490 
THE EARL OF BERKELEY AND MR. E. G. J. HARTLEY ON THE 
The brass tube serves to protect the glass from the pressure exerted by the 
dermatine ring L, when it is compressed by the nut and washer. 
It will be observed that by this arrangement the outside surface of the porcelain can 
be subjected to pressure while the inside is only under atmospheric pressure, and if the 
wrought-iron vessel be connected to one terminal of the battery and the platinum 
wire to the other, a current can be passed through the porcelain and the solutions. 
Pressures of about 130 atmospheres were used. 
The Operation of Determining the Equilibrium Pressure. 
There are three separate operations required for the purpose of determining the 
equilibrium pressure. 
(1) Guard-ring Leak. —-As explained in our preliminary paper, the semi-permeable 
membrane is never quite on the surface of the tube, and, therefore, it is impossible to 
get perfect contact between the dermatine rings (see fig. 1) and the membrane. 
The result of this is that a small quantity of solution is forced out at this point. 
In the old apparatus there was always a noticeable part of the membrane not 
subjected to pressure, over which this solution flowed and, therefore, abstracted water 
from the interior. In the new apparatus, by arranging that the dermatine rings 
overlap the ends of the tube, we have reduced the exposed membrane to a minimum; 
there* are, however, minute cracks in the glaze at the ends of the tubes, and the 
solution, flowing over these, draws water from the interior. To ensure that the rate 
at which the water is abstracted is constant during an experiment, the metal 
tubes YV (see fig. 1) were provided and were filled with enough solution to cover 
the ends of the porcelain. 
The amount abstracted, that is, the guard-ring leak, was determined before every 
observation in the following manner. The tube was set up in the osmotic apparatus 
and the latter was filled with water, filling the porcelain tube as well; the pressure 
tube T (see fig. 1) was replaced by an open glass tube, which was filled with water 
to the same level as that of the water in the capillary gauge ; the whole apparatus 
was then placed in ice, and some of the solution whose equilibrium pressure was 
afterwards to be determined was poured into YV. 
When the temperature, indicated by a thermometer placed in mercury in a hole 
bored some distance into the osmotic apparatus, had fallen nearly to 0° C., and was 
constant at that point, repeated measurements of the rate at which the water in the 
szauo’e fell were made until constant rates were obtained; the latter were taken to be 
due to the guard-ring leak. 
The guard-ring leak, which in our previous work had been equivalent to a rate 
such as would be produced by a pressure of two or three atmospheres on the solution, 
was now found to be very nearly negligible and corresponded to about 0T5 of an 
atmosphere ; it was also nearly constant for any one tube, as will be seen on reference 
fo the table of results, 
