21 



of age and rate of shear. An illustration of this general and well- 

 known beha\aour is the case of sodium behenate, where occasionally 

 a perfectly clear and apparently homogeneous solution can be divided 

 by pourmg into two test tubes, one of which will then be found to 

 contain practically all of the behenate, the other containing almost 

 only water. 



Freundlich in his study of certain vanadium pentoxide sols which 

 had been aged during many years, found that at boundaries or 

 throughout the sol, whenever the sol was set in movement, it was 

 anisotropic and exhibited all the behaviour of a crystalline liquid, 

 a behaviour which is likewise now attributed to the presence of long 

 molecules, both on physical and chemical grounds. This explanation 

 therefore links up with Bose's theory of swarms of long molecules 

 and with Vorlanders' observations that long molecules are required 

 for the formation of Uquid crystals. Sandquist's bromophenanthrene- 

 sulphonic acid is obviously a colloidal electrolyte, analogous to soap ; 

 and solutions can be obtained under certain conditions as crystalline 

 liquids. 



These strings of molecules which we contemplate may be microns 

 or millimetres long, in other words, they consist of innumerable 

 molecules placed length-wise and their formation would, of course, 

 be ascribed to residual affinity. The very strong tendency of soaps 

 to form such structures is proven by the manner in which they form 

 filamentous and fibrous curds (see below). We investigated several 

 sulphonates containing up to 10 carbon atoms, but as these are ring 

 compounds, the development of ionic micelle was only very slight, 

 and the solutions crystalhsed without tendency to form fibres. 



The ultramicroscopic investigation of soap gels presents some 

 difficulty, since it is difficult to know when it is a real gel that is being 

 observed. Probably all of Zsigmondy and Bachmann's-^ work (like 

 our own) refers to curds. A gel is probably optically clear apart 

 from the particles which result from hydrolysis and which are, of 

 course, devoid of motion on account of the high viscosity of the 

 soap solution. Nevertheless in one undoubted instance, we observed 

 and photographed a cloudy gel of sodium oleate which, in addition 

 to curd fibres, contained an exceedingly fine and deHcate filamentous 

 network. The quartz surfaces in contact with these sodium oleate 

 gels often exhibit indefinitely long and exceedingly fine filaments 

 just on the limits of visibility and just capable of being photo- 

 graphed with long exposures. Their regularity of form and texture 

 is astounding. They simulate living matter in their appearance; 

 they may take the form of a simple sine wave, or of regular waves 

 with higher harmonic series superimposed on them. Any one part 

 of such a filament is identical m form and structure with any other 

 part. They are probably derived from originally straight just 

 resolvable translucent tubes containing very regularly spaced whitish 

 dots or lengths. 



The beautiful pioneer observations of Zsigmondy and Bachmann 

 on soap curds and our own studies, in which the cinematograph was 

 combined with the ultramicroscope in order to follow the genesis 



