Ammonia emissions during composting of hog manure mixed with sawdust were studied in four runs comprising a total of 22 pilot-scale reactor vessels. These four runs extended previous work and both verified and extended the previous conclusions. The pilot-scale vessels were 205 L, insulated, stainless steel drums that were aerated either continuously (high/low thermostatically controlled fans) or intermittently (5 min high fan, 55 min off). Temperatures, ammonia emissions, air flow rates, carbon dioxide production and oxygen utilization, moisture and dry matter reduction, and initial and final chemical compositions were measured. Ammonia emissions from the intermittently aerated vessels were only about 50% as great as those from the continuously aerated ones, but this was found to be a result more related to total air flow than to aeration technique. All of the data for total ammonia emissions versus total air flow were fitted with a linear regression line, y=0.1309x+29.835 where y is ammonia expressed as g of N and x is air flow in kg, with R2=0.6808. This general trend indicates that about 50% reduction in ammonia emissions can be achieved with 75% reduction in air flow. For the aeration techniques used, the minimum oxygen level in the exhaust gas from the vessels was 5%, and this is probably a reasonable lower limit constraining air flow reduction. However, within this constraint, lower air flow now appears to be a technique that can reduce odorous ammonia emissions.