69.58 per cent. of quercetin, and had properties indicating the above triacetylquercetin. Again, if triacetylquercetin is suspended in acetic anhydride, and a little sulphuric acid added, tetracetylquercetin separates from the solution after a time.

Quercetin itself is not appreciably attacked by cold acetic anhydride if left in contact with it for several days, although in two months at the winter temperature partial acetylation had occurred. The product, which was neither of the above derivatives, was not closely examined. Although it seems probable that, in the acetylation of quercetin, alizarin, and similar substances, the hydroxyl which decomposes potassium acetate is first attacked by a mixture of acetic anhydride and sodium acetate, the behaviour of the sodium acetate in this reaction is in no way clear. For instance, alizarin methyl ether [OH: OCH, 1:2] does not decompose alcoholic potassium acetate, and is not acetylated by boiling for 4 hours with acetic anhydride. If sodium acetate, however, is present, the latter reaction rapidly takes place.

=

It is possible that the employment of this method of acetylation in the cold may be useful for the identification of colouring matters which yield no definite acetyl compound when treated in the usual

manner.

THEORETICAL.

The results of this investigation indicate that all the colouring matters of the anthraquinone, and many of those belonging to the flavone, xanthone, and ketone groups, have acid properties which enable them to decompose the alkali salts of various acids. With the excep

tion of morin and rhamnazin, no colouring matter of known constitution is markedly acidic unless it contains two hydroxyls in the ortho-position relatively to one another, and consequently has strong dyeing properties.

The acid nature of these members of the flavone or quercetin group is very interesting, for previously (Trans., 1895, 67, 644, and 1896, 69, 1439) it has been shown that they also possess marked basic properties, forming compounds of the colouring matter with one molecule of a mineral acid. On comparing the results of the former investigations with those of the present work, one is at once struck with the fact that colouring matters of the flavone series which possess basic properties, and those only, have a marked acid nature. Thus chrysin and apigenin yield neither acid compounds nor metallic derivatives by the methods described, and differ in this respect from the other colouring matters of this group which have been examined. Again, rhamnetin and rhamnazin, the methyl ethers of quercetin, form acid compounds with difficulty, and, on the other hand, their behaviour

towards potassium acetate indicates but feeble acid properties, 2 mols. of each being required for the decomposition of 1 mol. of the acetate. Finally, water exerts a decomposing influence on all acid and basic compounds of the members of this group.

In the so-called flavone derivatives here examined, as many as four to six hydroxyl groups are present, and in each case but one of these has the power of decomposing, under the conditions described, the alkali acetates. That this hydroxyl, per se, is strongly acid seems very doubtful, and it is much more probable in view of the decomposition effected by hot water, that the acid reaction is the result of a change in the molecule. In other words, the displacement of the hydrogen in this one hydroxyl group by a metal admits of a change of structure which cannot occur in the colouring matter itself. In former communications dealing with the acid compounds, the formation of these was explained according to the quinonoid theory of coloured compounds (H. E. Armstrong, Proc. Chem. Soc.), and this is also applicable to the metallic derivatives under discussion. Since the publication of this paper (loc. cit.), further work has shown that only those colouring matters (morin and rhamnazin excepted) which contain ortho-hydroxyls have acid and basic properties, and these results have suggested another constitutional formula; the latter (II) and that formerly considered most applicable (I) are here given, and represent monopotassium quercetin.

According to the formula I, the acid compounds can also be obviously represented, quercetin hydrochloride containing the group

and to some extent this is preferable to the first,

as it tends to explain the influence of the methoxy-groups on the acidity and basicity of rhamnetin and rhamnazin.

Feuerstein and Kostanecki (Ber., 1898, 31, 1760) have shown that

flavone, the parent substance of this group, is colourless, and it is probable that its yellow derivatives already have the quinonoid constitution I. In the formation of their metallic derivatives and acid compounds, the change represented by formula II occurs, and the production of these is thus assigned to an alteration in the quinonoid form of the dyestuff, which is again reversed by the action of water. The lower members of this series, for instance, chrysin,

OH

он со

are but feebly acid, and poor dyestuffs, properties which apparently go hand in hand, and these are considered as the effect of the quinonoid constitution, I. Although possible, it appears unlikely that the addition of two hydroxyls to chrysin in the positions in which they occur in quercetin would so strongly develop its acid function without their apparently taking part in the effect.

Morin, to which I have provisionally assigned the constitution of a flavone derivative (Trans., 1896, 69, 798), contains no orthohydroxyls, but to its metallic derivatives a quinonoid form can be given, as in the case of quercetin, for it must be remembered that, although in anthraquinone derivatives ortho-hydroxyls confer dyeing property with mordants, it is not so in the flavone group, for chrysin, apigenin, and isorhamnetin are colouring matters. Morin is colourless (Trans., 1895, 67, 650), and this property is considered by Herzig * (Monatsh, 1897, 18, 700) to throw doubt on its flavone constitution (loc. cit.); to this, however, the arguments above employed are considered sufficient answer.

