this may not be a sharp line of division, it is highly probable that typical members of two classes exhibit these properties of true solutions in such a different degree as to make the differentiation an important one. Unfortunately, however, colloidal mixtures have not yet been satisfactorily enough investigated with respect to these properties to enable a classification to be based exclusively upon them.
Colloidal Solutions.
Let us now consider the characteristics of the two classes as manifested by typical representatives, beginning with the colloidal solutions. These substances are, for the most part, obtained directly from animal or vegetable sources and are purified by dialysis. Among the most important are gelatine, agar-agar, unheated albumen, caramel, starch, dextrine, and many natural gums.
A number of the important properties of these colloidal solutions have already been alluded to, but some of them deserve further consideration. The contrasts and similarities between them and ordinary solutions may first be mentioned. Such colloids possess a much slower rate of diffusion, a much smaller osmotic pressure, and a much slighter influence on the vapor-pressure, freezing-point, and boiling-point of the solvent than do corresponding weights of crystalline substances. So small are these effects that whether they exist at all is a question to which much investigation and discussion have been devoted. The now existing experimental data seem to show, however, that the gelatinizing non-coagulable colloids do possess these properties and influences in an appreciable degree. The results of the osmotic pressure determinations in the cases where it has been measured against a parchment or animal membrane, which would not retain the mineral impurities, are especially significant. Thus by this method it has been found that a 6 per cent, glue solution exerts a pressure of about one third of an atmosphere, and that a 10 per cent, solution of the colloids of blood-serum produces one of 40 mm. of mercury. Further investigations in this direction would be of great value. The results of Graham, too, seem to leave no doubt as to the existence of diffusion; he found, for example, that albumen diffused one seventh as fast, and caramel one fourteenth as fast as cane-sugar. Thus, these colloids exhibit the same properties as ordinary dissolved substances, but in a lesser degree—a fact which is explained in accordance with the modern theory of solutions by the simple assumption that they are true solutions, but that their molecules consist of aggregates of the ultimate chemical molecules and are, therefore, of much greater weight and complexity than those of non-colloidal substances.
This assumption seems, however, of itself alone, scarcely sufficient to account for the abnormal viscosity of these colloids, their power of
