(2) Lange's Test.--This is a modification of the well-known Legal test. It is more sensitive and gives a sharper end-reaction. To a small quantity of urine add about one-twentieth its volume (1 drop for each 1 c.c.) of glacial acetic acid and a few drops of fresh concentrated aqueous solution of {86} sodium nitroprussid, and gently run a little ammonia upon its surface. If acetone be present, a purple ring will form within a few minutes at the junction of the two fluids. (3) Trommer's Test.--This new test has proved very satisfactory in the hands of the writer. The urine need not be distilled. Alkalinize about 10 c.c. of the urine with 2 or 3 c.c. of 40 per cent. caustic soda solution, add 10 or 12 drops of 10 per cent. alcoholic solution of salicylous acid (salicyl aldehyd), heat the upper portion nearly to the boiling-point, and keep at this temperature five minutes or longer. In the presence of acetone a purplish-red color appears in the heated portion. * * * * * (2) Diacetic acid occurs in the same conditions as acetone, but is less frequent and has more serious significance. In diabetes its presence is a grave symptom and often forewarns of approaching coma. It rarely or never occurs without acetone. _Detection_.--The urine must be fresh. * * * * * (1) Gerhardt's Test.--To a few cubic centimeters of the urine add solution of ferric chlorid (about 10 per cent.) drop by drop until the phosphates are precipitated; filter and add more of the ferric chlorid. If diacetic acid be present, the urine will assume a Bordeaux-red color which disappears upon boiling. A red or violet color which does not disappear upon boiling may be produced by other substances, as phenol, salicylates, and antipyrin. (2) Lindemann's Test.--To about 10 c.c. of urine add 5 drops 30 per cent. acetic acid, 5 drops Lugol's solution, and 2 or 3 c.c. chloroform, and shake. The chloroform does not change color if diacetic acid be present, but becomes reddish-violet in its absence. This test is claimed by its advocates to be more sensitive and more reliable than Gerhardt's. * * * * * {87} (3) Oxybutyric acid has much the same significance as diacetic acid, but is of more serious import. There is no satisfactory clinical test for it.