230 Proceedings of Royal Society of Edinburgh. [sess. 
a pupil of Wagner, has found that some organic substances in 
organic solvents do also exhibit it, e.g. cyanobenzol in ethyl alcohol. 
In the known cases of group (a), increase of concentration 
raises the transition temperature : there is very little to show in 
what way concentration affects the transition temperature of 
solutions in class ( b ), whether decrease of concentration will 
lower it or not, but measurements by Rudorf* on aqueous solutions 
of carbamide indicate that at 25° C. the relative viscosity decreases 
with dilution, and even becomes “ negative,” e.g . — 
— but the viscosity is so nearly the same as that of water that it is 
not safe to base any conclusions on these data. 
Increase of molecular weight, in the known cases of class (a), 
raises the transition temperature, and this affords another means 
of bringing it within the range of experiment. 
The general case, where the viscosity curves of solution and 
solvent intersect twice, is of some interest. According as the one 
curve or the other represents the solution, there will be a transition 
from “positive” to “negative” viscosity, or vice versa , , at both 
high and at low temperatures. It may not be possible to realise 
this case, except perhaps with a very soluble substance, and a 
solvent which permits of a wide range of temperature, but there 
should not be much difficulty in realising the particular case of it 
where at one extreme of temperature and concentration the one 
part of the curve is obtained, and the other part at the other 
extreme. 
I hope to commence experiments, in the near future, with a 
view to verifying these conclusions. 
Concentration. 
r] (Water at 25° = 1). 
0-937 n 
•469 
•234 
•117 
•058 
1-010 
1-002 
0-996 
•993 
•995 
" Zeit. f. Phys. Cliem., 43, p. 257 (1903). 
( Issued separately June 16, 1904.) 
