Several Zn-deficient soils from the major rice growing areas in the USA were characterized with respect to extractable and labile Zn and other parameters including available P, pH, organic matter, and cation exchange capacity. A greenhouse experiment was conducted to determine whether the above factors influence the response of several rice cultivars to ZnSO4 or ZnEDTA applied as a mixed preplant treatment or to the floodwater at panicle initiation. Response of the rice to Zn varied widely among soils with labile Zn (L value) being a poor indicator of Zn availability and soil pH, and DTPA or 2 N MgCl2 extraction of Zn being the most reliable indices. Soil incorporation prior to flooding was more effective than floodwater application for the initial crop, but both methods of placement were comparable for a subsequent ratoon crop. In either situation, ZnSO4 was superior to ZnEDTA. Recovery of Zn from ZnSO4 was generally less than 5 percent where Zn response was obtained and up to 14 percent from nonresponsive soils.