(2) A violet color was obtained when a solution of the acid was treated with a drop of sodium carbonate solution and then with a drop of ferrous sulphate.

(3) It reduced ammoniacal silver nitrate.

(4) It did not reduce Fehling solution.

The filtrate supposed to contain phloroglucinol was treated with hydrogen sulphide to remove lead, filtered, and shaken with ether. The residue left on evaporating the ether was taken up in water. This solution gave the following reactions characteristic for phloroglucinol:

(1) It reduced both silver nitrate and Fehling solution.

(2) It colored pine wood moistened with hydrochloric acid red.

(3) It gave a red color with vanillin and hydrochloric acid, and

(4) A deeper red color with oil of cloves and hydrochloric acid, becoming purple on standing.

(5) It gave a violet color with ferric chloride.

The substance is then, without doubt, fisetin. The formula[26] of fisetin is supposed to be C_{15}H_{10}O_{6}.

RHAMNOSE.

It was stated above that Schmid obtained a sugar solution by the decomposition of a fisetin-glucoside from _Rhus cotinus_, and Perkin obtained the same from a glucoside in _Rhus rhodanthema_. These investigators thought that the sugar was isodulcite or rhamnose, but they did not isolate it on account of the small quant.i.ties of material at their disposal. Moreover, the sugar is very hard to crystallize in the presence of other soluble substances and is not found in large quant.i.ty in plants. Maquenne[27] could obtain only 15 to 20 gm. of rhamnose by working up 1 kilogram of the berries of _Rhamnus infectorius_. a.s.suming that the free fisetin found in poison ivy leaves had its origin in the decomposition of a fisetin-glucoside by natural processes, it was reasonable to suppose that the sugar would also be found in the free state, although, according to Roscoe and Schorlemmer:[28] "Isodulcite does not occur in the free state in nature, but is found as a peculiar ethereal salt belonging to the cla.s.s of glucosides. On boiling with dilute sulphuric acid, this splits up into isodulcite and other bodies...." The more recent works on the sugars and on plant chemistry[29] mention the occurrence of rhamnose only in the glucoside form, with one possible exception. The exception referred to is the occurrence of a free sugar, supposed to be rhamnose, in a certain palm-wine.[30] Czapek says:[31] "The well-known methyl pentoses do not occur in the free state in plant organisms so far as we know."

Since rhamnose forms a lead compound, the sugar, if present, should be found in the first lead precipitate, A, and also in filtrate A in case it is not completely precipitated in the presence of acetic acid and alcohol.

The filtrate A (about two liters) was examined first. It had a light yellow color, contained an excess of lead acetate, and was acid from the acetic acid liberated in the precipitation of the lead compound A.[32]

This filtrate was evaporated to dryness under diminished pressure to remove alcohol, water, and acetic acid. The clear distillate had a peculiar odor suggesting both tea and amyl formate. It was saved for examination and was found to be not poisonous. The residue in the dish after evaporation was a tough reddish brown, gummy ma.s.s which could be drawn out into fine threads. It had a pleasant sweet odor. It was extracted several times with hot water, each portion being filtered. A brownish yellow powder remained undissolved and was saved. The combined filtrates deposited more of the yellow solid on standing. This powder will be referred to later as "P." The filtered liquid was freed from lead by hydrogen sulphide. The solution then had a lemon yellow color, a sweet odor and was acid from acetic acid. On concentrating the solution by evaporation and making a small portion of it alkaline with sodium hydroxide, the yellow color came out very intense[33]. The alkaline solution reduced Fehling solution and ammoniacal silver nitrate, indicating the presence of a sugar. Another portion of the solution gave a slight precipitate with phenyl hydrazine in the cold. The remainder of the solution was evaporated to dryness, extracted with water, filtered, and again evaporated. A dark sticky syrup was left which was only partly soluble in water. This was treated with water, filtered, and the filtrate was evaporated, the water being replaced from time to time to remove acetic acid. Finally the liquid gave the following tests for rhamnose, besides those already mentioned:

(1) With alpha-naphthol[34] and sulphuric acid, a purple violet color.

