Broad- host-range pathogens are appealing as candidates for commercial development as bioherbicides because of the wider market potential of a product that is effective against a range of weeds. But when these pathogens are able to spread in space (or time), a risk analysis is necessary. Here we test the hypothesis that a safety zone around a bioherbicide application site is adequate so long as it is wide enough to ensure that dispersing inoculum has diluted sufficiently that the density of inoculum occurring naturally in a susceptible crop is no more than doubled by the influx of bioherbicide spores. To this end the plant disease–pathogen inoculum density relationship using data from nine published experiments was modeled using the logistic equation. This revealed that a doubling of the natural spore density of a plant pathogen in the range of 10^3.4 to 10^6.7 spores/ml may generally be expected to result in unacceptable increases in disease in a susceptible crop. A doubling outside this range (< 10^3.4 or > 10^6.7) is less likely to do so. Therefore when the natural density of inoculum in a crop's environment occurs outside this range, an “acceptable” safety zone for the pathogen's use as a bioherbicide can in most cases be defined by the 1 ∶ 1 ratio of added ∶ natural inoculum. However, if a more “risk averse” safety zone is desired, it can be defined using a 1 ∶ 10 ratio of added ∶ natural inoculum.