Forty soils from top layer (0-20 cm) were sampled in different regions of China and Kjeldahl, HF-Kjeldahl and double treatment
methods were used to determine total N, total N plus fixed ammonium, and total N and the residual fixed ammonium left in soil after determination of total N, respectively, to evaluate if Kjeldahl’s
method could include the fixed N by soil
minerals. The fixed N by soil minerals was measured by Silva-Bremner procedure to make comparison. Results showed that total N determined by Kjeldahl’s method averaged 1.622 g kg-1, while that by HF- Kjeldahl’s method 1.633 g kg-1, and that by double procedure 1.666 g kg-1. Obviously results obtained by the last two methods, particularly the double treatment method, were higher than Kjeldahl’s, showing that Kjeldahl’s method could not or not fully release N fixed by 2:1 minerals in soil, and therefore the determined results would not be the true “total N” for soils that contained large amount of the fixed N. The mineral fixedN averaged 166 mg kg-1, accounting for 10.1% of the total N while the residual fixed N amounted to 30.4 mg kg-1, equivalent to 1.9% of the total N or 18.3% of the total fixed N. The residual fixed N was correlated neither to organic matter nor to total N, but closely related to the total fixed N with a correlation coefficient of 0.598 (n=40), showing that the fixed N was the sole source of the residues. Soils having high residues of the fixed N were just those containing high fixed N, and soils containing high fixed N were just those containing high amount of 2:1 minerals. As a result, Kjeldahl’s method could not give a true value of the total N for such soils. However, for those containing small or little amount of 2:1 minerals, there was no significant difference in results measured by these methods.