expressed in the formulas : / COC„H 1 OH /COC.H,. C.H.< and C.H/ V). \C0C 6 H 4 0H XX)C.H/ Phenolphthalein. Phenolphthalein anhydride. Fuorescein would, therefore, be a dihydroxyl derivative of trie above anhydride form, E. Fischer 1 made the phthalein of orcin, /CH,(i) C 8 H— OH (3) , X OH (5) and found it to resemble fluorescein very closely. He made a series of derivatives similar to those made from fluorescein by Baeyer. As a result of his study of this new phthalein he gave it a formula in harmony with that of Baeyer for fluores- cein, as follows : /CH, COC.H— OH C 6 H 4 <^ \»0 . COC.H— OH X CH a Baeyer' s view of the structure of fluorescein, (and of the phthaleins in general), was modified in some essential points as a result of his further study of phenolphthalein. 2 A difficulty arose in explaining by the accepted diketone formula the action of melted alkalies on phenolphthalein. Such action resulted in the formation of dioxybenzophenone, /C„H 4 OH CO< X C,H 4 OH and benzoic acid. According to the diketone formula, this would require a shifting of one of the C,H,OH groups. It oc- curred to him that this decomposition could easily be explained by assuming that it is one of the carbonyl oxygen atoms and 1 Ann. Chem (Liebig), 183, 63. 8 Ibid., 202, 36. not the anhydride oxygen of the phthalic anhydride that is re- placed by two phenol residues. The difference is seen in the following structural formulas : •CO carbonyl oxygen C„HX >0 — anhydride oxygen. x CO n ^C,H 4 OH COC e H 4 OH . C 0, >0. | | X C„H 3 OH/ CO— O A precedent for this he found in the formation of cedriret, C 6 H,(OCH,),0 I I, C e H,(OCH 3 ),0 from pyrogallol dimethyl ether by oxidation. To hydrogallein he assigned the formula : / C.H,(OH),v C„H 4 C< No, | | \c.H,(OH)/ CO— o as it is formed from gallein by the assumption of hydrogen ; and to gallin the formula : ,C e H,(OH), ^C 6 H,(OH), No. P TT fCiTX\ / COOH And gallol was represented thus : /C 6 H a (OH), v C„H 4 CH< No. | \C.H,(0H)/ CH.OH The fact that gallein forms a tetr&cetate and a tetrabenzoate he says, might be used as an argument against the formula as given by him, since one would expect a compound with such a formula to yield a ^'acetate and a akhenzoate. Analyses of the 15 compounds, however, leave no doubt that four acetyl, or ben- zoyl, groups have been introduced. The properties of these derivatives, also, preclude the possibility of their containing a quinone group. He adds, ' ' ' Man muss daher, wenn man an der Annahme einer Chinonbildung im Gallein festhalten will, eine Sprengung dieser Bindung bei der Aetherbildung voraus- setzen." Such a reaction, he continues, is not without prece- dent. Sarauw 2 has made hydroquinone diacetate from quinone by action of acetic anhydride and sodium acetate upon it. Graebe 3 has also made the diacetate of tetrachlorhydroquinone by the action of acetyl chloride upon chloranil. In 1892 Herzig* repeated a part of Buchka's work and ob- tained results which do not entirely agree with Buchka's state- ments. He was led to this investigation by a study of certain other compounds of the phthalein series. Graebe and Caro 6 had found that rosolic acid heated in a sealed tube with acetic anhydride forms an acetate, not of rosolic acid, but of its leuco derivative. This suggested the idea to Herzig that the action of acetic anhydride upon the phthaleins is not one of simple acetylization but of reduction as well. This idea was supported by the fact that the melting-points of the acetates of the phthaleins and their corresponding phthalins are very close to- gether. The question which presented itself was, do the phthaleins and the corresponding phthalins give the same prod- uct when acetylized ? To answer it he prepared carefully some fluorescein acetate and also some fluorescin acetate, and subjected them not only to the melting-point test but to chemical tests as well. He found that while these two acetates have almost exactly the same melting-point (flu- orescein acetate 20o°-20i" C, fluorescin acetate 2oo°-202° C. ) , they react very differently towards a dilute solution of alkali. Fluorescein acetate dissolves very slowly — only as it is saponified— and after several hours unchanged fluorescein acetate may be filtered out. Fluorescin acetate, on the other hand, dis- ^nn. Chem. (Liebig), 209, 255. " Ber. d. chem. Ges., 12, 680. 3 Ann. Chem. (Liebig), 146, 13. 4 Monatsh. Chem., 13, 425- ' Ann. Chem. (Liebig), 179. 198. i6 solves in dilute alkalies at once, and, if acidified, unchanged fluorescin acetate will be precipitated. They are clearly differ- ent substances. Inasmuch as Buchka had not subjected his hydrogallein acetate to such a test as this but had relied upon a melting-point test to prove its identity with gallein acetate, it occurred to Herzig to prepare gallein and hydrogallein ace- tates in order to determine whether they agreed in chemical conduct as they did in melting-point according to Buchka. As a result of this work he found, in the first place, that gallein acetate melted at 237°-Z38° C, instead of at 247°-248° C, as given by Buchka. In the next place, Herzig' s efforts to make hydrogallein and its acetate did not succeed. He purified his gallein carefully by converting it first into the ace- tate, recrystallizing this, and saponifying with potassium hy- droxide. This purified product he dissolved in potas- sium hydroxide solution and reduced in the cold by means of zinc dust. The reduced product thus made he acetylized with acetic anhydride and obtained a product that melted at 211 "-213° C. and dissolved in potassium hydroxide solution without saponification. It could not, therefore, have been identical with gallein acetate. Herzig as- sumed that it was the same compound as Buchka' s gallin tetracetate, which, according to Buchka, melts at 220 , though he did not make the latter compound in order to compare the two side by side, nor did he give any analyses. He concludes the account of his investigation by sasdng : ' ' Buchka scheint nach seiner Beschreibung eine grossere Menge des Hydrogal- leins in der Hand gehabt zu haben, und es konnen daher meine beiden negativen Befunde vorlaufig die Existenz des- selben nicht in Frage stellen. Allerdings glaube ich, dass die Darstellung des Hydrogalleins nicht so einfach ist, wie sie Buchka beschreibt und dass noch andere Vorsichtsmassregeln nothwendig sind, die der genannte Autor entweder nur unbe- wusst eingehalten oder in seiner Arbeit zu erwahnen vergessen hat. Ebenso halte ich es vorlaufig nicht fur ausgeschlossen, dass die Identitat der Acetylproducte des Galleins und Hydro- galleins nur auf Grund der Schmelzpunkte ausgesprochen wurde, wahrend sie sich in Bezug auf die I,6slichkeit in Alka- 17 lien von einander so unterscheiden konnen, wie Acetylfluores- cei'n von Acetylfluorescin. " Gallein diphenylsulphonate and tetraphenylsulphonate have also been made by M. Georgescu. 1 The gallein diphenylsul- phonate was made by treating an alkaline solution of gallein with benzene sulphonchloride until the color of the solution be- comes greenish-yellow. In solution the compound is dichroic, green by reflected light, reddish-yellow by transmitted light. He assigned to it the formula : /OSO s C 6 H 6 CO— C.H,< c„h, yo \ //OH COC.H/ X)SO„C„H« 'a^e* Gallein tetraphenylsulphonate was made by the same inves- tigator by treating an alkaline solution of gallein with benzene sulphonchloride in excess. A brown precipitate, insoluble in alkalies, was thrown down, which, when purified, was nearly colorless and melted at i87 u -i88° C. To this he gave the formula : /OSO s C 6 H 6 CO— C e H 2 < / x \0S0 3 C 6 H 6 C 6 H, >0 \ / /OS0 3 C 6 H 6 CO-C 6 H a / OS0 3 C 6 H 6 The period between 1891 and 1897 was one of great activity in researches upon the constitution of the phthaleins. This was stimulated by a suggestion as to the structure of fluores- cein on the part of Bernthsen, who first gave the quinoid struc- ture to this compound instead of the lactoid structure proposed by Baeyer. The relation of the two may be seen in the follow- ing formulas for fluorescein : 1 Buletinul Societ&tii De Sciinte Fizice, ±, 215 (1892). i8 /OH /OH /C.H S < / C . H >< C.H 4 C< >0 C,H 4 C4 >0 CO— O x OH COOH ^O factoid structure. Quinoid structure. Bernthsen 1 suggested that in free condition as well as in such derivatives as have no color the structure is to be represented by the lactoid formula, while in such derivatives as possess color the structure is best represented by the quinoid arrange- ment. The discussion for the next few years centered largely around this question — the lactoid versus the quinoid arrange- ment — and one other, the relation of the hydroxyl groups to the methane carbon atom. Richard Meyer 2 showed the connection between fluorescein and fluoran (phenolphthalei'u anhydride) — that fluorescein is in reality dioxyfluoran. This he did in two ways. First, by treating fluorescein with phosphorus pentabromide he obtained fluorescein dibromide : /Br C. :.h 4 c< >o :0— O x Br CO— O N Br This he reduced by nascent hydrogen and obtained fluoran : /Br /C,HX /C,H 4 v C, H 4 C< >0 + 4 H = C, H 4 C< >0 + 2HBr. I l X C.H,< I l X C„H/ CO— O x Br CO— O Fluorescein dibromide. Fluoran. Second, by action of phosphorus pentachloride on fluorescein he made fluorescein dichloride : /CI C e H 4 C/ >0. i iVtt/ CO— o \ci ' Chem. Ztg., 16, 1956. 2 Ber. d. chem. Ges., 25, 1385. 19 This, in turn, when treated with phosphorus tribromide gave him monobromfluorescei'n dibromide : y Br /C 6 H,Br< C 6 H 4 C< >0. CO— O ^Br When this is treated with alcoholic caustic soda and zinc dust it is reduced to an acid to which he assigns the formula : C e H t CH< >0 COOH and names hydrofluoranic acid. The same acid is formed when fluoran is heated with alcoholic caustic soda and zinc dust There could be no doubt, therefore, that the relation of fluorescein to fluoran was expressed in the formulas : /OH >0 C e H 4 C/ >0 i , C.H/ I | X C H S < CO— O » CO— O X)H Fluoran. Fluorescein. Meyer next applied himself to the task of establishing the structure of fluoran. 1 He found that fluoran, when heated with lime, decomposes into benzene, carbon dioxide, and xanthone : H, C 6 H 4 + I O CO— /C 6 H 4 \ /CjH^ C < >0 = C 6 H + CO, + CO< >0. I X C e H/ X C e H/ O Fluoran. Xanthone. Again, he distilled fluoran with soda-lime and zinc dust 2 and found the products to be water, carbon dioxide, and diphenyl- ene phenylme thane. ■Ber. d. chem. Ges., 25, 2118. 2 Ibid., 25, 3586. 20 C.H 4 C0. I I x c,h/ CO— o Fluorescein, as a dioxy-derivative, would be : /OH /C;h 3 < C e H 4 C/ >0 . I I X C 6 H 3 < CO— O X)H Friedlander 1 contributed some observations which seemed to confirm the views of Baeyer as to the relation of the hydroxyl groups of phenolphthalein to its methane carbon atom. This he did in the preparation of the so-called phenolphthalein oxime and a study of its decomposition-products. Phenolphthalein in alkaline solution reacts with hydroxylamine as a quinone or a ketone would act, forming an oxime according to the reaction : /C 6 H 4 OH /C 6 H 4 OH C„H 4 C< + H 2 NOH = H a O + C 6 H 4 C/ | ^C 6 H 4 = | X C 6 H 4 = NOH COOH COOH This oxime is a light-yellow product, dissolves in alkalies with yellow color, reprecipitated by acids. Its decomposition- products were especially interesting. Fused with caustic soda, it yielded benzoic acid, /-oxybenzoic acid, and />-amidophenol : /C 6 H 4 OH C H 4 C4 + 2 H 2 = C„H 6 COOH + | ^C„H 4 = NOH COOH /OH /NH, C e H 4 < + C H 4 < XX)OH(^) NDHO) 1 Ber. d. chetn. Ges., 26, 172 and 2258. The oxime, when boiled with dilute sulphuric acid, forms /-oxy-0-benzoylbenzoic acid and /-amidophenol : /C„H 4 OH C.H,C< + H,0 = | ^C 8 H 4 = NOH COOH yCOOH(o) /NH, C.H t < + C„H 4 < \COC„H,(OH)(/) \OHO) From these results it would seem that the hydroxyl groups are para to the methane carbon atom, as first suggested by Baeyer. Friedlander, in testing the oxime reaction with other phthaleins, found that it took place most readily in those cases in which the quinoid structure was made known by color. This applied to derivatives as well as to the phthaleins them- selves. For example, the dark-blue or violet alkaline solution of tetrabromphthalei'n reacted easily with hydroxylamine, but not so the light-yellow solution of the di- or tetranitrophenol- phthalein. It was of special interest to observe that the solu- tion of fluorescein in dilute alkalies — which is reddish-yellow — did not react with hydroxylamine, while the deep-violet solu- tion of fluorescein in hot concentrated alkalies combines with that reagent easily. These facts, he claimed, supported the lactoid formula for free fluorescein and its stable salts, at the same time pointing