Big city life: carnivores in urban environments

Abstract

Cities may represent one of the most challenging environments for carnivorous mammals. For example, cities have a dearth of vegetation and other natural resources, coupled with increased habitat fragmentation and an abundance of roads as well as altered climate (e.g. temperature, light, rainfall and water runoff). It is therefore intriguing that several carnivore species have become established in cities across the globe. Medium-sized carnivores such as the red fox, coyote, Eurasian badger and raccoon not only survive in cities but also have managed to exploit anthropogenic food sources and shelter to their significant advantage, achieving higher population densities than are found under natural conditions. In addition, although they may not live permanently within cities, even large carnivores such as bears, wolves and hyaenas derive significant benefit from living adjacent to urbanized areas. In this review, we examine the history of urban adaptation by mammalian carnivores, explore where they are living, what they eat, what kills them and the behavioural consequences of living in urban areas. We review the biology of urban carnivores, exploring traits such as body size and dietary flexibility. Finally, we consider the consequences of having populations of carnivores in urbanized areas, both for humans and for these charismatic mammals. In conclusion, in a time of massive environmental change across the globe, the continuing encroachment of urbanization upon wilderness areas is substantially reducing the availability of natural habitats for many species; therefore, understanding the biology of any taxon that is able to adapt to and exploit anthropogenically disturbed systems must aid us in both controlling and developing suitable conservation measures for the future of such species.

Introduction

Wild carnivores have doubtless been entering human settlements for millennia, either by mistake, as scavengers or as predators, or through deliberate encouragement by humans to control pests or aid hunting. For example, grey wolves Canis lupus started developing a close association with humans ∼100 000 years ago (Vilà et al., 1997) with a ‘formal’ domestication of dogs Canis familiaris around 12 000–14 000 years ago (Savolainen et al., 2002). Similarly, cats Felis catus may have started to feed upon rodents dwelling around human food stores around 9500 years ago (Driscoll et al., 2007) and thus become habituated to people. At this time, human settlements may have represented an altered but perhaps not significantly challenging habitat. As human societies have grown, however, landscapes have been increasingly altered through anthropogenic activities (Baker & Harris, 2007; Gehrt, 2010).

For the first time in history, the majority of the human population resides within urban areas, with over 3 billion people living in cities across the world (UNFPA, 2007; Gehrt, 2010). Gehrt (2010) defines ‘urban’ as an area of human residence, activity and associated land area developed for those purposes, usually defined by a threshold human density. These large groupings of people and associated structures comprise at least one town or city (Gehrt, 2010) and include a wide range of anthropogenic disturbances, including buildings and associated infrastructure, for example, gardens, roads, waste ground and parkland (Baker & Harris, 2007). However, the definition of what is classified as ‘urban’ varies greatly depending on geographic location, which, in part, may reflect population density present in the country. Furthermore, while city centres may represent the extreme of anthropogenically altered environments, city suburbs, villages and small towns or even rural farmland also represent challenges in terms of altered landscapes (Fig. 1).

Figure 1. Anthropogenically altered landscapes vary along a gradient from rural farmlands through to the centres of cities, the ultimate urban habitat. Urban environments have significantly altered cover and food resources, and also present challenges in the forms of obstructions (shown by increasing number of • symbols). Diet is listed as O: omnivore C: carnivore (l: live prey, c: carrion). Urban environments may offer substantial anthropogenic food sources for carnivores, and although there is likely to be little natural habitat available for cover, there are plenty of artificial structures, for example, road culverts, abandoned buildings, etc. While domestic cats and dogs have become feral in undeveloped landscapes, most other carnivore species illustrated demonstrate increased use of anthropogenically altered landscapes. The degree to which individual carnivore species are found living among human built-up areas is demonstrated by the relative length of the arrows for each species (e.g. red foxes are present within the centres of cities across the globe). *Skunks: Principally striped skunk, Mephitis mephitis, also less commonly spotted skunk, Spilogale putorius, hog-nosed skunk, Conepatus leuconotus, and hooded skunk Mephitis macroura.

