FILTRATION. 



281 



many hours before the observations are conducted, and the stage of 

 relatively constant rapidity of filtration (a uniform rate is never actually 

 attained) is reached quicker at higher than at lower pressure. This 

 phenomenon is not due to any stopping of pores by particles suspended 

 in the fluids, since it is not noted when filtration is effected through 

 unglazed porcelain, and is probably simply a result of compression of 

 tortuous channels. 



The quantity of filtrate rises with the pressure, but in lower ratio. 



Thus, in an experiment by Tigerstedt and Santesson l of filtration of 

 distilled water through goldbeater's skin (serosa of ox-gut), which had 

 previously been exposed for 95 hours to a pressure of 80 cms. of water, 



The experiments of Wilibald Schmidt 2 showed a contrary result, 

 i.e. that the filtration rapidity increased at a higher ratio than the 

 pressure (possibly due to using dried membranes, the pores of which 

 were opened during experiment), as also did those of v. Kegeczy. 3 



A period of rest, interpolated between two filtration experiments at 

 the same pressure, is found to often cause an increase of the permeability 

 of the membrane above the value it possessed at the time of dis- 

 continuing the first experiment (Eckhard, Euneberg, Tigerstedt and 

 Santesson). 



Thus Tigerstedt and Santesson, 4 in a filtration of distilled water 

 through gold-beater's skin at 40 cm. pressure, observed a filtrate of 

 490 grm. per minute, but after a resting period of 530' the filtrate at 

 the same pressure was - 577 grm. per minute. But whether or not this 

 phenomenon is observed, is probably due to whether or not the elastic 

 limits of the fibres have been passed, " recovery " not being possible if 

 the membrane has been excessively stretched. The interpolation of a 

 period of filtration at lower pressure of course produces the same effect. 5 



The rapidity of filtration rises with the temperature, 6 and, according 

 to Schmidt, the temperature coefficient is nearly that of Poiseuille, for 

 the flow of fluids in capillary tubes. 



The nature of the solution to be filtered must obviously affect 

 both the rapidity of filtration, from differences in viscosity, and also 

 the quantitative composition of the filtrate in relation to that of the 

 original solution. 



The following experiment from Tigerstedt and Santesson 7 may be 

 quoted in evidence of the first point : 



1 Loc. ciL. p. 31. 



2 Ann. d. Phys. u. Chem,, Leipzig, 1856, Bd. xcix. S. 337 ; 1861, Bd. cxiv. S. 337. 



3 Loc. cit. 4 Loc. cit. , p. 30. 5 Runeberg, loc. cit. 



6 Schmidt, loc. cit. ; Eckhard, loc. cit. ; Lowy, Ztschr. f. physiol. Chem., Strassburg, 

 1885, Bd. ix. S. 537. 



7 Loc. cit., p. 42. 



