This is of some historical interest, as having been used by Pasteur in showing, for example, that certain torulæ produce no alcohol in the course of their development.1

1 Ann, Chim. anal, 1914, 19, 169; Pharm. J., 1914, 39, 47.

Minute traces of alcohol were detected by distilling the liquid under examination from a long-necked retort connected with a Liebig's condenser. "A certain sign of the presence of alcohol is contained in the first few drops distilled; these always assume the form of little drops or striae, or, better still, oily tears, when alcohol is present in the distillate." In this way, taking 100 c.c. of the liquid, the presence of 0 1 per cent. of alcohol can be detected, and with care and practice even 0 01 per cent. Distilling one-third of the liquid, and redistilling, much smaller proportions can be detected.

Klocker has modified this process, so that it is possible to detect as little as 0.002 or 0.001 per cent. of alcohol in liquids such as yeast-water or wort.2 Five c.c. of the liquid are placed in a test-tube, 18 cm. long by 2 4 cm. diameter, fitted with a cork through which passes a glass tube of length 80 cm. and external diameter 3 mm. The lower end of the tube extends only to the bottom of the cork. The test-tube is fixed vertically over a piece of wire gauze, and is heated by a small flame. To promote regular ebullition, a small spiral of thin copper wire may be placed in the test-tube. Frothing may be minimised by adding 3 drops of strong nitric acid.

The presence of alcohol in the liquid is shown by the formation of drops having a characteristic oily appearance (" tears ") in the long tube. The higher these drops appear in the tube, the smaller is the quantity of alcohol present.

This reaction is not shown by any substance, except alcohol, likely to be present in fermenting liquids.