Data on the prevalence and populations of pathogens in individual foods are critical to the development of product-specific quantitative microbial risk assessments. An outbreak of salmonellosis associated with the consumption of raw almonds in 2000 to 2001 provided an opportunity to evaluate the levels of Salmonella in the recalled product. Duplicate 100-g samples from each of fifty 22.7-kg boxes of recalled almonds were enriched by one of two methods. Salmonella was isolated by at least one method from 42 boxes (84% positive). The levels of Salmonella determined by a three-tube most-probable-number (MPN) method were 8.5+/-1.3 MPN/100 g. In a subsequent study, raw almonds that arrived at almond processors were sampled from 2001 through 2005 to determine the overall prevalence and levels of Salmonella and to characterize the Salmonella isolates obtained. Aerobic plate counts, coliform counts, and MPN levels of Escherichia coli were also determined on positive samples. An isolation frequency for Salmonella of 81 (0.87%+/-0.2%) of 9,274 samples tested (100 g) was determined for raw almonds sampled from throughout California over the 5-year period. Salmonella was not isolated upon retesting in 59 of 65 positive samples. When detected, levels were 1.2 to 2.9 MPN/100 g. Of the 81 total isolates, 35 different serotypes of Salmonella were represented. Aerobic plate counts, coliform counts, and E. coli levels did not correlate with the presence of Salmonella.