In a round-bottomed, short-necked flask of 300 c.c. capacity place 200 c.c. of a mixture containing, per litre, 90 grams of potassium dichromate and 85 c.c. of sulphuric acid (D = 1.845). Add 10 c.c. of the alcohol to be tested, at 90 per cent. strength, or else such a quantity as will contain this amount of alcohol. Side by side with the main experiment carry out also control experiments with (1) pure ethyl alcohol and (2) ethyl alcohol containing 0 2 per cent. of methyl alcohol.

Mix the alcohol thoroughly with the dichromate solution, let the mixture stand thirty minutes, and then distil it slowly. Reject the first 20 c.c. of distillate, as this portion contains acetaldehyde and ethers which mask the final reaction. Collect the next 50 c.c, distilled slowly, in a small stoppered flask of 51-52 c.c capacity, and add 1 c.c of pure dimethylaniline (b.p. 192°). Close the flask, which should now be quite full, hermetically; and keep it during eighteen to twenty hours at a temperature of 15-18° to effect the condensation.

Pour the contents 'into a round-bottomed distillation flask (120-125 c.c), add 3 granules of pumice and 4 or 5 drops of phenolphthalein solution (1 gram per litre); then from a burette run in sodium hydroxide solution (160 grams per litre) until a permanent rose coloration is just obtained. Now distil off 30 c.c. to eliminate the excess of dimethylaniline, cool the residue in the flask, and add to it 30 c.c. of water and 1 c.c. of glacial acetic acid. Shake well, and remove 10 c.c. of the mixture to a small test-tube. Make a suspension of lead dioxide in water (4 grams per litre), and add a few drops (the same number in the controls as in the main experiment) to the contents of the test-tube, which then cloe with the thumb and invert a time or two. If methyl alcohol was present a fine blue colour is produced. The pure ethyl alcohol should show no blue tint; or in any case, if a slight coloration should appear, too much importance must not be attributed to it. After the tubes have been compared they may be heated to the boiling point of the liquids and again examined at the end of five minutes; the differences of colour are thus much accentuated.

1 J. Chem. Soc. (Abst.), 1915, 108, ii, 588. 2 Compt. rend., 1898,127, 232.

3 Ann. Inst. Pasteur, 1902, 16, 8.

H. Scudder1 remarks that this method is delicate and fairly rapid (time required, about five hours). But it must be used with care, as the test may be spoiled by slight variations in procedure, or even sometimes by causes of unknown origin when the process has been strictly carried out. The chief objection is that all the acetaldehyde must be removed, otherwise an interfering colour will result. But some methylal distils over with the acetaldehyde, and hence, since the test depends upon the methylal, if only a small quantity is present the result may be vitiated.

J. Wolff2 has found that spirits containing caramel or sucrose, when treated with chromic acid and distilled in a current of steam, yield polymerisation products of formaldehyde which give the strong blue colour as above described on treatment with dimethyl-aniline and lead peroxide. It was found that lævulose, whether added directly or arising from the hydrolysis of cane-sugar, was the only sugar giving this reaction. Hence spirits containing cane-sugar, invert-sugar, or caramel should always be re-distilled before they are examined by the foregoing process. (Distillation is, of course, a general precaution which should be taken whatever process is used.)