be found that great changes have occurred in the constitution of the two liquids. At the commencement, although their total tensions were equal, the proportions in which the various salts were distributed in A, and therefore their partial tensions, were very different to their proportions and partial tensions in B. At the end of the experiment each of the several salts is equally divided between A and B, supposing the volume of A to equal that of B. This experiment shows that the molecules of substances in solution are free to move. They behave like gases. Gases diffuse through a membrane until their partial tensions are the same in the two spaces which the membrane separates. The æther in which physicists picture gases as dissolved offers no resistance to the migration of their molecules; neither does the solvent—water, for example—prevent the movement of salts which are distributed through it. One other illustration of the phenomena of osmosis will suffice to give an idea of the laws by which they are governed. In the case just cited the membrane was permeable to all the salts in solution. When the phenomena of osmosis were first investigated, a distinction was drawn between substances which will pass through membranes—crystalloids—and substances which cannot pass—colloids. We have already had occasion to note that, whereas albumin is a colloid which does not diffuse, its hydrate, peptone, is a crystalloid which does. The term “crystalloid” indicates that substances which can be crystallized are diffusible. Substances which are diffusible are therefore allied to those which crystallize. The nature of the membrane used to test diffusibility was not at first taken into account. Now a distinction is drawn between membranes which are permeable to all diffusible substances, and membranes which are permeable to the solvent, but impermeable to the substances which it dissolves. The latter are termed “hemipermeable.” Imagine now that water is separated from a solution of sugar by a membrane which stops sugar, but is permeable to water. Water will pass through the membrane into the solution of sugar. The level of the solution will rise. Pressure will be needed, and a very considerable pressure, to prevent its rising—to prevent endosmosis, that is to say. The force needed to resist osmosis is directly proportional to the degree of concentration of the solution. If the solution contain 1 per cent. of sugar, a pressure of 500 millimetres of mercury is needed; if it contain 2 per cent., a pressure of 1,000 millimetres; if 6 per cent., of 3,000 millimetres.