Salmonella Enteritidis is a leading cause of gastroenteritis associated with consumption of contaminated poultry meat and eggs. Because pulsed-field gel electrophoresis (PFGE) has limited utility in distinguishing between clonal Salmonella Enteritidis isolates, random amplified polymorphic DNA (RAPD) PCR has been recommended as an alternative molecular fingerprinting tool. This study's objective was to determine whether increasing PCR stringency would improve the repeatability of RAPD DNA patterns based on assessment of target sites within the genome. An in silico PCR was performed to predict amplification products from an Salmonella Enteritidis genome sequence for three different RAPD primers (1247, 1283, and OPA4) and to determine whether any primer would be more likely to amplify variable regions within the genome. A comparison of within- and between-isolate similarities in RAPD patterns was performed using primer 1247, which was predicted by in silico analysis to yield a variable size range of amplicons. In order to reduce artifactual variability associated with the method, three different methods for template preparation were evaluated. All were found to provide comparable results with respect to the similarities observed with repeated analyses of the same Salmonella Enteritidis isolates (n  =  18, P  =  0.91). Although the median within-isolate similarity (76.0%) was significantly greater than the median between-isolate similarity (66.7%; P  =  0.001), duplicate RAPD-PCR runs of the same Salmonella Enteritidis isolates produced DNA patterns that ranged in similarity between 61.5 and 100%. These results indicate that the repeatability of RAPD-PCR is insufficient to distinguish genetic differences among related and unrelated Salmonella Enteritidis isolates.