Spread of mummy berry disease of blueberry, caused by Monilinia vaccinii-corymbosi, occurs in two discrete monocycles; primary infection by ascospores results in shoot blight, while secondary infection of open flowers by conidia leads to fruit mummification. Gradients of primary and secondary infection from point sources of ascospores and conidia placed in separate plant rows were recorded in each of 2 years at two sites with no history of the disease. Primary infection gradients were longer downwind than upwind, with 95% of blighted shoots occurring within 30 m of the ascospore point source. This observation, along with a positive correlation (r = 0.852, P = 0.0072, n = 8) between the distance over which shoot blight occurred and wind speed parallel to the row, supports the role of wind as a key factor in ascospore dispersal. By contrast, secondary infection gradients were shorter downwind and longer upwind, with 95% of infected fruit occurring within 20 m of the conidial point source. The shorter downwind spread of secondary infection, along with a nonsignificant correlation (r = -0.649, P = 0.0812, n = 8) between the distance over which infected fruit occurred and wind speed, suggests that factors other than wind are important in the transfer of conidia to open flowers; this could include conidial dispersal by bee pollinators, which have been shown previously to forage primarily upwind. Exponential and Pareto cumulative distribution functions were fitted to cumulative counts of blighted shoots and infected fruit to model spread of primary and secondary infection. The Pareto model, which is characterized by a longer tail and predicts more infection farther from the inoculum source, better fits the observed disease gradients in most cases.