Blackleg disease ( Leptosphaeria maculans) causes severe yield losses in oilseed rape ( Brassica napus) worldwide from the development of cankers in the crown of the plant and, in Western Australia, also from seedling death following early infections. Although most aspects of the epidemiology of this disease are relatively well described, the role of soilborne ascospores and pycnidiospores in the disease epidemic remains unknown. Three separate experiments were undertaken to de. ne the role for soilborne ascospores and pycnidiospores of the blackleg pathogen in the disease epidemiology of oilseed rape. Both spore types were able to cause seedling death, even after the spores had remained in a plant growth medium for up to 21 days before sowing. The most severe seedling death occurred where seeds were sown 7 days before infestation ( 59 and 40% seedlings dead for pycnidiospores and ascospores, respectively) or at 14 days before infestation ( 53 and 32% seedlings dead for pycnidiospores and ascospores, respectively). For pycnidiospores in moist sand, there was significantly more seedling death for sowings 7, 14 or 21 days after infestation compared with concurrent sowing and infestation. For pycnidiospores, allowing the sand to dry between infestation and sowing, significantly reduced mortality for seedlings, but not for ascopsores. For ascospores, adding a 20-mm layer of uninfested sand over the top of the infested sand reduced seedling death. A buried slide assay indicated that both ascospores and pycnidiospores were capable of germination in the sand with lengths of germ tubes similar to those observed on control slides incubated in a sterile moist chamber. To our knowledge, this is the first study to indicate a clear role for soilborne inoculum in the epidemiology of blackleg disease. It is postulated that this role will be of high relative importance in the disease epidemiology in situations where the ascospore showers have been low or restricted to a short period at the beginning of growing season. However, soilborne inoculum could also play a significant role in the disease epidemiology in situations where one or more major ascospore showers have occurred just before or during the early stages of sowing and/or seedling development. The findings, in relation to soilborne pycnidiospore infestations, are further evidence that blackleg, at least in Australia, is a polycyclic disease. Our findings may also explain the responses of canola seedlings to the application of fungicides to soil or seed in reducing blackleg severity and improving seedling survival in blackleg-affected fields in previous studies in Western Australia. The implications for field management of this disease in Australia include a greater need for the application of fungicidal seed or fertiliser treatments and/or utilisation of higher levels of seedling and/or adult host resistance in situations where soilborne inoculum is present.