150 PROTOPLASM 



filters. In both of these cases, the ultramicroscopic pores of the 

 fine colloidal structure of a gel serve as the ultrafilter. 



Most livmg membranes are ultrafilters, the Bowman capsule 

 of the kidney being an example. It serves in permitting certain 

 substances of low molecular weight (uric acid) to pass from the 

 blood into the convoluted tubules of the kidney yet holds back 

 the larger colloidal particles of the blood proteins. It is difficult 

 to escape the opinion that colloidal jellies used as ultrafilters 

 owe this property to a fine porous, or sievelike, structure. 

 This structure may be due not to granules, as in the case of 

 silica, but rather to the intermeshing of long and fibrous crystal- 

 line units. That this is true is indicated by that extraordinary 

 behavior of gels known as thixotropy. 



Thixotropy. — Thixotropy is a most interesting and little under- 

 stood property of gels, first thoroughly studied by Schalek and 

 Szegvary in the laboratory of Freundlich in Berlin. An iron 

 oxide sol is prepared after the manner of Graham (by adding a 

 solution of iron chloride to one of ammonium carbonate until the 

 flocculation is no longer redissolved; the mixture is then dialyzed 

 and a 6 per cent FezOs sol, ferrum oxydatum dialysatum, results). 

 To the ferric oxide sol a few milligrams of sodium chloride are 

 added, which causes the sol to set to a gel in a short time. This 

 gel may be broken down simply by mechanical agitation, by 

 shaking or stirring, and it then has the remarkable capacity to 

 regelatinize into a gel as firm as the original one; the process 

 may apparently be repeated any number of times. The gel is 

 so firm that the beaker containing it may be turned upside down 

 and the gel remain within. The sol — the liquid form — is so 

 fluid that it flows like a thin oil. The phenomenon has been 

 named thixotropy. Borjeson, working in Svedberg's laboratory 

 in Sweden, observed the similar behavior of an alcoholic cadmium 

 sol of only 0.2 per cent concentration. The simplest experiment 

 in thixotropy is that done with bentonite or colloidal clay. If a 

 few milligrams (the amount need not be exact) of sodium chloride 

 is added to a test tube full of 10 per cent bentonite and vigorously 

 shaken, a colloidal dispersion results which will set to a gel in a 

 few minutes and can then be broken down by shaking, after 

 which it will again set. The process may be repeated any 

 number of times. The mechanism by means of which a thixo- 

 tropic gel may repeatedly reform after breaking down is not 



