X new leucoanthocyanidin, isomelacacidin, has been found with its \$-epirner, melacacidin, in the heartwood of three species of Acacia. Extraction and separation of the epirners is described, and melacacidin obtained pure and crystalline. Isomelacacidin behaves as a reactive p-hydrosybenzyl alcohol and readily forms an ethyl ether by reaction with ethanol, and this distinction from melacacidin facilitates its separation therefrom. The comparative inertness of melacacidin is attributed to an unexpxted conformational stability in its Z(eq),3(as),4(eq)-half-~hair conformation, which inhibits resonance stabilisation of the 4-carbonium ion and so reduces the benzylic character of the 4-hydroxyl group. HEARTWOODS of three Australian Acacia species (A. excelsa, A. hai.pophyZZa, and A. melanoxylon) have each been found to contain melacacidin (0-2-1.0%) and a new leucoanthocyanidin , isomelacacidin ( >0.20/o). Melacacidin was first isolated, in amorphous form, from A. nzeZaizoxyZo?z by King and BottomleyJ2 who proved its structure (I). It was later shown to have the geometrical configuration (11); from comparison of the specific rotations of the free phenol, its 7,8,3',4'-tetramethyl ether, and the methyl ether diacetatewith corresponding values in the (--)-epicatechin series, (-)-melacacidin is more probably represented by (11) than by its mirror image." Melacacidin has now been extracted by a different method, purified by counter-current distribution, and obtained crystalline from both A. excelsa and A. harpophylla; the latter appears to be the best source. Crude extractives from Acacia heartwoods were shown by paper chromatography to contain melacacidin and two other monomeric leucoanthocyanidins, isomelacacidin and O-ethylisomelacacidin. The three compounds were separated by counter-current distribution, but their isolation was simplified after recognition that O-ethylisomelacacidin was an artefact very readily formed from isomelacacidin and the ethanol which had been used in manipulating the extractives. The crude leucoanthocyanidin mixture was boiled with ethanolic 1 % acetic acid, to convert isomelacacidin into O-ethylisomelacacidin, and this was easily separated from melacacidin by counter-current procedures because of its enhanced distribution ratio ; the separatory-funnel procedure described by Bush and Densen proved useful for preparative separations. Hydrolysis of crystalline O-ethylisomelacacidin gave isomelacacidin, which has not yet been induced to crystallise and also hinders crystallisation of melacacidin. The structure of isomelacacidin (111) was inferred largely from the properties of its ethyl derivative, O-ethylisomelacacidin ( IT7), which crystallised as a hydrate but yielded the anhydrous compound C,,H180, when dried, and gave the expected values in the Zeisel and Kuhn-Roth determinations. O-Ethylisomelacacidin crystallised unchanged from aqueous methanol, and on paper chromatograms it was separable from isomelacacidin and O-methylisomelacacidin ; moreover, O-ethylisomelacacidin was separated froin isomelacacidin by counter-current distribution between solvents, so that it is clearly an ether and not merely a solvate. O-Ethylisomelacacidin gave an alkaliinsoluble tetramethyl ether which yielded a toluene-\$-sulphonate, thus disclosing the presence of four phenolic groups and an alcoholic hydroxyl group. The anthocyanidin formed from O-ethylisomelacacidin and hot %+hydrochloric acid was chromatographically indistinguishable from 3,7,8,3',4'-pentahydroxyflavylium chloride similarly derived froin melacacidin, so that O-ethylisomelacacidin is either (IV) or the analogous 3-ethoxy-lhydroxy-compound. The extreme lability of the ethyl group in O-ethylisomelacacidin excludes this alternative and requires the benzylic ether structure (IV). The close relation between O-et hylisomelacacidin, isomelacacidin, and melacacidin was confirmed by acidcatalysed hydrolysis and epimerisation. Thus O-ethylisomelacacidin was rapidly hydrolysed to isomelacacidin by hot water or dilute acetic acid (see p. 4111). Conversion of isomelacacidin into melacacidin was very slow but melacacidin was epimerised rapidly to isomelacacidin (ca. 90?/0) by hot dilute hydrochloride acid. The equilibrium between the two leucoanthocyanidins in aqueous acids was therefore greatly in favour of isonielacacidin, but conversion of melacacidin into O-ethylisomelacacidin was negligible in ethanol containing acetic acid. Isomelacacidin was further characterised as the sulphone (V), a derivative suggested by the investigations of reactive benzyl alcohols by Kenyon and his co-workers5 who found them to undergo reversible formation of sulphones by reaction with sulphinic acids. This sulphone (V) crystallised readily and was easily purified; it gave a penta-acetate, a tetramethyl ether, and a tetramethyl ether acetate. O-Ethylisomelacacidin and toluene-\$ sulphinic acid gave the sulphone (V) when heated in weakly acidic solution, but melacacidin did not react until the acidity was increased to that already known to cause isomerisation. The product from melacacidin was the same sulphone (V), and this establishes the steric identity in melacacidin and isomelacacidin of those positions not affected by acid. The properties of isomelacacidin accord well with those of other benzyl alcohols activated by o- and fi-hydroxyl groups,6 and the reactivity of these alcohols is attributed to resonance stabilisation of the related benzyl carbonium ions. This indicates that the 4-hydroxyl group in isomelacacidin is axial (or can easily become asial) and that the 2(eq),3(ax),4(ax) conformation is preferred-it permits maximum resonance stabilisation of the 4-carbonium ion through coplanarity of the attached groups. The preferred conformation of melacacidin is, however, 2(eq) ,3(ax) ,4(eq) and it is probable that intramolecular hydrogen bonding between the heterocyclic oxygen atom and the 3(ax)-hydroxyl group stabilises both epimersin the conformations mentioned. The comparatively unreactive nature of melacacidin is a consequence of its conformational stability in the 2(eq),3(ax) ,4(eq)-conformation which is unfavourable for resonance stabilisation of the 4-carbonium ion (attached groups not coplanar with the benzene ring). The rapid conversion of m elacacidin into isomelacacidin indicates that epimerisation is probably the first step in formation of the related anthocyanidin from melacacidin and aqueous acid. Formation of the 4-carbonium ion from isomelacacidin occurs very readily and would undoubtedly account for one process of the polyrnerisation to phlobaphenes which accompanies formation of anthocyanidins from leucoanthocyanidins.2 Analogous polymerisations through the 2-carbonium ions formed by ring fission of flavans have been discussed by others. Pel togynol and peltogynol B occur in Peltogyne porphyrocardia,* and teracacidin and isoteracacidin are found together in Acacia interte~ta,~ so that flavan-3,4-diols may frequently, or even generally, occur as mixtures of 4-epimers.