The lability (mobility and bioavailability) of metals varies significantly with soil properties for similar total soil metal concentrations. We studied desorption of Cu, Ni and Zn, from 15 diverse, unamended soils. These studies included evaluation of the effects of soil:solution extraction ratio and the roles of soil properties on metal desorption. Dcsorption was examined for each metal by computing distribution coefficients (Kd) for each metal in each soil where Kd = [M]soil/[M]solution, Results from soil:solution ratio studies demonstrated that Kd values for the metals tended to increase with increasing soil:solution ratio. This result also held true for distribution of soil organic matter (SOM). Because the soil:solution ratio has a significant effect on measured metal distributions, we selected a high soil:solution ratio to more closely approach natural soil conditions. Copper showed strong affinity to operationally defined dissolved organic matter (DOM). In this study, DOM was operationally defined based on the total organic carbon (TOC) content in 0.45-microm or 0.22-microm filtrates of the extracts. The Kd of Cu correlated linearly (r2 = 0.91) with the Kd of organic matter (Kd-om) where the Kd-om is equal to SOM as measured by Walkley-Black wet combustion and converted to total carbon (TC) by a factor of 0.59. These values representing solid phase TC were then divided by soluble organic carbon as measured by TOC analysis (DOM). The conversion factor of 0.59 was employed in order to construct Kd-om values based on solid phase carbon and solution phase carbon. SOM plays a significant role in the fate of Cu in soil systems. Soil-solution distribution of Ni and Zn, as well as the activity of free Cu2+, were closely related to SOM, but not to DOM. Kd values for Ni, Zn and free Cu2+ in a particular soil were divided by the SOM content in the same soil. This normalization of the Kd values for Ni, Zn, and free Cu2+ to the SOM content resulted in significant improvements in the linear relationships between non-normalized Kd values and soil pH. The semi-empirical normalized regression equations can be used to predict the solubility of Ni and Zn and the activity of free Cu2+ as a function of pH.