Sampling.—Samples should always be drawn from at least 10 sacks or separate parts of the bulk, and, in the case of valonia, special care should be taken to have a fair average quantity of "beard." No attention is usually paid to this point by merchants, and the proportion varies greatly in different parts of the same cargo. If several sacks are spread in layers on a level floor, and then portions going quite to the ground are taken from several parts of the floor, this will be accomplished. Where samples must be dealt with which have not been specially drawn, it might be safest to weigh out from each the same proportion of beard and whole cups, bearing in mind that the beard is always the richest part of the valonia. In sampling myrobalans, it should be remembered that the poor and light nuts will rise to the top, and hence the hand should be plunged well into the sack. Grinding of valonia and myrobalans when practicable is probably best done in a small disintegrator, fitted with gratings. The material, of which some pounds must be used, is screened over a sieve of say 15 wires per in., and all coarser parts are returned to the mill till they will pass. The mill must grind into a close box, that no dust may be lost. Bark may be reduced to fine saw-dust by cutting a portion of each piece in the sample with a circular saw or rasp driven by a lathe. The advantage of these methods is that samples can be ground without previous drying, and thus in many cases time may be saved and separate determination of moisture avoided. When this is not practicable, the sample of some lbs. at least is ground in an ordinary bark-mill, well mixed, spread out flat on a floor or table, and several portions are taken as already described, say 50-100 grm. in all, and dried in a water- or air-oven at 212° F. (100° C.). The moisture is best determined, to save time, in a small separate portion of 10 grm., which must be dried till it ceases to lose weight, and the loss taken as moisture. It must be weighed in a covered capsule, as it is very hygroscopic. When the larger portion of the sample has been dried some hours, it is passed twice through a good coffee-mill, and then returned to the oven till thoroughly dried, for which, 12-24 hours is generally sufficient. Another method sometimes convenient is to take each acorn, or each piece of bark of the sample to be tested, and snip a piece from it with a pair of tinners' shears, taking care that in the case of valonia the section runs right to the centre of the cup; and in bark, that fair shares of the outer and inner layers are taken. The reason for drying before grinding is, that unless hard dried, tanning materials cannot be passed through a small mill. Bark and valonia usually contain 12-16 per cent. of moisture.
Exhaustion.—10 grm. of valonia, 20-30 grm. of bark, or corresponding quantities of other material, are boiled briskly for half an hour with 1 litre of distilled water, a funnel being placed in the neck of the flask, and great care being taken at first to avoid frothing and boiling over. The flasks used should have a capacity of at least 11/2 litre. The whole contents are finally rinsed into a gauged flask, allowed to cool to 59° F. (15° C.), and made up to 1 litre. In the case of sumach, a little more boiling even than this is desirable. This method has been found by the writer to give better results than boiling with successive portions of water. Another method is to boil for 1/2 hour with 250 c.c. of water, then pour the whole on a filter, wash with boiling water so long as a drop of the filtrate blackens paper moistened with a dilute solution of ferric acetate, and finally make up to 1 litre. Many materials, however, clog the filter to such an extent that washing is almost impossible. Kathreiner has used 15 litres of water, and corresponding quantities of material, in a large steam-jacketed copper pan, for each exhaustion, making the weight up finally to 15 kilos., with very uniform and excellent results. (See also [p. 130].) With all materials which deposit ellagic acid or other insoluble derivatives, on cooling and standing, considerably higher results will be obtained if the titration be made as soon as the liquor is cold, than if it be allowed to stand 24 hours; in this respect, a uniform practice should be adhered to. Addition of 1/2 c.c. of glacial acetic acid renders the infusions less liable to change.
