306 N. L. BOWEN 
agreement with theory if the variation of the coefficient of diffusivity 
is taken into account. It is possible, too, that from the observed 
values a definite quantitative relation between the coefficient of 
diffusivity and the concentration could be calculated, but the 
writer has not been able to find any attack upon a problem of this 
Initial Condition 
Final po ZOU 
Catulated 1» k=02 
AbAn,/0 20 30 40 50 60 70 80 90 TOPSWE 
Fic. 5.—Diffusion experiment No. 25 
kind in the various treatments of diffusion of concentration or of 
temperature. All of these adhere to a constant value of the coefhi- 
cient of diffusivity. While in some respects it is highly desirable to 
check the present results more thoroughly along the lines indicated 
above, yet, for the purposes of the present paper, this is unneces- 
sary. The results afford us very definite information as to the 