to the quinoid structure for such derivatives as possess marked color. O. Fischer and Hepp 1 made fluorescein anilide by boiling to- gether fluorescein, aniline, and aniline hydrochloride. The ex- cess of aniline was removed by adding excess of alkali and dis- tilling with steam. The residue was crystallized from alcohol. The crystals were colorless, dissolving without color in alcohol and glacial acetic acid, in alkalies with green fluorescence. On account of its lack of color they concluded that it could not have the quinoid structure and wrote the equation expressing its formation as follows : ' Ber. d. chem. Ges., 26, 2236. /C a H 3 OH /C 6 H 3 OH C H 4 C< >0 + H 2 NC e H B = H,0 + C a H 4 C< >0. CO— O CO— N— C„H B By a reaction similar to that used in the preparation of fluorescein anilide, Albert 1 made phenolphthalein anilide : C„H 4 C< + H,NC„H = H,O + C 8 H 4 C< I I X C 6 H 4 OH | | N:„H 4 OH CO— O CO— N— C„H 6 This anilide is also colorless and agrees in its properties gen- erally with the anilide of fluorescein. Again, by the same reaction, Albert 8 made gallein anilide and represented the structure in this way : C,H,C(C,H,OH),0,0. I I 0=C NC 6 H 6 He found it to be colorless, soluble in alkalies without color, and with methyl iodide in a sealed tube at ioo°, formed a di- methyl derivative. A year later O. Fischer and Hepp 1 made the anilide of fluores- cein monomethyl ether. In their study of this compound they were led to modify their view of the structure of phthalein anilides. Up to this time these had been represented as phthal- eins in which the anhydride oxygen of the phthalic acid resi- due was replaced by an aniline residue. The authors now represented them as phthaleins in which the carbonyl oxygen is replaced by the aniline residue. Thus, for the anilide of fluo- rescein monomethyl ether they gave the formula : /C.H.OCH, C 6 H 4 C< >0 / X C 6 H s OH O / C = NC 6 H 6 1 Ber. d. chem. Ges., 26, 3077. 2 Ibid., 27, 2793. slbid., 28,398. 23 And for fluorescein anilide : C e H, C< >0 O / C = NC„H 6 They assign no reason for either formula. That the anilides above referred to have not the quinoid structure was shown by converting them into methyl ethers, splitting off the aniline residue, which resulted in the formation of a colorless ether, as shown by the following formulas : ,C„H s OH / X C.H s OH O / C = NC 6 H 6 Fluorescein anilide. /C 6 H 8 OCH, C, H. C< >0 O / C = NC.H. Dimethylfluorescein anilide. . H, C< >0 ,OCH„ / o / c = o Dimethyl ether of fluorescein. (Colorless.) Haller and Guyot 1 submitted evidence for the lactoid struc- ture of both phenolphthalei'n and fluorescein in the free state by making diphenyl carbamate derivatives of each. This was done by simply heating the phthalei'n with phenyl isocyanate to 130 , and recrystallizing the product from benzene. There resulted a colorless compound, stable with boiling water, but easily decomposed by alkalies, regenerating the corresponding phthalein. They represented the reaction for phenolphthalei'n thus : 1 Compt. rend., 116, 479. 2 4 /C,H 4 OH C„H ( C< +20CNC 6 H S = I I X C 6 H 4 OH CO— O /C t H 4 OCONHC,H, C H C' T " ' | \c,H 4 OCONHC.H.' CO— O Phenolphthalein bis-phenylcarbamate. Fluorescein reacted similarly, yielding fluorescein bis-phenyl carbamate. Now, since phenyl isocyanate reacts with hy- droxyl groups but not with carbonyl groups, and since two phenyl isocyanate residues are introduced into the molecule, it is shown that mfree phenolphthalein, as well as in free fluores- cein, there are two hydroxyl groups. A valuable contribution to the question of the structure of the phthaleins was made by Nietzki and Schroter 1 in their work on the ethyl ethers of fluorescein. They succeeded in preparing diethyl ethers of fluorescein by reactions that leave no doubt as to their structure. Fluorescein was first reduced to fluorescin, and this converted into the ethyl ester in the usual way. The relation of these compounds is shown in the following formulas : /C H 3 OH /C.H.OH C„H,C< >0 — C 6 H 4 CH< >0 — | | ^C.H.OH | \C 6 H 3 OH CO— O COOH Fluorescein. Fluorescin. ;,H 4 CH< I \P T ,C„H,OH >0 ^C 6 H 3 OH ..COOC.H, Fluorescin ethyl ester. Fluorescin ethyl ester, by oxidation, goes easily into fluores- cein ethyl ester : /C„H s OH C.H.C^ >0 COOC.H, ■ Ber. d. chem. Ges., 28, 44. 25 This compound has color. When treated with sodium ethylate and ethyl bromide, it undergoes the changes shown in the following equations : /C.H.OH C e H,C< >0 + C,H 6 ONa = I ^C 6 H s =0 COOC 2 H B /C„H 3 ONa C,H 6 OH + C 6 H 4 C<; >0 ; 1 ^C 6 H=0 COOC,H 6 /C 6 H s ONa C.H,^ >0 + C 3 H B Br = COOC,H NaBr + C 6 H,c/ >0 . I ^C 6 H s =0 COOC 2 H t Fluorescein diethyl ether. (Colored.) That the colored fluorescein diethyl ether has the structure assigned to it above — the quinoid structure — is shown not only by the method by which it is prepared, but also by the fact that when carefully saponified it loses the ethyl group com- bined with the carboxyl group and is converted into a mono- ethyl ether, which, also, has color. / C„H 3 OC,H 6 C.H 4 C£ >0 1 ^C e H s = / C 6 H 3 OC 2 H l - C 6 H 4 C/ >0 1 ^C 6 H 3 = O COOC,H s COOH The same authors (loc. tit. ) give valuable evidence as to the structure of the colorless, or lactoid diethyl ether of fluorescein. They reduce fluorescein to fluorescin, as in the previous case, esterify, and from the ester, by means of sodium ethylate and ethyl bromide, in the correct proportions, obtain fluorescin triethyl ether, which, on saponification and oxidation, yields colorless fluorescein diethyl ether. The relations of these products to each other are shown as follows : 26 /C 6 H s OH /C e H 3 OH C e H 4 C< >0 — C 6 H,CH<( >0 | | \C.H 3 OH | X C,H 3 OH CO— O COOH Fluorescein. Fluorescin. /C.H.OH C,H 4 CH< >0. | ^C.H.OH COOC.H, Fluorescin ethyl ester. C.H.OC.H. /C.H,OC,H C 6 H t CH0 C„H.C >o P TT or T ^C.H 3 OC a H B | | NC.H.OC.H. COOC„H B CO— O Fluorescin triethyl ether. Fluorescein diethyl ether. (Colorless.) From the methods of preparation, there can be little doubt of the correctness of the above formulas for the two diethyl ethers of fluorescein — the colored, or quinoid, and the color- less, or lactoid. It is interesting to note, in this connection, as showing that fluorescein reacts as a tautomeric substance, that the same authors 1 succeeded in making both ethers in one operation. The potassium compound of fluorescein was heated with ethyl bromide in a closed vessel on the water-bath for several days. The two ethers thus made were separated by fractional crystallization from alcohol, the colored ether being more soluble in that liquid than the colorless ether. O. Fischer and Hepp 2 made the quinoid (colored) dimethyl ether of fluorescein by action of methyl iodide upon an alkaline solution of fluorescein in methyl alcohol. It was given the formula : /C.H.OCH, C.H.C^ >0 . COOCH, That this is correct was shown by its conduct when treated with an alkali — it saponified and yielded monomethyl fluores- cein, also colored : = Ber. d. chetn. Ges., 28, 49. 2 Ibid., 28, 396. 2 7 /C.H.OCH, C B H 4 C4 >0 . COOH In the year 1895 two investigations were reported which proved the relation of the hydroxyl groups to the methane car- bon atom. Graebe had already pointed out that there were only three possible arrangements — remembering that resorcin is a meta derivative : HO o-^-Unsy mmetrical . Matras' made dinitrofluorescein, decomposed this with potas- 1 Chem. Ztg., 19, 408. 28 sium hydroxide and obtained a volatile nitroresorcin which melted at 85 ° C. From this it appears that the product was that derivative of resorcin, in which the groups have the relation OH : NO, : OH = 1:2:3, The formula for the dinitro- fluorescein, then, would be : NO, O NO, HO OH Since the nitro group is inserted between the hydroxyl groups, formulas II and III are excluded, since in them that position is already occupied. According to this, therefore, the hydroxyl groups are para to the methane carbon atom. To the same conclusion, led an investigation on eosin begun by A. Baeyer 1 and continued by Heller. 2 Baeyer heated eosin with concentrated solution of sodium hydroxide and obtained dibromdioxybenzoyl benzoic acid : :.H 4 C<( 1 N C.HBr 3 OH >0 + H a O C.HBr,OH CO— O Eosin. C 9 H,Br 2 (OH) 2 + C.H 4 COC.HBr,(OH) t . Dibromresorcin. COOH Dibromdioxybenzoylben- zoic acid. Heller treated this last compound with warm sulphuric acid 1 Ann. Chem. (Liebig), 183, 56. * Ber. d. chem. Ges., 28, 312. 2 9 and found that it condensed by loss of water to dibromxantho- purpurin, thus : >C.Br,(OH), COOH ,CO s Since this is a derivative of xanthopurpurin it must have the structure : CO OH OH CO Br For its hydrated form (see above) there are two alternatives I. II. CO OH COOH OH OH OH COOH Br CO Br But formula II is excluded, since it has been already shown' that one of the hydroxyl groups is ortho to the phthalic acid residue. It will be seen from formula I that the remaining hydroxyl group is para to the phthalic acid residue. This established the relation of the hydroxyl group to the methane carbon atom so far as one-half of the molecule was con- cerned. It only remained to show that fluorescein is symmetri- cal, that the two resorcin residues sustain the same relation to the molecule. This fact was first established by Baeyer. 2 He found that the dioxybenzoylbenzoic acid, obtained by the de- composition of fluorescein, when heated again, loses water with formation of fluorescein and phthalic anhydride : 1 Cf. R. Meyer's work on relation of fluorescein to fluoran, p. 18. 2 Ann. Chem. (Liebig), 183, 25. 3° 2C.H / COC„H a (OH), 4 \COOH Dioxybenzoylbenzoic acid. 2 H,0 + C.H / >0 + C, H 4 C < >0. CO— O Phthalic anhydride. Fluorescein. Afterwards it was shown 1 that dibromdioxybenzoylbenzoic acid also loses water when heated, forming phthalic anhydride and eosin : XOC„HBr,(OH), 2C 6 H 4 < ^COOH /CO. /C e HBr,OH 2 H a O + C 6 H,< >0 + C.H 4 C< >0. X5(K | | N^HBr.OH CO— o Eosin. These syntheses of fluorescein and eosin from one product of their decomposition, show clearly that they have a symmetri- cal structure, and constitute the last link in the chain of evi- dence for the para relation of the two hydroxyl groups to the methane carbon atom. Baeyer 2 found that when phenolphthalein was decomposed by fusing with caustic soda one of the products was dioxyben- zophenone. It was afterwards shown that this was para dioxybenzophenone : HO -CO- OH This proved that in phenolphthalein, also, the hydroxyl groups are para to the methane carbon atom as in fluorescein. Herzig and H. Meyer 3 advocated the lactoid structure of 1 Ber. d. chem. Ges., 28, 1576. 2 Ann. Chem. (Liebig), 202, 38. 3 Ber. d. chem. Ges., 28, 3258. 3i salts of phenolphthalei'n by reason of a colorless dimethyl ether made by them. An alkaline solution of phenolphthalei'n treated with methyl iodide, gave a yield of 85-90 per cent of the colorless, or lactoid, dimethyl ether — a white crystalline substance, insoluble in alkalies. To this they assign the for- mula : /C.H.OCH, C.H 4 C< I I ^C.H.OCH, CO— o With this as the structure of the dimethyl ether, the sodium salt, which is an intermediate product in its preparation should, they claim, have a similar structure : /C 6 H,ONa C„H 4 C< I I \C H 4 ONa CO— o That the colorless dimethyl ether really has the structure given above cannot be doubted in view of its preparation by E. Grande 1 by a different method. He made it by the action of anisol upon phthalic anhydride in the presence of aluminum chloride, thus : / C 6 H,OCH 3 C,H 4 C = C.H 4 C< +H g O. I j + 2C.H 6 OCH, = I I \C.H 4 0CH S CO— O CO— O This product is identical in every respect with that obtained by Herzig and Meyer as given above. In making the dimethyl ether of fluorescein, however, by the action of methyl iodide upon its alkaline solution, Herzig and Meyer 4 obtained a colored product to which the quinoid arrangement was assigned : /C 6 H 3 OCH 2 C 6 H 4 C/ >0 . I ^C,H,=0 COOCH, 1 Gazz. chim. ital., 26, 1 and 222. 2 Loc. cit. 32 This would seem to favor the quinoid structure of the alkali salts of fluorescein, just as in the case of phenolphthalein the fact that in alkaline solution it yields with methyl iodide the lactoid dimethyl ether was used by the authors as an argument in favor of the lactoid structure of its alkali salts. The same authors 1 criticize Friedlander's interpretation of the results he obtained with the oxime of phenolphthalein. 