One of the oldest and most trustworthy processes for the detection of methyl alcohol in the presence of ethyl alcohol is that described by Riche and Bardy.1 It is based upon the conversion of the two alcohols first into the corresponding iodides and then into the alkyl anilines, by the oxidation of which a violet dyestuff is produced from the dimethyl-aniline, but not from the ethyl compound. As given by M. Louis Calvet,2 the process is as follows: -

(1) Conversion Into The Iodides

Place 10 c.c. of the alcohol to be tested in a flat-bottomed flask of 90 c.c. capacity containing 10 grams of iodine, add 2 grams of red phosphorus, attach the flask to a condenser, and warm gently to distil the iodides. Receive the latter in a vessel containing 20 c.c. of water.

(2) Combining The Iodides With Aniline

Pour the iodides and the water into a separating funnel, and draw off the iodides into a flask similar to that used in (1); add 10 grams of pure aniline, and heat under a reflux condenser.

(3) Conversion Of The Hydriodides Of The Ethyl- And Methyl-Anilines Into The Free Bases

When the above-mentioned reaction is complete, dissolve the resulting crystalline mass in about 50 c.c. of hot water, and decompose the hydriodides with an excess of sodium hydroxide solution (D = 133). The free bases separate as an oily layer which is allowed to collect in the neck of the flask.

(4) Oxidation Of The Methyl- And Ethyl-Anilines

Take 1 c.c. of the liberated bases and mix it intimately with 10 grams of oxidising powder composed of: - Sand, 100 grams, copper nitrate 3 grams, and sodium chloride, 3 grams. Place the mixture loosely in a test-tube, and heat it during six hours in a water-bath kept at a temperature of 90°. Then fill the tube with alcohol, pour the whole contents into a nickel or porcelain basin of about a litre capacity; add 300 c.c. of water, boil to expel the alcohol and to precipitate resinous matters formed during the oxidation, filter, and make up the volume of the filtrate to 500 c.c. with water.

(5) Dyeing

Measure out 20 c.c. of this filtrate into 350 c.c. of water, contained in a porcelain basin of 500 c.c. capacity and heated on a water-bath, and place in it a square of cashmere of 1 dcm. side. Heat till the dye is completely extracted from the bath, stirring the fabric from time to time with a glass rod. Compare the colour thus obtained with that given by a control experiment on ethyl alcohol containing a known quantity of methyl alcohol.

The process has the advantage of yielding dyed material which can be preserved and produced as evidence of the presence of methyl alcohol. In the form just described, however, it is stated (Calvet, loc. cit.) not to be capable of detecting methyl alcohol unless the latter is present to the extent of 1 1/2 to 2 per cent. of the mixture of alcohols. Moreover, it is somewhat lengthy.

1 Compt. rend., 1875, 80, 1076. 2 " Alcools," p. 109.

Much smaller proportions can be detected by the following modification of the method used in the Government Laboratory, London. The process is also much shortened.

Ten c.c. of the alcohol, of strength not less than 85 per cent., are placed in a flask of about 75 c.c. capacity, containing 10 grams of powdered iodine. 1.5 Grams of red phosphorus are then added, and the flask connected at once with a reflux condenser. The reaction commences almost immediately. When it has ceased, and after reversing the condenser, the mixture is slowly distilled, and 0 5 c.c. of the mixed iodides collected under water in a very small separating funnel marked to show this volume (0.5 c.c). A Cribb's condenser about 6 inches long is convenient for the distillation. With small proportions of methyl alcohol such as the method is specially devised for detecting, most of the methyl iodide will be present in this 0.5 c.c. fraction.

The mixed iodides thus obtained are then run into a small wide-necked flask (50 c.c. capacity) containing 0.5 c.c. of recently distilled aniline. The flask is loosely corked, and allowed to stand for thirty minutes or so on a sand-bath at a temperature of about 60°. The crystalline mass of alkylaniline hydriodides thus obtained is dissolved in 30 to 50 c.c. of warm water, and decanted from any small globule of unchanged aniline into a separator of about 100 to 150 c.c. capacity. Two c.c. of sodium hydroxide solution (D = 1.3) are now added, and 20 c.c. of petrole'um ether. After shaking the mixture and allowing it to separate, the lower layer is run off and discarded, and the ethereal solution washed twice with a little water. It is then poured into a wide-necked flask (100 c.c.) containing 5 grams of oxidising mixture (sand, 100 grams, copper nitrate, 3 grams, sodium chloride, 2 grams). The petroleum ether is evaporated off, and the flask heated in a steam oven at approximately 100° for 1 1/2 hours.

The dye thus formed is extracted by. digestion with 50 c.c. of boiling alcohol for two or three minutes under a reflux condenser. When cold, 5 c.c. of the alcoholic extract are filtered off and made up to 100 c.c. with water.

Of this diluted dye 5 c.c. are added to 400 c.c. of hot water, and a piece of clean Berlin wool, 18 inches long, is immersed in the solution. The latter is then kept at about 60° for twenty minutes or so, with occasional stirring, in order to fix the dye on the wool.

After rinsing the wool in water and allowing it to dry, it may be coiled into a flat coil and attached to gummed paper for comparison of its depth of violet colour with that given by control experiments, carried out in the same way with pure ethyl alcohol and known mixtures of the two alcohols.

As little as 0 1 per cent. of methyl alcohol in mixture with ethyl alcohol can be thus detected. With much larger proportions, the whole of the dyestuff will not be extracted by the wool, in which case a smaller quantity than 5 c.c. of the diluted dye - say 1 or 2 c.c. - should be taken to make the dye-bath.

With a little practice the method can be made approximately quantitative for small proportions of methyl alcohol. It is a useful confirmatory process in specially-important cases.