Graham found [Footnote: _Phil. Transact._, 1861, 183.] that the _tannin
molecule_ is of considerable size, since its diffusion velocity is 200
times less than that of common salt. Paterno [Footnote:
_Zeits. phys. Chem._, 1890, iv. 457.] was the first to determine the
molecular weight of tannin, employing Raoult's method; he found that
tannin in aqueous solution behaves like a colloid and that hence
Raoult's method is not applicable. When, on the other hand, he dissolved
tannin in acetic acid, results concordant with the formula of
C_14H_10O_9, corresponding to a molecular weight of 322, were
obtained. Sabanajew [Footnote: _Ibid._, 1890, v. 192.] later determined
the molecular weight of tannin in aqueous solution as 1104, in acetic
acid solution as 1113-1322, Krafft [Footnote: _Ber._, 1899, 32, 1613.]
as 1587-1626 in aqueous solution. Walden [Footnote: _Ibid._, 1898,
3167.] determined the molecular weight of tannin-schuchardt as
1350-1560, tannin-merck as 753-763, digallic acid as 307-316 (calculated
322). Feist [Footnote: _Chem. Ztg._, 1908, 918.] determined the
molecular weight of tannin as 615 and one of his own preparation as 746,
Turkish tannin as 521 and Chinese tannin as 899. In this connection it
should be noted that the calculated molecular weight of pentagalloyl
glucose, which in E. Fischer's opinion forms a substantial part of the
tannin molecule, is 940, but Fischer also thinks that this compound
possesses a much higher molecular weight.

STRUCTURE OF TANNIN--The oldest structural formula of tannin is Schiff's
digallic acid formula:--[Footnote 1: _Ber_., 1871, 4, 231.]

---------CO.O.----------
^ ^ OH
| | | |