2 Instead of proving, as Friedlander supposed, the existence of a quinoid oxygen by forming the compound | ^C e H 4 = NOH COOH it in reality points, they claim, to a very different conclusion. They base this criticism upon facts brought out by Friedlan- der, himself ; viz. , the oxime has slight color, it forms with acetic anhydride a colorless diacetate, and dissolves in acids. Further, in acid or alkaline solutions it is reduced by zinc dust, forming a compound insoluble in acids. These facts, they say, can best be explained on the assumption that the oxime has the structure /C.H.OH ! I N C c H 4 OH CO— NOH agreeing with the analogous formula first given for the anilide. 3 The product obtained by reduction and found to be insoluble in acids would be represented thus : /C.H.OH C 6 H 4 C< I I X C,H 4 OH CO— NH Nietzki 4 made the quinoid diethyl ether of tetrabromphenol- phthalein by a reaction similar to that used by him in making ' Ber. d. chem. Ges., 28, 3258. 2 Cf. p. 20. 3 Cf. p. 22. 1 Chem. Ztg., 20, 806. 33 the colored diethyl ether of fluorescein. The phthalein was reduced to the corresponding phthalin, this was esterified, then oxidized to the ester of the phthalein, the silver salt of this with ethyl iodide yielded the colored (quinoid) diethyl ether of tetrabromphenolphthalein. The steps are indicated in the following formulas : /C.H 2 Br 2 OH /C 6 H,Br a OH C.H 4 C< — C H 4 CH< ; | | \c„H. i Br a OH | \C a H a Br,OH CO— O COOH /C„H Q Br a OH / C.H 1 Br,OH C„H 4 CH<( — C 6 H 4 C4 | \C e H a Br,OH | ^C 6 H 2 Br s = COOC s H 6 COOC,H 6 ^C.H.Br.OAg / C.H 1 Br,OC,H. C 6 H 4 C< — C 6 H 4 C4 |° ^C H,Br a = O | ^C H,Br 9 =O COOC,H s COOC,H B A. Bistrzycki and K. Nencki 1 have made a contribution to the question of the structure of phenolphthalein. They call attention to the two structures for alkali salts of phenol- phthalein : /C,H 4 OK /C.H.OK C.H 4 C/ C 6 H 4 C4 1° |\C 6 H 4 0K ! ^C„H 4 = CO-0 COOK i. n. and claim that if formula I is correct, it should yield by the Baumann-Schotten reaction a rfz-benzoate insoluble in alkalies and colorless. On the other hand, if formula II be correct, then it should yield a mono-benzoate under the same condi- tions— possibly a to-benzoate by the breaking up of the qui- noid arrangement and becoming 1 Ber. d. chem Ges., 29, I3 1 - 34 /C.H.OK | \ X C t H 4 OK COOK OH The authors found that a afe-benzoate was formed under the conditions mentioned, colorless and insoluble in alkalies. If the compound were derived from formula II, whether it were the monobenzoate or the tribenzoate it should be soluble in alkalies. From this they advocate the lactoid structure of salts of phenolphthalem. The action of phenylhydrazine upon fluorescein and phenol- phthalem was investigated by Gattermann. ' He found that when the phthalei'n was heated with phenylhydrazine, combi- nation took place, crystals being formed that had no color and were soluble in alkalies. Not only did fluorescein itself react with phenylhydrazine but also fluorescein dichloride. He found further that the simple hydrazide of both fluorescein and phe- nolphthalein could be converted into dimethyl and diethyl ethers by the action of methyl iodide upon their alkaline solu- tions. The structure assigned by him to these hydrazides is shown in the formulas : /C„H 4 OH /C 6 H,OH C„H 4 C< C H 4 C< >0. I I X C 6 H.OH | | \C 6 H s OH CO— N— NHC„H B CO— N— NHC.H, Phenolphthaleiin hydrazide. Fluorescein hydrazide. It will be readily seen how in alkaline solution these com- pounds would yield, with methyl or ethyl iodide, diethers. From the foregoing discussion it would seem to be clearly established that fluorescein and phenolphthalem really react as tautomeric compounds, in some cases giving lactoid deriva- tives, in others quinoid derivatives. In regard to gallein, however, it cannot be said that its structure has been determined with the same degree of cer- tainty.. Buchka, who has done most of thewoik on gallein, gave it the formula : 1 Ber d. chem. Ges., 32, 1132. I i CO— o 35 C„H,OH >o,\o. C 8 H 2 OH / This he did by reason of his study of hydrogallein, a compound which Herzig, who repeated a part of Buchka's work, did not succeed in making. Schultz and Julius in their ' ' Tabellarische Uebersicht der kiinstlichen organischen Farbstoffe," edition of 1 89 1, page 98, give it the formula: C„H a OH— O >o I C„H„OH— O . C.H.CO— O 1 The same structure is given it in L,eon I,ef evres ' ' Matieres Colorantes Artificielle", 1896, page 1190. Schultz and Julius, in the 1897 edition of their work, page 142, give its structure more in detail as follows : HO representing it, however, as orthoquinone. Nietzki,in his "Chemie der organischen Farbstoffe" (1897), gives it the structure : 36 HO OH He adds, however, ' ' Wahrscheinlich ist es fast, dass diese bei- den Sauerstoffatome in einem Kern eine Orthochinongruppe bilden, auch diirfte das Gallein zu den Chinoiden Carboxyl- derivaten gehoren." This idea of Nietzki's regarding thequinoid condition is con- tained in the structure suggested in Krafft's Organische Chemie, 1897, P a S e 595 > as follows : ,OH CO < X > O OH C.H, % ■O O According to this gallein contained a quinoid, carboxyl, and quinone group. With this diversity of opinion, it would seem important to have additional experimental evidence as to the correct formula for gallein. This is presented in the following pages. II. Coerulein. This compound also was discovered by Baeyer in 1871, 1 and was the first of the class of phthalideins — compounds derived from phthaleins by loss of water. He made it by the action 1 Ber. d. chem. Ges., 4. 457. 37 of concentrated sulphuric acid upon gallein at 2oo°C, the color of the solution changing from reddish brown to greenish brown. Coerulein was precipitated by pouring into water. When washed carefully and dried, it appeared as a blue-black mass. It chars when heated giving at the same time a colorless sublimate. It is soluble in concentrated sulphuric acid with olive-brown color, from which it may be crystallized. Slightly soluble in alcohol, water, or ether, easily in glacial acetic acid with dirty green color. In hot aniline it is easily soluble with deep blue color from which it may be precipitated in blue flakes by addition of alcohol or acetic ether. Alkalies dissolve coerulein with green color. By reduction it yields coerulin. Though the first phthalidein discovered, its structure was not investigated until after that of phenolphthalidein had been determined. Baeyer, 1 in making his investigation of phenol- phthalidein found that when heated with zinc dust, there dis- tilled a product which proved to be phenylanthracene. He had two points to guide him in assigning to it a formula : It was derived from phenolphthalein by loss of water, and it gave an anthracene derivative on reduction. In view of these facts he gave it the formula : HO C.H.OH \ / C C 8 H 4 <^ )>C,H 3 OH CO Though containing three hydroxyl groups, he found that it gave a ^-acetyl derivative. He explained this by saying that the anthranol hydroxyl is peculiar in being very difficult to acetylize, in this case not reacting at all. In 1 88 1 Buchka 2 took up the study of coerulein in connection with his work on gallein. He found that it could be made not only from gallein and sulphuric acid but also from gallin, con- verting it first into coerulin by sulphuric acid and oxidizing this to coerulein. Its properties were found to be those already assigned to it. ' Ann. Chem. (Liebig), 202, 90. 2 Ibid., 209, 258. 38 He prepared coerulein triacetate by action of acetic anhy- dride upon it, also by gentle oxidation of coerulin tetracetate. In this latter reaction he suggests that coerulei'n tetracetate is first formed which immediately loses one of the acetyl groups. He found that the acetate could be easily saponified by alkalies and by concentrated sulphuric acid. Its solution in acetic acid is decolorized by zinc dust, assuming, at the same time, a greenish yellow fluorescence. When diluted, yellow flakes separate which are so unstable as to make analysis impossible. Reduction of coeruleiin in solution in ammonium hydroxide by zinc dust gives coerulin, which is easily oxidized again, even by the air, to coerulei'n. He prepared coerulin acetate by using coerulein and acetic anhydride with a little zinc dust. His analyses led him to pronounce the product a tetracetate. He found its melting-point to be 256°. When gently oxidized it yields coerulein triacetate. He showed the connection between coerulein and phenylan- thracene by mixing with zinc dust and distilling in an atmos- phere of hydrogen. A yellow product was obtained which crystallized from alcohol in yellow plates and when dissolved in glacial acetic acid, some potassium dichromate added and boiled, gave a product which he identified as phenyloxanthra- nol. Buchka accepted the conclusion of Baeyer as to the structure of phenolphthalidein and phenolphthalidin and assigned cor- responding formulas to coerulein and coerulin, with some mod- ification as to coerulein. Thus for coerulin he adopted the formula : C.H,(OH)„. C„H Though this formula contains five hydroxyl groups, he found that it formed a tetracetate. He agreed with Baeyer in say- ing that the anthranol hydroxyl group was unaffected by acetic anhydride. 39 The formula for coerulein, if it be analogous to phenolphthali- dei'n, should be, he concludes, HO. /C.H.OH. C„H 4 ^ ^>C HOH / II O But it forms an anhydride, as does rosolic acid, forming a com- pound to which he gives the structure : On. >C.H„ • c< >cQo. C.Hr >C e HOH / o Though containing only one hydroxyl group, he found that it gave a triacetyl derivative. He assumes in this case, as in the case of gallein, that the bond between the two quinone groups is broken. The formation of coerulein triacetate from coerulin tetracetate, he says, supports the formula given above for coerulein. Prud'homme,' in his study of the reaction between coerulein and aniline hydrochloride, came to a different conclusion as to the number of hydroxyl groups in coerulein. When these two substances were heated together, condensation took place giving a derivative of coerulein containing two aniline residues. This product was found to have basic properties, and its ace- tate, with proper mordants, could de used as a dye, although an unsatisfactory one, since it changes color in the light. The fact that the free base is not fixed at all by mordants of iron, chromium, or aluminum, is used by the author to prove that it contains no hydroxyl groups. Coerulein, however, must contain at least two such groups in order to harmonize with the law of I,iebermann and Kostanecki as to the coloring quali- ties of oxyketone dyes, that they must contain two hydroxyl groups in the ortho or the para relation to each other. If, then, the coerulein-aniline condensation product has no hy- 1 Bull. Soc. chimique (Paris) (3) xi, 1136. 4-C droxyl group, and coerulem has at least tpo, it follows that the aniline residue is connected with the coerulei'n molecule through its hydroxyl groups. And if the aniline residues are introduced through the hydroxyl groups and only two residues are so introduced, it follows that coerulei'n contains two and only two hydroxyl groups. In view of this fact, as well as its connection with gallein, he gives coeruleiin the formula : | >C,HOH— O C e H 4 < >C HOH-O O This formula also represents coerulei'n as a derivative of phenyl- oxanthranol. In Schultz and Julius' ' 'Tabellarische Uebersicht der kiinst- lichen organischen Farbstoffe," edition of 1891, page 98, the formula for coerulein is given thus : f C-C 6 HOH— o I > C.H 4 < -O C--C.H,. \/ o In the 1897 edition of the same work, page 142, its struc- ture is given more in detail, as follows : Ov OH O 1 0= — o 4i Nietzki in his "Chemie der organischen Farbstoffe," 1897, page 175, gives coerulei'n the same structure as that assigned to it by Schultz and Julius in the 1891 edition of their work quoted above. He adds, however, in a foot-note the following remark : ' ' Eine Controlle dieser Formel ware wohl zeitgemass. Die vielen Sauerstoffbindungen lassen sich kaum mit den heutigen Anschauungen iiber derartige ringformig constituirte Molekule vereinigen. Auch sind die beizenfarbenden Eigen- schaften bei Vorhandensein nur einer Hydroxylgruppe kaum zu erklaren. ' ' From the foregoing historical review may be gathered the development of ideas in regard to the structure of the phthal- eins and phthalideins down to the present time. In the fol- lowing pages is given an account of an investigation under- taken with a view to throwing additional light upon the sub- ject so far as two of these are concerned — gallein and coerulein. EXPERIMENTAL. I. GALLEIN. The gallein used in this investigation was supplied by the Badische Anilin und Soda Fabrik, Eudwigshafen am Rhein, to whom grateful thanks are herewith tendered for the generous interest they have manifested in the work. It was an unusually pure product and beautifully crystallized. By transmitted light it was red, by reflected light green, with a bronzy luster. When acetylized it left no residue other than gallein acetate. Gallein dissolves in an aqueous solution of pyrogallol, more easily in boiling solution, from which, after several hours, it crystallizes out. That gallein, in addition to its marked acid properties, has slight basic properties is shown by the fact that its solubility in water or alcohol is considerably increased by the presence of an acid. This same conduct on the part of fluorescein was noted by Baeyer' and by Waddell. 