392 THIRD REPORT — 18SS. 



Strongly attracted fluid in contact with it, is supposed to be 

 greater than the attraction of the latter particles for themselves. 

 By reason of this last supposition the fluid will, according to La 

 Place, form an internal coating of the capillary tube, Avhich in- 

 ternal coating will itself act as a capillary tube upon the fluid 

 next in contact. This new capillary tube, attracting the par- 

 ticles of the opposite fluid more strongly than its OAvn, will in 

 turn be coated by a thin stratum of that fluid, which may again 

 be regarded as a fresh capillary tube ; and this alternation may 

 extend itself to the very axis of the original tube, the opposite 

 fluids thus forming a series of interlacing cylinders one within 

 another. 



That cylinder of fluid which lies in immediate contact with 

 the tube, must be supposed to become mixed with the opposite 

 fluid, either before it issues from the tube, or at a distance be- 

 yond it less than the sphere of sensible attraction. Were this 

 not the case, the tube acting equally at its two extremities 

 would produce no effect ; and the action of endosmose (so far 

 at least as it depends on the tube,) would cease, allowing the or- 

 dinary mixing process time to extend itself within the tube ; 

 after which the motion would recommence. 



The maximum force is obtained by supposing this mixture to 

 be completed within the tube, at a distance from the issuing ex- 

 tremity not less than the sphere of sensible attraction ; and this 

 is the force Avhich it is our object to compute, for it is plain that 

 at the moment when the endosmose ceases from the increase of 

 pressure within the vessel where the fluid accumulates, the or- 

 dinary mixing process will be at liberty to extend itself within 

 the tube ; and, in experiment, it is by means of these ultimate 

 pressures that the forces are compared. 



The principle of the computation is this : — the process being 

 supposed to have gone on for some time, the effective attractions 

 are found by taking the sums of the attractions due to the dif- 

 ferent substances regarded as coexisting (each with a dimi- 

 nished density,) within the same identical space. Expressions 

 for the forces are thus obtained in terms of these partial densi- 

 ties. The fluids being supposed to suffer no penetration of di- 

 mensions in consequence of mixture, the partial densities, which 

 are four in number, are readily expressed in terms of the tvfo 

 actual or total densities, and the constants depending on the 

 initial state of the fluids ; and thus we are able to express the 

 forces in terms of the actual densities and the constants above 

 mentioned. 



The expression for the maximum force of endosmose is thus 

 found to be composed of two terms ; the first being propor- 



