In tomato glasshouses, the population structure of airborne inoculum of Botrytis cinerea depends on the production of endogenous inoculum on diseased plants as well as on incoming exogenous inoculum. Both types of inocula may contribute differently to the development of epidemics. Two strains of B. cinerea were introduced in each of four separate compartments of an experimental tomato glasshouse. We monitored their impact on disease development and on the genetic diversity of B. cinerea populations using microsatellite markers. The naturally occurring airborne inoculum of B. cinerea displayed a high level of genetic diversity and was rapidly displaced in the glasshouse, as isolates with microsatellite profiles identical to the introduced strains amounted to 66% of the inoculum sampled from the air 14 days after inoculation and 91% of those collected from stem lesions 60 days after inoculation. This suggested an important role of secondary inoculum in disease development, which is compatible with the hypothesis of a polycyclic development of gray mold epidemics in tomato glasshouses. In controlled-environment tests on tomatoes, a wide range of aggressiveness levels was observed, both for isolates sampled from the air and from lesions on plants. Hypotheses are proposed to explain the negligible impact of naturally incoming isolates on the epidemics observed inside the four glasshouse compartments