Twenty-four pasteurized infant feeds, prepared in a Glasgow hospital, were examined microbiologically. All produced a satisfactory total aerobic mesophilic count of < or = 1.0 x 10(4) cfu/g (mean 6.3 x 10(1) cfu/g) within 1 h of preparation. Bacillus cereus was detected in two infant feeds immediately after preparation and one of these had a B. cereus count of 1.4 x 10(3) cfu/g exceeding the recommended safety limit of < or = 1.0 x 10(3) cfu/g. Subsequent storage over a 14 h period at 25 degrees C or greater resulted in the appearance of B. cereus in a further eight feeds, the majority of which exceeded the safety limit of 10(3) cfu/g. The microbiological quality of each infant feed depended on the type and number of organisms initially present, and on the temperature and duration of storage. Incubation of feeds at < or = 10 degrees C for 14 h did not alter the microbiological quality (P = 0.05). While Bacillus licheniformis and Bacillus subtilis were the predominant organisms isolated within 8 h of incubation (45.8 and 20.8% of feeds, respectively), additional storage resulted in the emergence of B.cereus I (25%) and II (20.8%) as dominant Bacillus spp. The addition of glucose polymers and other supplements to infant formulae did not affect the type and number of organisms present (P = 0.05). Diarrhoeal enterotoxin was detected in three of the five formulations which supported the growth of B. cereus II via the B. cereus enterotoxin reverse phase latex agglutination test BCET-RPLA system. Although the infant feeds were of similar microbiological quality (P = 0.05), the majority of Bacillus spp. isolated have been previously implicated in either foodborne illnesses and/or opportunist infections.