2 In the case of aurin, which resembles gallein to a marked degree, 1 Ann. Chem. (Liebig), 183, 6. ' J. Phys. Chem., 2, 173. 4 2 Dale and Schorlemmer' have isolated a hydrochloric acid salt. 2 When gallein is dissolved in alcohol and an alcoholic solution of potassium hydroxide added, the potassium salt of gallein is precipitated. It has a blue color. Attempts were made to reduce gallein by means of hydrogen sulphide. Solutions in ammonium hydroxide, in potassium hydroxide, sodium carbonate, and in alcohol remained un- changed after passing the gas through them for some time. Sulphur dioxide was also tried as a reducing agent, but with negative results. An alcoholic solution of gallein remains un- changed when treated with sulphur dioxide. Gallein in dilute caustic potash solution, when treated with sulphur dioxide, is thrown down as a dark-red precipitate, due to the neutraliza- tion of the potassium hydroxide. An aqueous solution of gallein, made by boiling gallein with water and filtering the solution, treated with a solution of lead acetate or lead nitrate in water, gives first a deep-blue solution from which, finally, all the gallein is precipitated as the basic lead salt, with indigo-blue color. This reaction is quite deli- cate and might be used as a test for either gallein or lead. Gallein does not melt or undergo any change when heated to 300° C. Air passed through a solution of gallein in caustic potash causes the color to change from deep-blue to brown. Hydro- chloric acid added to this brown solution produces vigorous effervescence but no precipitate. It is probable that oxidation takes place here as with pyrogallol itself. This tendency of gallein to oxidize is seen further in its action on silver nitrate, which it reduces to metallic silver. Gallein Tetracetate. Gallein tetracetate was made according to Buchka's direc- tions—by boiling gallein and an equal weight of sodium ace- tate in four times the weight of acetic anhydride for four hours. After a number of trials it was found that the acetyli- 1 Ann. Chem. (Liebig) ig6, 88. s Cf. Collie and Tickle on "Salts of Dimethylpyrone and the Quadrivalence of Oxygen." J. Chem. Soc, 75 and 76, 710. 43 zation was complete in one hour. The excess of acetic anhy- dride was distilled off and the residue poured into a large amount of water. The acetate appeared on the bottom of the vessel, part as crystalline particles, part as an oil that soon solidified. The product thus obtained was filtered and when dry had a gray color ; yield almost quantitative. Crystallized from benzene, it was obtained in colorless needles. Recrys- tallized from benzene, acetone, alcohol, and acetic acid in suc- cession till no further change in its melting-point could be ob- served, it was found to melt at 241° C The melting-point as given by Buchka is 247"-248°, by Herzig 2 236°-237° '. Gallein acetate in a cold 5 per cent solution of sodium car- bonate is insoluble ; the solution remains perfectly colorless, though allowed to stand two or three days. When boiled for some time there is partial saponification, as seen by the slight purple tinge of the carbonate solution. When treated with a few drops of concentrated sulphuric acid in the cold, galleiin acetate dissolves, giving the acid solu- tion a deep-red color. When diluted with water the solution retains its red color, and after standing some time deposits red crystals. These crystals, when filtered out and washed with a little water, dissolve in ammonium hydroxide with the purple color and in caustic potash solution with the deep blue color so characteristic of galleiin, and give also the unmistakable test for galleiin with lead acetate. There can be no doubt that sulphuric acid saponifies the acetate and the crystals that separate are those of galleiin. The determination of the exact constitution of gallein ace- tate was an urgent matter. The difficulty of deciding it by a combustion analysis will be seen at once by a comparison of the percentages of carbon and hydrogen in gallein acetate of different content of acetyl groups. No. of CH3CO groups. C. Per cent. 4 63.16 H. Acetyl groups. Per cent. Per cent. 3-78 32-34 3-84 37-48 3-92 41.89 5 62.72 6 62.34 1 All melting-points were determined by means of a thermometer compared with one standardized by the Physikalisch-Technische Reichsanstalt, Charlottenburg, Germany. 2 Monatsh. Chem., 13, 426- 44 From the above it will be seen that the percentage of carbon differs by about 0.4 per cent, hydrogen by about 0.07 per cent in the several cases. Buchka found C = 63.52 and 62.72, H = 4.44 and 3.82, results that agree as well with the assump- tion that this is a fieniacetate as that it is a tetracet&te. There is a much wider margin in the percentages of acetyl residues in the three supposed possible formulas, as shown in the last column in the above table. Four C 2 H,0 groups would equal 32.34 per cent ; five would equal 37.48 per cent, a difference of 5. 14 per cent ; six would equal 41.89 per cent of the whole molecule, a difference of 4.41 per cent. Several attempts were made to determine the number of acetyl residues in the acetate by saponification by means of different alkalies, but without success, since the galleiin residue itself undergoes some decomposition when heated with alkalies giving volatile acid products that have not been further inves- tigated. It was found, however, that a modification of Wen- zel's method 1 could be used satisfactorily. The distinguishing points of this method are : First, saponification by means of strong sulphuric acid ; second, removal of free sulphuric acid by means of secondary sodium phosphate ; third, distillation of the liberated acetic acid in vacuo. The three important modifications of the Wenzel method, suggested by this work, are : First, the substitution of distillation in steam for distilla- tion in vacuo ; second, the abandonment of one of the con- densers ; third, the use of dimethylamidoazobenzene to detect and determine mineral acid in the distillate. The arrangement of the apparatus may be understood from the accompanying sketch : A is a 300 cc. distilling flask with delivery tube in- clined upwards ; B is a 100 cc. distilling flask with delivery tube inclined down so as to connect easily with C, an Allihn condenser. D is a 750 cc. Erlenmeyer flask, F a small sepa- ratory funnel, 5" a screw clamp. A weighed amount of material is placed in A and 3 or 4 cc. of sulphuric acid diluted with water, two of acid to one of water by volume, run in through the funnel F, screw clamp S being closed. An excess of tenth-normal sodium hydroxide is 1 Monatsh. Chem., 18, 659. 45 measured into flask D, the whole apparatus properly connected, and water started through the condenser. The flask A is then heated for one hour to the temperature of a boiling water-bath, when 3 cc. of distilled water are added and the heat continued for half an hour longer. There are now introduced 20 cc. of a solution made by dissolving 45 grams of secondary sodium phosphate, and 10 grams of glacial phosphoric acid in 100 cc. of water. Finally, add 50 cc. of boiled distilled water, heat with a burner, and pass in steam. liquid will soon begin to collect in flask B, which serves as a trap to prevent acid from being carried over mechanically. When flask B is one-fourth full of liquid, heat it with a burner, the heat being so regula- ted as to keep the liquid at a constant level. When the dis- tillate has collected in the receiver to the amount of 600 cc. , 4-6 stop the heat and transfer the distillate to a liter measuring flask. Dilute to the mark and mix thoroughly. Measure out 500 cc. of the solution and titrate with tenth-normal HC1, using phenolphthalein as indicator. In order to tell whether any mineral acid (H,SO s , etc.) has been carried over, the re- maining portion is titrated with tenth-normal HC1, using di- methylamidoazobenzene as indicator. This is sensitive to free mineral acids but not to dilute organic acids. Of course the amount of the difference between the two titrations, i. e. the amount of mineral acid, is to be deducted from the total amount of acid found. The process of distillation is not yet complete. One or two cc. of tenth-normal NaOH are again measured into flask D, and the heating continued till 300 cc. are collected. This is titrated directly with tenth-normal HC1, using phenolphthal- ein as indicator. The total amount of acid in this distillate is so small as to make the detection of mineral acid impossible — none was discovered in it in any determination made. To make sure that all volatile acid has been driven from flask A, a third distillate is collected which rarely showed more than 0.2 cc. tenth-normal NaOH neutralized. The net amount of tenth-normal NaOH neutralized multiplied by 0.0043 gives the weight of acetyl groups present. This method was used in all the acetyl determinations reported in this paper. In order to test the apparatus and chemicals, preliminary de- terminations were made using a compound of known composi- tion — acetanilide. Two analyses of acetanilide were made by the method given above with the following results : Weight of substance. Gram. N/10 NaOH. cc. Mineral acid. Net n/io NaOH. C3H3O. Per cent. O.5472 O.2260 40.25 16.85 O.O 40.25 O.O 16.85 3I-63 32.05 Calculated percentage of C 2 H„0 in acetanilide = 31.85. Galleiin acetate analyzed according to this method gave the following results : 47 Weight of substance. n/io NaOH. Mineral acid. Net n/io NaOH. C=H 3 0. Gram. cc. cc. Per cent. O.1946 i5-i o-3 14.8 32.70 O.2584 19.7 0.0 19.7 32.74 Calculated 1 for C^O^C^O),, 37. 4 8 per cent C 3 H a O. Calculated for C 20 H 8 O,(C 3 H 3 O) 4 , 32.34 per cent C,H 3 0. Calculated for C 20 H 9 O 7 (C 2 H 3 O) 3 , 26.32 per cent C 3 H 3 0. There can be no doubt from these results that this is a tetracetate of gallein. GAI.LEIN TETRABENZOATE. Buchka made this compound by use of benzoyl chloride. It was deemed advisable to try making it from benzoic anhy- dride. Gallei'n (10 grams) and benzoic anhydride (50 grams) were heated to 190° C. on a metal-bath for seven hours. The melt was then transferred to a 2 -liter round-bottomed flask and the excess of benzoic anhydride removed with steam. The black residue was filtered out and, after draining thoroughly, dissolved in acetone and boiled with boneblack. The acetone solution was concentrated and a little water added. In a short time yellow crystals appeared which, by repeated crystalliza- tion from acetone, alcohol, and acetic acid, became white and melted at 226 . Melting-point given by Buchka is 231 °. Efforts were made to determine the number of benzoyl resi- dues in this compound by the method used in analyzing the acetate, but failed on account of the difficulty of completely saponifying the benzoate by sulphuric acid. But that it con- tains four benzoyl groups can scarcely be doubted from its analogy to the tetracetate. HYDROGALLEIN TETRACETATE. The preparation of this compound from the so-called hydro- gallem, and its identity with gallein tetracetate, constitute the basis for Buchka' s formula for gallei'n. It will be remembered, 1 In all calculations in this paper the following atomic weights have been used: O = 16, H = 1.008, C = 12.01, N = 14.04, Ag = 107.92, 1 = 126.85. 