An enzyme immunoassay (EIA) was developed to detect Salmonella in foods. The following variables were examined: appropriate enzyme-antibody conjugates, the use of modified commercial antisera, and optimum test protocols to insure specificity and sensitivity. Alkaline phosphatase was the best enzyme probe; cellular peroxidases interfered when peroxidase conjugates were used. Commercial Spicer-Edwards polyvalent H antiserum was a satisfactory source of specific IgG after purification by staphylococcal protein A affinity chromatography to remove IgM. Purification of commercial antisera was necessary because IgM is the primary immunoglobulin class elicited by cross-reactive 0 antigens. An indirect test protocol was developed for use with microtitration plates or Gilford microcuvettes. Samples from enrichment cultures were mixed with H-specific IgG and allowed to react; unbound antibody was then removed by three 5-min centrifugation washes. Goat anti-rabbit alkaline phosphatase antibody conjugate was added and allowed to react; unbound conjugate was removed by centrifugation washing as before. Positive samples were indicated by the production of a chromogenic reaction product following the addition of alkaline phosphatase substrate. The color could be read visually or quantified by using an absorbance colorimeter. Ninety-eight food samples were examined to compare the new EIA procedure with three other salmonellae screening methods, i.e., enrichment serology, immunofluorescence, and the FDA standard pure culture technique. The EIA was sensitive and specific and possessed certain advantages over other methods currently in use. These included the ability to quantify data, simultaneous multisample analysis, and the requirement of little technical expertise to perform the assay. Furthermore, when the EIA was applied to the screening of pre-enrichment media, the results indicated that it might be decidedly more sensitive than the "standard" pure culture analysis. The feasibility of using the EIA to screen pre-enrichment media and hence, negating the requirement of selective enrichment, shows great potential and warrants further investigation.