PROFESSOR W. C. ROBERT S-AUSTEN OR THE DIFFUSION OF METALS. 401 
In the same manner the area c x y g, would represent the distribution of gold at 
one particular stage of the diffusion. 
The greater the diffusive powder of the metal represented by the area c xy g i the 
sooner wall it become identified with the area c e f g, and consequently in experi¬ 
ments which have lasted for equal times, the flatter the line x, y, the greater is the 
diffusivity of the metal it represents. 
The diffusion of chloride of sodium in water at 18° C. has been plotted in Plate 8 (as 
a dotted line) for comparison. It will be seen that (at 18° C.) common salt is much 
less diffusive in water than the metals as yet examined are at 500° in the solvents 
lead or bismuth. 
Fig. 5. 
a b 
The conditions under which the experiments on the diffusion of metals w>ere con¬ 
ducted differed so much, that it would have been meaningless to have plotted the 
actual distributions in a single plate. The time and temperature were different in 
each set of experiments, and the value of h w r as also variable. Plate 8 accordingly 
represents the theoretical distribution calculated from the experimental value of k 
corrected to 500°, and for the assumed time, 7 days, and the value of h, shown in this 
figure, 1'27 centims. ; the tube being divided into 14 sections. The small extent to 
which these theoretical distributions differ from those actually measured, may be seen in 
figs. 3 and 4, and in tables A, B, C, and D. The curves were drawn, for convenience, by 
plotting the mean concentration of each section at its mid point. The error so intro¬ 
duced is almost invisible on the curves and does not affect their relative position. No 
scale of the initial concentration is attached because the distribution of the diffusing 
metal is but slightly affected by small differences in concentration. 
The concentration of the diffusing metal, and as affecting this the duration of the 
experiment are, nevertheless, of much importance. The more rapid diffusivity in 
the alloys, poor in precious metal, is extremely marked. As this was especially the 
case in experiments which only lasted for a short period, where many of the segments 
contained very dilute solutions, it may be in part an effect requiring careful study, 
since it points to a possible simplification of the gold molecule when the osmotic 
MDCCCXCVI.—A. 3 F 
