S. The reason for the reported difference in spoilage behaviour of skim and whole pasteurised milks was investigated. The rates of growth of psychrotrophic bacteria were not significantly different in the two milks and the bacterial types, all pseudomonads, present at spoilage were also similar. However, when representative spoilage organisms were cultured into freshly pasteurised skim and whole milks, skim milks exhibited predominantly bitter flavours while whole milk showed mostly sour flavours. The different spoilage behaviours can be largely explained by greater proteolysis in skim milk than in whole milk, caused by higher production of protease and greater susceptibility of the protein to protease attack. In addition, lipolysis in whole milk, caused by the substantial quantities of lipase produced by spoilage pseudomonads in this milk, also contributes to the different flavours produced during cold storage of these milk types.K : Spoilage, skim milk, proteolysis, protease, lipase Skim milk has gained a reputation for having a shorter shelf life than whole milk. This is evident in reports from various countries (Langlois et al. 1966 ;Janzen et al. 1982 ;Brown et al. 1984 ; Pieper & Tims, 1987 ;Griffiths, 1989 ;Heo, 1989), and in complaints received by some dairy companies. Janzen et al. (1982) reported that the average shelf lives of skim milks stored at 4n5 and 7 mC were less than those of the corresponding whole milks by 2n3 d and 1 d, respectively, although at 10 mC, there was no difference. Pieper & Tims (1987) found that skim milk spoiled 3-4 d sooner than whole milk at 4n5 mC and, in one of two experiments, at 7 mC also. Heo (1989) examined commercial milk samples held at 7n2 mC and found that after 10 d storage, 91 % of both whole and skim milks were acceptable, whereas after 14 d, 82 % of whole milks and only 55 % of skim milks were still acceptable.The reason for this difference is unknown. Some workers have assumed that the difference in spoilage rates is due to different rates of growth of psychrotrophic bacterial contaminants in the different milks. However, it has been shown that the bacterial numbers in the different milks after incubation under identical conditions are not significantly different (Janzen et al. 1982 ;Brown et al. 1984 ;Allen et al. 1989 ;