DIFFUSION. 263 
where a is the amount of substance, passing in time z, through section q, 
from an infinitely long cylinder of solution of concentration c, into 
another such cylinder of pure solvent. 
This formula was experimentally verified by Voigtlander l with 
cylinders of agar jelly, in which diffusion occurs as easily as in water. 2 
He further investigated the temperature coefficient (a) for k, and found 
that it is not a linear function of the temperature, as stated by Weber,"' 
but stands in the following relation 4 : — 
K t = K (l + aty 
In the case of an electrolyte in solution, the diffusion must be con- 
sidered as that of the ions into which it is dissociated on passing into 
solution. The velocity of the separated ions may be very different, but 
since in solution by virtue of their opposite charges they cannot part, 
the more rapidly moving ion must be retarded by the more slowly 
moving, and the more slowly moving accelerated by its more active 
fellow. The diffusion of an electrolyte may also be accelerated by 
the presence in the liquid into which it is diffusing, of ions charged 
oppositely to those forming the more active partner in the diffusing 
substance. Thus hydrochloric acid diffuses faster into a solution of 
sodium chloride than into water. 5 As a rule, those electrolytes which 
are the best conductors, are the most diffusible in solution. The pre- 
sence of a substance that is not an electrolyte in the fluid into which 
diffusion is taking place may slow the diffusion of an electrolyte. 
Thus sodium chloride diffuses more slowly into sugar solution than 
into water, and the presence of ethyl alcohol also retards its diffusion. 7 
In the case of non-electrolytes in solution, diffusion must concern the 
" molecules " of the dissolved substance, and the " aggregates " of colloids 
will find their way with greater difficulty than the " molecules " of 
crystalloids. 
No definite rule can be stated as regards the effect of concentration 
of the solution upon the rapidity of diffusion of the dissolved substance. 
With sodium chloride the coefficient of diffusion is practically unaltered 
by change in concentration of the solution. In the case of magnesium 
sulphate the coefficient falls with the concentration of solution, while 
with hydrochloric, nitric, and sulphuric acids the coefficient rises with 
the concentration. 8 
The simultaneous diffusion of two salts, studied first by Graham, has 
been since more completely investigated by Marignac. 9 In general the 
rapidity of diffusion of the more diffusible of a pair of salts diffusing 
simultaneously is found to be increased, that of the less diffusible 
diminished. 
In the following table the diffusions of five pairs of salts, separately 
and simultaneously, are contrasted. 
1 Ztschr.f. physical. Ohem., Leipzig, 1889, Bd. iii. S. 316. 
2 Graham, Ann. d. Chem., Leipzig, 1862, Bd. cxxi. S. 5, 29. 
3 Ann. d. Phys. u. Chem., Leipzig, 1879, Bd. vii. S. 536. 
4 For the values of a, which vary slightly with different substances, see Voigtlander's 
original paper, loc. cit. 
5 Arrhenius. Ztschr.f. physikal. Chem., Leipzig, 1892, Bd. x. S. 51. 
6 Long, Ann. d. Phys. u. Chem., Leipzig, 1880, Bd. ix. S. 613 ; Lenz, Mem. Acad. imp. 
d. sc. de St. Petersbourg, 1882, tome vii. p. 30. 
7 Arrhenius, loc. cit. 
8 Scheffer, Ztschr. f. physikal. Chem., Leipzig, 1888, Bd. ii. S. 390. 
9 Ann. de chim., Paris, 1874, Ser. 5, tome ii. p. 546. 
