The predominant oaks of the Maryland Piedmont (Quercus alba and Q velutina) do not occur on serpentine soils. On serpentine sites they are replaced by Q. stellata and Q marilandica. Soil and plant water relations were measured to evaluate water as a limiting factor to distribution of oaks on serpentine soil. May through August soil water potentials of serpentine soils were not lower than adjacent nonserpentine soils. Dawn xylem water potentials of oaks on serpentine soils were similar to nonserpentine oaks during early and midsummer. By late summer, water potentials of the oaks on serpentine sites were higher. Dawn xylem water potentials and soil water potentials were significantly correlated for all species. A significant correlation between xylem water potential and stomatal conductance was found for the oaks on the serpentine soils. No such correlation was found for oaks on nonserpentine soil, but a xylem water potential threshold for stomatal closure was observed. Despite differences in their responses to moisture availability, water does not appear to be the environmental factor responsible for the distributional pattern of these species. INTRODUCTION The occurrence of serpentine soils with their atypical vegetations is a worldwide phenomenon (Walker, 1954). Serpentine soils vary widely in their mineral composition and hence in the edaphic problems encountered by the associated vegetation. Many serpentine soils are high in Mg and have a low Ca:Mg ratio. Toxic levels of Ni, Cr and Co may be present. In addition, one or more mineral nutrients (N, P, K, Mo) may be limiting to plant growth (Proctor and Woodeil, 1975). The physical properties of serpentine soils may also be adverse to plant growth (Proctor and Woodell, 1975). Frequently the soils are severely eroded, shallow and stony. These conditions often result in low soil water availability, which has been hypothesized (Monteferrante, 1973; Miller, 1977) to be an important influence on the vegetation. Whittaker (1954) observed that along a moisture continuum, plant communities of serpentine soils are more xeric than the vegetation of other soils. Few studies have examined the availability of water to plants on serpentine soils. Whittaker (1960) measured soil water in northern California and southern Oregon and found no significant differences between serpentine and nonserpentine soils. Miller (1977) measured soil water parameters (capillary water, permanent wilting percentage and available water) and found no intrinsic differences between a serpentine soil of southeastern Pennsylvania and an adjacent nonserpentine site. However, when he measured the available water in June 1964, there was less available water in the serpentine soil. He concluded that soil moisture and nutrients may have a synergistic effect in limiting the vegetation on serpentine soils. Monteferrante (1973) suggested that, in addition to soil minerals, the shallowness of the soils with their lowered water storage capacity was responsible for the patterns of vegetation. S. N. Turitzen (pers. comm.) studied the diurnal water relations of populations of Smilax rotundifo1ia on and off serpentine soil in southeastern Pennsylvania. He found similar predawn xylem water potentials in both populations, but lower midday xylem water potentials in the serpentine populations. He suggested that differences in exposure accounted for the differences in diurnal water relations. The purpose of this study was to evaluate water relations of oak species found on and off serpentine soils in order to infer if plant moisture stress may be a limiting factor for the oaks on serpentine soils.