A series of experiments was conducted using faecal samples collected from commercial swine farms to evaluate the effects of variation in methods used for the detection of Salmonella bacteria. The primary objective of the studies was to compare the protocols routinely used in two laboratories in the USA. The studies included five experiments comparing the enrichment protocols used routinely in the respective laboratories (Method 1:10 g faeces-buffered peptone water (BPW) pre-enrichment-selective enrichment in Rappaport/Vassiliadis (RV) broth; Method 2: similar to 1g faeces-primary enrichments in tetrathionate and Hajna GN broths-secondary enrichment in RV broth). The effects of enrichment temperatures (37 vs 42 degrees C) using RV broth (two experiments) and delayed secondary enrichment (four experiments) were also evaluated. Direct comparison of Method 1 and Method 2 indicated comparable results. However, when compared using faecal samples of equal weight, the Method 2 enrichment protocol was more sensitive for detecting Salmonella bacteria than the Method 1 protocol. Enrichment in RV at 42 degrees C was superior to 37 degrees C, particularly for samples that were pre-enriched in BPW. Delayed secondary enrichment increased detection of Salmonella bacteria in swine faeces. These results highlight the imperfect sensitivity of culture methods, and the need for researchers to consider the sensitivity of bacteriological methods in the design and interpretation of the results of epidemiologic studies based on faecal culture.