The non-random distribution of herbivorous insects on plants is often assumed adaptive. However, many studies fail to demonstrate this adaptiveness. To test the predictions of the preference–performance and optimal foraging hypotheses, while addressing adaptiveness of oviposition site selection for progenies (larvae) and females, respectively, we examined larval mine distributions (a proxy for egg distribution) of the pygmy moth Stigmella sorbi across shoots, compound leaves and leaflets of rowan (Sorbus aucuparia). The study was conducted near the subarctic town of Kirovsk, in northwestern Russia, from 1992 to 2007. Larvae survived best on long vegetative shoots, yet S. sorbi females showed no preference for shoot type, even though generative shoots are a potential carbohydrate source. They preferentially oviposited on basal leaflets, despite offspring mortality from both bottom-up and top-down factors being unaffected by leaflet position. This preference may benefit females rather than offspring by reducing sunlight and predator exposure. The presence of multiple S. sorbi mines on the same leaf did not affect parasitism rates or cocoon weight compared to solitary mines. However, clumping significantly increased the mortality of newly hatched larvae and explained the more even mine distribution among leaflets observed in high-density years. Our findings do not unequivocally support the preference–performance or apparency hypotheses. However, some patterns may appear adaptive when viewed through the lens of an optimal foraging strategy, proposing that S. sorbi maximizes overall fitness through oviposition behaviour that prioritizes female performance over offspring performance.