It may also be prepared by evaporating the alcoholic extract, and extracting with water, which leaves the phlobaphenes, or higher anhydrides undissolved. The first anhydride, which is partially soluble, may be precipitated by the addition of salt, and the quercitannic acid extracted by shaking the filtered solution with acetic ether. In very dilute alcohol it yields a pure yellow precipitate with lead acetate. In aqueous solution the precipitate is light-brown. It gives a blue-black with ferric salts. When pure, quercitannic acid yields nothing to pure ether or to benzene. If quercitannic acid be heated to 266°-284° F. (130°-140° C.) it loses water, and yields a red anhydride slightly soluble in water, which constitutes the red colouring matter of oak bark, and which has also been called "difficultly soluble tannin." It precipitates gelatin and is one of the class which Eitner well designates "tanning colouring matters." It gives a brownish red with lead acetate, and a blue-black with ferric salts; it is difficultly soluble in water and ether, but readily so in alcohol of all strengths. Together with other anhydrides, it exists naturally formed in the bark. At higher temperatures, or by boiling with acids, a series of higher anhydrides may be obtained which are quite insoluble in water, but are soluble in alcohol and caustic alkalies. No glucose is produced by treatment of pure quercitannic acid with acids, that formed by so treating oak-bark extract being due to the alteration of the lævulose present. If oak-reds are fused with potash they yield, according to Johansen (Chem. Soc. Jour., xxxii. 721), protocatechuic and butyric acids. If heated in sealed tubes with dilute hydrochloric acid, gallic acid only is formed, with evolution of methyl chloride. The constitutional formula of quercitannic acid is as yet very uncertain: it is probably a methyl derivative of digallic acid. Its formula is variously given as C_{17}H_{16}O_{9} (Etti, Chem. Soc. Jour., xliv. 994), C_{28}H_{26}O_{15} (Ibid. xl. 901, Löwe), C_{19}H_{16}O_{10} (Böttinger, Berichte, xvi. 2710). For further details the original memoirs must be consulted. For the reactions of oak-bark and valonia infusions the Table, p. 113, may be consulted. Before describing the catechol-tannins it will be necessary to speak of a group of compounds of which it is probable that these tannins are decomposition products. These are the catechins. It is as yet by no means certain how far they should be considered a group, some chemists holding the opinion that there is only one catechin, of which the rest are merely impure preparations.