With the spread of urban environments (e.g. McKinney, 2002; Radeloff et al., 2005), many terrestrial species have withdrawn into reduced ranges; this response is particularly noticeable in mammalian carnivores (Woodroffe & Ginsberg, 1998; Woodroffe, 2000; Cardillo et al., 2004). Many carnivore species actively avoid urban areas, rapidly disappearing from encroaching urban spread (‘urbanophobes’, sensu Witte, Diesing & Godde, 1985, ‘urban avoiders’, sensu McKinney, 2006). A number of other species, however, can be described as truly urban dwellers, maintaining varying levels of intimacy with humans, residing within cities and built-up areas across the globe, despite the significantly artificial environment. For some, cities have grown up around their preferred habitat; their presence close to human societies therefore represents continuation of a somewhat altered lifestyle (e.g. Radeloff et al., 2005), and they usually do not make extensive use of anthropogenic resources, largely still relying on natural resources (‘urban adapters’, sensu McKinney, 2006). By contrast, fully synanthropic species (‘urban exploiters’, sensu McKinney, 2006) may actively invade city environments, make use of anthropogenic food and shelter, and often attain population densities far above those found for rural habitats. In this paper, we have reviewed available information on carnivores dwelling in urban environments (either as ‘urban adapters’ or ‘urban exploiters’) and compare these with species that have not successfully adapted to the urban environment (‘urban avoiders’).

Why review the biology and ecology of urban carnivores? Firstly, as cities grow, we are removing alternative habitat for these animals. With increasing loss of undeveloped landscape, urban resources are likely to become more important for conservation of wild animals. City-dwelling is becoming more important for wildlife as the global human population grows (Baker & Harris, 2007). Secondly, it is evident that increasing numbers of carnivores are using urban areas. For example, during the 1990s, there was a 15-fold increase in the numbers of nuisance coyotes Canis latrans removed annually from the Chicago metropolitan area (Gehrt, 2011); similarly, there has been a 10-fold increase in complaints about black bears Ursus americanus in urban Nevada (Beckmann & Lackey, 2008). Thirdly, it is biologically interesting that some species, but not others, do so well in urban environments. Understanding more about the biology of these animals is likely to aid our management and conservation of carnivores as a group. Finally, as Gehrt, Riley & Cypher (2010) pointed out, carnivores elicit strong feelings in people (e.g. fascination, admiration, fear and hate), which may be a manifestation of our ancestral predator–prey relationships and which certainly mould and direct our interactions with these animals.

In this review, we summarize the history of terrestrial mammalian carnivore species as urban dwellers in a taxonomic framework (section: ‘History of carnivore urban adaptation’). We then examine how the ecology of carnivores is influenced by urban living, addressing their habitat utilization and diet (section: ‘How is the ecology of mammal carnivores influences by urban living?’). We explore the causes of mortality and the effects of increased density on carnivore behavior and sociality. In the following section (‘Which species make the best urban adapter/exploiter?’), we investigate features that may allow a species to become adapted to urban environments (i.e. taxonomy, body size, diet and phylogenetic history). Finally, we explore the consequences of carnivore presence within cities for humans, and, in turn, what a future in urban areas may hold for carnivores.

History of carnivore urban adaptation

Carnivores have demonstrated a range of adaptation to living with humans (Fig. 1). Dogs and cats have lived in close association with humans for millennia, and as human populations have spread, these animals have travelled with them, gaining access to some of the most remote locations on the globe. Their extremely adaptable nature has allowed dogs and cats to move out from human habitation to exploit new environments. A key example of this has been the establishment of the dingo Canis lupus dingo in Australia. Dingoes entered the continent with human settlers some 3500–4000 years ago (Corbett, 1995) and have since become established over the entire continent. Arguably, the cat is even more successful in its exploitation of habitats. Its high mobility and flexible biology makes the cat robust to habitat fragmentation (Crooks, 2002) and, coupled with transportation by humans (McKinney, 2006), these animals have spread largely unchecked over new landscapes (e.g. islands such as Australia and New Zealand) where they are an important and numerous predator (Schmidt, Lopez & Pierce, 2007). While domestic dogs and cats have moved out from human settlements to become feral in wild areas (Fig. 1), other carnivore species have encroached to varying degrees into human habitation (Fig. 1).