4 3 however, that Herzig tried to make it from Buchka's hydro- gallein, but did not succeed. Gallein was dissolved in slight excess of potassium hydroxide solution, zinc dust added cautiously, and the temperature of the contents of the flask prevented from rising by a stream of cold water around the flask. When the blue color of the alka- line solution disappeared, becoming more or less brown, dilute sulphuric acid was added slowly and in such a way as to pre- vent an appreciable rise in temperature. The acid liquid was extracted with ether, the ether solution rapidly filtered to re- move suspended particles, and dried over calcium chloride in an atmosphere of carbon dioxide. The ether solution was then distilled in a stream of carbon dioxide. An oil with a slight red color remained and on standing solidified. Acetic anhydride in excess was added and the liquid boiled, on a metal- bath, for four hours, in a flask connected with a reflux condenser. A portion of the excess of acetic anhydride was distilled off and the residue poured into water. The acetate collected on the bottom of the vessel as an oil, which after- wards solidified. The yellow product was crystallized from benzene and gave a product melting at 2i3°-2i4°. It was re- crystallized from alcohol and acetic acid in succession till the melting-point was constant. The final product 'was colorless and melted at 21 6°. The difference between this compound and gallein acetate is seen in its conduct with sodium carbonate and concentrated sulphuric acid. This compound, when treated with a 5 per cent solution of sodium carbonate in the cold, dissolves, at once imparting to the solution at first a greenish-yellow fluorescence ; the color rapidly deepens, becoming later a decided purple. The acetate, when treated with a few drops of cold concentra- ted sulphuric acid, dissolves with red color. When this is di- luted with water it forms a green solution which, after stand- ing, deposits a green precipitate. This, filtered off and washed with little water, dissolves in ammonium hydroxide with green color and in aniline with blue color, both characteristic of coerulein. Sulphuric acid saponifies the so-called hydrogallein acetate and condenses one of the products (gallin) even in the 49 cold, to coerulin, which is easily oxidized in alkaline solu- tion to coerulei'n, forming a green solution with ammonium hydroxide. Determination of the acetyl groups in this compound gave the following results : Weight of Mineral Net n/io substance, n/io NaOH. acid. NaOH. C 2 H 3 0. Gram. cc. cc. Percent. O.2328 21.5 4.0 17.5 32.32 0.1690 15.7 2.75 12.95 32.95 Calculated for C, H 11 O,(C J H s O)„ 26.14 per cent C 5 H 3 0. Calculated for C, H 10 O 7 (C 9 H s O) 4 , 32.21 per cent C,H 3 0. Calculated for C M H,O t (C,H,0)„ 37.32 per cent C a H.O. Clearly the compound is a tetracetate, but it has none of the characteristics of gallein tetracetate but has all the properties of gallin tetracetate described below. GALUN TETRACKTATE. This compound was made by the method given by Buchka. Gallein dissolved in ammonium hydroxide was reduced to gal- lin by means of zinc dust, the solution being kept at the boil- ing-point. After boiling thus for two hours, the purple color having disappeared, dilute sulphuric acid was added little by little through the condenser- tube till in slight excess. The acid solution was filtered while hot and cooled in an atmos- phere of carbon dioxide. It was then shaken with ether to dissolve the gallin, the ether solution separated and filtered, dried over calcium chloride in an atmosphere of carbon dioxide, and finally the ether distilled off in a current of carbon diox- ide. The residue, gallin, appeared as a red oil which solidified on standing. This was converted into the acetate according to details given under hydrogallein acetate. The crude material was recrystallized from alcohol and acetic acid and gave a color- less product melting at 21 6° C. Its conduct with a 5 per cent solution of sodium carbonate and with concentrated sulphuric acid was identical with that of the so-called hydrogallein ace- tate detailed above. The difficulty of deciding the composition of hydrogallein (gallin) acetate by combustion analysis will appear from the 5o following table, which gives the percentages of carbon and hydrogen according to the different formulas : No. of c=h 3 o groups. C. H. Per cent. Per cent. 3 63.41 4.06 4 62.92 4.12 5 62.50 4.15 Buchka found C = 62.72 per cent ; H = 4.37 per cent. Determinations of the number of acetyl groups in gallin ace- tate resulted as follows : Weight of substance. Gram. n/io NaOH. cc. Mineral acid. cc. Net N/io NaGH. cc. CH3O. Per cent. 0. 2490 O.1728 20.3 14.4 1-5 1.2 18.8 13.2 32.46 32.84 Calculated for C^H.ACC^O),, 26.14 per cent C„H s O. Calculated for C a0 H 10 O 7 (C 5 H s O)„ 32.21 per cent C,H s O. Calculated for C, H„O,(C i H,O) b) 37.32 per cent C,H a O. In view of these results this compound is correctly called gallin tetracetate. The residue in flask A (see sketch), when treated with an excess of caustic potash gives a blue color, as would be ex- pected of gallin in presence of air. GAIiOI, ACETATE. Buchka claims to have obtained a lower reduction-product of gallei'n by the use of zinc dust and acid solution. Having the formula as given by him, C 9 H 4 CH< >0, I \C 6 H,(OH)/ CH a OH gallol would be expected to yield a penta.ceXa.te. According to Buchka' s method for making this substance, gallei'n (5 grams) was placed in a liter flask with 700 cc. of water, a little dilute sulphuric acid added, and some zinc dust. The solution was kept boiling for ten hours, adding from time to time small 5i amounts of dilute acid and zinc dust. The solution lost its deep red color and became straw-yellow. It was then filtered and allowed to cool. A small deposit of crystals appeared after awhile on the bottom of the vessel. The solution was shaken with e,ther and the ether dried and distilled off with the same precautions as noted under hydrogallein acetate. A residue was left as a red oil, to which were added the crystals which had separated and the whole acetylized. The acetate was re- crystallized from alcohol, yielding colorless crystals melting sharply at 216 . Buchka's gallol pentacetate melted at 230 . The product made as above described reacted with 5 per cent solution of sodium carbonate and with concentrated sulphuric acid in every respect as the acetates of hydrogallein and gallin and could not be distinguished from them in any way. A de- termination of the number of acetyl groups resulted thus : Weight of Mineral Net n/io substance. n/io NaOH. acid. NaOH. C 2 H 3 0. Gram. cc. cc. cc. Per cent. O.2237 20.5 3.46 17.04 32.75 Calculated for C 50 H 10 O,(C 2 H 3 O) 4 , 32.21 per cent C a H s O. A/>e«fecetate, according to Buchka's formula, would have 38.25 per cent C a H s O. This compound not only shows all the characteristics of gallin acetate but the results obtained on analysis prove the two compounds to be identical. An attempt was made to reduce gallein in acetic acid solu- tion to gallol by action of sodium amalgam, with negative re- sults. An effort was made to prepare gallol acetate directly from gallein acetate. Gallein (5 grams) is mixed with an equal weight of sodium acetate and 25 grams of acetic anhydride and boiled with a return-ccndenser for one hour. Zinc dust in small amounts is added from time to time, while the boiling is continued for three hours. A small quantity of the material is taken out, the acetate formed is crystallized once from alco- hol and found to melt at 21 5 , and to dissolve in dilute ammo- nia with purple color. The remainder of the mixture is boiled seven hours longer, zinc dust in small quantities being added from time to time. The product crystallized once from alco- 52 hoi melts at 214.5" an & dissolves in dilute ammonia with pur- ple color. Thus it would appear to be impossible to obtain gallol by reduction of gallei'n or its acetate by zinc dust in acid solution, even with prolonged heating. As a result of this work, however, a method for the prepara- tion of gallin acetate was found which is very much shorter and simpler than that described by Buchka, and gives a yield almost as great as that required by theory. To 5 grams of gallein add 5 grams of fused sodium acetate and 20 cc. of acetic anhydride. Boil with return-condenser for fifteen minutes, add 2 grams zinc dust and boil fifteen minutes longer. Allow the excess of zinc dust to settle to the bottom of the flask and pour off the solution carefully into a half liter of water. A small portion of the precipitate is granular; the bulk of it is an oil which soon solidifies, especially when agitated. Recrystallize from acetic acid, using boneblack to decolorize, afterwards from alcohol slightly diluted with water. SILVER SALT OF GALLIN TETRACETATE. That gallin tetracetate is an acid was shown not only by its action with dilute sodium carbonate solution but also by mak- ing its silver salt. Gallin acetate was dissolved in 95 per cent alcohol and an alcoholic solution of silver nitrate added in ex- cess. The mixture was heated to boiling, filtered, and left in the dark for several hours. There separated a white precipi- tate resembling in appearance silver chloride. It was prac- tically insoluble in alcohol when once precipitated, did not melt sharply, but decomposed when heated, and turned dark when exposed to sunlight. The so-called gallol acetate gave rise to the same compound under the same conditions — another proof of the identity of gallin and gallol acetates. On analysis the compound gave the following results for silver : Weight of Weight of substance. Ag. Ag. Gram. Gram. Per cent. I O.IO44 O.OI72 16.48 II 0.1072 0.0179 16.69 Calculated for a, H„O,(CH,,CO) 4 Ag ( 16.82 percent. Ag. 53 The analysis was made by placing the well-washed and dried product in a small porcelain crucible and burning out the or- ganic matter by careful ignition. The first analysis above was made from a product obtained from the so-called gallol acetate, the second from gallin acetate. GALLEIN METHYL ESTER. This compound was made by one of the usual methods for preparation of esters. Gallei'n(5 grams) , methyl alcohol (30CC) . , and concentrated sulphuric acid ( 10 cc. ) were mixed, forming a deep red solution . This was boiled with return-condenser for one and one-quarter hours when there separated a good quantity of crystals. The mass was then poured into 500 cc. water, fil- tered, the residue washed and dried, at first in the air, finally in an air-bath at 70°. To free the ester from any gallein which might have been present, the ester was dissolved in ether, in which gallein is insoluble, the solution filtered and concentrated. The ester came down in granular masses of dark red color, with bronzy luster. It dissolves in alcohol, acetone, ether, chloroform, acetic ether, with red color, but is almost insoluble in benzene. It dissolves at once in alkalies and alkaline carbonates with the colors characteristic of free gallein in these solvents— probably owing to saponification. It did not melt when heated to 280 C. A determination of the methoxy group by Zeisel's method in a sample dried at ioo° C. resulted as follows : Weight of ester. Weight of Agl. CH3O. Per cent. 0.1819 0.1011 7.34 Calculated for one methoxy group, 8.22 per cent. GALLEIN ETHYL ESTER. Gallein (5 grams), absolute alcohol (30 cc), concentrated sulphuric acid ( 20 cc. ) were heated with return-condenser for two and three-quarters hours. The gallein was entirely in solution at the beginning of the operation and, while still heat- ing, crystals in good quantity were deposited. The mixture was poured into 500 cc. water, filtered, and washed. The es- 54 ter appeared as beautiful crystals, red by transmitted light, green by reflected light. It dissolves in alcohol, acetone, ether, chloroform, and acetic ether with red color. Quite a number of efforts were made to crystallize it from different sol- vents, but the best method of purifying it was found to be solution in anhydrous ether, filtering to remove gallei'n, and concentration of the solution almost to dryness. The ester then appeared as granular masses. Analysis of sample dried at ioo° C. to constant weight gave, by Zeisel's method, the following results : Weight of ester. Weight of Agl. C 3 H s O. Gram. Gram. Per cent. O.0811 O.0518 12.25 Calculated for one ethoxy group, 11.48 per cent. The results obtained with the methyl and the ethyl esters of gallein indicate very, clearly that gallein is an acid and contains one carboxyl group. It is probable that what Baeyer consid- ered alcohol of crystallization was in reality due to the pres- ence of this ester. GALLEIN PHENYIXARBAMATE. By action of phenyl isocyanate upon galleiin, combination took place, giving rise to a new compound. Gallein dried at ioo° C, and a little more than the calculated amount of phenylisocyanate were heated quickly to the boiling-point of the mixture and kept at this temperature for ten minutes — agitating from time to time. On cooling, the residue appeared as a black tar. This was taken up in benzene, the solution fil- tered, boiled with boneblack, and allowed to stand. A small amount of a crystalline deposit appeared which was filtered out and found to consist largely of diphenyl urea. To the fil- trate, petroleum ether was added, when there appeared a bulky yellow precipitate. This was filtered out and dissolved in lit- tle chloroform. On standing, some white crystals appeared which did not saponify with caustic potash solution, and melted at 235 C, and was probably diphenyl urea. This was removed by filtration and the gallein phenylcarbamate precipi- tated from the chloroform solution by petroleum ether, filtered 55 out, and the process of dissolving in chloroform and precipita- ting with petroleum ether continued till no further signs of di- phenyl urea appeared. The gallem carbamate thus made is a solid, with light yellow color, does not melt sharply on account of decomposition, dis- solves in alcohol and acetone with green fluorescence. It is an acid, as is shown by the fact that it is soluble in cold dilute sodium carbonate solution with purple color, which is changed to red by acids. The red color of the acid solution is not changed to purple by ammonia solution as would be the case were gallem formed here. The carbamate dissolves also in cold dilute solution of ammonia with color, green by reflected light, red by transmitted light. In caustic alkali it dissolves (apparently with saponification) giving a deep-blue solution. The nitrogen in this compound was determined by the Kjel- dahl method, using Dafert's modification. A weighed amount of carbamate was placed in a 300 cc. digesting flask, 10 cc. of alcohol added to dissolve it, 30 cc. dilute sulphuric acid (1:5) added, and a small amount of zinc dust. The mixture was heated with return-condenser till the color disappeared (nearly) . The flask was then disconnected from the condenser, and the alcohol and water evaporated till the organic matter showed first signs of charring. The remainder of the process was that of the ordinary Kjeldahl method. The following results were obtained : Weight of carbamate. N/IO NH3. Amount of N. N. Gram. CC. Gram. Per cent. I O.4274 18.0 O.O25272 5-9* 2 O.271I 11. 