Summary. Microbial generation of toxic gases from antimony, arsenic, or phosphorus in compounds used as fire retardants in cot mattresses has been proposed as a cause of sudden infant death. To test this hypothesis, 23 polyvinyl chloride mattress samples from cot death cases were incubated on malt agar plates until good microbial growth was obtained. Silver nitrate and mercuric chloride test papers were then inserted and the colour reactions recorded. The predominant organism, recovered from all mattresses tested, was not, as claimed in earlier work, the fungus Scopulariopsis brevicaulis, but a mix of common environmental Bacillus spp. Test paper colour changes occurred whenever bacterial growth was present, but these reactions also occurred in control tests in which no mattress material was present on the plates. Chemical and instrumental analyses of exposed test papers showed that the colour reactions were not due to deposits of antimony, arsenic, or phosphorus. Our findings do not support the hypothesis that toxic gases derived from antimony, arsenic, or phosphorus are a cause of sudden infant death. More sulphur was found in test papers exposed in plates containing bacterial growth than in those without such growth. This result suggests that the test paper reactions were due to the generation of sulphur-containing compounds during bacterial growth on the agar medium.