The filtration of a fluid, by virtue of its gra vity, through a porous lamina, the capillary canals of which are very minute, is not readily appreciable, unless the inferior or outer surface of this porous plate is kept plunged in or moistened by the same fluid. It is in this way only that the filtration of fluids through animal membranes, the texture of which is dense (a piece of bladder for example,) becomes appre ciable. It is essential that the inferior aspect of the membrane be bathed with the same fluid as that which rests on its superior aspect, in order that no foreign cause modify its filtration. We know in fact that the heterogeneousness of two fluids, by producing endosmosis, would completely mask the effects of simple filtration. Would I, then, try the filtration of water through a membrane, I apply this membrane to an en dosmometer, which I fill with water to a certain height in the tube of the instrument; I next apply the lower surface of this membrane to the surface of a body of water placed below it. The water contained in the endosmometer filters through the membrane and mingles with the water in the vessel below ; the amount of this filtration in a given time is indicated by the fall of the column in the graduated tube of the instrument. Would I essay comparatively the filtration of any watery solution, I place this solution in the same endosmometer, and taking care to keep the exterior of the membranous part of the instrument in contact with a'solution of the same nature, situated below it, I observe the degree to which the depression of the co lumn in the tube takes place in a space of time equal to that which was taken by the filtration of the water. It is necessary to begin by proving the filtration of water; after this the filtration of the watery solution may be tried ; but it is always to be borne in mind that the membrane of the endosmometer must have been kept plunged in the watery solution about to he experimented on for at least a quarter of an hour, in order that it may become tho roughly impiegoated with the solution, and to secure that this should take the place of the water which the membrane had formerly con tained in its pores. Without this measure of precaution, the results of the second experi ment would be faulty. It is also indispensa ble that the circumstances under which the two experiments are performed are in all re spects exactly alike. It was in this way that I proceeded to ascertain comparatively the capa city of filtration of water to that of a watery solution of oxalic acid through a piece of blad der. I found that the filtrating power of rain water, at the temperature of + 21 cent. being denoted by 24, the filtrating power of a watery solution of oxalic acid of no greater density than 1.005, (1.2 of acid to 100 of solution,) was denoted by 12. A solution of the same acid, of the density of 1.01,beingtried, its filtratingpower was found to be represented by 9. By these ex periments it is therefore proved that water tra verses an animal membrane more readily than a solution of oxalic acid. Why then does the latter solution traverse an animal membrane more readily and in greater quantity than water, when it is water which is in contact with the surface of the membrane opposite to that which is in contact with the acid ? This is a question which I find it impossible to answer in the present state of our knowledge.
The discovery of this singular property of the oxalic acid to cause the endosmotic current to flow towards the water when separated from the latter fluid by a lamina of animal mem brane, led me to imagine that all the acids would be found to possess a similar property. And this I ascertained, in the first instance, to be the case in regard to the tartaric and citric acids. Both of these acids are much more so luble in water than oxalic acid. The saturated solution of oxalic acid at + 25 cent. has no higher a density than 1.045 (11.6 acid to 100 of the solution.) But the solubility of the tar taric and citric acids is such that their watery solutions may have a density of far greater amount. I tried the endosmotic effects of the tartaric and citric acids in watery solution of various density, and I discovered, not without surprise, that very dense solutions of them and solutions of inferior density exhibited endos motic phenomena in inverse ratios. Thus, when a solution of tartaric acid was of a den sity above 1.05, (t1 crystallized acid in 100 of solution,) and it was divided from water by an animal membrane, the temperature being + 25 cent. the endosmotic current is directed from the water towards the acid ; but when, under the same circumstances, the density of the acid solution is below 1.05, the current of endos mosis is directed from the acid towards the water, just as we have found it to be with refe rence to the oxalic acid. Consequently, ac cording to its greater or less density, tartaric acid presents the phenomenon of endosmosis in two opposite directions. At the mean density of 1.05, at a temperature of + 25° cent. it exhibits no obvious endosmotic phenomena whatever; not that there is not reciprocal pe netration between the acid and the water, which are divided by the animal membrane ; but this reciprocal penetration takes place so equally on either side, that there is no increase of bulk of the one fluid at the cost of the other—there is no endosnzosis. The citric acid exhibits pre-. cisely the same phenomena; the point of mean density, which divides its two opposed endos motic capacities, is also very nearly the same, namely, 1.05 at a temperature of + 25° cent. These facts induced me to imagine that if the oxalic acid alone presented the endosmotic cur rent directed from the acid towards the water, this arose from the fact of its solution at + 25° cent. falling short of the density necessary to
permit the acid solution to cause the endosmo tie current to flow from the water towards the acid.