Analysis.—Of all the methods which have been proposed for the estimation of tannins, the only one which has met with any general acceptance is that of Löwenthal, and indeed it is the only one which in rapidity of execution and constancy of results is fitted for general use. The method, as originally proposed, depends on the oxidation of the astringent solution by permanganate in presence of indigo, which not only serves as an indicator, but controls the oxidation, limiting it to those bodies which are more oxidisable than indigo. As, however, these include gallic acid and other substances which are useless to either tanner or dyer, it is necessary to remove the tannin, and by a second titration to obtain its value by difference. This Löwenthal (Zeitschrift f. Anal. Chemie, 1877, p. 33) accomplished by a solution of gelatin and common salt, to which, after mixture with the tannin infusion, a small quantity of sulphuric or hydrochloric acid was added. It was necessary to let this stand at least some hours before a clear filtrate could be obtained, and the gelatin remaining in solution had a slight though generally negligible effect on the permanganate. In some cases, even after long standing, perfect filtration was extremely difficult and tedious, and it was also clearly proved by Simand (Ding. Polyt. Jour., ccxliv. 400) that a certain proportion of the tanno-gelatin precipitate, varying with the acid present, and with the species of tannin, remained in solution, and thus gave too low a result. He therefore proposed to revert to the old method of separating tannin with hide raspings, or, as an improved substitute, with the gelatinous tissue of bones, and this is probably the most accurate method, but has the disadvantage of requiring considerable time for its execution. (See also [p. 130].) The writer has therefore tried, and he thinks successfully, so to modify Löwenthal's original method as to increase its accuracy, and at the same time to make it more rapid and easy of execution. It was found that by saturating the clear filtrate with salt, a further precipitate containing tannin was formed, but unfortunately, it was so finely divided that no amount of standing, or even of warming, and repeated passing through the paper, would obtain a clear filtrate. Finally, he hit on the device of mixing with the liquid, before filtration, a portion of the pure kaolin used by photographers. The effect was instantaneous and complete. A perfectly clear filtrate was obtained without any of the tedious waiting which before was necessary, and it was not only free from tannin, but also nearly so from gelatin, so that it only gave the faintest cloudiness with tannin solution. Gelatin gives a more considerable precipitate, but this is simply due to its insolubility in the saturated salt solution, and it is redissolved on dilution with water.[L]
[L] Hunt (Jour. Soc. Chem. Industry, April 1885) states, that saturation with salt causes partial precipitation of gallic acid when present, and that results agreeing more closely with those obtained by absorption with hide are obtained by employing a mixture of 50 c.c. liquor, 25 c.c. 2 per cent. gelatin solution, and 25 c.c. saturated solution of salt containing 50 c.c. of concentrated sulphuric acid per litre and a teaspoonful of kaolin. This approaches very nearly to Löwenthal's original method, but with the addition of the kaolin, and as in it, it is to be feared that a portion of the tannate of gelatin will remain in solution. For accurate work, therefore, absorption by hide-raspings is preferable, though even that has been shown by the writer to remove gallic acid and other matters beside tannin. Hunt states that raw hide also absorbs catechin.
A slight error is introduced by the presence of a trace of oxidisable matter in the gelatin, and when very great accuracy is required, it is well to make a blank estimation of "not-tannin" without tannin infusion, and deduct 1/2 of the permanganate consumed as a correction from the not-tannin; but this may usually be disregarded. Each titration should be made twice, and successive tests should not differ by more than 0·1 c.c. of permanganate.
Reagents.—Solutions are required of (1) Pure potash permanganate, 1 grm. per litre. (2) Pure soda or potash sulphindigotate, 5 grm., and concentrated sulphuric acid, 50 grm. per litre. (3) Pure oxalic acid, 6·3 grm. per litre (decinormal). The sulphindigotate (indigo carmine), must be filtered, and when oxidised by permanganate, should give a pure clear yellow, free from any trace of brown or orange. Any contamination with indigo-purple, which gives brown oxidation-products, is quite fatal to the accuracy of the analysis. The permanganate solution is standardised by measuring 10 c.c. of the (decinormal) oxalic acid solution, adding a little pure sulphuric acid and distilled water, warming to 136° F. (58° C.), and running in the permanganate till a faint permanent pink is produced, for which about 32-33 c.c. should be required. The indigo-carmine solution should be of such strength that 14-16 c.c. of permanganate are required to bleach the quantity employed, which may be 20-25 c.c., as convenient. (4) Gelatin solution: 2 grm. of Nelson's or other good gelatin are allowed to swell in distilled water for two hours, melted by setting the glass in a pan of boiling water, and made up to 100 c.c. This will not keep. (5) Dilute sulphuric acid: 10 c.c. of pure concentrated acid are added to 90 c.c. of distilled water. (6) Good table salt. (7) Purified kaolin.
