J. Avian Biol. 39: 663671, 2008 doi: 10.1111/j.1600-048X.2008.04359.x, # 2008 The Authors. J. Compilation # 2008 J. Avian Biol. Received 20 August 2007, accepted 13 March 2008 Common cuckoo Cuculus canorus parasitism, antiparasite defence and gene flow in closely located populations of great reed warblers Acrocephalus arundinaceus Csaba Moska ´t, Bengt Hansson, Lilla Baraba ´s, Istva ´n Ba ´rtol and Zsolt Karcza C. Moska´t (correspondence), Animal Ecology Research Group of the Hungarian Academy of Sciences, c/o Hungarian Natural History Museum, Budapest, Ludovika te´r 2, H-1083, Hungary. E-mail: moskat@nhmus.hu.  B. Hansson, Department of Animal Ecology, Ecology Building, Lund University, S-223 62 Lund, Sweden.  L. Baraba´s, Zoological Department, Hungarian Natural History Museum, H-1066, Budapest, Baross u. 13, Hungary. Present address of LB: West Hungarian University, Faculty of Forestry, Wildlife Management Institute, Sopron, Ady E. u. 5, H-9400, Hungary.  I. Ba´rtol, Directorate of the Kiskunsa´g National Park, Kecskeme´t, Liszt F. u. 19, H-6000, Hungary.  Z. Karcza, Ringing Center of BirdLife Hungary, Budapest, Ko¨lto¨u. 21, H-1121, Hungary. In Hungary an unusually high rate of parasitism on the great reed warbler Acrocephalus arundinaceus by the common cuckoo Cuculus canorus has been maintained for at least the last one hundred years. We evaluated parasitism rate, antiparasite defence and genetic differentiation among Hungarian great reed warblers at three sites located 40130 km from each other, where hosts suffered from a high (4168%), moderate (11%), and almost no ( B1%) parasitism. We were especially interested in whether the level of antiparasite defence was related to the local parasitism rate, and, if not, to understand why. There was no difference among the three sites in the responses to experimental parasitism by non- mimetic model cuckoo eggs (rejection rate 7182%), which can be explained by strong gene flow between populations: there was low level of philopatry and no genetic differentiation in the region. Reproductive success of the host in the heavily parasitised site was about 54% of that in the unparasitised site, indicating that long-term persistence of host populations in highly exploited areas depends on continuous immigration. Obligate brood parasites take no care of their own offspring, and depend entirely on the host species to hatch and care their young (Payne 1998). The common cuckoo (hereafter ‘‘cuckoo’’) Cuculus canorus is a widespread avian brood parasite exploiting several passerine bird species in the northern Palearctic (Wyllie 1981). A cuckoo chick normally evicts the host’s eggs or nestlings from nest after hatching (e.g. Wyllie 1981, Honza et al. 2007), reducing the reproductive success of the host (e.g. Kleven et al. 2004). For this reason cuckoo parasitism poses a severe cost for the host (Davies and Brooke 1988, Øien et al. 1998), which in turn selects for antiparasite defence, like egg rejection behaviour (e.g. Davies and Brooke 1988, Moksnes et al. 1991). On the other hand, cuckoos adapt to the hosts, e.g. by rapid egg-laying (Wyllie 1981) and refined egg-mimicry (e.g. Davies and Brooke 1988, Moksnes and Røskaft 1995, Honza et al. 2001, Moska ´t and Honza 2002, Avile ´s and Møller 2004, Avile ´s et al. 2006, Cherry et al. 2007a,b). The coevolutionary process between a host species and its brood parasite is often regarded as a continuous arms race (e.g. Dawkins and Krebs 1979, Davies and Brooke 1989, Rothstein 1990, Moksnes et al. 1991, Davies 2000), which is generally limited by the adaptability of the cuckoo (Honza et al. 2004, Lova ´szi and Moska ´t 2004). Cuckoo parasitism rate is generally low: often less than 10% (Brooke and Davies 1987, Moksnes and Røskaft 1987, Davies 2000), rarely between 2030% (Lotem et al. 1995, Rutila et al. 2002, Antonov et al. 2006a, 2007a), but may occasionally go up to 50% (Schulze-Hagen 1992). Such high levels can occur when parasitism has recently started in a population, but often drop to low levels after a period of a few years (Takasu et al. 1993, Nakamura et al. 1998, Takasu 1998). In very rare cases, high levels of parasitism have been maintained for longer time periods, as in some populations in Hungary (see below). The often relatively low parasitism rate slows down the coevolutionary process between the host and the brood parasite, as Røskaft et al. (2002, 2006) suggested in the spatial habitat structure hypothesis. Fluctuations in the level of parasitism in spatially separated host populations could be expected on the metapopulation level, where the host’s adaptations to parasitism may also vary locally (Soler et al. 1998, Grim 2002, Røskaft et al. 2002, 2006). The coevolutionary process can also be slowed down by high costs of defence against parasitism relative to the benefits obtained (e.g. Lotem et al. 1992, Lotem 1993, Davies et al. 1996). Density of the host population proved to be a key factor in comparison of parasitised and non-parasitised reed warbler 663