Adsorption and desorption of picloram (4-amino-3,5,6-trichloropicolinic acid) in soils was studied utilizing both batch technique and soil columns. The five soils from the states of Utah and California ranged from 0.6 to 18.7 percent in organic matter content, 5 to 21 percent in sesquioxides and 6.45 to 7.55 in pH. In batch studies, the factors affecting picloram adsorption-desorption included time, soil type, organic matter, herbicide concentration, temperature, pH, and inorganic electrolytes. The soils adsorbed 10.8 to 58.2 percent of the added picloram from 0.05 to 10,0 ppm of picloram aqueous solutions. Adsorption was highly correlated with organic matter (r • 0.99), pH (r •- 0.98), and aesquioxides content (r • 0.96) of the soils, Adsorption was significantly correlated with cation exchange capacity of the soils at the 0.05 level. Equilibration timo rongod from 4 to 120 hours. An incraaa in the temperature from 17.7 C to 25 C cauaad 1ncroaa•d adsorptin; raising the temperature further resulted in a decrease in adsorption. The adsorption followed the Freundlich model, but not the Langmuir, Tempkin, or B.E.T. models. Adsorption was inversely related to induced pH changes from pH 3.6 to pH 9.2. From 41.0 to 71.8 percent of the initially adsorbed picloram could be eluted with two successive extractions of deionized water from the three soils at 25 C. The desorption thereafter was slow. No evidence was found of picloram precipitation with the addition of inorganic electrolytes in aqueous solutions without soil. Divalent inorganic cations were , generally, more effective in increasing adsorption than monovalent cations. Studies of the effect of CaC12 concentrations (from 5 to 500 me/ 1) on picloram adsorption by four soils suggested an exchange type reaction, picloram acting as a cation. These interpretations were not conclusive. Column studies with two of the soils, at 1.0 and 10.0 ppm picloram concentrations and in the natural state as well as after calcium saturation, corrobated the findings of batch studies; inorganic salts did impede picloram movement in soils, presumably by increasing picloram adsorption. Applicability of a mathematical model, based on chromatographic theory, to predict picloram movement in two soils was tested. The model overestimated adsorption. However, when a lower retardation factor was used, the prediction was reasonable on the adsorption side of the herbicide distribution curve, but not the desorption side. This suggested that a distribution coefficient, based on its change with soil depth, may improve the predictability of the model.