Now if we select a few simple figures which are in accord with the above considerations, we can examine the value of the factor a{1}c^(1/n{1}) - a{2}c^(1/n{2}) in a purely illustrative and typical way, and at any rate form some idea as to the manner in which it is likely to vary. The figures might be:-- Ion. | _n_. | _a_. -------------------------+---------+--------- Hydrion _or_ hydroxylion | 20 | 10 Kation of a metal | 15 | 7 Organic anion | 10 | 8 Inorganic anion | 6 | 6 For the sake of simplicity we can assume that these ions are all monovalent. The ions adsorbed by unit mass will then be 10c^(1/20), etc. If these hypothetical adsorption isotherms be plotted as usual we get the fairly typical curves shown in Fig. 1. Now in practice there are always two of these ions, each giving its own specific effect in opposite senses, and the difference ( a{1}c^(1/n{1}) - a{2}c^(1/n{2}) ) represents the nett charge adsorbed. Hence we have the following combinations:-- Inorganic acid 10c^(1/20) - 6c^(1/6) Organic acid 10c^(1/20) - 8c^(1/10) Alkali 10c^(1/20) - 7c^(1/15) Inorganic salt 7c^(1/15) - 6c^(1/6) If we plot these values of nett adsorption against the concentration we obtain the curves shown in Fig. 2. [Illustration: FIG. 2.] On the assumption that the nett charge adsorbed is the dominant factor in determining the maximum swelling at equilibrium, one must therefore regard the curves of Fig. 2 as representing the changes in volume of the swollen gel as the concentration is increased. Now in _type_ these curves correspond to those obtained by experiment from hydrochloric acid, acetic acid, caustic soda, and common salt. The maximum swelling with hydrochloric acid increases rapidly with the concentration at first and then rapidly decreases, though not at such a great rate. The swelling with acetic acid increases less rapidly and to a less maximum, but decreases more slowly. With common salt there is a slight swelling followed by contraction. Caustic soda gives a rapid increase in volume at first, afterwards much less so, and finally yields an exceedingly slow decrease. The correspondence of these facts with the type-curves inevitably suggests that the phenomenon of swelling might be accounted for, in part at least, along these lines.