the estimation of small quantities. In the present work neither of the permanganate methods to which reference is made (14, 15) has been tested.

Allen (12) gives a method using a solution of potassium dichro-mate in approximately 40 per cent. sulpuric acid. The standard solution is made up by dissolving 16.97 g. pure potassium dichromate in 200 c. c. of distilled water and then making up to 1000 c. c. with 50 per cent. sulphuric acid. The excess of dichromate after oxida-tion of the alcohol is estimated by titrating with standard (N/10) thio-sulphate the iodine liberated on the addition of potassium iodide to the reaction mixture after dilution with 200 c. c. of water. In this method oxidation proceeds to the acetic acid stage, so that 1 c. c. of the dichromate is equivalent to 0.005305 g. of alcohol.

A large number of experiments have been carried out using a standard alcohol solution of such a strength as is expected to be obtained in the later work. The heating of the solution has been varied from slow to rapid, from heating only to 80° C. or to boiling for varying periods of time, but satisfactory results have not been obtained. Allen merely states that the solution is to be heated until a green but not a greenish-yellow tint is reached, but the difficulty is to decide when the solution has reached the right stage since the experiments described here have shown that alhough visually the solutions are at the same stage, the oxidation has not proceeded to the same degree; indeed, in many cases, oxidation must have gone beyond the acetic acid stage to explain the large consumption of dichromate. The results are very erratic, sometimes low, sometimes high, and sometimes approximately correct, although the tendency is distinctly towards high results.

In all cases the total volume of reaction mixture was about 150 c. C. Heating must be carried out carefully, especially in the early stages, otherwise loss of alcohol occurs.

Table 2 contains some of the results obtained when the various factors governing the oxidation were altered.

In the other method (13) using potassium dichromate as oxidizing agent, the whole reaction is carried out in concentrated sulphuric acid. The solution is made up by dissolving 4.4128 g. of pure dichro-mate in as small a volume of water as possible and then making up to 1000 c. c. with concentrated sulphuric acid, being careful to add the acid slowly, so that at no time shall the temperature rise above 100° C. At a certain stage the chromic oxide. precipitates out but readily redissolves when more acid is added. From a burette a measured quantity of the solution is run slowly into a cold mixture of a known volume of acoholic solution with three times its volume of concen-trated sulphuric acid, which has been added slowly, keeping the flask cool under the water tap. The mixture is then slowly heated to 98° C. (not above) and the temperature held there for exactly five minutes. After cooling the liquid is diluted with 200 c. c. of cold water, and after cooling again the excess dichromate's is destroyed by excess of standard ferrous ammonium sulphate solution, this excess in turn being estimated with standard permanganate.

The method sounds long and involved, but in practice it is rapid, and gives accurate and reproducible results. In very few cases did the quantity of alcohol present in the volume of liquid taken exceed 10 mg., and in the great majority of cases it was considerably less than this.

In Table 3 are given the results of experiments made to test the accuracy of the method using alcoholic solutions of known strength. Solution A was made up from industrial absolute alcohol which was taken to be 99.5 per cent. pure, but in reality may have been only 98 per cent, which would explain the low results of the first three experiments as compared with the experiments using solution B for the making of which pure alcohol was used.

[The section below cannot be correctly rendered as it contains complex formatting. See the image of the page for a more accurate rendering.]

It is necessary to point out that this method cannot be used in the presence of chlorides as evolution of chlorine occurs and high results are obtained. Thus a given quantity of solution contained 10.56 mg. of alcohol; after adding 1 c.c. of N/10 sodium chloride and treating as above an apparent result of 12.27 mg. was obtained. Silver sulphate may be used to remove chloride.

An alcoholic soda solution, prepared for use in the main series of experiments, was then analyzed by the above method. Into a weigh-ing bottle were rapidly placed 10.2940 g. of this solution which were