SUMMARYWhen the natural incidence ofCladosporium fulvumCooke, occurring at a relatively late stage of cropping, was controlled with fungicides, comparisons with unsprayed control plants showed that there was an interval ofc.6 weeks between the incidence of severe infection, when the fungus was colonizing at least 50 % of the host's leaf area, and fruit yield decreases.Eye estimates made on separate leaves usually over‐estimated the percentage of infection. Correction factors were derived from the regression of eye estimates on infected areas calculated from leaf tracings. Whole‐plant infection indices were obtained by averaging corrected values of percentage infection separately assessed on leaves immediately above each fruit truss.Although dichlofluanid and tank‐mixed zineb decreased the incidence ofC. fulvummore than captan sprays, tomato yields (including green fruit) were increased similarly by all three fungicides from 2·99 to an average of 3·51 kg/plant duringc.3 months' picking. Decreasing the intervals between successive sprays from 21 to 7 days greatly decreased leaf mould infection but did not significantly affect yields. Restricting applications to plant tops (namely foliage from the fourth‐youngest truss to the mainstem apex) gave yield benefits equal to those gained from sprays applied to who