The Penman-Monteith formulation of evaporation has been criticised for its reliance upon point estimates so that areal estimates of wetland evaporation based upon single weather stations may be misleading. Typically, wetlands comprise a complex mosaic of land cover types from each of which evaporative rates may differ. The need to account for wetland patches when monitoring hydrological fluxes has been noted. This paper presents work carried out over a wet grassland in Southern England. The significance of fetch on actual evaporation was examined using the approach adopted by Gash (1986) based upon surface roughness to estimate the fraction of evaporation sensed from a specified distance upwind of the monitoring station. This theoretical analysis (assuming near-neutral conditions) reveals that the fraction of evaporation contributed by the surrounding area increases steadily to a value of 77% at a distance of 224 m and thereafter declines rapidly. Thus, point climate observations may not reflect surface conditions at greater distances. This result was tested through the deployment of four weather stations on the wetland. The resultant data suggested that homogeneous conditions prevailed so that the central weather station provided reliable areal estimates of reference evaporation during the observation period March-April 1999. This may be a result of not accounting for high wind speeds and roughness found in wetlands that lead to widespread atmospheric mixing. It should be noted this analysis was based upon data collected during the period March-April when wind direction was constant (westerly) and the land surface was moist. There could be more variation at other times of the year that would lead to greater heterogeneity in actual evaporation.