Cot mattress materials were investigated as potential reservoirs of bacteria in relation to the sudden infant death syndrome (SIDS). The sleeping position of the infant significantly influenced bacterial population density of cot mattress polyurethane foams (p<0.0000001) and their covers (p<0.004). Staphylococcus aureus was isolated at significantly higher frequency (p<0.03) from the infant's head region of cot mattress materials. Significantly higher bacterial population densities (p<0.001) were associated with polyurethane foams from non-integral mattresses (exposed polyurethane foam), when compared to those from mattresses completely covered by polyvinyl chloride (integral type mattress). The frequency of isolation of S. aureus from polyurethane foams from non-integral mattresses was also significantly higher (p=0.03) than from foams from the integral type. The following factors were significantly associated with increased frequency of isolation of S. aureus: from the polyurethane foam, previous use of non-integral mattresses by another child (p=0.03 for all sample sites, p=0.01 for torso region); from the covers, sleeping in the prone position (p=0.003 head region, p=0.001 torso region). Prone sleeping was also significantly associated with increased bacterial population levels (p=0.01) and increased frequency of isolation of Escherichia coli (p=0.02) from the torso region of cot mattress covers. These findings could explain some recently identified risk factors for SIDS associated with type and previous use of cot mattresses. Clostridium perfringens was isolated at very low frequency and Streptococcus pyogenes was not isolated from any cot mattress materials tested.