4 O.O160056 5-90 3 O.1966 7.8 O.OIO9512 5-57 4 O.2116 8.3 0.01 16532 5-5i 5 O.2204 8.7 O.OI22148 5-54 6 O.2372 9-35 O.OI31274 5-53 Calculated for C„H t O,(OCNHC.H,)„ 5.83 per cent N. Calculated for C„H.O,(OCNHC.H § )„ 6.68 per cent N. Calculated for ChACOCNHC.H.), (the formula for this compound on the basis of Buchka's formula for gallem), 4.67 per cent N. 56 The analyses above recorded were made from two samples which were prepared at different times — Nos. i and 2 from one, Nos. 3, 4, 5, and 6 from the other. The results can lead to but one conclusion, that there are three phenyl isocyanate residues introduced and that the compound is gallein tri- phenylcarbamate. GALLEIN BASIC LEAD SALT. A sodium carbonate solution was boiled with an excess of gallein till there was no further evolution of gas. This solu- tion was then filtered from the excess of gallein and an excess of lead acetate solution added. The precipitate was filtered out and washed till the filtrate no longer gave a test for lead, and dried at 120 C. The lead compound was a dark blue powder. On analysis for lead the following results were found : Pb. Per cent. 05-4 65.O Calculated for C ao H 8 0,(PbOH)„ 65.92 per cent Pb. GALLEIN TRIMETHYL ETHER. Five grams of gallein were dissolved in 200 cc. of methyl alcohol, 4 grams of caustic potash in 50 cc. methyl alcohol, the two solutions united, and a portion (100 cc.) of the solvent distilled off. To the residue was added methyl iodide (15 grams), and the whole heated to boiling on a water-bath for twenty to thirty hours. At intervals of about eight hours, 1 or 2 grams of solid caustic potash were added, heated for one hour, and about 5 grams methyl iodide added. The solution, at first blue, quickly becomes red. After sufficient heating the methyl alcohol and excess of methyl iodide were distilled off on a water-bath, the residue treated with 500 cc. of a 1 per cent solution of potassium hydroxide to dissolve any partially methylated products, and filtered. The residue will be con- sidered in connection with the tetramethyl ethers. The alkaline filtrate was acidified with a little concentrated hydrochloric acid. A yellow precipitate was thrown down Weight of substance. Gram. Weight of PbS0 4 . Gram. 0.9343 O.2014 O.8960 O.1880 57 which was filtered out and washed. The dried residue was dissolved in acetone, boiled with boneblack, concentrated, a little water added, and allowed to stand. After a few hours there was a good yield of crystals which had a golden yellow color. These were then dissolved in alcohol, boiled with bone- black, concentrated, and allowed to crystallize — little water be- ing added. Additional crops of crystals were obtained by further dilution of the mother-liquors. By repetition of this treatment crystals were finally obtained which were colorless and melted at 229 . A determination of methoxy groups was made by Zeisel's method. Apparatus and chemicals were first tested by ma- king a determination of the methoxy group in anisic acid, with the following result : Weight of Weight of anisic acid. Agl. CH 3 0. Gram. Gram. Per cent. O.1905 0.2910 20.18 Calculated for C 6 H 4 (OCH s )COOH, 20.39 per cent CH.O. The following results were obtained with the galleiin ether dried at ioo° C. : Weight of ether. Weight of Agl. CH 3 0. Gram. Gram. Per cent. O.1866 O.3100 21.96 O.2376 O.3966 22.o6 Calculated for C 30 H 8 O 6 (OCH 3 ) 3 (according to Buchka's for- mula), 1 13.95 per cent CH.O. Calculated for C.H.O/OCH,)., 22.91 per cent CH s O. Calculated for C, H 8 O 5 (OCH s ) 4 , 29.55 per cent CH,0. The substance is hence a trimethyl ether. Galleiin trimethyl ether dissolves in alcohol and acetone with a delicate pink tint ; in benzene its solution is colorless. It re- sembles phenolphthalein quite closely. In dilute sodium car- bonate solution, in dilute ammonia, as well as in caustic alka- lies, it dissolves at once, even in the cold, with deep-red color. 1 The solubility of this ether in alkaline solutions is a further argu- ment against the correctness of Buchka's formula for gallein. 58 From such solutions, by the addition of acids, it is precipitated again as the colorless trimethyl ether. This gallein trimethyl ether was also made by saponification of the colored tetramethyl ether as stated below. TRIMETHYL GALLEIN MONOACETATE. Gallein trimethyl ether was mixed with an equal weight of fused sodium acetate and five times its weight of acetic anhy- dride, and the mixture kept at the boiling-point for one hour. The excess of acetic anhydride was distilled off and the residue poured into a large bulk of water. The precipitate was at first oily, but soon solidified. This was filtered off t and washed. The residue thus obtained was recrystallized from alcohol. The crystals were needle-shaped and colorless, showed double refraction, and were either monoclinic or triclinic. Melting- point 197 . They dissolve in alcohol, acetone, and chloro- form, are insoluble in cold aqueous caustic potash, but after prolonged boiling they dissolve, giving a red color to the solu- tion (apparently after saponification of the acetyl group). Saponification takes place more readily in alcoholic caustic pot- ash. The product dried at ioo° C. and analyzed for methoxy groups gave the following results : Weight of substance. Gram. Weight of Agl. Gram- CH30. Per cent. 0.2004 O.169O 0.3028 O.2570 19.96 20.08 Calculated for CH 3 0. C, H 8 O ,(OCH 3 ) s (C,H 3 0), 20 .76 per cent The results of these analyses show that this derivative of gallein contains three methoxy groups and one acetyl group. Its properties also harmonize with this view. GALLEIN TETRAMETHYL ETHER— COLORED. The residue insoluble in alkalies, obtained in making tri- methyl gallein, was dissolved in 50 per cent alcohol, boiled with little boneblack to remove gummy material, concentrated, 59 and allowed to crystallize. Two distinct sets of crystals were deposited. One set consisted of pyramidal crystals belonging to the monoclinic system. They were red, the larger ones ex- hibiting a bronzy luster. By far the larger proportion of the crystals were of this sort. The others were much lighter in color and were needle-shaped. These last will be discussed under the head of gallein tetramethyl ether — colorless. The dark red crystals mentioned above melt sharply at 199 , dissolve in alcohol, acetone, chloroform, ether, and acetic ether, with red color. The ether is insoluble in cold alkalies, in hot alkalies it forms a red solution — after partial saponification. To an alcoholic solution of the ether a solution of sodium car- bonate is added, the solution evaporated to dryness on a steam- bath, the residue dissolved in water and filtered, the filtrate acidified, there appears a precipitate with a yellow color. This is filtered out, recrystallized from alcohol, and found to be col- orless and to melt at 229 . The product shows also all the other characteristics of gallein trimethyl ether already de- scribed. The number of methoxy groups in the colored ether was determined with the following results : Weight of Weight of substance. Agl. CH3O. Gram- Gram. Per cent. O.177O O.3828 28.56 O.2140 O.4656 28.74 Calculated for C, H,O,(OCHs) 4 , 29.55 per cent CHsO. Calculated for C 30 H„O 4 (OCH3) 3 , 22.91 per cent CHsO. This ether is unquestionably a gallein tetramethyl ether, and since one of the methyl groups may be split off by the action of sodium carbonate solution, leaving a product with three methyl groups (see gallein trimethyl ether above), one of the methyl groups must replace a hydrogen atom of a carboxyl group in the form of an ester. GALLEIN TETRAMETHYL ETHER — COLORLESS. The light-colored needle-shaped crystals referred to above were separated from the deep red ones, dissolved in acetic acid, boiled with bone-black, and filtered. The slightly red 6o solution was diluted with water till a precipitate appeared, then heated till all dissolved, and allowed to cool slowly. The crystals which separated still had some color. To remove this they were dissolved in alcohol, a little potassium hydroxide was added, and the solution heated to boiling. This saponified the colored ether, forming the potassium salt of the trimethyl ether which is easily soluble in water. The alcoholic solution diluted with water yields colorless crystals of the tetramethyl ether. These are needle-shaped, show double refraction, and are either monoclinic or triclinic. The same compound was obtained by heating gallein tri- methyl ether with excess of methyl iodide in a sealed tube to 125 C. for five hours. The colorless ether melts at 195 , is soluble in alcohol, ace- tone, and chloroform, but to a less degree than the colored ether. It is insoluble in aqueous caustic potash (laboratory reagent) even on boiling. Analysis of a sample dried at ioo° C. gave the following re- sults : Weight of substance. Gram. Weight of Agl. Gram. CH3O. Per cent. O.0804 O.I751 28.77 Calculated for C J0 H 6 O,(OCH 3 ) 4 , 29.55 per cent CH s O. GALLEIN TETRAETHYL ETHER — COLORED. Gallein (5 grams) was dissolved in 200 cc. methyl alcohol, 1 caustic potash (4 grams) in 50 cc. of the same solvent and the two united. A portion (100 cc.) of the solvent was distilled off, and to the residue was added an excess of ethyl iodide. The color changes from blue to red. The mixture kept at the boiling-point for five to ten hours. The methyl alcohol and the excess of ethyl iodide were then distilled off on a water- bath. The residue was treated with 500 cc. of a 1 per cent solution of potassium hydroxide to dissolve any partially ethylated products, and filtered. The filtrate was used in the preparation of the triethyl ether as described below. 1 Methyl alcohol was used in this case because it is a better solvent than ethyl alcohol for the potassium salt of gallein. 6i The residue insoluble in alkali was deep red, at first a viscous mass, which slowly hardened. This was dissolved in methyl alcohol, boiled two or three times with boneblack to remove gummy impurities, concentrated, and allowed to stand several days. There appeared first needle-shaped crystals which were filtered out and are described below as the colorless tetraethyl ether. After these, in greater amount, came tabu- lar, monoclinic crystals of a deep red color, very much like those of the colored tetramethyl ether, and melted at 155 . This ether dissolves in methyl alcohol, ethyl alcohol, acetone, chloroform, acetic ether, and acetic acid. In cold alkalies it is insoluble. If some of the ether be dissolved in alcohol, a little sodium carbonate or potassium hydroxide be added, and the mixture heated to dryness on a water-bath, it undergoes partial saponification. This was shown by dissolving the residue in water, filtering, acidifying the filtrate, when a yellow precipitate was thrown down. This product, on recrystallization, proved to be colorless, and showed the characteristics of the triether described below. The following is the result of a determination of ethoxy groups : Weight of Weight of substance. Agl. C2H5O. Gram. Gram. Per cent. O.2015 O.3788 3 6 - 11 0.1 199 0.2356 37-70 Calculated for C.HACOCJHJ,, 37.82 per cent C,H 6 0. Calculated for C ao H 9 4 (OC 2 HJ 3 , 30.59 per cent C,H b O. From the above analyses, as well as from the method of its preparation and its properties and reactions, this ether is the analogue of the colored galleiin tetramethyl ether. GALLEIN TETRAETHYL ETHER— COLORLESS. The needle-shaped crystals referred to in connection with the colored tetraethyl ether after recrystallization from alcohol, using boneblack to decolorize, were obtained almost colorless, and found to melt at 144 . They dissolve in alcohol, acetone, acetic acid, and chloroform, but less easily than the colored variety. This ether is insoluble in aqueous alkalies, even on boiling. Analysis gave the following result for ethoxy groups : 62 Weight of ether. Weight of Agl. C 2 H s O. Gram. Gram. Per cent. O.0682 0. 