employed in preparing diethyl-lophinium iodide, which Kühn obtained by heating lophine with ethyl iodide at 100°.

The following simple reaction appeared calculated to decide between the two formulæ. Radziszewski's formula contains, as already pointed out, a benzaldehyde-residue (benzylidene); the anhydro-base formula contains a benzoic acid-residue (benzenyl). Fischer and Troschke have shown that lophine may be heated with hydriodic acid and amorphous phosphorus to 220° withont change. It seemed to me that if it were possible, by the action of the acid at a still higher temperature, to split up the lophine, a compound of Radziszewski's formula ought to yield benzaldehyde, which would then be reduced to toluene; whilst a compound of the anhydro-base formula would yield benzoic acid, which would not undergo further change. As benzoic acid could not, in presence of a powerful reducing agent like hydriodic acid, be furnished either by the dibenzyl-residue of Radziszewski's formula, or by the stilbene-residue of the anhydro-base formula, the formation of benzoic acid under these circumstances might be taken as deciding in favour of the latter formula.

As a fact, I find that lophine when heated with hydriodic acid and amorphous phosphorus to a temperature a little over 300°, splits up, yielding benzoic acid. As the pressure with strong hydriodic acid proved unmanageable, a mixture of one volume of the strongest hydriodic acid with four volumes of fuming hydrochloric acid was employed instead. That a sufficiency of the reducing agent had been employed, was evident from the fact that, on cooling, the upper part of the tubes contained crystals of phosphonium iodide. The pressure on opening, in spite of the dilution with hydrochloric acid, was very great. The main portion of the lophine was recovered unchanged, the difficultly soluble lophine salt fusing together and thus escaping further action. No resinous products are formed.

Had the above reaction occurred at a lower temperature, I should have regarded it as absolutely conclusive against Radziszewski's formula. As it is, I think the probability that the benzoic acid can have been formed from anything else than a benzoic acid residue very slight. It is to be borne in mind that, as the temperature rises and the danger of bye-reactions increases, the power of the reducing agent also increases. Further, the temperature, though high, is at least 100° lower than that at which lophine boils without decomposition. According to a determination made by means of a Geissler high-temperature mercurial thermometer registering to 450° (with internal pressure to prevent the boiling of the mercury), lophine boils at 415° (uncorr.). Of course the indications of such an instrument areonly approximate.*

* Radziszewski also makes some highly ingenious suggestions concerning the con

In fulfilment of the promise made in the former communication on the above subject, I have studied the action of other aldehydes, together with ammonia, upon benzil. Parahydroxybenzaldehyde yields,

stitution of glyoxaline, which he regards as "lophine, in which the three phenylgroups are replaced by three hydrogen-atoms." He considers that when glyoxaline is formed by the action of ammonia upon glyoxal, a portion of the glyoxal first takes up the elements of water, yielding formic acid (the production of which in the reaction has been observed by Ljubavin) and formaldehyde; and that this last substance then reacts with glyoxal and ammonia to form glyoxaline, a reaction which would correspond with that in which lophine is formed from benzaldehyde, benzil, and ammonia. I do not at present propose to go into this subject further than to point out that if, while accepting the above analogy, we formulate glyoxaline on the basis of the anhydro-formula of lophine, thus:

we arrive at a formula which, I consider, accounts better for the reactions of this compound than any of the formula which have as yet been proposed for it. The only fact connected with glyoxaline which this formula does not readily explain, is the identity of methylglyoxaline with oxalmethyline.

Addendum. Since the above was written, Radziszewski has published a second paper (Ber., 15, 2706), in which he describes the synthesis of Wallach's paraoxalmethyline by the interaction of glyoxal, acetaldehyde, and ammonia-here again employing a reaction belonging to the class of condensations discovered by me. He does not appear to have expected to obtain paraoxalmethyline, but only some homologue of glyoxaline. He formulates paraoxalmethyline on the type of his lophine formula. I intend discussing this reaction more fully elsewhere, but in the meantime desire to take this opportunity of putting on record the following formule, which are founded upon the glyoxaline formula given above :

These formulæ, which furnish a consistent account of the reactions of Wallach's oxalines, were constructed by me in August last, at which time I suggested to Dr. F. E. Matthews, who was then working with me, that he should attempt the synthesis of paraoxalmethyline from glyoxal, acetaldehyde, and ammonia. I thus predicted the result now obtained by Radziszewski. I wish, however, expressly to state that this train of thought was mainly suggested by the above speculations of Radziszewski on the formation of glyoxaline. But, at the same time, I do not think that it would have been possible for Radziszewski, holding the views which he does concerning the constitution of glyoxaline, to predict the formation of paraoxalmethyline in the above reaction.

VOL. XLIII.

C

as already described, parahydroxylophine. With salicylaldehyde a different reaction occurs: 2 mols. of aldehyde, together with 2 of ammonia, react with 1 of benzil, yielding a compound totally distinct in its properties from parahydroxylophine; thus:

=

24

C1H10O2+2C,H2O2 + 2NH, C2H2N20, + 20H2.
New compound.

Benzil,

Salicylaldehyde.

Furfuraldehyde acts in a similar manner.

