56o SCIENCE PROGRESS 



cylindrical platinum kathode inside and a cylindrical copper 

 anode outside. Simultaneously, the interior was rilled with 

 N/io ferrocyanide and the outer space with N/io copper sul- 

 phate. An electric current under a pressure of no volts was 

 applied during two or three hours until a maximum resistance 

 was reached, the interior being rinsed out with fresh ferro- 

 cyanide every two or three minutes to remove the alkali set 

 free by the electrolysis. The cell was then rinsed and soaked 

 during one to three days and the process repeated. It was 

 found essential to deposit the membrane at a temperature not 

 lower than that at which the cell was to be used ; similarly 

 it was advisable to measure the osmotic pressures first at 

 higher and afterwards at the lower temperatures. The mem- 

 branes were tested with weight-normal sugar solutions, with 

 membrane-formers of N/io concentration, the course of the 

 meniscus in the manometer being carefully watched to detect 

 irregularities of motion due to the breakage and repair of the 

 membrane in the pores. The electrolytic treatment and tests 

 were repeated over and over again until the behaviour of the 

 cell was satisfactory, the membrane-formers in the osmotic 

 tests being finally reduced to o'oi osmotically normal concentra- 

 tion. The minimum time required to form a cell was a month 

 but the operation often occupied three or four months, all the 

 essential operations being carried out in thermostats. In testing 

 the membranes it was not sufficient to secure steady pressures : 

 no reliance was placed upon a cell in which the maximum 

 pressure for the given concentration was not developed. When 

 the cell had been passed as satisfactory, no experiment was 

 accepted in which the concentration of the solution was not 

 perfectly maintained. This was found to be no mere ideal but 

 a test that could be applied rigidly to every measurement 

 recorded. In one experiment, a cell, not specially selected, 

 a constant pressure of 12*522 atmospheres was maintained during 

 sixty days, the range of fluctuation being almost exactly equal to 

 the range of atmospheric pressures during this period. 



In a new cell the maximum osmotic pressure might be 

 reached in as little as six hours ; in an old cell, with a greatly 

 thickened membrane, as much as ten days might be required. 

 The old membranes were perfect in their osmotic qualities but 

 were rejected because of their slow action. This rendered 

 them tedious to use and greatly increased. the lag in recovering 



