Physical Cause of Endosmosis. 285 



taneous. This fact proves evidently that the capillary canals, 

 which transmit the water impelled towards the acid, or the 

 acid impelled towards the water, do not contain a current con- 

 stantly moving in the same direction. The fragments of gold- 

 leaf, when they fall after their movement of ascent, are natu- 

 rally directed towards the canals, which serve at the moment 

 for the current descending from the acid towards the water ; 

 they remain motionless on this spot until the canals upon the 

 orifices of which they are placed change their descending into 

 an ascending current, when they resume their upward motion 

 under the influence of the ascending liquid. Without these 

 alternations of the two opposing currents in the same capillary 

 canal, it is not easy to conceive how the endosmosis can be 

 equal to the excess of the capillary action of the opposite 

 liquid. In order to explain the phenomenon, we should have 

 to admit either that the two opposing currents existed simulta- 

 neously in the same canal, which is impossible, or that the num- 

 ber of the capillary canals of the membrane is equally divided 

 between the two liquids, and consequently between the two cur- 

 rents, and there appears no reason to suppose the existence of 

 this equal division. But, as experience proves that the capillary 

 canals do riot transmit the same liquid without discontinuance, 

 it becomes necessary to admit that the same canal serves alter- 

 nately for the two opposite currents : this is the only manner 

 in which we can understand the exact relative proportion which 

 exists between the quantity of liquid which is accumulated and 

 the excess of the capillary action of that liquid over that of its 

 antagonist. If, in opposition to what appears to be proved by 

 experience, we assume that there is but one current through the 

 membrane, and that the opposite current is but an optical de- 

 ception, the result of the affinity of mixture of the two liquids, 

 then the effect of endosmosis would be explained by the oppo- 

 sition of two unequal forces of capillary impulsion ; part of 

 the greater force would be employed to counterbalance the 

 smaller force, the effect of which it would thus suspend or 

 neutralise, and there would only remain the excess of the 

 greater force of capillary impulsion over the smaller force to 

 produce endosmosis. In this view of the case, it would not 

 be necessary to admit that the same capillary canal serves 



