Sections of dead, dried watermelon vines naturally infected with Didymella bryoniae were placed in nylon mesh bags, covered with soil, and placed in a field at 0-, 12.5-, or 25-cm depths in the November or December prior to 1997, 1998, and 1999. The percentage of 1-cm vine segments that yielded D. bryoniae and the number of segments retrieved intact declined over time. Vine segments cultured on semiselective medium yielded D. bryoniae for 30, 24, and 21 weeks after placement in 1997, 1998, and 1999, respectively. D. bryoniae was not recovered 32 and 29 weeks after placement at any depth in 1998 and 1999, respectively. A hypocotyl-infection assay using watermelon seedlings was developed to detect D. bryoniae conidia produced on recovered vine sections. Using known concentrations of conidia, the percentage of seedlings exhibiting gummy stem blight symptoms increased with the logarithm of the inoculum density. Viable conidia were produced on retrieved vine sections for up to 32 weeks after burial, based on culturing on semiselective agar, but conidia caused disease for only 16 weeks after burial in the hypocotyl-infection assay. Conidia of D. bryoniae were recovered 8 weeks longer at 0 cm than at lower depths in 1999, but not in 1998. Incorporating infested crop residue into soil reduced the number of infective propagules and survival of D. bryoniae.