In the next experiment separate two solutions, A and B, by a hemipermeable membrane. Let A contain one salt only—X; let B contain several salts—X, Y, Z. Water will pass from A to B, or _vice versa_, unless the osmotic pressure of the salts which the solutions contain is the same. The osmotic pressure will be found to be the same if the total number of molecules dissolved in A equals the total number of molecules dissolved in B. If in A there be N molecules of X (per unit volume), and if in B there be nX, n′Y, n″Z, the osmotic pressure will be the same provided n + n′ + n″ = N. This, it will be seen, is a very different matter from equality of percentage composition. Some molecules are light; others are heavy. The percentage weight of X + Y + Z in B may be very different from the percentage weight of X in A. To estimate the osmotic pressure of a mixed solution, it is not sufficient to add together the percentages of the various salts which it contains. “Concentration,” in the sense in which it was used in regard to blood and urine, refers to the number of molecules of dissolved substances in a given volume, not to their weight. It would be undesirable to attempt in this place to enter upon the theory of osmosis. Enough has been said to suggest to the reader that he should, when endeavouring to apply its laws to the explanation of physiological phenomena, bear the following facts in mind: Some membranes are permeable to water and to the crystalloids which it dissolves; others, although permeable to water, are impermeable to substances in solution. Some substances are diffusible through permeable membranes; others are not. Osmosis of water occurs from the solution of lower to the solution of higher concentration. Diffusion of crystalloids is their escape, owing to their own molecular movements, from a situation in which they are denser to a situation in which they are less dense. It must be added, however, that various circumstances prevent the reduction of the laws of osmosis to simple terms—the tendency of salts to dissociate when in solution, their bases and acids acting as independent “ions,” is an example of the complications which produce apparent departures from these laws. It must further be added, and with emphasis, that, important though it be that anyone who attempts to explain the interchanges which occur between the various fluids of the body should be conversant with the laws of osmosis, it is impracticable, and in some cases misleading, to rigidly apply them. Living membranes and dead membranes do not necessarily control diffusion in the same manner. Still less do the laws which govern diffusion through dead membranes hold good, without qualification, to living cells.