In November of 1984 and 1985, Euphorbia lathyris was planted into a field naturally infested with Macrophomina phaseolina located at the Campbell Avenue Farm in Tucson, Arizona. Plants without foliar symptoms and rhizosphere soil were sampled regularly from emergence until the following May or June. Soil rhizosphere populations ranged from 0.7-3.0 cfu/g soil in 1985 to 8.0-24.1 cfu/g soil in 1986, and did not change significantly over either growing season (P > 0.05). Both the incidence of disease and the number of infection sites per cm of root increased significantly (P < 0.05) over each growing season and were not related to rhizosphere soil populations of M. phaseolina (P > 0.05). The distribution of infection sites along the tap root over both growing seasons remained the same in that most were located in the top 0-7 cm of tap root. Infected E. lathyris without apparent symptoms were subjected to low-water and high-temperature stress treatments in growth chambers. Root infection was not found to be dependent upon any stress. Lesion development was significantly dependent upon the imposition of any stress treatment, and further root colonization was significantly dependent upon low-water stress (P < 0.05). M. phaseolina was consistently recovered from asymptomatic roots. A consistently lower leaf water potential was measured on infected E. lathyris than from non-infected controls when no stress treatment was applied. Polyclonal antisera made against hyphae and microsclerotia of M. phaseolina was not successful in detecting this pathogen in E. lathyris by I-ELISA. Antisera applied to fresh thin sections of infected plant tissue was effective in staining hyphae of M. phaseolina when used with a second antibody conjugated to fluorescence isothiocyanate or to an enzyme (to which a substrate was added to "stain" hyphae).