The method proposed by Koroleff of oxidizing total nitrogen in natural waters to nitrate by means of peroxodisulphate in alkaline medium at 120°C is studied in some detail applying earlier experiences concerning the reduction of nitrate by cadmium and the determination of nitrite formed as an azo dye. The choice of the proper concentrations of peroxodisulphate and sodium hydroxide for the digestion is critically discussed as well as the choice of the proper pH and buffer system for the reduction. As a result a new reagent is proposed whose concentrations of peroxodisulphate and sodium hydroxide are founded on the stoichiometric relations at the autodecomposition of the peroxodisulphate; the proper pH of 8.0 for the reduction is achieved by addition of the calculated amount of Tris(hydroximethyl)aminomethane hydrochloride. A method for the preparation of potassium peroxodisulphate of a very low nitrogen blank is given. The digestion is performed at 100° or at 120°C. The requisite digestion time, a function of the autodecomposition of the peroxodisulphate, is found to be 90 min at 100°C and 15 min at 120°C. The yields obtained by oxidation of various nitrogen compounds are studied. The same yields are obtained at both temperatures but are dependent on the form of the nitrogen linkage. Thus, for example, nitrate, nitrite, ammonia and urea give a yield of >99%, the yield for some aliphatic amino acids varies from 92 to >99% and the yield for some proteins from 92 to 95% while a double bond between nitrogen atoms seems entirely to prevent their oxidation to nitrate. The requisite excess of peroxodisulphate is dependent not only on the amount of nitrogen in the sample but rather on the total amount of oxidizable substances, primarily organic matter; the COD of the sample portion should not appreciably exceed 10% of the oxygen available from the peroxodisulphate. A study of the concentration dependence at constant ratio reagent to nitrogen compound shows that the yield as a rule decreases with increasing dilution (exceptions are glycine and ammonia) thus often preventing the use of large sample volumes at low nitrogen content in order to increase the accuracy. Aspects on the practical use of the method are presented together with some illustrations, also including its application on sea water which can be analyzed by the method if only care is taken to keep up the hydroxide ion concentration which is lowered by precipitation of magnesium hydroxide. The expected precision and accuracy are discussed. While the precision by analysis of clear waters of not too low nitrogen content need not be poorer than 1%, the accuracy will be dependent on the nature of the nitrogen compounds and their relative concentrations and a negative error varying from 0 to 8%, all according to these circumstances, seems probable.