. JACQUES LOEB 397 



mm, in diameter. One platinum electrode was put into this glass 

 tube and one into the beaker. The distance between the electrodes 

 was approximately the same in all experiments; namely, 6.7 cm. The 

 electrode in the manometer was usually that pole towards which the 

 transport of Hquid occurred so that the rise of level in the manometer 

 could serve as a measure for the volume of liquid transported. In 

 this case the transport occurred against a hydrostatic pressure and it 

 was necessary to compare the rise of level at about the same pressure 

 head in different experiments. 



The volume of liquid transported is a function of the external po- 

 tential difference which was either 50 or 40 volts in our experiments. 

 It turned out that in these experiments a disturbing variable entered 

 inasmuch as with constant voltage the intensity of the current rose 

 slowly and with the rise in intensity the amount of liquid flowing to 

 one of the poles also rose gradually. The writer is inclined to inter- 

 pret this gradual increase in intensity of current as being due to a 

 gradual increase in the number of interstices through which the current 

 can flow; and this means also an increase in the number of capillary 

 spaces through which electrical endosmose can occur. Hence, for 

 measurements of the transport of liquid only those changes in level 

 could be used which occurred after the current had become approxi- 

 mately constant. In order to accelerate this process at the begin- 

 ning of each experiment, a p.d. of 200 volts was used for 2 minutes or 

 less until the intensity of the current was sufficiently high (above 1 .0 or 

 2.0 milKamperes), and then the p.d. was lowered to the 40 or 50 volts 

 desired. The voltage was then kept constant. As a measure of the 

 effect of an electrolyte on the volume of liquid transported we used 

 the rise in the level of liquid in the glass tube during the first 15 min- 

 utes after the current had become fairly constant. The following 

 records will illustrate the way our figures were obtained. We omit 

 the preliminary short treatment of the solution with a current of 200 

 volts, and give only the records for the 50 volt effects. The solutions 

 used were m/512 KCl, m/512 K2SO4, m/512 and m/1,024 K4Fe(CN)6, 

 m/512 CaClo, and m/512 BaCl2. The solutions were almost neutral 

 but slightly on the acid side of neutrality, the pH being about 6.2, 

 and the transport of liquid occurred towards the cathode which was 

 put into the glass tube. Table I gives the time in minutes, the in- 



