The effect of components of primary inoculum dispersal in soil on the temporal dynamics of Phytophthora blight epidemics in bell pepper was evaluated in field and growth-chamber experiments. Phytophthora capsici may potentially be dispersed by one of several mechanisms in the soil, including inoculum movement to roots, root growth to inoculum, and root-to-root spread. Individual components of primary inoculum dispersal were manipulated in field plots by introducing (i) sporangia and mycelia directly in soil so that all three mechanisms of dispersal were possible, (ii) a plant with sporulating lesions on the soil surface in a plastic polyvinyl chloride (PVC) tube so inoculum movement to roots was possible, (iii) a wax-encased peat pot containing sporangia and mycelia in soil so root growth to inoculum was possible, (iv) a wax-encased peat pot containing infected roots in soil so root-to-root spread was possible, (v) noninfested V8 vermiculite media into soil directly as a control, or (vi) wax-encased noninfested soil as a control. In 1995 and 1996, final incidence of disease was highest in plots where sporangia and mycelia were buried directly in soil and all mechanisms of dispersal were operative (60 and 32%) and where infected plants were placed in PVC tubes on the soil surface and inoculum movement to roots occurred with rainfall (89 and 23%). Disease onset was delayed in 1995 and 1996, and final incidence was lower in plants in plots where wax-encased sporangia (6 and 22%) or wax-encased infected roots (22%) were buried in soil and root growth to inoculum or root-to-root spread occurred. Incidence of root infections was higher over time in plots where inoculum moved to roots or all mechanisms of dispersal were possible. In growth-chamber studies, ultimately all plants became diseased regardless of the dispersal mechanism of primary inoculum, but disease onset was delayed when plant roots had to grow through a wax layer to inoculum or infected roots in tension funnels that contained small volumes of soil. Our data from both field and growth-chamber studies demonstrate that the mechanism of dispersal of the primary inoculum in soil can have large effects on the temporal dynamics of disease.