13OI 36.61 Calculated for C, H B O,(OC 2 H 6 ) 1 , 37.82 per cent C 2 H 6 0. GAWvEIN TRIETHYI, ETHER. The alkaline nitrate obtained in the preparation of gallei'n tetraethyl ether was acidified with concentrated hydrochloric acid. A heavy yellow precipitate was thrown down. This was filtered out, washed well, and recrystallized from alcohol, using boneblack to decolorize. Colorless crystals were obtained which, on analysis, gave the following result : Weight of Weight of substance. Agl. C2H5O. Gram. Gram. Per cent O.IO57 O.1765 32.04 Calculated for C, H a O 4 (OC J HJ 1) 30.59 per cent C 2 H 6 0. Calculated for C !0 H e O s (OC 2 H B ) 4 , 37,82 per cent C 3 H 6 0. Gallein triethyl ether dissolves easily in most of the organic solvents. In cold dilute sodium carbonate solution, in dilute ammonia, as well as in the caustic alkalies it dissolves at once with deep red color, just as phenolphthalein does. On acidi- fying these solutions the colorless triethyl ether is reprecipitated. . GAIJJN PENTAMETHYI, ETHER. Gallein was reduced by zinc dust in alkaline solution, and the resulting gallin extracted with ether in the usual way. After complete removal of the ether by distillation in carbon dioxide, the flask containing the residue was connected with a return condenser and also a hydrogen generator. After re- moving as completely as possible all air from the apparatus by means of hydrogen, a slight excess of potassium hydroxide dissolved in 95 per cent alcohol was added through the con- denser tube. The solution turned light blue— due to slight oxidation. After heating five minutes an excess of methyl iodide was added and the mixture heated to boiling on a water-bath for ten hours. The alcohol and excess of methyl iodide were distilled to dryness and the residue treated with 63 500 cc. of a i per cent solution of caustic potash. The residue was filtered out, washed, and recrystallized from alcohol, using boneblack to decolorize. The crystals at first were more or less yellow, recrystallizing from alcohol and acetic acid in succes- sion they became almost white and melted at 127 . Analyses gave the following results : Weight of Weight of substance. Agl. CH 3 0. Gram. Gram. Per cent. O.1596 O.4167 34.49 O.1877 O.4887 34.39 Calculated for C^H^^OCH,),, 35.50 per cent CH s O. Calculated for C 30 H 10 O 3 (OCH s ) 4 , 29.38 per cent CH s O. The compound is therefore the fientamethyl ether of gallin. The ether dissolves readily in alcohol, acetone, and acetic acid. It is insoluble in cold alkalies, but may be saponified by the method already outlined in connection with the gallein col- ored tetra ethers, giving a product that is colorless and dis- solves in alkalies without color. 11. coerulein. The coerulein used for the following investigation was also furnished by the Badische Anilin und Soda Fabrik, and was of exceptional purity. It was found that if coerulein be pulverized and suspended in water, it could be dissolved by saturating the water with sulphur dioxide. This solution when boiled for some time to remove the gas, was decomposed and coerulein precipitated. When filtered out and dried, it had a beautiful bronzy lustre and seemed to be crystallized. One marked difference between gallein and coerulein is that while the former dissolves easily in solution of sodium carbon- ate, the latter is practically insoluble in such a solution, im- parting to it only a faint yellow color even after boiling for some time. In aniline, pyridine, and quinoline, coerulein dissolves, more readily if heated. In aniline it forms a deep blue solution, the color remaining unchanged in excess of acetic acid. The solu- tion in pyridine is colored greenish blue, which turns purple 6 4 in excess of acetic acid. On boiling with quinoline the color is purple, which becomes greenish blue on cooling, and green when diluted with more quinoline. COERULEIN TRIACETATE. Coerulein acetate was made according to the method given by Buchka — by boiling coerulein with sodium acetate and ex- cess of acetic anhydride. After completion of the reaction a part of the acetic anhydride was distilled off and the residue poured into water. The red precipitate was filtered out and washed. Unsuccessful attempts were made to crystallize the acetate from glacial acetic acid, alcohol, acetone, chloroform, benzene, and acetic ether, in each of which it dissolves with red color. It was finally purified by dissolving it in acetone, filtering, evaporating to small bulk and precipitating the ace- tate by adding to the acetone solution a saturated solution of sodium chloride. After washing till the filtrate gave no reac- tion for chlorides and drying coerulein acetate appeared as a reddish brown powder. That it is coeruiei'n frzacetate cannot be doubted from the following determinations of the number of acetyl groups present : Weight of acetate. n/io Mineral Net n/io Per cent Gram. NaOH. acid. NaOH. C„H,0. cc. cc. cc. 0.2I53 16.79 2.55 14.24 28.43 o.i753 i3-5o 2.05 11.45 28.08 Calculated for C S0 H 7 O 6 (C a H 3 O)3, 27.33 per cent C^HsO. Calculated for C^HeCMC.H.O),, 33.47 per cent C.HaO. The difficulty of deciding by combustion analysis how many acetyl groups this compound contains, will be seen from the following table : No. of acetyl groups. Per cent C Per cent H. 3 66.10 3.39 4 65.37 3-50 Buchka found C = 65.7 and 66.71, and H = 3.75, but never- theless calls the compound the triacetate. COERUUN PENTACETATE. Buchka' s directions were followed in the preparation of this 65 compound, using 5 grams of coeruleiin, 25 grams of acetic an- hydride, and a little zinc dust. After heating for twenty min- utes the mass began to solidify and had a dark green color. The contents of the flask were transferred to a porcelain evap- orator and heated on the water-bath till nearly all the acetic anhydride disappeared. The residue was then put into 1200 cc. water to remove all zinc salts, etc. , filtered and washed. The mass was dark green. Recrystallized from glacial acetic acid, the acetate appears in greenish yellow needles which show parallel extinction and form a felted mass. Solutions of coeru- lin acetate in acetic acid, alcohol, acetone, and chloroform have a greenish yellow fluorescence. Buchka gives the melting-point of this compound as 256 C This could not be verified — the compound decomposed when heated and did not melt at all. Determinations of the number of acetyl groups in this com- pound resulted as follows : Weight of acetate. N/io Mineral NetN/lo Percent Gram. NaOH. acid. NaOH. C a H 3 0. cc. cc. cc. 0.1837 J 9-9 3-°5 . l6 - 8 5 39-44 0.1991 20.55 2.25 18.3 39.52 Calculated for C J0 H 8 O 6 (C 2 H3O) 4 , 33-34 per cent C,H 3 0. Calculated for C 20 H,O 6 (C 2 H 3 O) 6> 38.54 per cent C 2 H 3 0. The product is therefore a /^/acetate. Calculated percentages of carbon and hydrogen are as fol- lows : No. of C,H s O groups. Per cent C Per cent H. 4 65.11 3-87 5 64.51 3-94 Buchka found C = 65.31 and 64.64, H = 4.60 per cent, and on the bases of these analyses pronounced the compound a tetracetate. In the determination of the number of acetyl groups, however, there is a much greater difference in the per- centages for four and for five groups, and according to results given above for these, it cannot be doubted that the compound is in reality coerulin fientaceta.t&. COERUEEIN MONOMETHYL ETHERS. Coerulein (7 grams) and potassium hydroxide (3.5 grams) 66 were dissolved in 150 cc. methyl alcohol and digested for one hour at the boiling-point of the mixture. There was then added an excess of methyl iodide and the whole boiled on a water-bath for twenty hours, adding at intervals of eight hours 2 grams of potassium hydroxide and boiling to make sure of solution, then slight excess of methyl iodide. The excess of methyl iodide and the methyl alcohol were then removed by distillation on water-bath, the residue treated with 500 cc. of a 1 per cent solution of potassium hydroxide, and filtered. The insoluble residue will be described below. The alkaline filtrate was acidified and a heavy dark precipi- tate was thrown down and filtered out. The precipitate was dissolved in boiling alcohol, filtered, and on standing there were deposited crystals of a dark color, with bronzy luster. The compound decomposes when heated and does not melt. It dissolves in alcohol, acetone, pyridine, aniline, and potas- sium hydroxide solution, imparting an olive-brown color to the solution in each case. A determination of methoxy groups gave the following re- sult : Weight of ether. Weight of Agl. CH s O. Gram. Gram. Per cent. O.2365 O.I703 9.5I Calculated for C 30 H„O s (OCH 3 ), 8.61 per cent CH s O. Calculated for C a0 H 8 O 1 (OCH 3 ) a) 16.57 per cent CH s O. The compound is therefore coerulein monomethyl ether. The mother-liquor from which this ether was obtained was allowed to evaporate spontaneously, when another crop of crys- tals appeared, This compound, when dried and analyzed for methoxy groups, gave the following result : :iglit of substance. Gram. Weight of Agl. Gram. CH,0. Per cent. O.I2I6 0.0755 8.20 Calculated for C 10 H,O,(OCH,), 8.61 per cent CH a O. It is clearly a coerulein monomethyl ether but has different properties from those of the coerulein ether described above. It dissolves in acetone and pyridine with a purple color, in 6 7 alcohol and aniline with a greenish-blue color, in potassiun hydroxide solution with a light-green color. It is uniform^ more soluble than the isomeric compound already described. COERULEIN METHYL ETHER — ALKALI INSOLUBLE. The residue insoluble in alkali (see above under coeruleii monomethyl ether) was dissolved in alcohol, the solution fil tered, and concentrated. After standing, a tarry substanc< was deposited. This was filtered out and the nitrate, whicl now had a beautiful green fluorescence, allowed to stand fo: some days. There appeared needle-shaped crystals which wer< red by transmitted light, green by reflected light. These dis solve with some difficulty in alcohol, giving it a light purpl( color, easily in concentrated sulphuric acid with red color, bu are insoluble in alkalies even on boiling. The quantity o crystals was so small as to make an analysis of the compounc impossible, but it is in all probability coerulein trimethyl ether COERULEIN MONOETHYL ETHER. This compound was made in the same way as the mono methyl ether, using ethyl iodide and boiling only nine hours It was likewise crystallized from alcohol and had all the char acteristics of the corresponding methyl derivative. It wai dried at ioo° C. and the ethoxy group determined which re suited as follows : Weight of ether. Weight of Agl. C,H,0. Gram. Gram. Per cent, 0.1313 O.0808 11.82 Calculated for C, H,O.