The expression of the acid property of colouring matters of the xanthone and benzophenone groups, as an assumption of or a property of an existing quinonoid arrangement can obviously be readily expressed by formula I. In the latter class, no colouring matters exist which do not contain ortho-hydroxyls.

. Alizarin Group.—The colour of the metallic derivatives of alizarin, anthragallol, &c., as described in this paper, differs more markedly

* In his paper, Herzig suggests that Bablich and I omitted to find the water of crystallisation in tetrabromomorin ethyl ether. This substance was merely examined as to the ethoxy-group, and the presence or absence of water of crystallisation was immaterial;

C17H10BrO requires C2H=4'48; C17H10BгO+H,O=4.36. Found 4:56.

I am frequently aware of the presence of water of crystallisation under certain conditions in the substances I examine, and if no special interest is thereby affected, ignore or remove this when necessary, before analysis.

from that of the dyestuff themselves than is the case with the yellow colouring matters, being almost as intense as the fully saturated salts. As these, when etherified, show that the acidic hydroxyl is in the paraposition relatively to one of the carbonyl groups, this supports the theory that the formation of these derivatives is accompanied by quinonoid change, again reversed on removal of the metal.

An essential feature of this group is the presence of a second hydroxyl in the ortho-position relatively to that necessary for the production of quinonoid change, and this also confers dyeing property. The red monobarium derivative of alizarin described in this paper indicates the occurrence of two classes of coloured compounds, one violet, and a second, which on its formation reverts to the original constitution of alizarin and is red. Quinizarin, C1H8O4 [OH: OH=1:4], contains parahydroxyls, but is soluble in alkalis with a deep blue coloration, if anything, more intense than that given by alizarin itself. This behaviour, which at first sight seems opposed to the explanation suggested above, is readily accounted for on the assumption that the parahydroxyls assume a quinonoid change, and the intense colour of the alkaline solution would thus be due to the presence in quinizarin of two such groups. Its lack of acidic and tinctorial properties points to the fact that this second change is not readily assumed and is of an unstable nature; this appears to be the case, for according to Schunck and Römer (Ber., 1877, 10, 555) the blue alkaline solution is readily decomposed by carbonic acid, whereas that of the alizarin derivative is not. In confirmation of this instability, it was noted that when crystals of dipotassium quinizarin and monopotassium alizarin were placed side by side with free access of air, the alizarin was not affected, but the quinizarin derivative became red from separation of free quinizarin.

The anthraquinone colouring matters have not yielded any acid compounds, but as orange red morindone (trihydroxymethylanthraquinone) and alizarin, for instance, dissolve in sulphuric acid with a deep blue and deep dark red colour respectively, it is likely that these solutions contain unstable acid compounds similar in type to the corresponding flavone derivatives.

Dyeing Properties of the Monopotassium Derivatives of Quercetin and Alizarin.

The employment of calcium acetate to assist in the dyeing of certain phenolic colouring matters is well known, and researches on this point have indicated that the calcium is taken up to form an essential part of the colour lake; the alkali acetates also, although not employed commercially, are known to produce a somewhat similar effect. The reaction here involved appears analogous to the production of the metallic derivatives described in this paper, indicating as probable that the calcium, for instance, in an alizarin lake has partly neutralised the hydroxyl which is in the meta-position relatively to the carbonyl group. Dyeing experiments with the potassium derivative of alizarin and quercetin on aluminium mordanted wool at first sight corroborated this view, in that the shades produced were very similar in character to those given by mixtures of the colouring matter with sodium or calcium acetates; on the other hand, these results were hardly conclusive, owing to the acid nature of the alum mordant, which tends to take up the metal, in part, at all events, during the dyeing operation. With mordanted calico, the effect was hardly comparable, as the percentage of metal in potassium alizarin is in excess of that necessary for successful dyeing, but in further experiments this may be obviated by employing mixtures of this compound and free alizarin. As previously stated, the alkali acetates are not decomposed by the colouring matter in aqueous solution, although as the liquid in the case of alizarin acquires a red tint, an unstable compound of the alizarin and the acetate may exist in the hot solution. The acetates of the alkaline earthy metals, however, are more active, for with a boiling calcium acetate solution, alizarin yields, apparently, first the monocalcium derivative and ultimately, after some hours, the well known calcium alizarin, C1HO,Ca; by this method, the latter was obtained in glistening needles having a metallic lustre, so far as I am aware, it has previously been obtained only in the amorphous condition. Found, Ca=14.13. Theory requires Ca = 14.38 per cent. This reaction will be further studied, together with the behaviour of other colouring matters towards the acetates of this group.

CLOTHWORKERS' RESEARCH LABORATORY,

DYEING DEPARTMENT,

YORKSHIRE COLLEGE.