59.]

DIFFUSION OF LIQUIDS.

89

at a constant temperature. At the end of the time the rate of diffusion was ascertained by drawing off successive layers of the liquid in the jar by means of a fine syphon. The open end of the short limb of the syphon was kept in contact with the surface of the fluid, and successive measures of 50 cub. centim. (or of the whole contents of the jar) were drawn off into separate vessels and evaporated or analysed. In this way the amount of diffusion into each successive layer of liquid was ascertained.

(59) Laws of Diffusion of Liquids. From experiments conducted upon this principle several important conclusions have been deduced :—

In

i. It is found that by employing solutions of the same substance, but of different degrees of strength, the quantities of the substance diffused in equal times, are, cæteris paribus, proportioned to the quantity in the solution. For example, four different solutions of common sait, in water, were prepared, containing respectively, 1, 2, 3, and 4 parts of salt to 100 parts of water. eight days' time the quantities diffused were, in the first solution, 2.78 grains; in the second, 5'54 grains, or just double the amount; in the third, 8:37 grains, or three times the quantity; and in the fourth, IIII grains, or almost exactly four times the amount diffused from the first solution.

ii. No direct relation is observable between the density of a solution and its diffusibility, but the quantities of the substance diffused from solutions containing equal weights of different bodies vary with the nature of the substance, as will be seen by reference to the following table. The solutions in each case contained 20 parts of the solid, dissolved in 100 parts of water, and were exposed for eight days at a temperature of 16° C.

The extreme slowness with which albumin becomes diffused is remarkable; and is no doubt connected with its functions in the

90

EQUI-DIFFUSIVE GROUPS OF SALTS.

[59.

animal system, where it is present so abundantly in the serum of the blood and in other important liquids.

On comparing together the times in which different substances are diffused in equal quantities, some remarkable numerical relations were discovered, and a close parallelism was observed to hold between the phenomena of liquid diffusion and those which accompany the diffusion of gases (67).

It has been found that saline substances may be arranged in groups, the members of each group being equi-diffusive, and the rates of diffusion in each group being connected with the rate of diffusion of the other groups by a simple numerical relation. Equi-diffusive groups coincide in many cases with isomorphous groups, but are often more comprehensive. The relations of the most important of these equi-diffusive groups may be pointed out, as follows:

The

The first group contains hydrochloric, hydriodic, and hydrobromic acids; perhaps also nitric acid. These acids are the most diffusible substances known. The second group contains potassic hydrate, and probably ammonia. The third group, nitrates of potassium and ammonium, chloride, bromide, iodide, and chlorate of potassium, and chloride of ammonium. fourth, sodic nitrate, chloride, bromide, and iodide. The fifth, potassic sulphate, carbonate, and ferrocyanide, as well as ammonium sulphate; probably also the normal and acid chromate, acid carbonate, acetate and ferricyanide of potassium. The sixth group contains sodic sulphate and carbonate; and the seventh, magnesic and zincic sulphates. The nitrates of barium, strontium, and calcium also form an equi-diffusive group.

On comparing together the squares of the times in which equal quantities of these different salts are diffused, these numbers exhibit a very interesting ratio to each other, which is illustrated by the following table. In the first column of figures the relative diffusibility of the different groups is given as compared with the hydrochloric acid group; the second shows the times required for the diffusion of equal weights of the individuals composing each group; and in the third is shown the ratio of the squares of those times of equal diffusion,

And

It has been observed that in the case of gases (67), the squares of the times required for the diffusion of equal volumes are to one another in the inverse ratio of their densities. hence it has been inferred by analogy that the molecules of these salts as they exist in solution, possess densities which are to each other as the squares of their times of equal diffusion: that, for example, the solution densities of hydrochloric acid, potassic hydrate, and potassic nitrate, are as 1 : 1·62: 3*24.

All experiments on the diffusion of liquids proceed with. greater regularity in dilute solutions: as the liquid approaches the point of saturation the uniformity of action is interfered with, by the tendency to cohesion of the particles of the solid.

iii. The quantity of any substance diffused from a solution of uniform strength increases as the temperature rises: for example, the rate of diffusion of hydrochloric acid increases as follows:

at

Graham supposed from his early experiments that the ratio of diffusion between different bodies, if compared at the same temperature, remains constant, whatever the temperature which the comparison is made; but subsequent experiments led him to the conclusion, that the more highly diffusive the substance, the less does it gain in diffusiveness by rise of temperature.

iv. It is found that if two substances which do not combine chemically, and which possess different degrees of diffusiveness, be mixed in solution, and be placed in a diffusion cell, they may be partially separated by the process of diffusion, the more diffusible one passing out the more rapidly; the salt which is least

92

OSMOSE.

