this composition, the formation of isobutyric acid would leave only one formula possible, namely,

The stability of the compound is remarkable, if it is the lactone of a B-hydroxymalonic acid; the formation of isobutyric acid, however, does not admit of the possibility of its being a y-lactone, for if that were the case, acetone should be formed on oxidation, whereas careful search failed to detect it.

The reactions described in the preceding paragraphs may be represented as follows:

Experiments were made with pilocarpine to obtain the bromoethyl ester, by treatment with phosphorus pentabromide, but the basic character of pilocarpine interfered with the reaction, and a negative result was obtained.

Fusion of Isopilocarpine with Caustic Potash.

Chastaing (Compt. rend., 1882, 94, 223) has described the fusion of pilocarpine with caustic potash, and stated that the products of the reaction were methylamine, carbon dioxide, butyric acid, and traces of acetic acid, and that the reaction may be represented by the equation

2011H16O2N2 + 2H2O= 2CO2+ 4C1H ̧O2+4CH ̧•NH2 + O2,

The analytical numbers quoted for the platinichloride * prove, however, that he was working with a mixture of ammonia and methylamine, and no details are given of the identification of butyric or acetic acids.

In the preliminary account of this reaction (loc. cit., 496), the bases formed were proved to be ammonia and methylamine, and the acids yielded a silver salt, which contained a higher percentage of silver than required for silver butyrate. The reaction has been further investigated. Five grams of pure isopilocarpine were fused with 50 grams of caustic potash, and the fused mass, after cooling, dissolved in water acidified with sulphuric acid, and distilled with

* Found Pt=42.5 per cent.

(NH3)2, H2PtCl requires Pt=43 9 and (CH, NH2)2, H2PtCl, requires Pt=41.3 per cent..

steam. The acid liquid was extracted with ether, the ethereal solution dried over calcium chloride, and distilled.

The residual liquid, which had a strong, rancid smell, was distilled at atmospheric pressure, when a small fraction was collected below 120°, but the greater portion came over between 120° and 155°, leaving no residue. On adding a little water to the higher fraction, oily drops separated, which dissolved on the addition of more water. From the aqueous liquid, a crystalline barium salt was prepared; this was extracted with boiling alcohol, in which it was very slightly

soluble.

(i) The portion dissolved by alcohol gave a white, silver salt, which was not reduced on standing. On analysis the following result was obtained:

0.0388 gave 0.0214 Ag. Ag=55.2.

(ii) The residue after extraction with alcohol gave a crystalline silver salt, which was recrystallised from hot water. On analysis, the following result was obtained:

0.0592

gave 0.0328 Ag. Ag=55·4.

CH,O,Ag requires Ag=55.38 per cent.

The calcium salt was soluble in warm water, but separated again on cooling. The analytical numbers, boiling point, immiscibility with water, and solubility of the calcium salt, prove conclusively that the acid formed is isobutyric acid.

The portion boiling below 120° gave a small amount of an amorphous barium salt, insoluble in alcohol, and an amorphous silver salt (Ag= 57.4 per cent.), which underwent reduction on standing. No trace of acetic acid could be found.

Further investigation of the subject is proceeding on different lines, and it is hoped, either by the use of larger quantities of material, or by synthetical methods, to bring forward further proof of the correctness, or otherwise, of the formula now proposed for the acid C7H1004

THE WELLCOME CHEMICAL RESEARCH LABORATORIES.

LXXX.-Isomeric Partially Racemic Salts containing Parts I-VI. Hydrind

Quinquevalent Nitrogen.

amine Bromocamphorsulphonates, Chlorocamphorsulphonates, and cis-π-Camphanates.

By FREDERIC STANLEY KIPPING, Ph.D., D.Sc., F.R.S.

THIS paper contains an account of the preparation and investigation of three pairs of isomeric partially racemic salts, the isomerism of which is apparently of a kind different from any yet recorded.

*

All these salts have been obtained from a-hydrindamine, a primary base easily prepared by reducing a-hydrindoxime (compare Revis and Kipping, Trans., 1897, 71, 250).

Hydrindamine contains an asymmetric carbon group, as will be seen from its formula,

and the synthetical base is therefore composed of equal quantities of enantiomorphous components.

When this base is combined with various optically active acids, it shows an altogether unusual behaviour, giving in some cases what appears to be a homogeneous partially racemic salt, in others a mixture of unequal quantities of two isomeric salts; the latter are not the ordinary isomerides which are obtained from most externally compensated bases under such conditions, that is to say, they do not represent salts of the separate enantiomorphously related components of the base, but are both comparable to partially racemic substances.

The results of the examination of these remarkable isomerides may be briefly summarised as follows.