(2) With thymol[35] and sulphuric acid, a red color.

(3) With resorcinol[36] and sulphuric acid, red color.

(4) With orcinol[37] and hydrochloric acid, red color.

(5) With ammonium picrate and sodium picrate, yellowish red color.

(6) With phloroglucinol and hydrochloric acid, red color.

(7) It decolorized an alkaline solution of pota.s.sium ferricyanide.

(8) It gave a white precipitate with lead acetate.

The filtrate B (p. 20) from which gallic acid was precipitated by sulphuric acid in four fractions was saved to examine for sugar. To remove gallic acid completely, and other vegetable matter, it was shaken out several times with ether, and was kept at a low temperature with salt and ice for a long time. It was left standing for several weeks, during which time more brown matter separated out and was filtered off.

The filtrate was evaporated to a small bulk, cooled, and filtered from crystals of pota.s.sium sulphate. The filtrate was evaporated to dryness, the residue taken up in water and filtered through bone-black. Addition of alcohol caused complete precipitation of pota.s.sium sulphate. The solution then gave the above mentioned characteristic tests for rhamnose.

All attempts to get the osazone of the sugar by the method of Fischer[38] failed, probably on account of the small quant.i.ty of the sugar present. The plant, it will be remembered, was originally extracted with ether in which rhamnose is practically insoluble. The above described tests, however, can leave no doubt as to the ident.i.ty of the sugar.

Additional evidence that the sugar is rhamnose was obtained by a method described by Maquenne[39] as follows:

"The production of methyl furfurol in the dehydration of isodulcite furnishes a very simple means of characterizing this sugar in mixtures which contain it; it is sufficient, for example, to distil 50 gm. of quercitron wood with as much sulphuric acid and about 150 gm. of water, then to rectify the liquid obtained in order to get several drops of the crude furfurol, which on addition of alcohol and concentrated sulphuric acid gives immediately the green coloration characteristic of methyl furfurol. This procedure is applicable to extracts as well as to entire plants, and has the advantage that it does not require the separation of isodulcite, the crystallization of which is often very slow and at times impossible when it is mixed with other very soluble substances."

The experiment was tried with the crude ether extract of the plant according to the directions of Maquenne, and the green color with alcohol and sulphuric acid was obtained from the thicker oily portion of the distillate. This test can be made with hydrochloric acid[40] as well as with sulphuric. Therefore the color test was tried with the ester mixture prepared in one of the early experiments by boiling the original plant material with hydrochloric acid and alcohol. Methyl furfurol was found here also, this method indeed giving better results than that of Maquenne.

The presence of free rhamnose has thus been shown in the original material, in the first precipitate by lead acetate, and in the filtrate from this precipitate. Experiments to be described under "The Poison"

showed that the ether extract from the Soxhlet apparatus contained a substance which yielded rhamnose when hydrolyzed by dilute sulphuric acid.

The presence of free gallic acid, fisetin, and rhamnose in the plant can be readily explained by a series of a.s.sumptions for which there is a considerable amount of experimental evidence. There is reason to believe that tannin-like bodies are formed at the expense of chlorophyll,[41]

that complex tannin bodies can be broken down by acetic acid (also found in _Rhus toxicodendron_) into a tannic acid and a glucoside (for example, the "fustin-tannide" mentioned above yields tannic acid and fisetin-glucoside); and finally that the glucoside can be hydrolyzed by acids or enzymes giving, in the sumach plants, fisetin and rhamnose.

Nitrogenous ferments which can effect the hydrolysis of glucosides and give rise to sugars are frequently found in plants, for example, emulsin in almonds, myrosin in mustard, and erythrozym in madder. Acree and Hinkins[42] found that diastase, pancreatin, and a number of other enzymes cause hydrolysis of triacetyl glucose with the formation of glucose and acetic acid. Stevens[43] obtained a nitrogenous oxidizing enzyme from _Rhus vernicifera_. The close relationship between the poisonous species of _Rhus_ would lead us to suppose that the same soluble ferment exists in poison ivy, though it was not detected in the original material used in these experiments, probably because the plant was extracted with ether in which the enzyme is insoluble. The existence of such a soluble ferment would explain the presence of free sugar and free fisetin.

