Three sample preparation procedures, soak, stomach, and blend, were evaluated using the Bacteriological Analytical Manual Salmonella culture method with eight types of leafy green produce. In the soak method, test portions were added to lactose broth without homogenization; in the stomach method, test portions were stomached with lactose broth; and in the blend method, test portions were blended with lactose broth. Twenty artificially contaminated test portions were analyzed with each procedure in individual experimental trials. The number of test portions identified as positive were compared among the procedures. Statistically significant differences were identified with Fisher's exact two-tailed F test (P < 0.05). Where differences did occur (P < 0.05), the soak procedure was the most effective or was at least as effective as homogenization by either blending or stomaching. Statistically significant differences most frequently occurred with romaine lettuce and cabbage; for these items, blending was significantly less effective than the soak procedure. Overall, for all of the produce types examined, results showed that the soak procedure was more effective than either of the homogenization procedures in recovering Salmonella from leafy green produce. Of the 540 test portions examined by each sample preparation method, 344 were positive for the presence of Salmonella by soaking, 293 by stomaching, and 232 by blending. We recommend that the soak procedure replace homogenization for the analysis of leafy green produce because the soak procedure is more productive than homogenization by either blending or stomaching of the leafy green produce types as reported herein.