144 
Walter Stiles 
among those examined, although the absolute coefficient of diffusion 
varies greatly. These results agree well with those obtained for zinc 
salts to which reference has already been made. 
On theoretical grounds Nernst (1888) propounded the formula 
D e = D 18 [i + a {d - 18)] 
to express the influence of temperature on the coefficient of diffusion, 
D e and D 18 being the coefficients of diffusion at 6° C. and 18 0 C. 
respectively, and a a constant having the value 0-026 for neutral 
salts and 0*024 f° r acids. 
Assuming the correctness of Nernst’s general formula, the value 
of the temperature coefficient a in Nernst’s equation has been cal¬ 
culated by Oholm from measurements of the coefficient of diffusion 
of a number of electrolytes (Oholm, 1902, 1905) and non-electrolytes 
(Oholm, 1910) at temperatures between o° and 20° C. The values 
found by him are summarised in Table VII. 
Table VII 
Temperature Coefficient of Diffusivity of a number of substances 
(Data from Oholm) 
Temperature Coefficient 
Substance 
HC 1 
NaCl 
KC 1 
LiCl 
KI 
NaOH 
KOH 
Acetic acid 
Sucrose 
Lactose 
Maltose 
Raffinose 
Arabinose 
Dextrin 
Nicotine 
0-027 
0-0235 
0-023 
0-021 
0-028 
0-032 
0-032 
0-032 
0-032 
0-044 
0-035 
0-016 
Although there are obvious exceptions, Oholm concludes that in 
general the temperature coefficient of diffusivity is less the greater 
the diffusivity. Thus hydrochloric acid, with its very high coefficient 
of diffusion, has a very low temperature coefficient of diffusivity, 
while the slow diffusing sugars have temperature coefficients con¬ 
siderably higher than those of neutral inorganic salts. 
It should be noted that it is assumed by all workers that the 
relation between diffusivity and temperature is a linear one. While 
this may be approximately true over the small temperature intervals 
with which Oholm worked, it is scarcely likely to be an exact 
