The widespread distribution in plant tissues of leuco-anthocyanins yielding cyanidin (II, R = OH) and delphinidin (I, R = OH) was demonstrated by Bate-Smith1, using paper chromatography for the identification of the anthocyanidins formed. In those tissues examined from among a large number of plant species, Bate-Smith found that leuco-compounds yielding cyanidin, and to a lesser extent delphinidin, predominate, while other patterns of hydroxylation were almost conspicuously absent. Pelargonidin (III, R = OH) has, for example, thus far been obtained only from the hot hydrochloric acid treatment of the reduction product of acetylated aromadendrin (2:3-dihydrokæmpferol) by Hillis2. The anthocyanidin, 3:7:8:3':4'-pentahydroxyflavylium chloride, generated from the naturally occurring flavane-3:4-diol, melacacidin (from the heartwood of Acacia melanoxylon), by King and Bottomley3, differs from those above in that in nucleus A the 8- in place of the 5-position is hydroxylated. A number of plants also contain flavonoid structures in which the 7-position only is hydroxylated in nucleus A; for example, robinetin, 2:3-dihydrorobinetin, fisetin, fustin, butin, butein, liquiritigenin and 7:4'-dihydroxyflavonol. Some of these are present in very minor proportion in the two economically most important vegetable tanning materials, namely, black wattle or 'mimosa' extract from the bark of Acacia mollissima, and quebracho extract derived from the heartwood of Schinopsis lorentzii4. The polyphenolic fractions of both give resorcinol in high yield from alkali fusions5. From this it is evident that chromatographic identification of anthocyanidins corresponding to (I), (II) and (III), but with the 5-position unhydroxylated, is of importance. Synthetic robinetinidin chloride or 3: 7: 3': 4': 5'pentahydroxyflavylium chloride (I, R = H)6, synthetic fisetinidin chloride or 3:7:3':4': tetrahydroxyflavylium chloride (II, R = H)7, and synthetic 3:7:4'-trihydroxyflavylium chloride (III, R = H)8, together with delphinidin, cyanidin and pelargonidin, were examined by means of paper chromatography, and the spots obtained also examined in the spectrophotometer directly on the paper, by the method of Bradfield and Flood9. Chromatograms were run in a constant-temperature room at 21°C. (Table 1). The 'Forestal solvent' mainly used by Bate-Smith1 (water/acetic acid/conc. hydrochloric acid, 10:30:3 v/v) was unsatisfactory for differentiating between pelargonidin (R_F = 0.71) and fisetinidin chlorides (R_F = 0.73). Their absorption maxima are close (530 and 525 mu, respectively) and their appearance on paper chromatograms very similar. A new solvent mixture consisting of 90 per cent formic acid and 3 N hydrochloric acid (1:1 v/v) differentiated these most effectively (R_F = 0.33 and 0.43, respectively) and gave good resolution of other constituents. 3:7:4'-Trihydroxyflavylium chloride remained at the origin in this solvent system, but moved with the solvent front in the 'Forestal solvent'. Robinetinidin chloride is again better separated from cyanidin and pelargonidin chlorides in the 'Forestal solvent' (R_F = 0.57, 0.50 and 0.71, respectively) than in the new formic/3 N hydrochloric acid mixture (R_F = 0.26, 0.22 and 0.33, respectively). 3:7:8:3':4'-Pentahydroxyflavylium chloride generated from Acacia melanoxylon showed similar R_F to robinetinidin in both solvent systems (see Table 1), but differs from the latter in its absorption maximum, and may further be distinguished by the abnormally large shift in its absorption maximum in the presence of 0.2 per cent aluminium chloride in an ethanolic solution of hydrochloric acid10, a method due to Geissman, Jorgenson and Harborne11. Anthocyanidins unhydroxylated in the 5-position have been generated by Freudenberg and Roux6 from 2:3-dihydrorobinetin (found in Robinia pseudacacia12) by its easy quantitative hydrogenation with platinum oxide in methanol solution to give 7:3':4':5'-tetrahydroxyflavane-3:4-diol, followed by hydrochloric acid/propanol treatment13 of the product under pressure to yield robinetinidin chloride (I, R = H). Fustin (2:3-dihydrofisetin) from Rhus succedanea14 undergoes the same conversions, finally yielding fisetinidin chloride (II, R = H). The synthetic flavane-3:4-diols obtained in both instances by the reduction of the corresponding 2:3-dihydroflavonols are crystalline, in contrast to that occurring naturally, for example, melacacidin3, and that obtained by the reduction of taxifolin (2:3-dihydroquercetin) with sodium borohydride15, both of which are non-crystalline. Sodium borohydride usually gives side-reactions during reduction of dihydroflavonols in aqueous and ethanolic solutions, but nevertheless affords an easy direct method of rapidly establishing the probable presence of 2:3-dihydroflavonols in polyphenolic mixtures, in the absence of leuco-anthocyanins of similar hydroxylation pattern. Thus the methanolic extract of the heartwood of Grevillea robusta contains no leuco-compound, but when preceded by a 5-min. reduction with sodium borohydride in aqueous medium, robinetinidin chloride results from treatment with hydrochloric acid. Quebracho extract ('ordinary'-unsulphited) affords anthocyanidins from hydrochloric acid/propanol treatment under pressure. The predominant one is easily observed as a light red spot on paper chromatograms, having an absorption maximum (523 mu) and R_F values (0.43 in formic acid/hydrochloric acid and 0.73 in water/acetic acid/hydrochloric acid). Comparison of these values with those of fisetinidin in Table 1 (λ_max = 525 mu and R_F = 0.43 and 0.73, respectively) indicates that a leuco-anthocyanin or leuco-anthocyanidin yielding fisetinidin is present in the heartwood of Schinopsis lorentzii.