Many of the tests which have been proposed for detecting the presence of methyl alcohol depend upon the oxidation of the alcohol to formaldehyde, and the recogni-ton of the latter by a colour reaction of some kind. Thus Deniges employs the well-known Schiff or Gayon reaction with decolorised fuchsine solution to show the presence of the formaldehyde, Bono (p.* 176) uses an alkaline solution of phenylhydrazine and sodium nitroprusside; Mulliken recommends the resorcinol-sulphuric acid test, and Hinkel uses the reaction with morphine and sulphuric acid. Gallic acid and phloroglucinol have also been employed.

1 Pharm. Zeit., 1913, 58, 850.

By far the most generally useful process is that devised by Deniges,1 which is carried out as follows: Into a test-tube, chosen of rather wide diameter to facilitate admixture of the various liquids, 0 1 c.c. of the strong alcohol is introduced, followed by 5 c.c. of a 1 per cent. solution of potassium permanganate, and 0 2 c.c. of strong sulphuric acid. The liquids are mixed and allowed to stand two or three minutes; then 1 c.c. of an 8 per cent. solution of oxalic acid is added and the tube again shaken. Rapid decolorisa-tion of the liquid ensues. When the solution has become of a sherry-yellow colour, 1 c.c. of strong sulphuric acid is added and the whole well mixed; the decolorisation now becomes complete. Five c.c. of decolorised solution of fuchsine are next added to the contents of the tube. In the course of a few minutes a violet colour appears, the depth of which depends upon the proportion of methyl alcohol in the liquid under examination. Except in considerable dilutions, the colour generally attains its maximum intensity by the end of fifteen minutes. With 10 per cent. of methyl alcohol it is extremely intense, with 1 per cent. very strong, and appreciable even at a dilution of 0.1 per cent. The test is not conclusive unless the alcoholic liquid has been properly purified, as glycerol and several other substances give a similar reaction.

The decolorised solution of fuchsine (Setoff's reagent, Gayon's reagent) employed by Deniges is prepared as follows: - To a litre of 01 per cent. solution of fuchsine add 20 c.c. of solution of sodium bisulphite at 36° Be., and after five to ten minutes pour into the mixture 20 c.c. of pure hydrochloric acid (D = 1.18). At the end of an hour or two the reagent becomes colourless, and is then ready for use.

A modified form of this method, adapted for the quantitative determination of methyl alcohol, is described a little further on (p. 183).

According to A. Bono,2 about 1 per cent. of methyl alcohol in ethyl alcohol can be detected by the following process.

1 Compt. rend., 1910, 150, 530, 831.

2 Chem. Zeit, 1912, 36, 1171.

Seventy-five c.c. of the sample, containing alcohol equivalent to 10 c.c. of strong alcohol, are distilled, and the vapour passed direct into a flask containing 50 c.c. of oxidising solution. The latter is a solution of potassium bichromate, saturated in the cold, and containing 60 c.c. of sulphuric acid per litre. The delivery-tube dips to the bottom of the solution, and the vapour after passing through the oxidising mixture is led into a condenser and collected. No formaldehyde is found, as a rule, in the first 25 c.c. of the distillate, which may be rejected. The next 10 c.c. are collected apart, and 2 c.c. of it are tested for formaldehyde by adding to it 10 drops of 0 5 per cent. solution of phenylhydrazine hydrochloride, 1 drop of 0 5 per cent. solution of sodium nitroprusside, and 10 drops of 10 per cent. solution of sodium hydroxide. If any formaldehyde is present, a blue coloration is obtained, changing to green and afterwards to a yellowish-red.

A more summary process is thus described by Mulliken.1

Dissolve one drop of the alcohol in 3 c.c. of water in a 6-inch test-tube. Wind a piece of rather light copper wire round a lead pencil, so that the closely-coiled spiral shall form a cylinder 2 cm. in length, while 20 cm. is left unbent to serve as a handle. Oxidise the spiral superficially by holding it in the upper part of the flame of a Bunsen burner; and then, while still at a red heat, plunge it into the alcoholic solution. (This treatment oxidises a portion of the methyl alcohol to formaldehyde.) Withdraw the spiral immediately and cool the test-tube with running water. Repeat the oxidation of the solution twice by the method given. Add one or two drops of 0.5 per cent. aqueous solution of resorcinol. Pour the mixture slowly into a second inclined test-tube containing 3-5 c.c. of pure concentrated sulphuric acid. If formaldehyde is present, a contact ring is formed, and on very gentle shaking, flocks of a characteristic red colour appear. A contact ring is also given by acetaldehyde, but the flocks are of a colour varying from yellow to brown. In this form, the test will show about 8 to 10 per cent. of methyl alcohol in ethyl alcohol.

A modified and more delicate form is due to Mulliken and Scudder.2 The oxidation is carried out in the same way, but the acetaldehyde is removed by boiling the solution down to about one-half its volume. Some of the formaldehyde is always lost in doing this; the best results are therefore obtained by evaporating down under diminished pressure at 25° to 30°. With very little formaldehyde, the formation of the characteristic flocks is slow; hence after addition of sulphuric acid the solution is allowed to stand for three minutes, and then shaken very gently for one minute, so as to cause the two liquid surfaces to mix, but slowly and not excessively. The formation of the red flocks is a necessary part of the test, as red contact rings are given by other substances, e.g., acetone and dimethyl ethyl carbinol. Methyl esters, methyl ethers, and secondary and tertiary butyl alcohols give the same reaction as methyl alcohol. This form of the test will detect about 2 to 3 per cent. of methyl alcohol in ethyl alcohol.