The Gaussian plume model (GPM) is considered as a valuable tool in predictions of the atmospheric transport of fungal spores and plant pollen in risk assessments. The validity of the model in this important area of application has not been extensively evaluated. A field experiment was set up to test and––if necessary––adapt the GPM, as applied to the dispersal of spores. Spores of the fern Lycopodium clavatum were released artificially over a period of 10 min from a source placed 70 cm above the surface in a potato crop. Spore catches were made with a network of Rotorod and Burkhard samplers, placed up to 100 m downwind from the source and at several heights and crosswind distances from the anticipated plume axis. The width and height of Gaussian plumes depend on atmospheric mixing, as affected by weather. Mixing parameters in risk assessments are commonly predicted on the basis of weather conditions. We found a low correlation (R=0.4) between measured spore concentrations and predicted spore concentrations, using a widely used prediction method (GPM Pasquill Atmospheric Diffusion, 2nd ed., Wiley, New York, 1974), based on cloud cover, wind speed, season and time of day. More precise methods for predicting the width and height of Gaussian plumes require detailed site-specific information (measurements of wind speed and temperature at two heights above the vegetation), and are therefore not readily applicable in risk assessments. An alternative that is often adequate is to use a worst case approach, in which the dispersal parameters are used that give the highest spore concentration at the location of interest. Predictability could be improved by measuring atmospheric stability during and just after weather conditions conducive to release of the pollen or spores of interest. We calibrated the model with a weighted least-squares method. Calibrating the model led to a more than 100-fold decrease in the sum of weighted squares. A comparison of estimated concentrations with the measurements confirmed that spore clouds originating from a point source take the form of a Gaussian plume: the coefficient of correlation between measured spore concentrations and fitted concentrations was 0.8. The fraction of spores that escaped the canopy and was available for long distance dispersal amounted to 64±17 percent. We found an 83 percent correlation between this so-called escape fraction and wind speed.