(OC,H,), 12.04 per cent C 2 H 6 0. Hence the compound is coerulein monoethyl ether. THEORETICAL. I. GALLEIN. The formula for galleiin, as given by Buchka, is .C.H.OH. C.H.C^ >cOo. I \C,H g OH/ CO— O 68 It is necessary to assume with this formula that oxidation takes place in the process of making, giving rise to a quinone group. This idea was first suggested to him by the supposed discovery of the so-called hydrogallein, which has been re- garded as a reduction-product intermediate between galleiin and gallin. Hydrogallein, according to Buchka, differs essentially from both galleiin and gallin, but gives the same product when acetylized as that obtained from gallein by acetylization. In order to account for this, he gave gallein the above formula and to hydrogallein the following formula : /C,H,(OH) lX C,H t C< >0. | | \c.H,(OH)/ CO— O The necessity for providing a formula for hydrogallein was the single consideration that led him to assign to gallein the above formula. Several attempts were made during the progress of this work to prepare hydrogallein and its acetate. The result was always the same — the hydrogallein acetate invariably proved to be identical in every respect with gallin acetate. Attempts were made also to make gallol acetate by Buchka' s method — reduction of gallein in acid solution by zinc dust, and acetyliza- tion of the product thus obtained. Here again the so-called gallol acetate was identical with gallin acetate. The identity of these three acetates is clearly shown not only by analysis but also by identity of melting-point and conduct with sodium car- bonate solution and concentrated sulphuric acid. So that no matter whether gallein be reduced by zinc dust in acid or alkaline solutions, hot or cold, the product is in all cases gal- lin, which, with acetic anhydride, yields gallin acetate. The existence of hydrogallein is, therefore, extremely improbable, and it is doubtful if gallol or its acetate has been isolated. The conclusions here stated with reference to hydrogallein is in en- tire accord with Herzig's work, to which reference has already been made. That gallein contains a carboxyl group is proved by the fact 69 that it forms esters in the ordinary way. Methyl and ethyl esters have both been prepared. Further, galleiin forms a stable ammonium salt, ' which suffers no loss of ammonia after pro- longed heating, in this respect showing a marked difference from phenolphthalein. Gallein also decomposes sodium, ba- rium, and calcium carbonates, driving out carbon dioxide and forming soluble compounds with the metals. The presence of the carboxyl group may explain, also, why gallein is converted more easily than other phthaleins into its corresponding phthalidein, since there is no necessity here for breaking the lactone ring. This will be referred to again in the theoretical discussion of coerulein. The formation of the colored tetramethyl and tetraethyl ethers easily saponified to the colorless triethers with alkalies also proves the presence of a carboxyl group in gallein. The existence of the blue basic lead salt and of the colored gallein triphenyl carbamate, which has acid properties, indicates the same thing. That gallein contains only three phenol h} r droxyl groups is shown by the introduction into its molecule of three phenyl isocyanate groups. It has been established, notably by Gold- schmidt, 2 that phenyl isocyanate does not react with quinone oxygen, but easily with hydroxyl hydrogen, except such as are present in carboxyl groups. The gallein triphenyl carbam- ate has a color and dissolves in cold solution of sodium carbonate with color. Hence it must contain a carboxyl group, but cannot be a derivative of gallin (formed by reduction). Since therefore Gallein, in free condition, contains one car- boxyl and three phenol hydroxyl groups, its structure must be : i Buchka : Ann. Chem. (Iyiebig), 209, 263. " Ber. d. chem. Ges., 22, 3105. 7 o OH O OH HO- =0 C 8 H 5 (OH) C.H.c/ | ^C„H -OH COOH % O -OH/ or —COOH The quinoid structure would explain the deep red color of galleiin and its solutions. According to this formula, the structure of a highly colored methyl or ethyl ester would be : OH O OH HO = o — COOR And for the colored triphenyl carbamate having acid proper- ties : 7i C,H B HNOCO r OCONHC 6 H 6 OCONHC.H, O =0 -C00H This formula for galleiin explains also the formation and properties of its colored tetramethyl and tetraethyl ethers. These would be represented thus : OR O RO OR = — COOR A compound with such a structure one would expect to be colored and to lose one alkyl group by saponification. As a matter of fact these ethers are colored (deep red) , and with so- dium carbonate or hydroxide saponify, leaving a trialkyl ether. The deep blue basic lead salt probably has the following for- mula : 72 OPbOH O HOPbO OPbOH O '\ COOPbOH and the blue potassium salt, which is precipitated when a satu- rated alcoholic solution of gallein and a saturated alcoholic so- lution of caustic potash are mixed, probably has a similar structure, the metal potassium replacing the PbOH group. There are compounds, however, which seem to be derived from a tautomeric form of gallein. This would have the lac- toid structure and be represented thus : OH O OH HO OH To this class belong the trimethyl and triethyl ethers which when uncombined, are colorless, but dissolve in sodium car- bonate with red color. The relation between the free ethers and their alkali salts appears to be the same as that between 73 phenolphthalein and its salts. This relation is shown in the following formulas : Galleiin trialkyl ether. (Colorless.) OR RO =0 COONa Sodium salt of gallein tri-alkyl ether. (Red.) That Formula I above correctly represents the structure of gallein trialkyl ethers is indicated further by the fact that it can be readily acetylized as well as transformed into the color- less tetralkyl ether. The products thus obtained are also colorless. Since the labile hydrogen by this process is replaced, 74 the trialkyl galleiin acetate is insoluble in alkalies except on prolonged boiling, which results in saponification of the acetyl group. These characteristics accord with the following struc- ture for the trialkyl gallei'n acetate : OR O OR RO OCOCH, The conception of gallein as a tautomeric substance finds sup- port also in the existence of two classes of tetralkyl ethers — one, the quinoid (colored) , already referred to ; the other, the lactoid (colorless) , which would have this structure : OR O OR RO OR With such a structure one would expect them to be colorless and unaffected by alkalies. Both of these characteristics they were found to possess. 75 The acetate and the benzoate of gallein would properly be derived from its lactoid modification, since they are color- less and (at least in the acetate) are known to contain four acid residues, getting Ac represent the acid residue, they would have the structure : OAc O OAc AcO OAc \ O . Nco Gallein anilide, a colorless compound, made by Albert, 1 should be a derivative of the lactoid modification also and have the structure : OH O OH HO OH The colorless methyl ether which he made from the anilide and called a rfnnethyl ether is very probably a tetramethyl 1 Ber. d. chem. Ges., 27i 2794- 7 6 ether. By combustion analysis he found carbon = 72.72 per cent, hydrogen = 4.58 per cent. Calculated for Per cent C. Per cent H C, 9 H I9 0,N C,.H„0 1 N 72.25 72.70 4.08 5.08 From the above it is seen that this compound is more probably a tetraether than a Aether. Another of the lactoid derivatives of gallein is the tetra- phenylsulphonate described by Georgescu. 1 He gives it the diketone formula which has been abandoned for all phthaleins. It more probably has the structure : OSO a C 6 H, C 6 H 6 O a SO OSO.C.H. Gallein, then, is tautomeric, yielding two classes of deriva- tives, those with color having the quinoid structure, and color- less derivatives with the lactoid structure. Free gallein has the quinoid structure. The structure of gallin is established in a measure by its re- lation to gallein, from which it is derived by addition of two atoms of hydrogen. This probably breaks the quinoid ar- rangement with simultaneous loss of color, and the product is doubtless properly represented thus : * Bulletinul Societ&tii De Sciinte, Fizice, 1, 215. 77 — COOH In support of this view of its structure is its easy transfor- mation into the colorless tetracetate, a compound with acid properties. This characteristic it shows not only in dissolving readily in sodium carbonate solution, but also in forming a silver salt. The structure of gallin acetate and its silver salt is shown in the following formulas : AcO OAc Gallin acetate. 78 OAc O OAc AcO OAc — COOAg Silver salt of gallin acetate. A further proof of the correctness of this view of the struc- ture of gallin is the formation of a fientamethyl derivative by action of methyl iodide upon an alkaline solution of gallin. It probably has the following structure : CH a O/ OCH. — COOCHs As was to be expected, this ether, though insoluble in cold alkalies, may be saponified by heating in alkali solutions, forming a product easily soluble in sodium carbonate without color. 79 II. COERTJXEIN. Coerulem is the phthalidein of galleiin and sustains to it the same relation as that between phenolphthalein and phenol- phthalidein. Baeyer, in his work on phenolphthalidein, found that when it was distilled with zinc dust, it yields phenylanthracene. Buchka obtained the same product by dis- tilling coeruleiin with zinc dust. Since coerulein is derived from gallein, probably by the loss of one molecule of water, and since, according to Buchka' s work, it contains the phenyl- anthracene group, it would seem that the hydroxyl group of the phthalic acid residue in gallein unites with one of the hy- drogen atoms of one of the pyrogallol residues to form water, the readjustment of affinities taking place as represented in the following formulas : OH O OH HO or 8o That the carboxyl group of gallei'n takes part in the reac- tion is indicated by the fact that, while galleiin is easily soluble in sodium carbonate solution, coerulem is practically insoluble in such a solution. The presence of a carboxyl group in gal- leiin explains the easy formation of coerulem, compared with other phthalidems. This formula is also in accord with the fact that when all the oxygen atoms are replaced in coerulem by hydrogen by means of zinc dust, phenylanthracene is formed. It easily explains how coerulein forms a triacetate easily re- duced with zinc dust in acetic acid solution. Such a deriva- tive, would be represented thus : CHsOCO O OCOCHs CH3OCO. =0 or Si CH 3 OCO O OCOCH3 0=^ v/ OCOCH CO With the formula given above for coerulein there is a possi- bility of having three varieties of monoalkyl ethers. Two monomethyl ethers are described in the experimental part of this paper. It is impossible at present to determine the exact position of the methyl group in these ethers. They are readily soluble in alkalies — a characteristic they could not show if only one hydroxyl group were present in coerulein as repre- sented in Buchka's formula for this compound. Again the fact that there are two monoethers requires that more than one hydroxyl group be present in the molecule. A methyl ether was made, which was entirely insoluble in alkalies, even boiling, but in such small amounts as to make analysis impossible. This was probably coerulein trimethyl ether. It would have the structure : OCH„ O OCEL CH3O ' \c // // CO =0 82 Coerulein thus appears as a derivative of anthragallol, CO OH \OH OH CO which, like coerulein, is soluble in alkalies with a green color. This will explain why coerulein resembles alizarin and anthra- gallol in its property of forming insoluble lakes with chromium, iron, and aluminium mordants. The name, alizarin green, by which coerulein is known, recalls this fact, which has long been known to the dyer. The above formula for coerulein also serves to recall the aurin group of dyestuffs, to which it shows certain resemblances. Coerulin is derived from coerulein by addition of two hydro- gen atoms. This would probably result first in breaking the quinoid group, thus : OH O HO' OH ^=0 + 2H 83 The reduced product would be expected to go over to the more stable form and become : OH O OH HO OH COH Coerulin is formed also from gallin by the action of concen- trated sulphuric acid in the cold. The transformation may be represented thus : 84 H,0 = or OH O OH HO OH COH 85 In favor of this view of the structure of coerulin is the fact that it forms a />e»/acetate with acetic anhydride. The more important facts brought out in this paper may be found in the following summary : SUMMARY. i. The only reduction product of gallem by zinc dust, whether in acid or alkaline solution, is gallin. 2. Gallem forms monoesters, showing that it contains one carboxyl group. 3. Three phenyl isocyanate groups may be introduced into a molecule of gallem, proving that it contains three phenol hydroxy 1 groups. 4. Two classes of tetralkyl ethers, colored and colorless, are derivable from gallem, indicating that it reacts in tautomeric modifications. 5. Colorless trialkyl derivatives of galleinare described. 6. A colorless trimethyl gallein acetate is described. 7. Gallin pentamethyl ether furnishes evidence as to the structure of gallin. 8. A silver salt of gallin acetate also furnishes evidence as to the structure of gallin. 9. Coerulein is shown to form a triacetate — by determination of its acetyl groups. 10. Coerulein forms two monomethyl ethers, and a methyl ether insoluble in alkalies. likewise a monoethyl ether. 11. Coerulin forms a pentacetate, as shown by a determina- tion of its acetyl groups.