EVIDENCE OF THE PRESENCE OF A FATTY ACID IN FILTRATE A.

The brown substance P, obtained from filtrate A by evaporation and extracting the residue with hot water, was suspended in warm water and dilute sulphuric was added. A white precipitate was formed and a strong fatty acid odor was developed. After the mixture had been heated for some hours on the water bath a small portion was made alkaline and it reduced Fehling solution. The main solution was filtered and the precipitate supposed to be a fatty acid was saved. The filtrate was neutralized with barium carbonate, filtered, evaporated, freed from caramel, and the solution then gave the tests mentioned above for rhamnose.

A portion of the precipitate supposed to be a fatty acid was ignited in a porcelain spoon. It fused, carbonized, and burned. The remainder was heated with alcoholic potash and reprecipitated with hydrochloric acid.

The precipitate was washed and heated with alcohol. Part of it dissolved. The insoluble part was found to be a lead compound. On boiling it with hydrochloric acid and cooling, lead chloride crystallized out. This was confirmed by dissolving the lead chloride in hot water and precipitating as lead sulphide. These experiments were not carried farther on account of the small quant.i.ty of material, but they show that the gummy substance obtained from filtrate A contained rhamnose (either as a lead compound of free sugar or as a lead compound of a rhamnoside), and also, most probably, the lead compound of an organic acid.[44]

THE FRAGRANT DISTILLATE.

Several times in the course of this work, extracts of the original plant material in alcohol and in water were distilled under diminished pressure for the purpose of concentrating the solutions. The distillate, in every case, had an ethereal odor suggesting amyl formate in very dilute solution, but was more fragrant. The distillate from a water extract was examined. It was a clear liquid, a little darker than pure water, was not poisonous, was neutral to litmus paper, gave no color with ferric chloride, reduced ammoniacal silver nitrate, but not Fehling solution, and gave a faint red color with dilute ammonium hydroxide and with sodium carbonate.

A small quant.i.ty of a finely divided black precipitate separated out from the water distillate on standing.

The substance with the fragrant odor was extracted by shaking the distillate with ether and letting the ether evaporate spontaneously. A very small quant.i.ty of a yellow solid was deposited on the sides of the dish. This substance had a strong and persistent odor, so sweet as to be almost nauseating. Not enough was obtained for examination or a.n.a.lysis.

This fragrant residue was difficultly soluble in water and the solution reduced silver nitrate in ammonia. A steam distillate of the original plant material had the same fragrant odor as the distillate from a water extract.

THE POISON.

288 grams of the original poisonous material were extracted with 50 per cent. alcohol, and this alcoholic solution was precipitated with lead acetate in the manner already described (p. 17). The lead precipitate so obtained was extracted with ether in Soxhlet extractors and after the extraction was found by test to be free from poison. Therefore the poison, if precipitated by the lead acetate, must have been extracted by the ether. This ether solution had a dark green color, and was acid from acetic acid brought down in the lead precipitate. The ether was evaporated in a vacuum desiccator without heat and there remained a small quant.i.ty of an acid mixture of water and a soft tar; the watery part was colored green, showing that the tar was soluble to some extent in dilute acetic acid. The mixture had the peculiar odor of the original material. A small drop of the green watery part was applied to the wrist, allowed to remain a few minutes and was then removed by absorbent paper, but the spot was not washed. Itching and reddening of the skin commenced within twenty-four hours. At the end of forty-eight hours, there was a well developed case of poisoning. How this was cured will be described in another place.

