The accuracy of particle size analyses depends on a complete dispersion of the soil sample and the maintenance of the stability of the dispersed phase. In some Latosols the dispersion is hindered by microaggregates of high stability. These clay particles are partially detached by chemical and mechanical dispersion and because of their size, account for a systematic overestimation of the silt fraction in the analyzed soil. In order to minimize the amount of pseudo-silt in particle-size analyses of Latosols, an experiment was carried out to test how the efficacy of dispersion was affected by the use of pre-treatments for the removal of cementing agents. The treatments were set up in a factorial arrangement of 7 x 2². Study material were samples of seven Latosols dispersed with NaOH 0.01 mol L-1, with or without pre-treatment for the removal of organic matter and poorly-crystalline Fe and Al oxides. The silt fraction was dispersed again, and clay and silt proportions were determined after the second dispersion. The removal of organic matter and poorly-crystallized Fe and Al oxides increased the clay fraction yield, indicating the breaking up of microaggregates in sizes that correspond to other particle size fractions, especially silt. Pseudo-silt was found in all studied Latosols, in greater amounts in the more oxidic types. Coarse and fine sand fractions also presented clayey microaggregates that are not easily dispersed. After the second dispersion, the silt fraction presented values from 0.199 to 0.537 kg kg-1 of clay. Pre-treatments are recommended to reduce the proportion of pseudo-silt, especially the one used to remove poorly-crystalline Fe and Al oxides.
dispersion; pseudo-particles; removal of organic matter; removal of poorly crystalline iron and aluminum oxides
