The efficacies of four systemic neonicotinoid insecticides applied to potted avocado trees at manufacturer-recommended rates were assessed against the avocado thrips, Scirtothrips perseae Nakahara. At the time of treatment, fully expanded first-flush young leaves were tagged for identification, and a proportion of these leaves was used in bioassays with second-instar thrips. At 7 weeks post-treatment, a second flush of leaves had fully expanded on the trees, and these leaves were included in additional bioassays comparing avocado thrips mortality on both first- and second-flush leaves. In bioassays with first-flush leaves, imidacloprid (273 mg AI pot(-1)) was the most effective insecticide, providing at least 70% mortality of thrips for 14 weeks. Thiamethoxam (137 mg AI pot(-1)), clothianidin (109 mg AI pot(-1)) and dinotefuran (241 mg AI pot(-1)) provided good control in bioassays that were conducted within 4 weeks of treatment, but thereafter their efficacies were inconsistent. In bioassays with second-flush leaves, imidacloprid provided at least 70% mortality up to 9 weeks after the insecticide application. Thereafter, mortality declined to 30% or lower. Bioassays with second-flush leaves collected from trees treated with thiamethoxam, clothianidin and dinotefuran resulted in unacceptably low thrips mortality. Monitoring of imidacloprid and thiamethoxam residues by ELISA showed that the greater persistence of imidacloprid in both first and second leaf flushes was due to a steadier uptake of this material. Although thiamethoxam residues rose quickly within the first leaf flush, levels had already begun to dissipate by the time the second leaf flush had started to develop.