Having described the products of decomposition, something must be said of the way in which these constituents are combined to form the unaltered tannins. The only tannin of which we have as yet anything approaching complete knowledge is that obtained from galls, sumach, and myrabolans, and which is thence called gallotannic acid. _Gallotannic, or digallic acid_ exists as the principal tannic acid of the galls of oak, tamarisk, &c.; and, in mixture with more or less ellagitannic acid, in myrabolans, divi-divi, sumach, pomegranate rind, and many other plant-products. It has also been formed by Schiff from gallic acid, by mixing it, after drying at 230° F. (110° C.), with phosphorus oxychloride to a thin paste, and heating, first to 212° F. (100° C.) and then to 248° F. (120° C ). Much hydrochloric acid was evolved, and the gallic acid was converted into a yellow powder, which after purification by washing with ether, dissolving in water, allowing the gallic acid to crystallise out, saturating with salt, washing the precipitated tannin in salt solution, and redissolving in alcohol and ether, had all the reactions of purified gall tannin, but was perfectly reconverted into gallic acid on boiling with hydrochloric acid, without the formation of any trace of ellagic acid, or glucose. By analysis of the tannin and its acetyl compounds it was shown to be digallic acid, and its constitutional formula is almost certainly as follows:-- {CO.OH {OH C_{6}H_{2}{ {OH {O } } {CO} {OH C_{6}H_{2}{OH {OH By boiling gallic acid with solution of arsenic acid, Schiff obtained a product which precipitated gelatin, and otherwise reacted like tannin, and he regards this as digallic acid, but other experimenters have failed to obtain digallic acid by this means, and have found that on complete removal of arsenic, the compound was reconverted to gallic acid. It therefore remains a moot point whether digallic acid is really formed, and reconverted into gallic acid by the prolonged action of hydric sulphide, which is necessary to remove the arsenic, or whether, as seems more probable, the supposed tannin is merely an arsenical compound of gallic acid.