One hundred and ninety-six boneless pork roasts were vacuum-packaged. Twenty-eight roasts remained vacuum-packaged to serve as controls. The remaining packages were divided into six groups of 28 packages each and injected with one of six gas mixtures: (a) 100% O2, (b) 20% CO2 + 80% N2, (c) 50% CO2 + 50% O2, (d) 20% CO2 + 80% O2, (e) 25% CO2 + 25% O2 + 50% N2 or (f) 51% CO2 + 30% O2 + 18% N2 + 1% CO. Five packages from each packaging treatment were removed from storage after 7, 14, 21, 28 or 35 days. Data collection included gas composition of the intact packages and off-odor, surface discoloration, overall appearance, retail caselife and palatability of cuts. The most prevalent gas in vacuum packages was CO2 with weight percentages of 62–88% during the 35-day storage period. Packages initially injected with an atmosphere containing 100% O2 showed a gradual decrease in O2 and an increase in CO2 with increased storage. Roasts stored in O2-containing atmospheres for 14 and/or 21 days had a higher incidence of off-odor and chops from such roasts had lower overall appearance ratings after 1 day of retail display and lower flavor and overall palatability ratings than comparable vacuum packaged meats. These differences were significant only for roasts stored in modified atmospheres containing high concentrations of O2 and only after extended periods of storage. Data suggest that a modified gas atmosphere of 20% CO2 + 80% N2 is a suitable alternative to vacuum-packaging.