4 '37 
Use of Collodion Osmometers. 
that the osmotic pressure of an electrolyte is not a linear function of the 
concentration, but falls short of a linear relation as the concentration rises. 
Thus the osmotic pressure of half-molar copper sulphate is less than twice 
that of a quarter-molar solution. The following determinations with this 
particular osmometer illustrate this point : 
Concentration 
inside. 
Concentration 
outside. 
Difference. 
Endosmosis. 
Endosmosis -f- 
concentration 
difference. 
n/2 
0 
n/2 
3*3 
3*3 
n /4 
0 
n/ 4 
1.8 
3.6 
n/2 
n/ 4 
n/ 4 
i *5 
3 *° 
Thus a given increase of concentration produces a greater effect on the 
osmotic pressure when the concentration is low than when it is high. 
Attention to this fact enables one to approach more rapidly the concentra- 
tion of an unknown solution. 
A particular determination will now be described. 
A solution, X) of concentration unknown to the experimenter, was pre- 
pared by dilution of the stock half-molar copper sulphate solution. The 
following readings were then obtained : 1 
Inside. Outside. Rise in i hour at 17 0 . 
m/2 water 3.3 cm. 
m/2 * 0.65, 0.7 
Assuming as a first approximation that osmotic pressure is directly propor- 
tional to concentration, we find that the value of x is in the neighbourhood 
of 0*40 molar (— - — x —V In view of the effect of ionization above 
3.3 2/ 
mentioned, this value is too high. A slightly weaker solution than 0*40 
molar was therefore set up, i.e. a 0-38 molar solution. The solution x was 
now put inside the membrane and the 0-38 solution outside, with the follow- 
ing result : 
Temperature at start = 17-0°. 
After 18 hours, rise = 0*3 cm. Temp. — 14-0°. 
The rise in 18 hours (corrected for temperature) was therefore approxi- 
mately 0-75 cm. Calculating as above, we find the approximation to the 
unknown solution to be 0*385 molar. This concentration was now made up 
and tested against the x solution. A slight drop, amounting to 0*2 cm. in 
24 hours, was recorded. Using this correction as before, the final value 
obtained was 0*384 molar. The correct value according to the original 
measurements was 0*383 molar. Thus with three trials the molecular 
strength of the unknown solution was fouqd correct to one unit in the third 
decimal place. In a second experiment, a solution was estimated as 0*368 
molar as against the correct figure of 0*366 molar. 
1 In any determination the whole of the membrane must be completely submerged in the 
outside solution. 