Red foxes Vulpes vulpes may be one of the most adaptable of the wild carnivores, inhabiting ‘the most expansive geographical range of any wild carnivore using habitats as varied as arctic tundra, arid deserts, and metropolitan centres’ (Macdonald, 1987; Voigt, 1987). The first unequivocal documentation of non-domestic predators dwelling in large cities is records of red foxes in British cities in the 1930s, although they may have been present much earlier (Teagle, 1967; Soulsbury et al., 2010). The urban red fox was regarded as a ‘British phenomenon’ for a long time, but subsequent records indicate significant numbers of red foxes residing within an estimated 114 cities across the globe, including 56 cities in the UK, 40 European cities, 10 North American cities and 6 Australian cities (reviewed by Soulsbury et al., 2010). Red foxes appear to actively colonize urban areas (Macdonald & Newdick, 1982; Harris & Rayner, 1986b; Wilkinson & Smith, 2001); this is particularly true for countries where this species is introduced, where there is a noted spread into a variety of habitats, including cities (Adkins & Stott, 1998; Marks & Bloomfield 1999b and references therein).

Raccoons Procyon lotor have been living in and around cities since the turn of the 20th century and are arguably one of the most common carnivores in modern North American cities (Seton, 1929; Hadidian et al., 2010). The raccoon was introduced into Japan where it is now regarded as a pest in both urban and rural areas (Ikeda et al., 2004) and has also spread in Germany where it was introduced ∼70 years ago (Frantz, Cyriacks & Schley, 2005).

Their ‘plasticity in behaviour, social ecology, and diet allows coyotes to not only exploit, but to thrive, in almost all environments modified by humans’ (Gese & Bekoff, 2004). Despite the success of coyotes in colonizing urban areas (Gese & Bekoff, 2004), little is known of their ecology in comparison with rural populations (Curtis, Bogan & Batcheller, 2007). This is partly due to difficulties inherent in such studies, but also because 20 years ago, there was little need for such studies (Gehrt & Prange, 2007), indicating a recent accession of coyotes to an urban-adapted niche. Coyotes may have always existed in and around cities in south-western North America, although their presence in midwestern and eastern cities has indicated their increases in population presence and size over the past ∼100 years (Gehrt & Riley, 2010). Sizable populations now exist in urban areas across North American cities (Gehrt, 2011). At least in some areas, coyotes became urbanized through the enclosure of undisturbed patches of environment within the urban matrix (Quinn, 1997b), although more recently they appear to be actively colonizing urban areas (Grinder & Krausman, 2001a; Gehrt, 2011). Andelt & Mahan (1980) provided one of the first descriptions of an urban coyote interacting with people and dogs in Lincoln, Nebraska, US, in 1975 before its death at the hands of a hunter. Coyotes have apparently increased in abundance, spreading across New York State at an estimated rate of 78–90 km decade⁻¹ over the past 60 years, culminating in the report of a coyote running through the streets of New York city in 2007 (Fener et al., 2005; Berchielli, 2007; Curtis et al., 2007).

Other carnivore species show less utilization of anthropogenic food sources and may still depend on expanses of native vegetation and resources. Often these species are found within suburban areas where the lower density of human living allows the retention of more natural environments compared with city centres.

Urban badgers Meles meles have been studied in several countries across Europe and Asia (reviewed in Roper, 2010). Badgers appear to have originally become urbanized through the enclosure of relicts of undeveloped habitat within an urban matrix, although there is also some evidence for active colonization (Harris, Baker & Soulsbury, 2010b). Teagle (1969) states that badgers in London, UK, ‘could still be found in Richmond Park and Wimbledon Common and in nearby parks, golf courses and other private property’ (emphasis added). Once established, animals will also spread within the urban matrix (Harris, 1984). Huck, Davison & Roper (2008a) and Delahay et al. (2009) note that complaints by people of damage to property by urban badgers is currently increasing in the UK, possibly implying an active increase in the badger population.

Striped skunks Mephitis mephitis and eastern spotted skunks Spilogale putorius are less well-studied as urban animals, but due to their defence behaviour of spraying, encounters with them can be dramatic and traumatic for humans and their pets (skunks represented 51% of total urban problem wildlife trapped in California up to 1990; Maestrelli, 1990). Reports of skunks in urban areas can therefore be out of proportion to their urban densities (Prange & Gehrt, 2004).

