OSMOTIC PRESSURES OF SOME CONCENTRATED AQUEOUS SOLUTIONS. 491 
(2) Determination of the Turning Point .—As already explained, a gradually 
increasing pressure is put upon the solution until the level of the water in the gauge 
ceases to fall and begins to rise. If there were no corrections to apply, the pressure 
at which this turning point takes place would give the equilibrium pressure ; but on 
account of the guard-ring leak, and the solution leak (see later), the turning point is 
not, in general', the pressure required. 
The method therefore adopted is to take rates of movement of the level in the 
gauge, corresponding to slightly differing pressures, both below and above the turning 
point; any small correction can then easily be applied. 
The operation, then, is first to determine the guard-ring leak ; after this the 
apparatus is withdrawn from the ice and the water outside the porcelain tube run out, 
and, after washing with a portion of the solution, it is filled with that solution—with 
very viscous solutions it is advisable to use a force pump. The washing and the 
filling should be carried out as rapidly as possible, for it must be remembered that 
until the equilibrium pressure can be put on the solution, the latter is in contact with 
the membrane and is diluting itself by abstracting water from the inside of the tube. 
When full, the apparatus is replaced in ice and connected to the beam pressure 
apparatus. Weights are then put on the pan, and at the same time the beam is kept 
horizontal by means of the wheel and screw and its plunger. The weights are put on 
the pan by increments of 2 and 3 kilos—this represents 8 or 12 atmospheres—at 
short intervals of time ; a very rapid addition of weights might strain the membrane, 
while the sudden application of the equilibrium pressure would almost certainly 
damage it. When within about 10 per cent, of the equilibrium pressure, the weight 
increment is reduced and the length of time between the changes is increased; for it 
is found that the water, which had been sucked into the solution during the filling, 
will, on account of imperfect diffusion, now be squeezed out again, and it was feared 
that a rapid flow of water through the membrane might be detrimental. The weights 
are added in this manner until the turning point is about reached ; then an interval of 
an hour is allowed to elapse to give time for the temperature throughout the 
apparatus to become steady at 0° C., and also for the more dilute stratum of solution 
near the membrane to diffuse. 
In the later experiments we proceeded as follows:—-The dead-weight gauge, 
previously adjusted to give a pressure close to that already on the solution, is turned 
on and the beam apparatus cut off. The weights are then rapidly adjusted until the 
rate of movement in the water gauge is as nearly as possible equal to the guard-ring 
leak. If the exact pressure cannot be applied by the weights we use (the increment 
of weight is equivalent to 0'34 atmosphere), a rate slightly above or slightly below 
the turning point is taken. The dead-weight gauge is then cut off and the beam 
apparatus turned on. 
At intervals of an hour the same process is repeated until we are satisfied that no 
further change in the rate corresponding to a given pressure will take place. Two 
3 R 2 
