For the estimation of amino nitrogen in complex fluid mixtures three typical methods are available. These are (
a
) the nitrous acid method (Van Slyke, 1911, 1912), (
b
) the formaldehyde titration (Sørensen, 1907; Malfatti, 1909; Lüers, (1920), and (
c
) the alcohol titration (Foreman, 1920; Willstätter and Waldschmidt-Leitz, 1921). Yet by general agreement none is entirely satisfactory. Method (
a
) is tedious in multiple determinations, is subject to fortuitous error from frothing, and from reducing agents present in solution. It is known to give faulty values for glycine, cystine, tryptophane, arginine, lysine, and glutamine (Schmidt, 1929; Plimmer, 1924; Chibnall and Westall, 1932), and for various peptides (see Thierfelder and von Cramm, 1919; Hopkins, 1929; Dunn, Butler, and Deakers, 1932). As discussed later, it gives a partial estimation of ammonia, urea and allantoin, creatinine, and the aminopurines. It is not influenced by proline (Van Slyke, 1911). Methods (
b
) and (
c
) have much in common. They are both rapid in use, but are of limited accuracy owing to a difficult end-point in complex solution, and of restricted scope since they must be confined to almost colourless extracts. Their quantitative defects are to be discussed, but the methods are best applied serially only to solutions of closely similar chemical composition. As a final option, colurimetry (Harding and MacLean, 1916; Riffart, 1922; Folln and Wu, 1922) is both too uncertain and too difficult to apply to chemically complex, coloured solutions.