Apart from the red fox and coyote, other canid species have been less successful in urban areas. Gray foxes Urocyon cinereoargenteus (Harrison, 1997; Iossa et al., 2010) and kit foxes Vulpes macrotis (Cypher, 2010) can be found in suburbs of some North American towns, but little is known about the biology of these urban populations at present.

Slender mongooses Galerella sanguinea have been observed in urban and suburban Pretoria and Johannesburg, South Africa, and small-spotted genets Genetta genetta have been observed in urban Johannesburg (PWB pers. obs.; R. Morley pers. comm.), around a town in Ethiopia (Admasu et al., 2004) and urban areas in southern France (Gaubert et al., 2008). Stone or beech martens Martes foina have been recorded as making extensive use of urban environments, at least in central Europe (Herr, 2008), and records date as far back as 1949 (Nicht 1969). More interestingly, the same species is noted as almost completely absent from urban areas in Iberia (Virgós & Casanovas, 1998; Virgós & García, 2002).

Of marsupial carnivores, Virginia opossums Didelphus virginianus are familiar urban animals over much of the US, both colonizing new areas and being introduced outside of their natural, increasing range (Maestrelli 1990 and references therein). Opossums also appear to show a preference for developed areas (Kanda, Fuller & Sievert, 2006; Markovchick-Nicholls et al., 2008). In Australia, southern brown bandicoot or quenda Isoodon obesulus and long-nosed bandicoot Peremeles nasuta populations have become enclosed by urban spread of a number of Australian cities (e.g. Dowle & Deane, 2009). Within this matrix, bandicoots may persist, benefiting from urbanization in terms of control of predators (e.g. red fox; Harris, Mills & Bencini, 2010a). In many cities, bandicoots become habituated to people (pers. obs.) and may benefit from deliberate or inadvertent feeding.

Finally, a number of carnivore species visit upon the fringes of cities or towns. Their home ranges may include some urban area or they may use urban areas for foraging, but they do not live exclusively within urban areas (Iossa et al., 2010).

Apart from domestic cats, very few felids can be considered established urban dwellers. Bobcats Lynx rufus (e.g. George & Crooks, 2006; Riley, 2006; Riley et al., 2010) and pumas Puma concolor (e.g. Beier, 1995; Markovchick-Nicholls et al., 2008; Beier, Riley & Sauvajot, 2010) have been reported from parks and large gardens in suburbs on the edge of the urban-undeveloped land interface in the US, but they do not appear to reside within built-up parts of the cities.

Grey wolves were persecuted by humans, resulting in their extermination from Britain and Ireland by 1773 and significant reduction in numbers on the European continent, driving the few survivors into remote areas far away from human settlement (Cosmosmith, 2011). However, protection of the species has led to increasing numbers of wolves over mainland Europe over recent years, and they are occasionally reported foraging on garbage dumps near towns (see the section: ‘Refuse’).

American black bears have been reported in urban areas of North America (Gunther, 1994; Beckmann & Lackey, 2008) and brown bears Ursus arctos will forage for foods in some European towns, notably Brazov, Romania (Pasitschniak-Arts, 1993; Quammen, 2003). Spotted hyaenas Crocuta crocuta famously enter the streets of Harar, Ethiopia (Kruuk, 2002), and striped hyaenas Hyaena hyaena forage in and around towns in Israel (Yom-Tov, 2003).

How is the ecology of mammal carnivores influenced by urban living?

Perhaps the best way of understanding how carnivores are influenced by living in an urban environment is to compare these animals with populations living in rural locations. This sounds simple in principle; however, it is remarkable how few comparative studies exist. Three key variables are believed to influence a species’ adoption of new environments (Shea & Chesson, 2002): resources, natural enemies and the physical environment. Cities may provide hospitable niches for carnivores due to reliable, non-seasonal food and water resources, reduced threat of natural enemies and/or altered physical environment (e.g. temperature, providing shelter) (Fig. 1). We discuss these aspects below.

Within cities, where are carnivores living?

