It may be well to remark here that the question of hydration plays an important and imperfectly understood part in the chemistry of tannin. When hemlock tannin is reduced to a thick extract by evaporation in the vacuum-pan, it appears to part with some of its combined water, and a portion only of the tannin is soluble in cold water, the remainder being precipitated as an insoluble anhy dride or phlobaphene, which cannot be dissolved even by subsequent boiling. If. on the other hand,: water of ae high a temperature is employed as has been used in the evaporation of the extract, the whole is brought again into solution, unlees the extract has been thickened at too high a temperature. This is true of most tanning extracts besitles hemlock.
It will thus be obvious that our knowledge of the ohetuistry of the various tannins is very imperfect, and it is useless to try to fill up the gaps by mere speculation.
As regards analytical examination of mixtures of tannins, a quantitative separation of the different species is as yet quite impossible ; but qualitatively, tannins may at least be detected when not in too complicated mixture. The following table gives distinctive reactions of some of the principal tannins and their derivatives.
Quantitative Determination.—Marly processes have been proposed for the quantitative estimation of tannins, but it cannot be said that any method yet known is wholly satisfactory. The oldest, that of Sir H. Davy, recently improved by Stoddart and others, consists in precipitating with gelatine, and drying and weighing the precipitate. This was almost impossible to filter off as directed by Davy ; but by the use of a little alum, and by pouring hot water on the precipitate, it becomes curdled into a mass which may be washed by decantation. As the precipitate contains varying quantities of tannin, according to the strength of solution employed ; aa it ia soluble in excess of gelatine solution, and as it is almost if not quite impossible to wash it free from gelatine an.1 alum, the method can hardly lay claim to much accuracy. A somewhat better one consists in the employment of a standard solution of gelatine with a little alum, determining the end of the reaction by filtering off a portion and ascertaining if another drop of the reagent produces a further precipitate. This method is very tedious, the end reaction is difficult to hit, the standard solution is very unstable, it is inapplicable to gambier and catch because the mixture will not filter clear, and its results are irregular, probably from the power of tannin to combine with various propor tions of gelatine. A third plan, which has a most seductive appearance of simplicity, is that of Hammer ; he takes the ap. gr., then absorbs the tannin with slightly moistened hide-raspinga,
again takes the sp. gr., and from the difference calcukttes the percentage of tannin, a difference of 5 per cent. of tannin corresponding to one of 1.020 sp. gr. (20° barkometer). Unfortunately the hide is more or lees soluble in the liquor, and absorbs aeids other than tannic with considerable energy ; the moistening of the rsspings introduces an error, snd the smallness of the quantity to be measured makes a slight error completely vitiate the reaults. With extreme care, due corrections for temperature, for the water introduced with the raspings, and for their solubility, and by sub stituting evaporation of the. infusions to dryness for mere calculation from their sp. gr., the method is useful as giving almost the only information obtainable as to the actual weight of tannin in any material capable of being absorbed by hide. It is, however, only suitable for use as a check on essier and more rapid methods, such as Lowenthal's, which give accurate relative results, but no information as to absolute weight of uoknown tannins. A modification of Hammer's method has been introduced by Miintz and Ramspacher, in which the liquor whence the tannin is to be removed is forced. through a piece of raw hide by pressure. This method, except that it is more rapid, has all the evils of Hammer's original in an intensified form, and gives such variable results as to be qu;te useless iu practice.
Several other methods have been proposed : such as Gerland's, with a volumetric solution of tartar emetic, used in presence of ammonia chloride ; Fleck's, by precipitation with acetate of copper, and subsequent washing with ammonia carbonate and gravimetrie estimation, either of the tannate dried at 100° (212° F.), or of the oxide of copper left on ignition ; and Carpcne's, by precipita tion with ammoniacal zinc acetate, and subsequent estimation with permanganate and indigo. These, though giving fairly accurate results on some tannins, are only of limited application. They may be passed over; as well as Jean's method with a volumetric solution of iodine in presence of °odic carbonate, and Allen's method with acetate of lead, which, though accurate, are somewhat tedious and difficult, ancl preseut no advantage over Lowenthal's improved process. This last is easy of execution, constant in results, and universally applicable. Before proceeding to describe it in detail, it may be well to give some hints as t,o the best modes of sampling and preparing tanning materials for analysis, since this is often more difficult and tedious than the actual analysis.