A small portion of the poisonous mixture was dissolved in alcohol, and this solution was divided into three parts. The first part was treated with ferric chloride, but it gave no color reaction. Another portion of the alcoholic solution was diluted with water. It became turbid. The third portion gave a dirty-green precipitate with lead acetate, which seemed to come down more readily when the solution was diluted with water. The main portion of the poisonous mixture was then dissolved in 95 per cent. alcohol and lead acetate in 50 per cent. alcohol was added.

The precipitate was filtered, washed, and decomposed by hydrogen sulphide in a mixture of water and ether. The ether solution was filtered and evaporated. The residue was a tar which, on standing in a desiccator for some time, became dry enough to break into sticky lumps.

An alcoholic solution of this substance gave a dark color with ferric chloride and a light colored precipitate with lead acetate.

To get more of the poisonous tar for study, 233 grams of original material were extracted with 95 per cent. alcohol. Strong alcohol was used in order to dissolve as much of the tar as possible. The solution had a dark greenish color, but was somewhat yellow in thin layers. The undissolved tar was filtered off and extracted twice again in the same way. The tar left after the third extraction was only slightly soluble in alcohol, and its solution was not poisonous. The three filtrates from these three extractions were precipitated separately by lead acetate in 50 per cent. alcohol. The first precipitate was largest, darkest in color, and carried down more tarry matter. The second was light green, and the third was quite small, black, and was not a lead compound at all, but some of the tar which separated out on diluting the strong alcohol with the weaker grade containing lead acetate. It was soluble in ether and less soluble in alcohol. The alcoholic solution of this third lot gave no precipitate with hydrogen sulphide. The first and second lead precipitates were filtered by suction and washed with water. They were kept a day or two in a desiccator over sulphuric acid, but did not become completely dry. The weight of these two moist precipitates together was 172 grams. They were combined and extracted with ether in Soxhlet extractors which were kept in operation during work hours for three days.

In the meantime the alcoholic filtrates from these lead precipitates were combined and concentrated on the water bath by distilling off two liters of alcohol. The alcohol obtained had the peculiar odor of the original material, but was not poisonous.

After a long extraction of the lead precipitate in the Soxhlet extractors, the green ether solutions were combined and washed by shaking them with water to remove lead acetate and acetic acid in case any should have been held in the lead precipitate. The ether was distilled off at a low temperature and there remained a soft tar, a portion of which was not completely soluble in 95 per cent. alcohol. The alcoholic solution had a greenish yellow color and was poisonous. The tar was also partly soluble in acetic acid, and this solution was found to contain lead. Thinking that the lead acetate had not been completely washed out, the main part of the tar was dissolved in ether and shaken with water. The wash water continued to give a test for lead as long as the washing was continued. This indicated probably the hydrolysis of an unstable lead compound. Hydrogen sulphide was pa.s.sed into the ether solution mixed with water to remove the lead. Lead sulphide was filtered off, and the ether was evaporated. A small portion of the tar residue in alcoholic solution gave a color reaction with ferric chloride. As this may have been due to traces of lead gallate dissolved in the extraction with ether and afterwards decomposed by hydrogen sulphide, the main portion of the tar was redissolved in ether and shaken with water until it no longer reacted with ferric chloride. The ether was then evaporated and a soft, black, poisonous tar or gum of uniform consistency was left which was shown by tests to be free from gallic acid and lead. These experiments showed that some of the poison was precipitated as a lead compound soluble in ether and some was brought down mechanically in the free state. To see if the extraction with ether in the Soxhlet apparatus was complete, the residue in the thimbles was decomposed by hydrogen sulphide and shaken with ether. The dark colored ether solution was freed from gallic acid by shaking with water and dilute sodium carbonate solution, and was evaporated. A small quant.i.ty of tar was obtained which was added to the main portion.

A solution of the poisonous tar in 95 per cent. alcohol did not reduce Fehling solution and did not give a precipitate with lead acetate except the separation of a small quant.i.ty of tar, which was not a lead compound. The lead compound of the poison was apparently soluble in 95 per cent. alcohol as well as in ether, for it would not precipitate in this medium, although it was found in the original precipitate by lead acetate. The alcoholic solution of the tar became turbid on diluting with water.