Oikos 119: 437–444, 2009 doi: 10.1111/j.1600-0706.2009.17848.x,© 2009 Th e Authors. Journal compilation © 2010 Oikos Subject Editor: Tim Benton. Accepted 28 July 2009437Does urbanization select for weak competitors in house sparrows?Veronika Bókony, Anna Kulcsár and András LikerV. Bókony ([email protected]) and A. Liker, Dept of Limnology, Univ. of Pannonia, Pf. 158, HU–8201 Veszprém, Hungary. – A. Kulcsár, Dept of Ethology, Eötvös Univ., Pf. 120, HU–1518 Budapest, Hungary.How urbanization aff ects animal populations is in the focus of current ecological research. Existing theory of this topic suggests that the cities’ more constant food supplies and lower predation pressure lead to a high proportion of weak com-petitors in urban populations. To evaluate this hypothesis, we tested whether competitive performance diff ers between diff erently urbanized populations of house sparrows Passer domesticus. We previously showed that wild urban sparrows are smaller and leaner than rural conspecifi cs, and this diff erence persists for months under identical captive conditions. Here we compared several aspects of their competitiveness (fi ghting, scrambling and searching for food) in captive mixed fl ocks of urban and rural birds. We found that sparrows exhibited consistent individual diff erences in competitiveness, but these diff erences were not related either to the degree of urbanization of their original habitats or to their body mass. Moreover, the variance in competitive abilities also did not diff er between birds from more and less urbanized habitats. Th us our results did not support the hypothesis that urbanization shifts population structure towards an over-abundance of weak competitors in house sparrows. We discuss possible explanations why sparrow populations may not diff er in competitive-ness despite the smaller body mass of urban birds.Urbanization and its eff ects on wildlife are topics of growing concern not only for conservation biologists but also for evo-lutionary ecologists (Marzluff 2001, Shochat et al. 2006). Besides re-structuring animal communities by decreasing diversity and increasing the density of a few, often non-native species (Marzluff 2001, Shochat et al. 2004), urbanization can induce substantial, rapid changes in the morphology (Beckmann and Berger 2003, Yeh 2004, Liker et al. 2008), physiology (Partecke et al. 2005, 2006, Bonier et al. 2007), behavior (Beckmann and Berger 2003, Slabbekoorn and Peet 2003, Yeh et al. 2007), and phenology (Fleischer et al. 2003, Partecke et al. 2004, Yeh and Price 2004) of wild animals. A recent theory (credit card hypothesis; Shochat 2004, Anderies et al. 2007) proposes that many of these changes are driven by two main factors: habitat productivity and predation pres-sure. Although direct evidence is scarce and controversial for both assumptions, urban habitats are thought to be character-ized by predictable food and water supplies on the one hand and reduced mortality by predation on the other (Shochat et al. 2004). Th ese two factors are predicted to lead to a “shift from small populations of mostly winners to large popula-tions of mostly losers” (Anderies et al. 2007), because predict-able food and low predation lead to high population densities and resource overmatching, so individuals eventually suff er reduced foraging success and maintain lower body reserves. Because these weak competitors can still ‘live on the credit’ of the predictable resource input in cities and are not selected out by predators, they are expected to constitute a high pro-portion of urban populations, while only a few individuals with superior competitive abilities are able to maintain high body condition (Shochat 2004, Anderies et al. 2007).Up to now, several studies compared urban and non-ur-ban populations of vertebrate species using various measures of body condition such as body mass (relative to skeletal size), plasma protein levels or white blood cell profi les. Th e results vary across species, seasons and the type of condition index used, some showing worse body condition in urban individuals (Ruiz et al. 2002, Partecke et al. 2005, Fokidis et al. 2008, Liker et al. 2008) while others showing the opposite or no diff erence (Gavett and Wakeley 1986a, Ots et al. 1998, Beckmann and Berger 2003, Schoech and Bowman 2003, Fokidis et al. 2008, French et al. 2008). However, the degree to which these studies support or confl ict the credit card hypothesis is diffi cult to assess, because (1) they did not directly measure the competitive abilities of urban and non-urban individuals, and (2) they did not compare the distribution or variance of the measured variables between habitats.To test one of the main predictions of the credit card hypothesis, in this study we aimed to investigate the competi-tive performance of individuals from diff erently urbanized habitats in house sparrows Passer domesticus. Th is species is one of the world’s most common urban birds, occupying a variety of habitats from heavily built-up city centers to small, isolated438MethodsStudy subjectsWe captured 60 house sparrows with mist nets between 1–18 October 2007 in four diff erently urbanized habitats in Hun-gary (Table 1, Appendix 1). Two urban sites are within the built areas of two cities (the capitol of Budapest and the town of Veszprém), whereas two rural sites are extensively sur-rounded by non-built, agricultural areas (at the edge of a village Nemesvámos and at a small, isolated farm Dórama-jor). We followed Liker et al. (2008) to quantify the degree of urbanization in each habitat by scoring the cover of vege-tation, paved roads, and buildings in a 1 km2 area around each capture site (for further details see Liker et al. 2008), and we also collected data on the density of multi-storey buildings and the residential human population for each settlement (Hungarian Central Statistical Offi ce, Population Census 2001, <www.nepszamlalas.hu/eng/index.html>) as suggested by Marzluff (2001). Th en we calculated an ‘urban-ization score’ for each site as the PC1 score from a principal component analysis of the seven measures of urbanization (Table 1). Th e analysis extracted one principal component that accounted for 85.1% of the total variance and correlated negatively with vegetation cover and positively with the den-sity of buildings, roads and humans (Table 1). According to this