S U M M A R YI n laboratory experiments germinating conidia penetrated lenticels and wounds but not the intact surfaces of apples. Date of harvest had no significant effect on the numbers of apples infected with Nectria galligena but the earliest picks rotted first in barn store. Inoculations of unpicked apples resulted in small arrested lesions which only developed into progressive rots after a considerable period in store. Rots developed most quickly from inoculations made between mid-August and mid-September. The size of rot increased with spore number and many inoculations with 10-100 conidia remained as arrested lesions.Arrested lesions developed 10-1 5 days after unripe apples were inoculated and consisted of a zone of fungal colonization surrounded by suberized, necrotic cells in which compounds toxic to both N . galligena and Penicillium expansum were detected. No antifungal compounds were found in progressive rots of mature apples or in healthy apples of any age. Antifungal activity, measured by inhibition of P. expansum, was greatest 15-20 days from inoculation of unripe apples with N . galligena but decreased after a total of 35 days incubation at 20 "C. Much less antifungal activity was produced in ripe or dessert apples.