Walter Stiles 
148 
rate of diffusion was practically independent of the concentration ox 
the gel. Sodium chloride, magnesium chloride and hydrogen chloride 
were among the substances examined, but the unpublished researches 
of Stiles and Adair have failed to confirm this conclusion in the case 
of sodium chloride, which appears to diffuse more slowly the greater 
the concentration of agar-agar, so that agar-agar behaves in the same 
way as gelatine in reducing the rate of diffusion. 
The consequence of Pick’s law in the case of linear diffusion in 
one direction, namely that x = CVDt where C is a constant (see 
p. 139), has been confirmed for diffusion into gels by Coleman (1S88), 
Chabry (1888), Voigtlander (1889), von Fiirth and Bubanovic (1918), 
Stiles (1920) and Adair (1920), these various authors using a great 
variety of diffusing substances and gels of gelatine, gelose and agar- 
agar. The law holds equally well when the gel contains an indicator 
to mark the progress of a layer of definite concentration provided 
the concentration of the indicator does not approach too near to that 
of the diffusing substance (Stiles, 1920). 
There are few observations on the influence of temperature on 
the coefficient of diffusion in gels. The results of Voigtlander on 
diffusion in agar-agar gels indicate a greater increase in diffusivity 
brought about by an increase in temperature from 20° C. to 40° C. 
than that brought about by an increase from o° C. to 20° C. Voigt¬ 
lander employs the equation 
d 0 = d 0 ( 1 + m 2 
to express the relation between the coefficient of diffusion and 
temperature, D e and D 0 being the coefficients of diffusion at 6 ° C. 
and o° C., and /3 a constant. It will be observed that as the values 
of jS are small (o-oi to 0-02) in comparison with unity, when 6 is also 
small then approximately 
D e = D 0 ( 1 + 2 
since the term [ 3 2 6 2 is small in comparison with 1 + 2 j 3 d. Consequently, 
when 6 is small the relation between the temperature and the 
coefficient of diffusion is approximately a linear one as assumed by 
de Heen, and by Oholm following Nernst. But when 6 is large the 
term j 6 2 6 2 is not negligible, and the coefficient of diffusion increases 
more rapidly with rise in temperature the higher the temperature. 
The unpublished results of Stiles and Adair on the diffusion of 
sodium chloride in agar-agar gels at temperatures from o° C. to 40° C. 
support this latter assumption rather than the opinion of a linear 
temperature coefficient. 
