An existing numerical model for particle diffusion into air from ground-based sources was recalculated for circular source plots of 7 m diameter. Results showed that a satisfactory relationship between horizontal flux of the diffusing gas at a single measuring height above the centre of the plot and flux density from the plot could be defined. This provides a simple technique for comparing volatilization rates of ammonia from treated and untreated manure in the field, under outdoor conditions. The technique was applied to the study of effects of monocalcium phosphate (MCP) addition on ammonia volatilization from pig manure. Mean ammonia flux densities during peak volatilization periods of the first 3 d after spreading were lower for MCP-treated plots by an average of 17%, with a maximum reduction of 30%. Variability of results from short-term sampling intervals (1-2 h) was high. This made short-term detection of differences of <20% difficult under the adopted experimental procedures. Observations for volatilization from source plots of equal source strength, as well as from source plots with different degrees of dilution of applied manure, permitted a further definition of the sensitivity of the technique. Suggestions are given for possible reductions in variability from improved experimental procedure.