Parasitological and selected microbiological safety aspects of Nham (Thai-style fermented pork sausage) were evaluated. Pork meat from pigs experimentally infected with Trichinella spiralis was divided into 4 portions; 3 portions were frozen at -23°C for 12, 18 and 24 hours before they were made into Nham. The formula for fermented sausage included 3% w/w NaCl and garlic, 0.015% each of NaN0₂ and NaNO₃, and either 1.0 or 1.5% w/w commercial starter culture. Fermentation was at 30°C for 96 hours. The higher level of starter culture resulted in faster acid production but no difference in the maximum number of lactic acid bacteria attained. Infected pork, after being frozen for 24 hours, was still found to cause a low level of parasitism. However, the corresponding fermented samples were noninfective. Use of 1.0% starter culture and 12 hours freezing of meat prior to grinding rendered trichinae larvae noninfective after 2 days of sausage fermentation. Staphylococcus aureus (10³ cells per g) and Escherichia coli (10⁴ cells per g) were introduced into the ground pork in a series of experiments. Starter culture levels used were 0.75% and 1.5% by weight. Fermentation without starter culture stabilized the numbers of E. coli but permitted slow multiplication of S. aureus. Adding starter culture at either level increased the rate of acid production during the initial days, resulting in a faster drop in pH and a decline in numbers of both S. aureus and E. coli. After 36 and 48 hours, viable S. aureus were not recovered from products with 1.5% and 0.75% starter culture, respectively. E. coli disappeared from sausages with 1.5% starter culture after 96 hours but persisted at low numbers when the lower level of starter culture was used. The rapid direct plating technique (RPT) for enumeration of E. coli was compared to the two standard methods generally used: plating on Violet Red Bile Agar (VRBA) and Host Probable Number (MPN). Total coliforms from 43 fermented pork sausages were enumerated. The RPT gave data highly correlated to those from VRBA but with a significantly lower correlation to MPN. The regression line of loglO coliform counts from the RPT and VRBA had a slope of 0.76 and an intercept of 0.98. The MPN assay gave significantly higher (P=0.05) numbers than either the VRBA or the RPT. / Graduation date: 1986