DIFFUSION IN DEFORMED GELS 



89 



made ; when this was the case, they never detached themselves 

 in the course of the experiment. Discrepancies in the results 

 obtained with the glass masks in the first experiments were 

 found to be due to a small air bubble adhering obstinately 

 to the wall of the hole ; a difficulty which, once diagnosed, was 

 easily overcome. 



If two preparations, e.g. one compressed and the other 

 free from strain, are to be compared, it is of course necessary 

 to employ not only orifices of the same size, but also equal 

 volumes of solution, so as to keep the diffusion " head " the 

 same. For this purpose rubber rings were cemented on the 

 shdes described, which held 0*5 c.c. of liquid. 



The prisms to be examined were cast in a brass tube of 

 square cross-section, corked at one end. They were left in 



Itt. 



rtifiKaii«««aiBiiaiii 



-CQx 



^ 



& 



Fig. I. 



TQr 



T^r 



the mould for at least twenty-four hours, so as to attain the 

 maximum modulus, and were then removed by dipping the 

 mould into hot water for a short time and withdrawing the 

 cork. To harden the surface again, they were then dipped 

 in cold water for a few seconds, dried rapidly with filter-paper, 

 and finally cut into pieces about 50 mm. long. The cross- 

 section of the prisms obtained in this manner was very uniformly 

 28 mm. square. 



The test pieces were always used immediately after prepara- 

 tion. They were compressed in the precision clamps illustrated 

 in Fig. I , made of gun-metal, with the exception of the screw 

 spindle, and of ample rigidity not to be deformed by the very 

 low stresses. The distance between the fixed and movable 

 jaw was measured by micrometer callipers to o-i mm. The 

 internal faces were coated with vaseline or liquid paraffin 