Hydrindamine and a-bromocamphorsulphonic acid unite to form unequal quantities of two hydrated salts, which contain different quantities of water of crystallisation; the anhydrous salts both have the molecular formula C,H1N,C10H14 BrO SO,H. The more sparingly soluble "a-salt" separates from water in opaque, bulky masses of minute needles, whereas the more readily soluble "B-salt" crystallises in well-defined, transparent prisms, absolutely different in appearance

* The letter a- which is used here to distinguish this base from 8-hydrindamine is not employed again, as to do so might cause confusion with the a-salts, which are described later.

from the crystals of its isomeride. The two salts have different specific rotations.

Hydrindamine and a-chlorocamphorsulphonic acid unite to form unequal quantities of two isomeric salts which are respectively analogous to those of the bromo-acid; the more sparingly soluble a-salt is isomeric with the corresponding bromo-salt, from which it cannot be distinguished by inspection. The ß-chloro-salt bears a close relationship to the B-bromo salt, but differs from the latter in at least one important respect; when crystallised from water, it is deposited either in what may be termed the normal form of transparent, well-defined prisms, closely resembling, and doubtless isomeric with, the crystals of the B-bromo-salt, or it is obtained in opaque masses indistinguishable by inspection from those of the a-chloro-salt; these two varieties, so different in appearance, both represent ẞ-salt, and although convertible one into the other by crystallisation, undergo this change in a rather bewildering manner, and under conditions which have not yet been established. The conversion of either of these a-salts (bromo- or chloro-) into the corresponding B-salt, or vice versa, by crystallisation only, has never been observed. The specific rotations of the a- and B-chloro-salts are practically the same in dilute aqueous or alcoholic solution, but different in other solvents.

Hydrindamine and cis--camphanic acid (Kipping, Trans., 1896, 69, 913) give apparently unequal quantities of two isomeric anhydrous salts of the composition C,H,N,C10H1404, which are indistinguishable in appearance, but differ in melting point and other properties; the more sparingly soluble a-salt melts at about 193°, the B-isomeride at about 173°, and the two compounds seem to be partially, but slowly, converted one into the other in aqueous solution. The two isomerides have widely different specific rotations in methyl alcoholic solution, but in dilute aqueous solution they give practically the same value ; the a-salt shows mutarotation in a very marked manner, but apparently the B-salt does not.

The question now to be discussed is, how to account for the existence of these isomerides. In doing so, it will be convenient to take a particular case, such as that of the hydrindamine bromocamphorsulphonates, with which most of the experiments have been made, the behaviour of the other salts being considered only when there is some important difference to note.

The most natural conclusion, namely, that the externally compensated base is resolved into its enantiomorphous components by fractional crystallisation of its salts with these optically active acids, will first be examined, and, as evidence against this view, the fact that the isomerides are obtained in very unequal quantities may first be quoted. Further, the base regenerated from the a-salt is optically inactive, and

is identical in this, and in every respect, with the base regenerated from the B-salt, and also with externally compensated hydrindamine prepared directly from hydrindoxime; this is proved by the fact that the salts obtained by combining the regenerated bases with optically inactive or active substances are identical. It may therefore be concluded, provisionally, at any rate, that the externally compensated base has not been resolved into its enantiomorphous components.

Granting this to be so, the formation of the isomerides might be explained by assuming that the salts of the sulphonic acids are merely differently hydrated varieties of one substance; this is not the case, however, because these salts are not convertible one into the other by crystallisation, they have different specific rotations in aqueous solution, and the anhydrous compounds have different melting points; moreover, such an assumption would not explain the existence of the isomeric anhydrous cis--camphanates. For similar reasons, the isomerides cannot be regarded as different crystallographic forms of one individual compound.

The view that the ẞ-salts of the sulphonic acids are derived from an acid physically or chemically isomeric with the acid of the a-salt might next be adopted, especially as these B-salts are formed in comparatively small quantities; it is shown, however, that the acid regenerated from the B-salt is identical with that of the a-salt; also that the behaviour of the salts themselves towards ferric chloride, phenyl isocyanate, piperidine, &c., is such as to practically exclude the possibility of the existence in the B-salt of a tautomeric form of the acid which might change when liberated from the base: the existence of the isomeric cis--camphanates, and many other facts, show that the isomerism is not due to the acid ion of the salts; further, the B-salt of any one of the acids is converted into the a-salt on repeatedly evaporating it with small quantities of externally compensated hydrindamine.

As these facts seem to prove that the isomerism is due to the hydrindamine, and as the base obtained from the a-salt is nevertheless identical with that regenerated from the B-salt, it might be assumed that the a- and ẞ-isomerides are actually the salts of enantiomorphously related bases, but that the latter undergo rapid racemisation at ordinary temperatures when liberated from the acid. That this explanation will not suffice is proved (a) by the fact that the salts are obtained in unequal quantities, (b) by the fact that the cis--camphanates afford optically inactive hydrindamine hydrochloride when decomposed directly with hydrochloric acid, and (c) by the values obtained for the molecular rotations of the salts in dilute aqueous solution. These three facts may be discussed a little, as they are all of considerable importance.

With regard to (a), it is quite possible that an externally com