The Na, K and P concentrations in soil extracts obtained by a conventional extraction method, at a soil:extractant (Mehlich-1) ratio of 1:10, were compared with those found at a soil:extractant (Mehlich-1) ratio of 1:5. Also, the data obtained by conventional quantification techniques, in which Na and K are quantified by flame photometry and P by UV-Vis molecular absorption spectrophotometry, were compared with those found by inductively coupled plasma optical emission spectrometry (ICP OES). Fifteen Brazilian soil samples were analyzed. In the repeatability study applied to all results, the coefficient of variation (CV) was highest for P and Na, mainly when the analyte concentrations were lower (< 9 mg dm-3 for P and < 10 mg dm-3 for Na). This fact was probably due to the heterogeneity of the extracts, which contained colloidal particles. Filtration or centrifugation instead of decantation of the extracts would probably result in lower CV. In the reproducibility study for three samples, non-reproducible results were only obtained for K in a single sample. All ICP OES results were similar to those found by UV-Vis molecular absorption spectrophotometry and flame photometry, indicating that the first technique was adequate for Na, K and P determination in the tropical soil extracts obtained with Mehlich-1 solution. The results for Na, K and P at a soil:extractant ratio of 1:5 were statistically different from those obtained at a ratio of 1:10. Most results obtained at a soil:extractant ratio of 1:5 were lower than at 1:10, mainly in samples with higher analyte concentrations. Therefore, for comparative studies involving macronutrients, the conventional method should be used at a soil:extractant (Mehlich-1) ratio of 1:10.
soil analysis; ICP OES; Mehlich-1
