1. The reliability of the Kjeldahl method for the determination of nitrogen in soils has been investigated using a range of soils containing from 0·03 to 2·7% nitrogen.2. The same result was obtained when soil was analysed by a variety of Kjeldahl procedures which included methods known to recover various forms of nitrogen not determined by Kjeldahl procedures commonly employed for soil analysis. From this and other evidence presented it is concluded that very little, if any, of the nitrogen in the soils examined was in the form of highly refractory nitrogen compounds or of compounds containing N—N or N—O linkages.3. Results by the method of determining nitrogen in soils recommended by the Association of Official Agricultural Chemists were 10–37% lower than those obtained by other methods tested. Satisfactory results were obtained by this method when the period of digestion recommended was increased.4. Ammonium-N fixed by clay minerals is determined by the Kjeldahl method.5. Selenium and mercury are considerably more effective than copper for catalysis of Kjeldahl digestion of soil. Conditions leading to loss of nitrogen using selenium are defined, and difficulties encountered using mercury are discussed.6. The most important factor in Kjeldahl analysis is the temperature of digestion with sulphuric acid, which is controlled largely by the amount of potassium (or sodium) sulphate used for digestion.7. The period of digestion required for Kjeldahl analysis of soil depends on the concentration of potassium sulphate in the digest. When the concentration is low (e.g. 0·3 g./ml. sulphuric acid) it is necessary to digest for several hours; when it is high (e.g. 1·0 g./ml. sulphuric acid) short periods of digestion are adequate. Catalysts greatly affect the rate of digestion when the salt concentration is low, but have little effect when the salt concentration is high.8. Nitrogen is lost during Kjeldahl analysis when the temperature of digestion exceeds about 400° C.9. Determinations of the amounts of sulphuric acid consumed by various mineral and organic soils during Kjeldahl digestion showed that there is little risk of loss of nitrogen under the conditions usually employed for Kjeldahl digestion of soil. Acid consumption values for various soil constituents are given, from which the amounts of sulphuric acid likely to be consumed during Kjeldahl digestion of different types of soil can be calculated.10. Semi-micro Kjeldahl methods of determining soil nitrogen gave the same results as macro-Kjeldahl methods.11. The use of the Hoskins apparatus for the determination of ammonium is described.12. It is concluded that the Kjeldahl method is satisfactory for the determination of nitrogen in soils provided a few simple precautions are observed. The merits and defects of different Kjeldahl procedures are discussed.