as it has been shown by Hermann and Long that exactly 3 volumes of carbon monoxide to 1 of ethylene are evolved. The residual potassium bromide is estimated by means of standard silver nitrate solution. Bromoform is specially suitable for this purpose for several reasons. It is very readily formed by the action of bromine and potash on acetone, and although very volatile in steam, it is not liable to loss due to its own evaporation. Further, its high molecular weight and large percentage of bromine conduce to accurate results, 58 grms. of acetone being responsible for the formation of 357 grms. of KBr. The method of carrying out the analysis is as follows:-- A known quantity of the solution to be tested, containing acetone to the extent of 0.1 to 0.2 grm., is pipetted into a 500 c.c. round-bottom flask, diluted with a little water, and mixed with 20 to 30 c.c. of a 10 per cent. solution of caustic potash. The flask is connected with a long reflex condenser, and is also fitted with a dropping funnel containing a solution of bromine in potassium bromide (200 grms. of Br and 250 grms. of KBr to 1 litre of water). The bromine solution is allowed to flow into the mixture until it has acquired a faint yellow tinge, the flask and its contents being then heated on the water bath at about 70 deg. C. for half-an- hour. Bromine solution is added drop by drop until the slight coloration is permanent, excess of bromine being got rid of by boiling for a minute or two with a little more caustic potash. The mixture is then distilled until the distillate is free from bromoform, halogen being tested for in the usual manner. Water is added to the contents of the flask if necessary. It may be here observed that no acetone can be detected in the distillate by means of the mercuric oxide test, and free bromine is also absent. The condenser having been washed out with a little alcohol, in order to remove any traces of bromoform which may have collected, the distillate and washings are mixed with 50 c.c. of alcohol and sufficient solid caustic potash to make an approximately 10 per cent. solution. The mixture is then heated on the water bath under a reflux condenser until the bromoform is completely decomposed. This generally occupies about three-quarters of an hour. The liquid is allowed to cool, evaporated to smaller bulk if necessary, and exactly neutralised with dilute nitric acid. It is then diluted with water to 500 c.c., and an aliquot part titrated with N/10 silver nitrate solution, using potassium chromate as indicator; 240 parts of bromine correspond to 58 parts of acetone. The complete analysis can be performed in one and a half to two hours. It is imperative that the bromine used should be pure, as crude bromine frequently contains bromoform. The method is suitable for the estimation of acetone in wood-spirit, the spirit being diluted to 10 times its volume, and 5 c.c. of this solution employed for the determination. For example--