The direct viable count method, used to detect viable but nonculturable bacteria in aquatic systems, was modified to examine epiphytic populations of Pseudomonas syringae. Viable-population sizes determined from the number of cells that elongated when incubated with yeast extract and nalidixic acid were compared with those determined by the conventional plate count method. The plate count method accurately determined the number of viable cells in epiphytic P. syringae populations in a state of active growth under conditions of high relative humidity. The plate count method also accurately determined the number of viable cells in P. syringae inoculum, or a growing P. syringae population, subject to desiccation stress under conditions of low relative humidity. In epiphytic populations of P. syringae older than 80 h, however, the plate count underestimated the viable-population size by about two- to fourfold, suggesting that up to 75% of the P. syringae population was nonculturable. These nonculturable cells may have entered a starvation-survival state, induced by low nutrient availability in the phyllosphere environment. Epiphytic P. syringae populations undergoing rapid size changes due to growth and death under fluctuating environmental conditions in the field should be accurately enumerated by the plate count method. However, the possible underestimation of viable-population size under some circumstances should be considered in epidemiological studies of phytopathogenic bacteria and when genetically engineered microorganisms in terrestrial ecosystems are monitored. (Copyright (c) 1992, American Society for Microbiology.)