Current sampling protocols for pest whitefly in greenhouse cropping systems are not adequate for detection of Bemisia tabaci when it is part of a mixed species population. Greenhouse vegetable and flower crops in the north, west and south of Sydney were sampled for fourth-instar whitefly nymphs in autumn, winter and spring of 2007. Bemisia tabaci was found in all regions: in the north in winter, the west in all seasons and the south in autumn and winter. A comparison of a Poisson model for pest abundance and a binomial model of presence/absence showed the same factors to be significant, indicating that it is sufficient to sample for presence/absence of the pest on individual plants, rather than to measure pest abundance per plant. A model using the binomial distribution determined that if more than 20% of plants are infested with B. tabaci, a sample size of 20 whole plants per greenhouse is sufficient to detect the presence of B. tabaci with greater than 95% confidence.