The analysis is performed in the following manner:—20 c.c. of indigo solution, and 5 c.c. of the infusion of tanning material is added, in a white basin as recommended by Kathreiner, to about 3/4 litre good water, which it is best to measure approximately, so that if it contains any impurity which affects the permanganate it should be constant, and thus be eliminated with the indigo. Permanganate solution is then allowed to drop in, with constant stirring till the pure yellow liquid shows a faint pinkish rim, most clearly seen on the shaded side. This end-reaction, which is of extraordinary delicacy, is due to Kathreiner, and is quite different to the pink caused by excess of permanganate, being an effect common to all pure yellow liquids. It is not needful to make the titration so slowly as has been advised—the permanganate may be dropped in steadily with vigorous stirring, so long as there is large excess of indigo, but as soon as the bottom of the basin can be seen through the solution, it must be added very cautiously, one or two drops at a time, and with occasional pauses, to allow time for its complete mixture through so large a mass of fluid. The titration is repeated twice, and the results added together and denoted by a. Then take 50 c.c. of the infusion, and add 28·6 c.c. of the gelatin solution of Nelson's gelatin of 2 grm. to 100 c.c. After shaking, the mixture is saturated with salt, which brings the volume up to 90 c.c., and 10 c.c. of the dilute sulphuric acid (containing 1 vol. of concentrated acid in 10) and a teaspoonful of pure kaolin are added. It is best to do this in a flask in which it can be well shaken, after which, filtration may be at once proceeded with, although it is safer to let it stand an hour or two: (the flask may be cleansed with caustic soda solution). 10 c.c. of this filtrate (= 5 c.c. of the original infusion) are employed for a second pair of titrations, which are added as before, and the result denoted b. If, further, c be the quantity of permanganate required to oxidise 10 c.c. of decinormal oxalic acid, and 10 grm. of the tanning material have been employed to make 1 litre of infusion, c : (a - b) :: 6·3 : x, where x is the percentage of tannin expressed in terms of crystallised oxalic acid. If it be desired to calculate the gallic acid and non-tannin substances contained in the infusion, the value in permanganate of the indigo alone must be determined. Calling this d, as c is to (b - d), so is 6·3 to the percentage of non-tannins in terms of oxalic acid, and for the present it is best invariably to calculate results in this way, since we do not actually know the relation of any single tannin to permanganate, even Neubauer's number for gallotannic acid being probably too high, according to the recent investigations of Councler and Schroeder,[M] and Oser's for quercitannic being at most only approximate. It happens, moreover, that this last equivalent (62·36 grm. of quercitannic acid = 63 grm. of crystallised oxalic acid) does not differ from that of oxalic acid more than the ordinary limits of error of such estimation, and the substitution is therefore of no commercial importance, while it is surely better to employ a standard which is easily and exactly verified than one which is certain to be modified by further research, and so to run the risk either of having our results made useless for future comparison, or of establishing a false or arbitrary equivalent. What is wanted for practical purposes is not the absolute weight of tannins in the various materials, but only a means for the relative comparison of two samples of the same material; cross comparisons of different tannins being simply delusive. If, however, it is necessary at any time to give actual percentages of gallotannic acid, it is probably best to stick to Neubauer's number for the present, as it is in general use. Neubauer states that 63 grm. of oxalic acid consume as much permanganate as 41·37 grm. of gallotannic acid. Tshekawa found 41·688 as the equivalent for tannin from Japanese gall nuts (Chem. News, xlii. 274). Councler and Schroeder on the other hand give only 34·3 grm. Simand gives 61·1 grm. as the equivalent of quercitannic acid. Commercial "pure tannin" always gives results higher than the truth, as the gallic acid which it contains consumes more permanganate than an equal weight of tannin, or even than the tannin which would yield it if boiled with acid. When this is done the equivalent used should be definitely stated, or it will certainly lead to confusion. Neubauer's equivalent is only properly applicable to gall nuts, and possibly to sumach and myrabolans. For oak bark Oser's number or that of oxalic acid is most likely nearly correct; and this may also be approximately true of oak wood and valonia, but as respects all other materials we have no information whatever, and the oxalic equivalent is as likely to be right as any other. (Compare note, [p. 128].)
[M] From researches by von Schroeder, published since the above was penned, it seems that the permanganate consumed by tannin is largely influenced by the way in which the titration is conducted, see [p. 128].
A few results are given below, not as showing the relative values of the materials, which, of course, cannot be directly compared by any analytical process, but for comparison with those obtained by other methods and modes of calculation:—