Sols and Gels — Helation to Protoplasynic Structure 89 



dry powder when crushed with a stout glass rod. Left to itself, the 

 powdered mass returns again to its original viscous liquid state. ^"^ 

 A technical product, Nuodex Calcium-S (a colloidal solution of 

 about 10 per cent calcium naphthenate in a petroleum distillate), 

 shows a similar behavior.^ "^ A complete curve of one of these 

 colloidal systems corresponding to curve 3 in Figure 1 has not yet 

 been measured. 



The behavior of coarse particles of suspensions under the micro- 

 scope gives us a clue concerning that property of the particles which 

 makes a suspension dilatant or thixotropic.^''- ^'' The particles of a 

 dilatant suspension are quite independent of each other; there is 

 not the least indication of coagulation. If they are settled on the 

 slide, they are all separated from each other; if dislodged and driven 

 into the liquid by a slight knock, they remain separated in Brownian 

 movement until they have settled again. On the other hand, the 

 particles of a thixotropic suspension are always found to be coagu- 

 lated to a certain degree, sticking together and forming clusters. 

 If brought into suspension in the liquid, they may be temporarily 

 separated from each other, but they always unite again to clusters. 



This distinctive behavior allows us to understand many results 

 obtained with thixotropic or dilatant systems. For instance, in 

 order to make a concentrated iron oxide sol thixotropic a small 

 amount of a coagulating electrolyte like NaCl must be added, an 

 amount much smaller than that needed for actual coagulation.-" 

 In agreement with the positive C-potential of the iron oxide par- 

 ticles and with the Schulze-Hardy rule, the anions are especially 

 effective. Smaller concentrations of polyvalent anions are necessary 

 to obtain the same state of thixotropy, which is characterized by the 

 same time of re-solidification to a gel after the original gel has been 

 liquefied to a sol by shaking. The thixotropic state, therefore, is 

 often considered to be a primary stage of coagulation. 



The following experiments done by W. Heller prove this concept 

 more quantitatively-^: If an iron oxide sol, which is just too dilute 

 in iron oxide to give a thixotropic gel on adding a certain amount 

 of electrolyte, is centrifuged after the electrolyte has been added, 

 a more concentrated, gelatinous sediment is accumulated at the 

 bottom of the vessel. This sediment is not formed on using the same 

 centrifugal force, if no electrolyte has been added. The sediment 

 can be reversibly redispersed in the liquid by shaking. If the sedi- 

 ment is separated from the less concentrated liquid on top, it is 

 found to be a normal, i. e., reversible, thixotropic gel. From the 



