A total of 228 intact male pigs form Duroc, Hampshire, Landrace, and Yorkshire breeds were used in the experiment. Samples of salivary gland and backfat were collected at slaughter for colorimetric assay of salivary and fat 16-androstene levels and fat skatole levels. Fat levels also were tested by a sensory panel using an R-index technique for detecting the presence of boar taint. The proportion of tainted carcasses determined by the sensory panel was 5.0% for androstenone and 11.4% for skatole, with a combined total of 15.0% tainted from either source. Sensory analysis of taint showed a lower proportion (P < .05) of tainted carcasses in Hampshire, with no difference in taint across the other three breeds. Analysis of taint compounds indicated that overall 14.5% of pigs had salivary gland 16-androstene levels and 20.9% had fat 16-androstene levels above acceptable limits. There was a higher (P < .05) proportion of Duroc pigs above the threshold levels for 16-androstenes in both salivary gland and fat. Landrace pigs had the lowest (P < .05) average tissue concentrations of steroids and skatole. Across breeds, only 1.8% of pigs had fat skatole concentrations above .25 ppm, which has been suggested as threshold levels of skatole for taint. The canonical correlation coefficient between fat compound levels and the R-indices of fat 16-androstenes and skatole was .40 (P < .001). Our results indicate breed differences in tissue levels of taint compounds and in taint assessed by a sensory panel. Levels of 16-androstene steroids were highly associated with taint, but more pigs had measured levels above the threshold than those identified as tainted by sensory analysis. Levels of fat skatole were low overall and did not account for all the pigs judged as tainted from skatole by sensory analysis.