156 SCIENCE PROGRESS. 



ning of the experiment was - o'i8q° ; at the end of the three 

 hours' experiment it was — ot JJ° C. — corresponding to a 

 loss of 6 per cent, of the dextrose. In the case of the 

 urea, the freezing-point at the beginning was "189°, and at 

 the end was- 0*154° C, a loss of 18 per cent. Here then 

 the initial rate of osmosis of the glucose was about five 

 times that of the urea ; the loss by diffusion of the glucose 

 was about one-third that of the urea. In the case of the 

 sodium chloride the loss amounted to 22 per cent. ; but 

 here the total difference of osmotic pressure was very 

 nearly double that of the other two solutions, and the result 

 is that the initial rate of osmosis of the sodium chloride takes 

 an intermediate place between that of urea and that of 

 glucose. 



In this paper the results of another experiment are 

 given to show that osmosis may occur from a fluid having 

 a higher final osmotic pressure towards a fluid having a 

 lower final osmotic pressure. If, for example, equimolecu- 

 lar solutions of sodium chloride and glucose be separated 

 by a peritoneal membrane, the osmotic flow will take place 

 from the fluid having the higher final osmotic pressure — 

 sodium chloride. We might compare with this experiment 

 the results of separating hydrogen at one atmosphere's 

 pressure from oxygen at two atmospheres' pressure by 

 means of a plate of graphite. In this case the initial result 

 will be a still further increase of pressure on the oxygen 

 side of the diaphragm — a movement of gas against pres- 

 sure taking place in consequence of the greater diffusion 

 velocity of hydrogen. 



So far we have only considered the behaviour of solu- 

 tions when separated by a membrane, the permeability of 

 which to salts is comparable to that of water ; so that the 

 passage of salts through the membrane depends merely on 

 the diffusion rates of the salts. There can be no doubt, 

 however, that we might get analogous movements of fluid 

 against total osmotic pressure determined, not by the 

 diffusibility of the salts, but by the permeability of the mem- 

 brane for the salts — a permeability which may depend on a 

 state of solution or attraction existing between membrane 



