The increasing use of prescribed fire in forest management and the continuing burning of agricultural crop residues creates problems in air pollution. More information is needed on yields of pollutant gases and particulates and how these emissions might be altered by varying burning practices. These factors are being evaluated in large scale laboratory type experiments and in field trials. Ten- to fifty-pound samples of several fuel types are burned under a cone-shaped tower which simulates open burning but permits continuous analysis of total hydrocarbons, CO, CO2, and O2. Grab samples are taken at two or more intervals during the fire for analysis of NO2 and individual hydrocarbons. Particulates are collected isokinetially throughout the fires. Rate of weight loss of the burning fuel, temperature and airflow are also recorded. From these data, yield of pollutant in pounds per ton of fuel burned are calculated. Field plots of forest floor fuels and straw and stubble from cereal crops are also burned. A ten-foot tower instrumented for temperature, CO2 and particulate isset up in plots in grain fields and samples the smoke as fires burn past the tower. A modified version of Boubel's portable pariculate sampler, which also provides for gas sampling, is being evaluated for possible use in forest plots. For both laboratory and field fires, variables being tested for effect on pollution are fuel type, loading, moisture level, methods of ignition, piled versus spread fuels, and head versus back fires.