Aldehydes and furfural are first eliminated from the spirit to be tested, as they are not without effect upon the colour produced with sulphuric acid by the higher alcohols.

For this purpose, 50 c.c. of the spirit at a strength of 50 per cent. are placed in a flask of about 250 c.c. content, and mixed with 1 gram of either metaphenylenediamine hydrochloride, or of aniline phosphate (equal volumes of aniline and phosphoric acid of sp. gr. 1.453). After addition of a few small pieces of pumicestone to regularise the ebullition, the alcohol is boiled gently for an hour under a reflux condenser, and then distilled off rather rapidly, and as completely as is practicable without running risk of charring the residue. The tube connecting flask and condenser during the distillation should be of relatively large diameter, and the limb leading from the flask should be short, in order to give the minimum reflux effect.

The volume of the distillate is made up exactly to 50 c.c., mixed, and 10 c.c. are placed in a clean flask. Into this, 10 c.c. of pure, colourless sulphuric acid (monohydrated) are run from a pipette down the side of the flask so as not to disturb the alcohol appreciably. The flask is then taken in a tube-holder and agitated briskly whilst holding it in the flame of a good Bunsen burner turned full on. When ebullition commences, which should occur at the end of fifteen seconds, the flask is removed, covered with a small watch-glass, and allowed to cool.

The colour produced is then compared in a colorimeter with that given by standards containing known quantities of isobutyl alcohol, treated in the same way. (In the absence of a suitable colorimeter the comparisons may, of course, be made in small "Nessler" tubes.)

The depths of colour produced are not exactly proportional to the amounts of higher alcohols present, so that comparisons should be made with standards not differing too much from the sample. In practice, a curve is constructed for use, showing the real content of higher alcohols corresponding with the apparent content given by comparison with the standard when the latter differs from the sample: this obviates the necessity for dilutions and the corresponding calculations, and lessens the number of standards required.

Saglier makes the process more sensitive for low proportions of higher alcohols by adding 20 drops or more of a 01 per cent. solution of furfural (in 50 per cent. alcohol) to both sample and standard before the treatment with sulphuric acid. This intensifies the colours produced, and as both sample and standard are treated alike, no error is caused.

Rocques has introduced some modifications into the method to render it more sensitive and more precise. He distils 100 c.c. of the 50 per cent. alcohol, after refluxing with 2 grams of meta-phenylenediamine hydrochloride, to a bulk of 75 c.c., and works with the spirit at this higher strength, namely, 66.7 per cent. instead of 50 per cent. Also the mixed acid and alcohol, instead of being heated in a Bunsen burner, are heated in a bath of calcium chloride (69 per cent. solution) to 120° for an hour.

Table for converting apparent content of higher alcohols into real content, when the standard used is 0 0667 gram of isobutyl alcohol in 100 c.c. of ethyl alcohol at 66.7 per cent. strength: -