The effects of temperature, soil moisture, soil type and fungal isolate on the longevity of Phoma exigua Desm. var. foveata (Foister) Boerema (the cause of potato gangrene) in soil were investigated by inoculating soils with pycnidiospores and incubating under constant conditions of temperature and moisture. There was an immediate, steady decline in fungal populations in all experiments, except in autoclaved soil where they initially remained constant for 3–4 months. Survival was prolonged by lower temperatures, populations in loam becoming undetectable within about 8 months at 15 °C, 2 years at 10° and 3 years at 5°. In general survival was better (i.e. populations were larger at any sample time) the drier the soil, but in very dry conditions ( < 20 % water holding capacity) it was as poor as in soils at or near field capacity. The fungus survived equally well in loam, clay, sandy-loam and peat soils at the same moisture contents, measured as percentage water holding capacity. One fungal isolate, which had been maintained in culture for several years, survived poorly but there were no consistent differences between five other isolates. The implication of these results is discussed in relation to survival in the field. The soil is unlikely to be an important reservoir of inoculum for potato crops, except possibly in the production of high-grade seed stocks, where even small populations could result in introduction of disease.