The logarithm of the mean number of greenhouse whitefly (Trialeurodes vaporariorum) puparia per leaf, when sampled from a single leaf stratum of a greenhouse tomato crop, had a linear relationship to the proportion of infested leaves, on a complementary log-log scale. Whitefly puparia had a clumped distribution with strong correlations between adjacent plant rows and plants within rows. In order to sample most patches of whitefly but to minimise the impact of correlation between adjacent plants on the sample, sample plants were selected from a rectangular grid covering either the whole greenhouse or an infested area of a large house. The rectangular grids consisted of plants at fixed intervals (every 7–10th plant) along the chosen rows, spaced on odd number of rows (3, 5, or 7) apart. Each grid had a random starting point. Decision charts incorporating action thresholds can be constructed based on the number of infested leaves and the total number of leaves examined. Whether parasitism byEncarsia formosais above or below any preset threshold can be determined by plotting the number of leaves with whitefly puparia against the number of leaves with fewer parasites than the threshold. Using recording charts, an advisor to the grower will obtain information on the location of whitefly and whether parasitism is above or below the preset threshold. This information can be used to determine whether action is required and to assist in deciding the type of action to be taken.