441 
1909-10.] Equilibrium in a Ternary System. 
ethyl salt. It was, however, found impossible to obtain a saturated solution, 
the liquid gradually assuming the consistency of a thick syrup in which 
solid particles were suspended in a fine state of division. A concentrated 
solution of the ester salt was then made and shaken up with a saturated 
solution of potassium carbonate. Mixture only took place to a limited 
extent, two distinct layers separating when the liquid was allowed to stand. 
Water was added drop by drop from a burette to the liquid, with constant 
stirring, until the two layers disappeared and one homogeneous solution 
was left. The disappearance of the two layers was quite sharply defined. 
The above experiment was then repeated quantitatively. 5 '5 gms. 
potassium ethyl di-propyl-malonate were dissolved in 2 gms. water and 
T05 c.cs. of this solution were put into a graduated tube. T05 c.cs. of a 
saturated solution (about 52 per cent.) of potassium carbonate were added, 
and thereafter water drop by drop. 
The following readings were taken : — 
Total Volume 
of Liquid. 
Vol. of K 2 C0 3 
Solution. 
Vol. of KEt. Salt 
Solution. 
2T0 
1-05 
105 
3-05 
1*55 
1-50 
3-45 
1-65 
1-80 
3-80 
1-45 
2*35 
3-95 
1T5 
2-80 
4-05 
homo 
geneous. 
From the above table it is seen that, while at first the volumes of both 
layers increase on addition of water, the potassium carbonate layer soon 
begins to diminish in volume while the volume of the ester salt solution 
increases steadily. This seems to indicate that it is the carbonate solution 
which gradually passes into that of the ester salt, one homogeneous liquid 
being obtained when the whole of the potassium carbonate solution 
disappears into the other layer. As will be seen later, on more accurate 
analysis this was found to be the case. 
The mixtures to be investigated were made up by putting weighed 
quantities of potassium ethyl di-propyl-malonate, potassium carbonate and 
water into a stoppered tube and thoroughly mixing in a thermostat at 25° C. 
The mixture was then allowed to settle into two layers at the temperature 
of the bath. 
Methods of analysing the two layers had next to be devised. The first 
method attempted was to find the total percentage of potassium in each 
layer, and then the percentage of carbon dioxide in the layers was found by 
