2 8o DIFFUSION, OSMOSIS, AND FIITRATION. 
noted, however, that animal membranes are known in which the 
resistance to the passage of fluid is quite different in opposite directions. 
The most familiar example is the shell membrane of the egg, which 
permits filtration far more easily from within outwards than in the 
reverse direction, 1 and the same is true of the skin of the frog. 2 
Filtration. 
By filtration is meant the passage of fluid through a membrane, as a 
result of a difference of hydrostatic pressure on the two sides. To filter 
water from a solution across a membrane into another solution, the 
difference between the hydrostatic pressures on the two sides must 
exceed the difference between the osmotic pressures of the solutions, in 
the case where the higher osmotic and hydrostatic pressures are on the 
same side. If a porous pot bearing a semipermeable membrane is 
filled with a solution and immersed in the pure solvent, the pressure 
necessary to produce filtration of the solvent from the solution is one 
just exceeding the full osmotic pressure of the solution ; but where we 
deal with permeable membranes, as those in the body, the necessary 
pressure is far less, because the difference of osmotic pressure on the 
two sides of the membrane is reduced by the diffusion of some of the 
dissolved substance. 
Experiments on filtration through animal membranes appear to have 
given very contradictory results, which seems to be due to the fact that 
not only does continued pressure upon such membranes vary their 
permeability, but a certain amount of " recovery " takes place in the 
intervals between use ; hence the conditions of the membranes have been 
by no means uniform in the experiments of different observers. 
If the passage of fluid through an animal membrane is by more or 
less tortuous paths, and if the walls of these are more or less elastic, it 
is obvious, not only that by continued pressure must the resistance to 
filtration rise, but that on removal of the pressure a slow " recovery " will 
take place. Deformation of the membrane, if simply tied over a tube, 
must also, unless it is properly supported, tend to distort channels and 
so increase resistance to filtration. To these sources of difference in the 
experiments must be added, previous drying of the membrane or not, 
the condition of imbibition of the fibrous tissue, and temperature 
variations. All these sources of differences are, moreover, accentuated 
by the fact that most of the experiments have been made with thick 
membranes, such as pericardium, bladder, intestine, and ureter. 
Nearly all who have studied filtration have found that at constant 
pressure the amount of the filtrate falls off with time, 3 but, as pointed 
out by Tigerstedt and Santesson, 4 this falling off is far more rapid in the 
earlier hours of an experiment than later, so that it is advisable in all 
such experiments on filtration to expose the membrane to pressure for 
1 Meckel quoted by Ranke, " Physiologie des Menschen," 1872, S. 122. 
- Cima, Mem. d. Accad. di Torino, 1853, vol. xiii. ; Reid, Journ. Physiol., Cambridge 
and London, 1890, vol. xi. p. 312. 
3 Liebig, " Untersuch. ii. einige Ursachen der Saftebewegungim thierisehen Organismus," 
Braunschweig, 1848, S. 7; Bcitr. z. Anal. u. Physiol. (Sekhard), Giessen, 1858, Bd. i. S. 
95 ; Runeberg, Arch. d. Heilk., Leipzig, 1876, Bd. xviii. S. 58 ; Gottwalt, Ztschr. f.physiol. 
Chem., Strassburg, 1880, Bd. iv. S. 423 ; v. Regeczy, Arch. f. d. ges. Physiol., Bonn, 1883 
Bd. xxx. S. 544. 
4 Bijhang. till k, Svens. Vct.-Akad., Stockholm, 1886, Bd. xi., No. 2. 
