[LLOYD] TANNIN IN THE LIVING CELL 7 



brown colouration and syneresis sets in (Figs. 1, 3, 7, PI. 1). This 

 always happens if accidentally, or otherwise, some of the tannin cells 

 have been burst. The emulsoid material then gushes forth into the 

 adjoining intercellular spaces, of which on coagulation it forms a 

 cast (Fig. 7, PI. 1) (Lloyd, 1910, 1916). If, on the other hand, these 

 cells are prepared for observation by setting them free in the sap of 

 the fruit (the sap may be more or less diluted with water), the emulsoid 

 swells, sufficiently it may be to rupture the cell wall, when a certain 

 amount of tannin escapes. This is due to the disturbance of the 

 adsorption equilibrium as it occurs in the fruit by the swelling of the 

 emulsoid. The escaping tannin forms with other substances present 

 in the sap, probably pectose, precipitation membranes which present 

 a most bizarre variety of forms (Figs. 4, 6, PI. 1).* Both the tannin 

 masses in situ and the precipitation membranes answer perfectly to 

 that item in the accepted definition of a tannin, which says that it 

 forms an imputrescible compound with albumin, save, of course, 

 that here albumin does not enter in. I have kept a bottle of the 

 separated tannin masses now for over ten years in water, where they 

 wholly retain their original character. 



Although a carbohydrate gel, it differs from agar in its swelling 

 responses to acids, bases and salts. Agar, as I have myself deter- 

 mined, and as amply shown by D. T. MacDougal (1921 and earlier), 

 shows maximum swelling in water as compared with nearly all ad- 

 mixtures of acids, bases and salts. The tannin mass emulsoid swells 

 less in solutions of salt than in water, more in solutions of bases and 

 of acids and less in bases than in acid solutions.^ Exact or even 

 uniformly correct measurements cannot be made with satisfaction, 

 since the cell wall always opposes itself more or less to the swelling 

 contents, but for general comparison it will serve to point out that 

 the tannin masses in the banana are held within relatively firm cell 

 walls, as compared with those in the persimmon. They, nevertheless, 

 swell in ammonium hydrate to 20 to 30 times their volume in water 

 (PI. 2). The swelling in acids and bases is accompanied by a greater 

 disturbance of adsorption equilibrium than occurs in water and by a 

 greater loss, therefore, of tannin. This is the explanation of the fact 

 that an edible fruit may appear quite non-astringent, unless the 

 material is held for a little while in the mouth so as to permit the 

 action of the alkaline saliva. The nonastringency of tannin-bearing 

 fruit at maturity is due to the adsorption equilibrium between un- 



^I have made these observations more recently also on the tannin cells of the 

 fruit of Musa (Lloyd, 1920). 



With regard to the last statement, I must express some doubt. 



