Abstract
It has been suggested that the socially parasitic butterfly Maculinea alcon detects ant odours before ovipositing on initial larval food plants near colonies of its obligate ant host Myrmica ruginodis. It has also been suggested that overcrowding on food plants near M. ruginodis is avoided by an ability to detect high egg loads, resulting in a switch to selecting plants near less suitable ant species. If confirmed, this hypothesis (H1) would have serious implications for the application of current population models aimed at the conservation of endangered Maculinea species, which are based on the null hypothesis (H0) that females randomly select food plants whose flower buds are at a precise phenological stage, making oviposition independent of ants. If H1 were wrong, practical management based upon its assumptions could lead to the extinction of protected populations. We present data for the five European species of Maculinea which show that (i) each oviposits on a phenologically restricted flower-bud stage, which accounts for the apparent host-ant-mediated niche separation in sympatric populations of Maculinea nausithous and Maculinea teleius, (ii) there is no temporal shift in oviposition by Maculinea arion in relation to host ant distribution or egg density, and (iii) oviposition patterns in 13 populations of M. alcon's closest relative, Macaulinea rebeli, conform to H0 not H1 predictions. It is concluded that conservation measures should continue to be based on H0.