RATE OF DIFFUSION OF IODINE IN KI 505 



except when specifically stated otherwise. As in the case of density, 

 viscositj^ is usually given in terms of percent concentrations, and has to be 

 changed to normal or molar terms before it is of real usq,for comparison. 

 The only data of interest for us that could be found when this work was 

 being done was that of Taylor and Rankin, who measured the viscosity 

 of KI solutions at the following concentrations, 1 n, 2 n, and 3 n (37). 



In view of the meagerness of the available data we determined the 

 viscosity of various KI solutions (see table VI ), and from this data the 

 curves on Plates I and II are drawn. If we examine Plate II, we see that 

 for most salts (ZnS04, LiCl, and NaCl are shown) the fluidity rapidhj 

 falls as the concentration increases, that both density curves and fluidity 

 curves are almost straight lines; and hence the one varies inversely as the 

 other. For KI the fluidity curve is pecuhar, it first rises and then falls, 

 being fairly sjinmetrical and having a maximum at about 2.5 normal 

 concentration. 



If we examine the diffusion curve (Plate I) in comparison with the 

 fluidity curve we see that it rises, as was to be expected, with the fluidity 

 curve; but 7nuch more rapidly. Further, at about the concentration at 

 which the fluidity curve shows a maximima, the diffusion curve becomes 

 almost flat. Thus we see that diffusion can not depend on viscosity alone 

 for our case, as was thought probable when the work was begun. It is 

 interesting to note that so long as the fluidity curve rises, the diffusion 

 curve rises more rapidly than either the fluidity curve or the density curve; 

 and that as these two curves begin to go apart the diffusion curve ceases 

 to rise and becomes flat. This seems to indicate some relation between 

 the densitj^ and the rate of diffusion similar to that between fluidity and 

 the rate of diffusion. (Of course here density and concentration go hand 

 in hand.) 



Degree of dissociation of KI3. Osaka has shown that KI3 is probably 

 less dissociated than KI (27). Hence, its dissociation would be very much 

 driven back in strong KI solution. This would mean a far larger proportion 

 of undissociated KI3 molecules in such solutions. If now we assume 

 that the effects of increasing viscosity is less on KI3 than on the ions of the 

 same salt, we have a possible explanation of the observed facts. This 

 explanation is, of course, very far from satisfactory, for we do not know 

 that the effects of increasing viscosity on the KI3 and its ions is as indicated 

 {Argumentum ad ignorantium) . Yet in the absence of something better 

 it may serve as a hint of a possible explanation. 



In general it is supposed that ions migrate faster than the correspond- 

 ing unchssociated salt. This would seem probable a priori as the undis- 