The presence of natural vegetation within cities is important for supporting significant numbers of carnivores (Baker & Harris, 2007). Therefore, proximity to large expanses of connected habitat (‘green zones’) within cities would provide refuge that may act as resources for animals. Garden size and garden structure are also important factors: Baker & Harris (2007) reported that urban carnivores in the UK are variously negatively affected by the increased fragmentation and reduced proximity of natural and semi-natural habitats, decreasing garden size and garden structure. The presence of flood channels or drainage lines, powerline corridors, beach strands and railroad corridors running through suburbs allow connectivity between habitat patches (Lewis, Sallee & Golightly, 1993) and would support populations of species that will not walk across open areas. The dispersal of food resources within a city is also likely to influence exploitation of these habitats by carnivores. Availability of soil types suitable for drainage and digging burrows is likely to limit utilization by burrowing species (see discussion by Kaneko, Maruyama & Macdonald, 2006). Finally, some urban carnivores make use of anthropogenic structures for shelter and do so even when natural alternatives are available, while other species appear to be completely adverse to using anthropogenic structures. For example, bandicoots show no obvious use of manmade structures, but are dependent on dense vegetation for cover: they are likely to withdraw from manicured or cleared urban gardens (Chambers & Dickman, 2002; FitzGibbon, Putland & Goldizen, 2007).

Foxes require both secure daytime rest sites and breeding sites (earths) to ensure their permanent presence (Baker et al., 2000). Even in urban environments, red foxes still seem to rely on areas to dig earths for denning, so that concentrated housing with small gardens discourages breeding (Harris & Rayner, 1986b). However, many British cities provide ideal habitat for red foxes, for example, inter-war housing with established gardens including hedges and shrubs for daytime cover, together with older residents, fewer children and hence less disturbance (Harris, 1981a; Harris & Rayner, 1986b). Harris (1981a) also recorded breeding foxes making earths under the floorboards of occupied houses and derelict buildings in Bristol, UK. In the US, small road culverts, old barns and other refugia are likely to provide important shelter for red foxes, particularly in the presence of coyote predators (Gosselink et al., 2007). In Australian cities, red foxes often reside in reserves or parklands (pers. obs.). Removing thickets of non-native plants (e.g. lantana, blackberry), which are preferred diurnal rest sites, has been proposed as one means of reducing red fox density in Australia (Marks & Bloomfield, 2006).

Coyotes do not appear to make direct use of buildings for shelter, but within built-up areas, patches of natural forest and scrub, even undeveloped plots amongst housing, are vital as protective cover (Atwood, Weeks & Gehring, 2004; Atwood, 2006; Baker, 2007). For example, all recorded dens in Cape Cod, US, were naturally dug and >300 m from houses (Way et al., 2001). Kit foxes also make use of undeveloped lands (e.g. vacant lots, fallow crop fields), industrial areas (e.g. manufacturing and shipping yards) and open spaces (e.g. parks, canals, railroad and powerline corridors), but will use manmade structures in addition to digging dens (Cypher, 2010).

In contrast with these species, many other carnivore species readily exploit anthropogenic structures for habitat. While badgers in Europe rarely seem to use buildings (Delahay et al., 2009; Roper, 2010), in the suburbs of Tokyo, Japanese badgers Meles anakuma make use of under-floor spaces of empty buildings as resting places (Kaneko et al., 2006). Where available, hollow trees seem to be preferred den sites for raccoons (Stuewer, 1943); however, in urban areas, raccoons favour parks and avoid major roads and the most built-up areas, but do enter houses and make use of sewers, chimneys and other structures as alternative denning sites where hollow trees are in short supply (Hoffmann & Gottschang, 1977; Prange, Gehrt & Wiggers, 2003, and references therein, Hadidian et al., 2010). Striped skunks survive in highly modified urban environments, including ‘single family homes on adjacent lots with manicured lawns and yards’ (Engeman et al., 2003) and can den in crawl spaces under houses (Clark, 1994), while eastern spotted skunks can enter attics (Maestrelli, 1990). Opossums find human habitation extremely suitable as shelter and ‘a penchant for building malodorous nests inside or beneath occupied buildings give the opossum an unwelcome reputation in urban areas’ (Maestrelli, 1990). Finally, according to Delibes (1983), European stone martens live ‘almost exclusively in the human dwellings and their immediate surroundings’ and they prefer inhabited buildings, particularly in winter, presumably because of warmth (Herr et al., 2010). They tend to be absent from grassland and large areas of arable land, probably due to the lack of tree-hollow shelters (Virgós & García 2002 and references therein).

