RESUMEN
This dissertation is composed of three papers. In the first paper, the prevalence of residential C. botulium in fish and other seafood sold at Davis supermarkets was examined. Prevalence ranged from 0-100%. Although C. botulium types A, B, E and F were detected, type A was predominant. Estimated MPN number in red snapper ranged from 0.9-2.4 and 0.03 to 1.2 organism/g fish, respectively. In the second paper, the suitability of using salmon tissue muscle homogenates (microsystem) as asubstrate for evaluation of nonproteolytic C. botulium growth risk was studied. Using factorial desing experiments and regression approach, the effect and interactions of variables such as temperature, storage time, spore inocolum level, initial microbial fish flora and type of modified atmosphere (MA) on the probability (P) of C. botulium growth and toxigenesis was examines. Several models for estimation of the length of the lag phase (LP) and the probability (P) of growth and toxigenesis by one nonproteolytic spore under a particular storage condition were developed. In the last paper, the risk of C. botulium growth and toxin production in fresh salmon fillets was evaluated. Regression equations were derived to estimate LP and P of toxigenesis. A close agreement between the LPS and Ps estimated in the microsystem and the fillets demonstrated the feasibility of the quantitative approach and the utility of the microsytem for risk analysis. A narrow margin safety existed in fish stored under MA when it was held at slightly abused temperatures