Holly leaf-miners on two continents: what makes an outbreak species? SABINE EBER, 1 HUGH P. SMITH, 1 RAPHAEL K. DIDHAM 2 and HOWARD V. CORNELL 2 1 NERC Centre for Population Biology, Imperial College at Silwood Park, U.K. and 2 Department of Biological Sciences, University of Delaware, U.S.A. Abstract. 1. Some herbivore species periodically undergo damaging, high-density outbreak phases followed by less damaging low-density phases. Others maintain steady, low to moderate density levels that do little damage to their hosts. 2. Two closely related holly leaf-miner species were compared that share many ecological traits and have very similar life cycles, but only one of which exhibits outbreaks. Phytomyza ilicicola in the eastern U.S.A. varied widely in mortality and infestation levels, reaching local densities of over 10 mines per leaf. In contrast, Phytomyza ilicis in the U.K. showed low infestation and high mortality at all sites. Using data from the literature and from ®eld studies, the factors responsible for these contrasting dynamics were sought. 3. Phytomyza ilicicola oviposits into the leaf lamina, and experiences weak larval competition only at high densities. Phytomyza ilicis oviposits into the leaf midrib, which leads to high mortality of young larvae before mine formation. Multiply mined leaves were therefore very common in P. ilicicola but rare in P. ilicis. 4. Differences in the parasitoid complexes of the two systems accounted for further differences in survival to adulthood. The main (larval) parasitoid, which was found to impose high, density-dependent mortality on P. ilicis, is missing on P. ilicicola. It is replaced by an egg±pupal parasitoid, which varies in its impact at different sites. Multiple emergence of adults from multiply mined leaves is therefore widespread in P. ilicicola but does not occur in P. ilicis. 5. The differences in oviposition behaviour and in the parasitoid complexes are likely to allow P. ilicicola to outbreak when habitat conditions are favourable, while P. ilicis is always tightly regulated. Key words. Density dependence, holly leaf-miner, larval competition, outbreak species, oviposition behaviour. Introduction Although herbivore species can vary widely in their population trajectories over space and time, two somewhat distinct dynamical patterns can be recognised (Price et al., 1990). Latent or non-outbreaking species typically maintain steady populations that ¯uctuate no more than two orders of magnitude (often less) and rarely do heavy damage to their host species. Eruptive or outbreaking species occasionally exhibit dense and damaging epidemic phases, alternating with less damaging low-density endemic phases. The epidemic density may exceed the endemic density by up to ®ve orders of magnitude. Outbreaking species comprise only a tiny propor- tion of the total herbivore fauna (Hunter, 1995) but because outbreaks can devastate whole plant populations at peak densities (Hill, 1997), they have been studied with an intensity that is disproportionate to their numbers (e.g. Schwerdtfeger, 1942; Barbosa & Schultz, 1987; Port & Guile, 1987; Price et al., 1990; Hunter, 1991, 1995; Swetnam & Lynch, 1993; Berger & Katzensteiner, 1994; Harrison, 1997). One approach to the question of why some species experience periodic outbreaks has been to compare the life histories and ecological traits of species groups with outbreak vs. non-outbreak dynamics (Hunter, 1995). For example, a Correspondence: Sabine Eber, Agroecology, Go Èttingen University, Waldweg 26, D-37073 Go Èttingen, Germany. E-mail: s.eber@ uaoe.gwdg.de 124 # 2001 Blackwell Science Ltd Ecological Entomology (2001) 26, 124±132 Ecological Entomology (2001) 26, 124±132