I hope to be able to lay before the Society, at an early date, the results of the investigation of this new class of compounds.

III.-CONTRIBUTIONS FROM THE JODRELL LABORATORY.

1.-Contributions to the Chemistry of Lignification.

By C. F. CROSS and E. J. BEVAN.

FOLLOWING the views advanced by physiologists on the chemical phenomena of lignification, we were led to forsake the incrustation theory, as not adequately expressing the facts established concerning the origin, properties, and decompositions of the lignified substance, and to adopt, as a working hypothesis, the alternative view of lignose, or bastose, as we ventured to call the jute-fibre substance, viz., that it is a chemical whole in the sense of presenting a true combination rather than a mixture of cellulose with its non-cellulose constituents. Subsequent observations have further justified this course. means of fractional solution in the ammonia-copper reagent, we uniformly obtained an amorphous modification of the fibre substance, exhibiting properties similar to the original as regards its behaviour both to chlorine and to acids.

By

In one particular, however, a difference is observed, in that the freshly precipitated amorphous modification gives only a slight reaction with aniline sulphate, and after a second solution and precipitation no coloration is obtained. That this reaction, supposed to be essentially characteristic of lignose, is in reality due to some product of change (probably of oxidation) is further shown by the fact that this property of giving a yellow colour with aniline salts is entirely lost after the substance has been boiled in a solution of sodium sulphite, the other properties remaining unaltered. We find moreover that a yellow reaction with aniline salts is characteristic of a number

of aromatic aldehydes. If, for instance, oil of cinnamon be shaken with a solution of the sulphate, the whole solidifies to a mass of bright yellow needles. Lignose we think, therefore, is to be considered apart from this property.

We have previously shown that jute is resolved in various ways, according to the methods or conditions brought to bear upon it, the cellulose for instance appearing either as cellulose or in the form of acids of the pectic class. So also the non-cellulose appears either as an astringent substance, or in the form of the chlorinated derivative previously described. In reference to the latter and its evident connection with the aromatic series, Dr. Armstrong directed our attention to the researches of Stenhouse and Groves on the chlorination of pyrogallol as probably bearing on the subject. We prepared mairogallol according to their method (Chem. Soc. J., 1875), and found that both it and the amorphous substances which constitute the chief portion of the product give, when treated with sodium sulphite solution, a colourreaction exactly resembling that which is characteristic of the freshly prepared lignose derivative. A close connection of these plant-constituents with the trihydric phenols, which can be seen to be suggested on grounds which are independent of this observation, we venture to think is thereby fairly established.

Following up this subject, we endeavoured to prepare a more highly chlorinated derivative of bastose. The derivative obtained by the action of chlorine gas upon bastose in presence of moisture is an amorphous yellow body, which, only when freshly prepared, gives the colour-reaction with sodium sulphite. Although this indicates the occurrence of molecular change during the process of purifying the body for analysis, and although its amorphous character places it in that much abused category of substances to which the ordinary criteria of purity are inapplicable, the numbers obtained in the analysis of preparations various in origin and differently prepared, were constant, and agreed with those required by the formula n(C19H18C1O9). In justification of the adoption of this formula we would state first that it was our only guide in investigating the constitution of lignified fibres, and secondly, that substances which go to build up living tissues are of very necessity colloïds, and their immediate derivatives also; but because colloïds they are none the less definite, and at all events the method of ultimate analysis must be applied to their investigation until it is shown to be nugatory.

The chlorinated compounds experimented upon were obtained, the one from jute and the other from the fibre of Musa paradisiaca, a monocotyledonous plant. The purified fibres were exposed in the damp state to an atmosphere of chlorine gas, and the reaction being complete, the products were severally dissolved away by means of

alcohol, precipitated with water, washed and dried first in a vacuum, and lastly at 100°. These were then separately dissolved in glacial acetic acid, and further chlorinated after the manner described by Stenhouse and Groves. The products were separated by pouring the acetic solution into water, whereby they were precipitated in the form of a yellowish-white substance resembling wax. After washing and drying, first in a vacuum, and lastly at 100°, they were analysed, with the following results :

-

The products are therefore identical. It is impossible to account for their derivation from the original tetrachlorobastin (which we may represent by the formula C3H36 ClO18) by a symmetrical equation. At present we cannot do more than record the results as they stand.

Starting indeed with a highly complex molecule, such as both bastose and the lower chlorobastin certainly are, and in view of the further complicating action of chlorine upon the trihydric phenols and their derivatives, which has been established by the work previously cited, we have no reason to expect a resolution into simpler molecules by means of this reaction. It would appear that only in the absence of oxidising conditions can this be effected, and it is from this point of view that we are following up the resolution of bastose, lignose, and the chlorobastins by means of the sulphites under extreme conditions of temperature and pressure.

Note on the Constitution of Lignose.

We would record two recent observations which bear upon the question of the mode of union of the constituents of lignose. (1.) Dry chlorine has no action upon this substance, whereas the presence of moisture determines instant combination, with evolution of heat. (2.) The furfural-yielding constituent survives exposure to chlorine, the chlorinated jute fibre giving an abundant yield of this aldehyde by distilling with hydrochloric acid.