Manure can provide valuable nutrients, especially N, for grass forage, but high NH, volatilization losses from standard surface-broadcast application limits N availability and raises environmental concerns. Eight field trials were conducted to evaluate the emission of NH, from liquid dairy manure, either surface broadcast or applied in narrow surface bands with a trailing-foot implement. Manure was applied using both techniques at rates of approximately 25 and 50 m3 ha(-1) on either orchardgrass (Dactylis glomerata L.) on a well-drained silt loam or reed canarygrass (Phalaris arundinacea L.) on a somewhat poorly drained clay soil. Ammonia emission was measured with a dynamic chamber/equilibrium concentration technique. High NH3 emission rates in broadcast treatments, especially at the high rate (2 to 13 kg ha(-1) h(-1)), occurred during the first few hours after spreading, followed by a rapid reduction to low levels (<0.5 kg ha(-1) h(-1) in most cases) by 24 h after spreading and in subsequent days. Band treatments often followed the same pattern but with initial rates substantially lower and with a less dramatic decrease over time. Total estimated NH3 losses from broadcast application, as a percent of total ammoniacal N (TAN) applied, averaged 39% (range of 20 to 59%) from the high manure rate and 25% (range of 9 to 52%) from the low rate. Band spreading reduced total NH3 losses by an average of 52 and 29% for the high and low manure rates, respectively. Results show that the trailing-foot band application method can reduce NH3 losses and conserve N for perennial forage production.