704 Proceedings of the Royal Society of Edinburgh. [Sess. 
attached the metal condenser cc. The tube dd is screwed in so that 
about 1 cm. of the end projects into the tube a. 
The condenser tube fits loosely into a glass siphon tube e whose 
capacity is about 150 c.cm. The lower end of the siphon tube passes 
through the double-bored cork g and is kept well below the level of the 
liquid in the flask. 
The siphon tube is open to the air, but all the other joints are air-tight. 
The tube a is loosely filled with lumps of quicklime. The end of the 
entry tube d and of the condenser tube c are guarded with wads of glass 
wool to keep back the lime, which falls to fine powder when it is slaked. 
The flask / is charged with acid to be esterified, mixed with excess of 
alcohol and about 100 c.cms. of benzene. A few drops of concentrated 
hydrochloric acid are added, and the apparatus is placed in position. 
Two small gas flames at hh are then lighted, and when the tube a is 
sufficiently hot to prevent condensation of alcohol or benzene the flask is 
heated on the steam bath. 
As the liquid in the flask boils, the vapour of the ternary mixture of 
alcohol, benzene, and water passes up the tube d and over the quicklime in 
the tube a, where the water is absorbed. 
The mixed vapours of benzene and alcohol pass on and are condensed in 
the condenser c. The condensed liquid collects in the siphon tube e until 
the pressure of the liquid in the siphon is greater than the pressure in 
the flask. 
The siphon then overflows, and the recovered benzene and alcohol flow 
back into the flask. 
The steam in the steam bath should be regulated so that the siphon 
overflows about four times in an hour. 
In one experiment the flask was charged with 100 gms. of tartaric acid, 
300 gms. of ethyl alcohol, 90 c.cs. of benzene, and a few drops of hydro- 
chloric acid. 
After boiling for five hours in the apparatus, the alcohol and benzene 
were distilled off and the ester was distilled in vacuo. 
125 gms. of pure tartaric ester were obtained, which is more than 
90 per cent, of the theoretical yield. Good yields of tartronic ester have 
also been obtained, and it is clear that the method is applicable to all cases 
in which Taylor's method can be used. 
(. Issued separately November 6 , 1908 .) 
