49(5 
PROFESSOR HENRY A. MIERS: AN ENQUIRY INTO 
from the saturated to a supersaturated solution; and (3) how the refractive index of 
supersaturated solutions varies with their strength. So far as I am aware, no 
previous observations have heen made upon the refractive indices of supersaturated, 
or even of saturated, solutions of the substances investigated in this paper. 
(B.) The Refractive Indices of Supersaturated Solutions. 
The observations were made in two ways ;— 
(1) By direct measurement of the minimum deviation in a hollow prism. 
containing the solution ; 
(2) By total reflection, by means of a glass prism of a known index immersed in 
the solution, using the new goniometer as described on p. 493, 
With the minimum deviation method a supersaturated solution contauiing a 
known weight of salt to a known weight of water was taken, and a sample of about 
G cub. centims. introduced into the prism; the solution was then diluted with a 
measured volume of water and another sample examined, and so on. The 
temperature was measured by a thermometer inserted in the prism. 
With the total reflection method it is easier than with the minimum deviation 
method to carry out a long series of observations upon solutions of slightly varying 
strength; for it is only necessary to add known volumes of water successively to a 
known solution. For example, 50 cub. centims. of a supersaturated solution of 
known strength having heen placed in the glass trough, its refractive index was 
determined by the totally reflecting prism; 1 cub. centim. of water was then added 
and the index re-determined; this process was repeated until the trough was nearly 
full; a measured volume of the liquid was then removed from the trough, and the 
process was repeated and carried on until the solution was quite dilute. The liquid 
was, of course, kept constantly stirred during these experiments. 
As regards the accuracy of the two methods, several examples will he found below 
of determinations of the same liquid by both ; these always agree in the third place 
of decimals, and sometimes in the fourth. 
As regards sensibility, an error of 1 minute in the reading for total reflection 
corresponds to about 2 in the fourth place of decimals, or to an eri'or of about 
2 minutes in tlie reading for minimum deviation. 
Althoimh tlie readings for total reflection cannot be made so accurately as those 
o o 
for minimum deviation, they could generally be relied upon to within 2 minutes, and 
sometimes to 1 minute. 
Preliminary observations were first made upon solutions of alum and of tartaric 
acid, chiefly with the object of plotting the results upon a curve in order to ascertain 
whether there is any sudden change in the curvature, especially on passing from a 
