A new, selective medium was used to determine the ability of different strains of Pseudomonas solanacearum to survive in the soil or in association with plant roots in the greenhouse. When survival of two strains of race 1, two of race 2, and one of race 3 in a soil–sand mixture (ψ = −2 bars (1 bar = 100 kPa), 28 °C, was determined, populations of each strain declined at different rates, but none of the strains survived beyond 24 weeks. To study survival of P. solanacearum in the rhizosphere, seedlings of resistant hosts (pepper, tomato, castor bean) and presumed nonhosts (bean, corn) were inoculated by dipping their roots in a bacterial suspension and then replanted. Bacterial populations in the rhizosphere were determined at various intervals up to 33 weeks. The pathogen was not able to survive in the rhizosphere; when high populations were detected, they were associated with localized or systemic infection of the roots of plants that did not show wilt symptoms. When the roots of inoculated plants were collected at different times and then surface sterilized by immersing them in 70% ethanol for 5 min, high populations of the bacterium were recovered from many of the plants in each treatment. That infection was not the result of the artificial inoculation procedures used was shown in experiments in which presumed nonhost plants, such as bean, peas, soybean, corn, sorghum, and rice, were grown from seed planted in soil infested with strains of race 1. High populations of the bacterium were recovered from surface-sterilized roots throughout the growth cycle of these plants. Thus, long-term survival of P. solanacearum appeared to be correlated with its ability to infect plant roots.