The principle underlying this process is the extraction of the higher alcohols with carbon tetrachloride, and their oxidation to the corresponding acids, which are then estimated by titration. The original description1 has been modified slightly by including suggestions made by Schidrowitz,2 and the method as now recommended is as follows.3

Two hundred c.c. of the spirit are taken, 1 c.c. of strong solution of potassium hydroxide is added, and the mixture boiled for an hour under a reflux condenser to hydrolyse esters. The liquid is then distilled from a flask fitted for distillation with steam, but the steam is not passed until only about 20 c.c. of the sample remain, and thereafter the operation is so conducted that by the time 300 c.c. of distillate are collected the residue in the flask is about 10 c.c.

The distillate is then mixed with saturated salt solution (acidified slightly with a few drops of sulphuric acid), until it has a specific gravity of 11. It is now divided into two equal parts, so that duplicate determinations may be made, and each part is extracted with 100 c.c. of purified carbon tetrachloride, using 40, 30, 20, and 10 c.c. successively.

Some ethyl alcohol is contained in the tetrachloride extract. To remove it, the extract is shaken first with 50 c.c. of the saturated salt solution, and separated, and then with 50 c.c. of saturated solution of sodium sulphate to remove remaining chloride.

For the oxidation of the higher alcohols a mixture of 5 grams of potassium dichromate, 2 grams of strong sulphuric acid, and 10 c.c. of water is added to the carbon tetrachloride extract in a flask provided with a reflux condenser, and the liquid boiled gently for at least eight hours on a water-bath. After the addition of 30 c.c. of water to the liquid in the flask, the contents are distilled until only about 20 c.c. remain; then this residue is steam-distilled as in the first operation until not much more than 5 c.c. remains when 300 c.c. of total distillate have been collected. The distillate is then titrated with decinormal barium hydroxide solution, shaking thoroughly after each addition, until it is neutral to methyl-orange; phenolphthalein is next added and the titration continued to the neutral point. Each c.c. of the N/10-solution required in the phenolphthalein stage of the titration represents 0 0088 gram of higher alcohols calculated as amyl alcohol.

1 Allen's "Commercial Organic Analysis," 1898, 1, 154.

2 J. Soc. Chem. Ind., 1902, 21, 815.

3 Allen's "Commercial Organic Analysis," 1909, 1, 188.

The barium hydroxide used in the methyl-orange stage was formerly ignored in the calculation, as the acidity here was supposed to be due to hydrochloric acid. This, however, appears to be unjustifiable, since mere traces of chlorides may be found, whilst the methyl-orange acidity is about 10 per cent. of the total.1 The recommendation made by Schidrowitz and Kaye is to calculate the whole of the acidity to amyl alcohol, but to estimate the chlorides gravimetrically if the methyl-orange acidity much exceeds 10 per cent. of the whole, and correct the calculation accordingly.