To 70 cc. of concentrated sulfuric acid (Note 1) in a 1-l. short-necked flask fitted with a mechanical stirrer and immersed in an ice-water bath is added, with vigorous stirring, 10 g. (0.145 mole) of solid sodium nitrite in small portions. The addition is made at such a rate that only a very small amount of nitrogen oxide is evolved. When solution is complete, 20 g. (0.145 mole) of salicylic acid is added in small portions with stirring; about fifteen minutes is required. The mixture is stirred at 20° until all the solid is in solution (Note 2). The mixture should then be light red to brown in color, very viscous, and quite homogeneous. It is surrounded by an ice-salt bath, and, when the temperature reaches 0°, 5 cc. (0.065 mole) of a 35–40 per cent solution of formaldehyde (formalin) (see Note 1 on p. 378) is slowly added with extremely vigorous stirring, at such a rate that the temperature does not rise above 5° (Note 3). The reaction is complete a few minutes after all the formaldehyde has been added. About 100 g. of finely crushed ice is then added, followed by 500 cc. of ice water; the stirring should be vigorous during the addition (Note 4). The contents of the flask are stirred until the aurin tricarboxylic acid has disintegrated into small pieces.

The solid is washed several times by decantation, using cold water, and finally filtered with suction. It is then dissolved in dilute ammonia (1 volume of concentrated ammonia with 2 volumes of water) while it is still on the filter paper in the suction funnel (Note 5). The filtrate is evaporated to dryness on a steam bath. The resulting glassy, light yellowish-brown ammonium salt, which weighs 19–22 g. (83–96 per cent of the theoretical amount) is sufficiently pure (Note 6) for use as a test for aluminum.¹