During a measurement period from June till November 2004, ammonia fluxes above non-fertilized managed grassland in The Netherlands were measured with a Gradient Ammonia—High Accuracy—Monitor (GRAHAM). Compared with earlier ammonia measurement systems, the GRAHAM has higher accuracy and a quality control system. Flux measurements are presented for two different periods, i.e. a warm, dry summer period (from 18 July till 15 August) and a wet, cool autumn period (23 September till 23 October). From these measurements canopy compensation points were derived. The canopy compensation point is defined as the effective surface concentration of ammonia. In the summer period (negative) deposition fluxes are observed in the evening, night and early morning due to leaf surface wetness, while in the afternoon emission fluxes are observed due to high canopy compensation points. The mean NH 3-flux in this period was 4 ng m −2 s −1, which corresponds to a net emission of 0.10 kg N ha −1 over the 28 day sampling period. The NH 3-flux in the autumn period mainly shows (negative) deposition fluxes due to small canopy compensation points caused by low temperatures and a generally wet surface. The mean NH 3-flux in this period is −24 ng m −2 s −1, which corresponds to a net deposition of 0.65 kg N ha −1 over the 31 day sampling period. Frequency distributions of the NH 3-concentration and flux show that despite higher average ambient NH 3-concentrations (13.3 μg m −3 in the summer period vs. 6.4 μg m −3 in the autumn period) there are more emission events in the summer period than in the autumn period (about 50% of the time in summer vs. 20% in autumn). This is caused by the high canopy compensation points in summer due to high temperatures and a dry surface. In autumn, deposition dominates due to a generally wet surface that induces low canopy compensation points. For our non-fertilized agricultural grassland site, the derived canopy compensation points (at temperatures between 7 and 29 °C) varied from 0.5 to 29.7 μg m −3 and were on an average 7.0 μg m −3, which is quite high for non-fertilized conditions and probably caused by high nitrogen inputs in the past or high dry deposition amounts from local sources. The average value for the ratio between NH 4 + and H + concentration in the canopy, Γ c, that was derived from our data was 2200.