Abstract
Considering the increasing need for reliable soil carbon (C) and nitrogen (N) determinations, 19 Brazilian soil samples were selected to compare the results of two methods (a) a modified dichromate wet digestion and Kjeldahl that has been in use by Embrapa in most of Brazilian soil surveys (Embrapa's method), and (b) an automated technique based on dry combustion (Perkin-Elmer 2.400 Series II–CHN Mode). Considering that the last method is very much influenced by sample size, two weights were tested (10 and 100 mg). The influence of two particle sizes for both methods (crushed and sieved to pass a 0.177-mm mesh) was also tested. A simple linear regression with the confidence limits for the intercept and slope were used to compare methods. Precision was evaluated by the coefficient of variation (C.V.). The results showed that simple crushing is inadequate for performance analysis by dry combustion mainly for small samples (10 mg), where the precision is also low (19.77% for N and 9.41% for C). The dry combustion method is significantly affected by the use of small samples (10 mg), mainly for N analysis. For finely ground samples (< 0.177 mm), a comparison of the results obtained by dry combustion (100 mg weight sample) and by Embrapa's method showed a significant 1:1 relation. Thus the use of correction factor is not necessary.