Tomato plants have their leaves, petioles and stems covered with glandular trichomes that protect the plant against two-spotted spider mites and many other herbivorous arthropods, but also hinder searching by phytoseiid mites and other natural enemies of these herbivores. This trichome cover creates competitor-free and enemy-free space for the tomato russet mite (TRM) Aculops lycopersici (Acari: Eriophyidae), being so minute that it can seek refuge and feed inbetween the glandular trichomes on tomato cultivars currently used in practice. Indeed, several species of predatory mites tested for biological control of TRM have been reported to feed and reproduce when offered TRM as prey in laboratory experiments, yet in practice these predator species appeared to be unable to prevent TRM outbreaks. Using the phytoseiid mite, Amblydromalus limonicus, we found exactly the same, but also obtained evidence for successful establishment of a population of this predatory mite on whole plants that had been previously infested with TRM. This successful establishment may be explained by our observation that the defensive barrier of glandular plant trichomes is literally dropped some time after TRM infestation of the tomato plants: the glandular trichome heads first rapidly develop a brownish discoloration after which they dry out and fall over onto the plant surface. Wherever TRM triggered this response, predatory mites were able to successfully establish a population. Nevertheless, biological control was still unsuccessful because trichome deterioration in TRM-infested areas takes a couple of days to take effect and because it is not a systemic response in the plant, thereby enabling TRM to seek temporary refuge from predation in pest-free trichome-dense areas which continue to be formed while the plant grows. We formulate a hypothesis unifying these observations into one framework with an explicit set of assumptions and predictions to be tested in future experiments.