Aims: Early detection of spoilage fungi (two Eurotium spp., a Penicillium chrysogenum species) in bread analogues over periods of 72 h at 25degreesC and 0.95 water activity was evaluated using volatile production patterns, hydrolytic enzyme production, and changes in fungal populations. Methods and Results: Using an electronic nose system it was possible to differentiate between uninoculated controls and samples contaminated with P. chrysogenum within 28 h. After 40-48 h it was possible to differentiate between the Eurotium spp., P. chrysogenum and the control using Principal Component Analysis (PCA). Cluster analyses could differentiate between the control, P. chrysogenum and the Eurotium spp. after 40 h. Of seven hydrolytic enzymes examined after 48 h, the specific activities of three were significantly different from uninoculated control bread. There were also differences between the mould species in production of three enzymes. Penicillium chrysogenum populations increased fastest, from about 10(3) cfu g(-1) to 10(6)-10(7) cfu g(-1) after 72 h. For the Eurotium spp. this increase was slower. Conclusions: Overall, this study suggests, for the first time, that an electronic nose system using a surface polymer sensor array is able to detect qualitative changes in volatile production patterns for the early detection of the activity of spoilage moulds in bakery products. Significance and Impact of the Study: Potential exists for application of such systems for microbial quality assurance in intermediate moisture food products.