228 
Proceedings of Royal Society of Edinburgh. [sess. 
tion, where all the experiments are made at one temperature, the 
general practice is not inconvenient, but it has several disadvan- 
tages : — 
1. It is possible, and maybe desirable, to determine the viscosity 
of a solution at temperatures below the freezing-point or above the 
boiling-point of the solvent ; in this case s 0 , t 0 cannot be deter- 
mined. 
2. It affords no good way of graphically representing the 
relation between viscosity and temperature. 
3. It may lead to misunderstanding. Most of the experiments 
on solutions have been made at 17° or 25° C., and a comparison 
of the relative viscosity of, e.g ., 1 n KC1 is as follows : — 
from which it appears that the relative viscosity of the solution in- 
creases with increase of temperature. In this connection it may be- 
remarked that Euler, * referring to the influence of temperature, 
says, — “whilst the specific viscosity of all solutions of non- 
electrolytes decreases with rise of temperature, the solutions of 
strongly-dissociated electrolytes are affected in the opposite- 
direction.” Without a definition of “specific” viscosity this- 
statement might be misunderstood. 
If the viscosities are referred to water at 0° as unit, it is seen 
that they do not increase with rise of temperature, but that they 
do not diminish so rapidly as the solvent ; in other words, the 
temperature coefficient of the solution is smaller than that of the 
solvent, but is of the same sign. Of course, there may still be 
a fundamental difference between the two classes of solutions. 
As to the unit, no maximum of viscosity for water is known 
(as there is of density at + 4° C.), and there is not much to choose 
between water at 0° and +4°; in either case, S 0 can be put =1 
without appreciable error in rj , which is ordinarily not more-, 
accurate than one in 500 or 600. 
Temp. 
15° 
Temp. 
25° 
1 
1-001 
Water 1 
1 n KC1 0-972 
Water 0*640 
1 n KC1 0-622 
* Zeit. /. Phys. Chem., 25, p. 536 (1898). 