[59.

soluble having, however, its diffusiveness somewhat reduced in proportion to the other. Upon this fact Graham observes, 'The mode in which the soil of the earth is moistened by rain is peculiarly favourable to separations by diffusion. The soluble salts of the soil may be supposed to be carried down together, to a certain depth, by the first portion of rain which falls, while they afterwards find an atmosphere of nearly pure water in the moisture which falls last, and occupies the surface stratum of the soil; diffusion of the salts upwards, with its separations and decompositions, must necessarily ensue. The salts of potash and ammonia, which are most required for vegetation, possess the highest diffusibility, and will rise first. The pre-eminent diffusibility of the alkaline hydrates may also be called into action in the soil by hydrate of lime, particularly as quicklime is applied as a top-dressing to grass lands.'

In some cases even chemical decomposition may be effected by the process of liquid diffusion. Thus, if a solution of ordinary alum (which is a compound of potassic and aluminic sulphates in fixed proportions) be placed so as to become diffused into water, the potassic sulphate will pass out more rapidly in proportion to the quantity present than the aluminic sulphate.

v. Provided that the liquids be dilute, it appears that one substance will become diffused into water already containing another body in solution, just as into pure water; but the rate is materially reduced if a portion of the diffusing substance be already present in the surrounding liquid.

In comparing with these the phenomena of gaseous diffusion (67), it will be seen how closely all these points coincide in the

two cases.

FIG. 35.

(60) Osmose.-Intimately connected with the process of liquid diffusion are the changes which occur when the two liquids are separated by the intervention of a porous diaphragm. The phenomena here are, however, more complicated, from the part exercised upon the result by the adhesion of the two liquids to the material of the diaphragm. The process of mixture will go on in this case notwithstanding the direct opposition of gravitation.

The following experiment exhibits this fact in a striking manner :-Provide a funnel, or a small jar (fig. 3, open at top and bottom, and furnished with a long, narrow stem: over the open mouth of the jar tie a piece of moistened bladder; fill the jar and a portion of the stem with spirit of wine (or with a solution of sugar in water), then place the jar, with its broad

61.]

ENDOSMOSIS AND EXOSMOSIS.

93

end downwards, in a shallow vessel containing water, noting the height at which the spirit or the solution stands in the stem. In the course of a few hours the column of liquid will be found to have increased in height, and if sufficient time be allowed, it will have risen to the top of the tube, and will at length overflow. This phenomenon has been explained in the following manner:

Owing to its greater adhesion to water than to spirit, the bladder is easily moistened by the water in contact with its lower surface, whilst the spirit above wets the bladder with difficulty; the water rises into the bladder by capillarity, and fills its pores; it thus reaches the upper surface, where it comes into contact with the spirit; a true liquid diffusion of the water. through the spirit then commences (owing to the combination between the two liquids); a fresh portion of water rises from below into the pores of the bladder to supply the place of that which has been removed, and thus the liquid within the funnel is constantly increasing in bulk, until at length, even in opposition to gravity, the liquid overflows; this flowing in of the liquid was termed by Dutrochet, who first particularly examined it, endosmosis (from evdov, inwards, and ouòs, impulse). At the same time that this action proceeds from without inwards, a very small quantity of spirit is passing out by a similar process into the water below, and this flowing out of the vessel is designated exosmosis. Upon this view the conditions essential to the phenomenon are the more complete adhesion of the bladder to one liquid than to the other, and the existence of a certain degree of chemical attraction between the two liquids. Whenever these conditions are realized, no matter what the liquids may be, the liquid which most freely wets the membrane passes out more rapidly than the other passes in. If a film of collodion, which is more easily wetted by alcohol than by water, be substituted for the bladder in the foregoing experiment, the direction of the osmose will be reversed, and the alcohol will pass into the water more rapidly than the water into the alcohol.

The foregoing explanation, although it is probably true for the particular experiment with alcohol and water, is however inadequate to explain the phenomenon generally, which is one of continual occurrence, and is of importance, especially when viewed in its physiological bearings: the investigations of Graham (Phil. Trans. 1854, 177) have also proved it to possess considerable interest in a purely chemical sense.

(61) Conditions of Osmose. The osmometer used in these experiments is represented in fig. 36. It consists of a bell-jar, ▲, of a capacity of 5 or 6 ounces (from 150 to 200 cub. centimetres), over the open mouth of which a plate of perforated zinc is placed, and over this is securely tied a piece of fresh