494 
PROFESSOR GRAHAM ON THE DIFFUSION OF LIQUIDS. 
eig-ht cells, 10*84, 10*77 j 10*81, 10*82; mean 10*81 grs. for two cells; calculated for 
2 per cent., 10*65 grs. of tartrate of soda in two cells. 
Dilfusion of 2 per cent, solutions of Tartrate of Potash in 8*08 days at 59°*9, and 
Tartrate of Soda in 9*9 days at 59°*6. 
Tartrate of potash .... 10*96 100 
Tartrate of soda .... 10*65 97‘18 
The tartrate of soda happens to correspond absolutely with the carbonate of soda, 
the diffusate of the latter salt obtained at the same temperature being also 10*65 grs. 
(p. 486). 
The double Tartrate of Potash and Soda afforded an interesting instance of decom- 
position produced by diffusion. Instead of the usual diffusion cells, a single plain 
cylindrical jar was employed about 5 inches in diameter and 10 inches in height. A 
portion of a 4 per cent, solution of Rochelle salt, amounting to 5000 grs. of liquid, was 
covered over by nine times as much pure water, in this jar, and the salt allowed to 
diffuse upwards for three weeks at 56°. The upper third of the fluid column, amount- 
ing to 17,000 grs. of liquid, was then drawn off, evaporated to dryness, and the salt 
ignited and converted, by the addition of hydrochloric acid, into chloride, which 
weighed 3*14 grs. The salt proved to be a mixture of the two chlorides in the pro- 
portion of 2*43 grs. of chloride of potassium and 0*71 gr. of chloride of sodium. 
This gives the following as the composition of the diffusate : 
Tartrate of potash 3*68 
Tartrate of soda 1*17 
4*85 
It hence follows that the proportion of potash, or of tartrate of potash, in the diffu- 
sate is nearly three times greater than existed in the original Rochelle salt. It will 
be recollected that the salt decomposed is, strictly speaking, a bibasic tartrate of pot- 
ash and soda, and not a double tartrate of potash and soda. The diffusate was ob- 
served to be exactly neutral to test-paper. This mode of diffusing upwards in water 
from the lower part of a deep jar obviously gives the greatest degree of separation 
attainable in consequence of unequal diffusibility. 
I may confine myself at present to the conclusion that of the nine pairs of potash 
and soda salts, of which the diffusion is compared in this paper, the potash salt 
uniformly exceeds in diffusibility the corresponding soda salt ; that the ratio between 
the two classes is always sensibly the same, or exhibits only a small range of varia- 
tion at the temperature of the experiments, which was near 60° ; with the peculiar 
exception of the hydrates of potash and soda. 
