AIM: To test the hypothesis that sudden infant death syndrome (SIDS) may be caused by toxins of commonly occurring bacteria in infants lacking developed immunity. METHODS: Nasopharyngeal microbial isolates from 22 pairs of SIDS cases and healthy infants matched for age (by month), sex, and sampling time (by month) were compared for lethal toxigenicity. Crude toxin preparations were made from isolates cultured on dialysis membrane overlaid on agar, and these preparations were then tested for lethality by intravenous injection into 11 day old chick embryos. RESULTS: Fifteen (68%) of the SIDS cases were each found to have at least one lethally toxigenic organism in their nasopharyngeal flora; only eight (36%) of the flora of normal infants included a lethally toxigenic species. CONCLUSION: Infants who have died of SIDS have a significantly higher (p less than 0.05) probability than matched healthy infants of having a lethally toxigenic bacterial species in their nasopharyngeal flora.