The preceding observations were made during the heats of summer. The centigrade thermo meter was standing at + when I determined the mean term if density of the solution of tar taric acid, above and short of which the endos mosis happening between this solution and water is directed towards the acid. It was of importance to know whether a depression of temperature would cause any modification in these phenomena. I therefore repeated the same experiments when the temperature was + 15° cent. and I was astonished to find that the mean term of density, of which we have spoken above, was considerably altered, being made to move in the direction of the increase of density of the acid solution. Thus the mean term of density being 1.05, (11 crystallized acid to 100 solution,) at a temperature of + 25° cent. it came to be 1.1, (21 acid to 100 solu tion,) at a temperature of + of the same scale ; that is to say, the solution of tartaric acid, which now occupies the mean term, con tains nearly twice as much acid as the solution which stood at the previous mean term, when the temperature was ten degrees of the centi grade scale higher. This first essay was enough to lead to the inference that the mean term o/ density, which we are now discussing, would undergo further alterations in the same sense with further depressions of temperature ; and this was actually found to be the case. At a temperature of sio cent. the solution of tarta ric acid, of the density 1.1, was no longer the solution of mean density dividing the two op posed endosmotic currents, as it was when the temperature was + 15° cent. This solution then caused the endosmotic current to flow freely towards the water. I had to increase its density to 1.15 (30 acid to 100 solution) to come to the new mean term, beyond which the current of endosmosis was directed towards the acid, and within which it was directed towards the water. With the temperature depressed to a quarter of a degree cent. above zero, the solution of tartaric acid, of the density of 1.15, no longer presented the mean term ; this solution now occasioned endosmosis towards the water, which indicated that the mean term was to be sought for in a more dense solution of tartaric acid, and this I actually found in a solution of the density of 1.21 (40 acid to 100 solution). Every solution of this acid of greater density than 1.21, at the temperature of *th of a degree above zero cent caused the endosmotie cur rent to flow from the water towards the acid, and every solution of the same acid, under the density of 1.21, caused the endosmotic current from the acid towards the water. From all these experiments it follows that a fall of tem perature favours the endosmosis towards the water, and that a rise of temperature favours the endosmosis towards the acid. fact, the same solution of tartaric acid occasions at one time endosmosis towards the acid, when the temperature is high ; at another, endosmosis towards the water when the temperature is re latively low. I t would appear from this, that a depression of temperature renders the solu tion of tartaric acid more apt than water to permeate animal membranes, and that there is a certain concordance between this capacity of permeation and the temperature and the den sity of the acid solution. This phenonemon, at first sight, appears analogous to that which N. Girard discovered,* in regard to the com parative flow of a solution of nitre and of pure water through a capillary glass tube. M. Girard found that, at a temperature of a solu tion of one part of nitrate of potash in three parts of water flows more rapidly than pure water through a capillary glass tube, whilst the same solution flows more slowly than water when the temperature is above + To discover whether this apparent analogy was well founded or not, I made an experiment to ascertain the relative duration of the flow through a capillary glass tube of a given mea sure of pure water, and a like measure of a solution of tartaric acid, the density of which was 1.05 (21.8 parts acid, 100 solution.) The temperature being + 7° cent. I found that fifteen centilitres of water flowed through a ca pillary glass tube in 157 seconds ; but the same quantity of the solution of tartaric acid required 301 seconds to pass through the same capillary tube. There is consequently no ac tual analogy to be established between the re sults of the experiments of M. Girard and the fact of the endosmosis towards the water, which takes place when at a temperature of +7° cent. a solution of tartaric acid of the den sity of 1.105, is separated from a volume of by a piece of an animal membrane. It may be as well if I here state that when a solution of one part of nitrate of potash in three parts of water was separated by a piece of bladder from pure water, I have always ob served the endosmotic current directed towards the solution ; the temperature might be at zero, nr 10°, or higher, the same phenome non always occurred. This is sufficient to prove that endosmosis is governed by laws en tirely different from those that preside over simple capillary filtration. I add, that the solution of tartaric acid, of 1.105 density, hay ing a viscidity nearly the double of that of water, and passing, nevertheless, by ciidesnio sis into the latter fluid, when it is separated from it by an animal membrane, and the tem perature is + 7° cent. also proves that endos mosis does not generally depend on the visci dity of fluids.