284 Dutrochet on the 



sion, in relation with the water. This phenomenon is, therefore, 

 unquestionably produced by the excess of capillary ascension 

 of one of the liquids separated by the partition of the endos- 

 mometer. 



Hence endosmosis is the result of the opposition of two 

 unequal capillary forces acting at the two extremities of the 

 same capillary tube. These two forces impel the two opposite 

 liquids towards each other in unequal quantities, so that one 

 of them (that which has the smallest force of capillary ascen- 

 sion) is gradually augmented in volume ; and it is this excess 

 of capillary force which produces the endosmosis. Having de- 

 monstrated that the endosmosing liquid is impelled towards the 

 liquid with which it unites, in a quantity proportionate to the 

 excess of its capillary ascension over that of the liquid towards 

 which it is so impelled, it remains to be explained by what 

 mechanism this phenomenon is produced. It appears certain 

 that, under these circumstances, there are two opposite cur- 

 rents, the one strong and the other weak : these two currents 

 may be seen by putting muriatic acid into a glass endosmome- 

 ter, and plunging it into a glass vessel filled with water. We 

 shall then see the water rise, forming striae in the acid, and the 

 descend, forming similar striae in the water. It cannot be 

 supposed that these two opposite and unequal currents pass 

 simultaneously through the same canals ; but it is possible 

 that each capillary canal serves alternately to transmit the two 

 opposing currents, and the following fact appears to prove it to 

 be so. I put nitric acid into an endosmometer closed with a 

 piece of bladder, and added to it some very small fragments of 

 gold leaf. I then plunged the apparatus into water. The en- 

 dosmosis was immediately produced, and I perceived that the 

 water which it introduced into the acid raised rapidly some of 

 the fragments of gold leaf, whilst others remained stationary 

 against the membrane. The fragments of gold leaf, which had 

 experienced an ascending impulse, fell again by their own 

 weight upon the surface of the membrane : they remained 

 motionless there for an instant, and were then again strongly 

 impelled upwards : all the fragments of gold leaf presented 

 these alternations of rapid ascension and fall, followed by a 

 short repose on the membrane, but without being at all simul- 



