It is well known that, in separating water from alcohol by an organized animal or vege table membrane, the endosmotic current flows from the water towards the alcohol. I employed oil-silk (taffetas gaming) or silk covered with a layer of caoutchouc, which may be regarded as equivalent to a thin lamina of elastic gum, as the medium of separation between these two fluids. During the first thirty-six hours of the experiment, 1 observed an extremely slow en dosmotic current from the alcohol towards the water. After this period the endosmosis, with the same direction, became very rapid. This increase in the rapidity of the endosmosis I considered due to some alteration in the caout chouc produced by the action of the alcohol, and in consequence of which it became more readily permeable. The endosmotic current, however, let it be observed, is always from the water towards the alcohol in this experiment, instead of being from the alcohol towards the water, as is constantly the case when the septum between the spirit and the water is formed by an organic, whether animal or vegetable, tissue. We have thus a clear demonstration of the great influence possessed by the septum upon the direction of the current of endosmosis. We have, also, in the instance just quoted, a proof that the different degrees of viscidity of two liquids plays no part in the production of this phenomenon. I would remark that the endos motic current carrying the alcohol towards the water athwart the septum of caoutchouc is ac companied by a counter-current, which carries the water towards the alcohol through the same septum. I assured myself that the alcohol had received some addition of water; and yet it is well known that caoutchouc is impermeable to water ; which would seem to say that the latter fluid could only have passed through the sep tum of caoutchouc by becoming mingled with the alcohol occupying the molecular interstices of that substance. Once within these intersti ces the alcohol attracts the water by the affinity of mixture, (a(nite de mistion) and enables it to penetrate the substance of the caoutchouc, which deities all access to water when it is pure. It is therefore to the state of commixtion within the capillary tubes of the septum that the two opposed fluids proceed the one towards the other with cross hut unequal motions. The means I took to ascertain the fact of water having become mixed with the alcohol was simple enough : I set fire to a quantity of the fluid which had served for the experiment, and found that, after all the spirit had burned out, a considerable quantity of water remained, whilst the alcohol, previously to being so em ployed, burned away entirely, leaving no water behind it.
The theoretical views of Magnus in regard to endosmosis have been adopted by Berzelius in his Chemistry, and the idea upon which they are based has been reproduced by Al. Poisson. To give a clear notion of this theory, let us sup pose that a measure of salt water is separated from a measure of pure water by a permeable septum, a piece of bladder for example ; the current of endosmosis, in this instance, will be from the pure water towards the salt, and for the following reason : in the salt water there are three attractions, namely, the attraction of the molecules of the water for one another ; secondly, the attraction of the molecules of the salt for one another; and thirdly, the reciprocal attraction of the molecules of the water and of the molecules of the salt. The pure water on the opposite side of the septum again has no more than a single form of attraction, to wit, that of its particles for one another. The salt water subjected to three attractions will be moved, it may be imagined, with greater diffi culty than the pure water, the molecules of which are obedient to but one attraction. Con sequently, in the reciprocal attraction of these two fluids, the one, the molecules of which are the least subjected to attraction among them selves, will make its way with greatest rapidity athwart the capillary conduits of the dividing membrane.
This theory has a seducing aspect, but we shall find immediately that it is inapplicable to certain endosmotic phenomena presented by acids.
I have shown above that it is not always to wards the denser fluid that the endosmotic cur rent is turned. Thus alcohol and ether are very much less dense than water, and yet it is towards these fluids of inferior density that water flows in endosmotic experiments. Alco hol and ether have this in common with dense fluids generally, that they rise to a less height in capillary tubes than water. From this ob servation I was led to imagine that the endos motic current was always from the fluid having the greatest power of capillary ascension, to wards the fluid having the least of this capa city. It is true, indeed, as we have already
seen, that alcohol proceeds by endosmosis to wards water when the medium dividing them is caoutchouc. This would seem to say that alcohol would rise higher than water in capil lary tubes of caoutchouc; and it is certain that caoutchouc has a greater attraction for alcohol than for water, inasmuch as the surface of India-rubber is much more readily wetted by alcohol than by water, which only adheres to it partially and imperfectly. This fact, there fore, would not be in contradiction to my theory ; although I must confess that it is not reconcilable with certain endosmotic pheno mena presented by the acids, as we shall imme diately have occasion to perceive. I n spite of this, however, I do not think 1 pass in silence all the proofs that seem to establish this theory upon a basis of sufficient solidity ; for I cannot but perceive that it is applicable to the most general phenomena of endosmosis, phenomena, too, which the acids, like all other fluid bodies, exhibit, although they also present endosmotic phenomena in addition of a diffe rent nature, and which belong to them exclu sively.
Inequality of density being one cause pf en dosmosis among fluids, it became a point with me first to ascertain what differences in power of capillary ascension resulted from determi nate differences of density among fluids ; and next, to discover whether the difference in power of capillary ascent of two fluids bore any constant ratio to the difference of endos mosis as it is proclaimed by experiment.
The height to which different fluids rise in capillary tubes depends on a variety of causes, in appearance very different, but which must have some fundamental analogy. Of all fluids water is that which rises highest ; and sub stances held dissolved in it which increase its density, lessen its power of capillary ascent, which is also diminished by increase of tempe rature : hot water ascends a less way in a capil lary tube than cold water. Combustible fluids, such as alcohol and ether, are like dense fluids in regard to power of capillary ascent ; so that combustibility acts in the same manner as den sity in this respect. The matter of which ca pillary tubes are formed is also endowed with the power of modifying the capillary ascent of fluids. Thus water, at the same temperature, will not rise to the same height in a series of equal capillary tubes made of different mate rials. These multiplied elements, which enter into the determination of the capillary ascend ing power of different fluids, render it an ex tremely complicated phenomenon. To simplify the study of this phenomenon in the greatest possible degree, let us confine ourselves to the use of two fluids, namely, water and a solution of the hydroehlorate of soda. It is easy to try the latter fluid of different densities, and to compare the power of capillary ascent pos sessed by each of these with that of pure water at like temperatures. The same glass tube will answer for these comparative experiments. Be fore detailing these experiments, however, I have one important remark to make, which is this ; that the layer of fluid which moistens, internally, the canal of a tube is one of the elements of the capillary ascension which this tube effects. Thus, water will rise to a de terminate height, in a tube interiorly moistened with water; but if the interior of the tube be moistened by a saline solution, or by any other watery fluid, or by alcohol, pure water will not again rise so high in this tube as when it was moistened by water only. It will be in vain to attempt to cleanse the tube by passing water repeatedly through it ; water will never detach the stratum of saline or other liquid which ad heres to it, and which diminishes its power of producing capillary ascension. To detach this stratum of fluid you must pass a !Worm body repeatedly through the tube full of water ; it is only by the rubbing of this body that the stratum can be detached. It must be evident after this observation, that in making experiments on the power of capillary ascension with various fluids and with the same tube, it is necessary to cleanse this tube with great care before each experiment ; without this we should have de fective results. We must also take care not to warm the tube by holding it between the fingers, for if the temperature be increased it will no longer exert so strong a capillary attrac tion. Let us now pass to the detail of these experiments.