What do they eat?

A diversity of food resources are available to urban carnivores and the majority of well-established urban carnivores include a wide range of items in their diet (see further discussion in the section: ‘Diet’). Food resources available in urban areas include human refuse, crops (i.e. fruit and vegetables), synanthropic rodents and birds, pets, livestock and road-kill, or food made available through deliberate feeding. For example, more than half of the stomach contents of red foxes in Zürich, Switzerland, was anthropogenic, and 85% of surveyed households provided food for foxes (through rubbish bins, compost heaps, garden fruit and food for pets and wild birds) (Contesse et al., 2004). Consequently, urban carnivores have access to an increased range of high nutrition food as well as a greater degree of seasonal food security than do their rural counterparts. With the exception of coyotes (which have been reported to hunt singly in urban environments, not in groups), urban carnivore species are not generally group hunters (Iossa et al., 2010). This may reflect the generalist nature of the successful urban dwelling species, as well as the rich, easily accessible anthropogenic food that does not necessitate cooperative hunting behaviour.

Refuse

Carnivores can benefit by utilizing sources of high-energy food from human refuse. Even in rural areas, there may be huge amounts of anthropogenic waste. For example, Yom-Tov, Ashkenazi & Viner (1995) estimated that 1208 tons of meat was disposed of by farmers in the Golan Heights, Israel, in 1 year; many carnivore species take advantage of such resources. In urban areas, carnivores may forage at tip sites and often turn over bins in backyards, streets and parks.

Red foxes are both generalist and eclectic in their diet choice, that is, they eat a broad range of food types and also can adapt to substantial local variation in food types available (Harris, 1981b; Reynolds & Tapper, 1995). In addition to items that rural foxes consume (i.e. small mammals, fruit, earthworms, etc.), urban red foxes may rely heavily on scavenged anthropogenic food (Baker et al., 2000), and scavenged food can make up to 60% of an adult urban fox's diet (Doncaster, Dickman & Macdonald, 1990; Saunders et al., 1993). For example, in Zürich, over 50% of fox stomachs examined contained anthropogenic food and this increased from suburban to city centre zones (Contesse et al., 2004). Refuse and discarded fast food is of such abundance that the Zürich fox population is still increasing (Contesse et al., 2004). In Orange County, California, remains of human food and food packaging were present in 62% of fox scats, particularly during winter months (Lewis et al., 1993).

The faeces of raccoons from a primarily urban site (Glendale, Ohio, US) contained seeds from 46 species of plants including human food (e.g. grapes, corn and watermelon) that probably came from raided bins (Hoffmann & Gottschang, 1977). Prange et al. (2003) recommended reduction or elimination of anthropogenic food as the best control method for problem urban raccoons. Yom-Tov (2003) reported that around Israeli settlements, badgers do not feed on garbage dumps but rather feed on vegetables in agricultural fields. However, of questionnaires returned by Bristol residents, 16.4% of complaints were due to badgers disrupting bins (Harris, 1984). Striped skunks will raid bins and bee hives in urban areas (Clark, 1994) with up to 18% of the diet of eastern striped skunks living near humans sourced from trash (Hamilton, 1936), while bin-raiding by opossums make them one of the most commonly reported pest species (Clark, 1994).

Inadvertently enticing animals closer to human settlements through the provision of refuse is likely to be the first step towards these animals becoming habituated to human presence. For example, banded mongooses Mungos mungo have been recorded feeding at tips in Uganda (Gilchrist & Otali, 2002) as have red foxes in Saudi Arabia (Macdonald et al., 1999) and brown bears in Europe (Quammen, 2003). Wolves make use of refuse dumps in Israel (Yom-Tov, 2003), Canada (Geist, 2007), Italy (Cosmosmith, 2011) and Romania (Promberger et al., 1998). Such feeding behaviour has resulted in increased habituation to humans to the extent that they have little fear of people. In Canada, wolves are reported to approach the dump truck carrying refuse to the tip (timing their arrival to that of the truck) and thus have come to associate human smell with the provision of food (Geist, 2007).

Animals that raid human refuse for food are likely to also ingest substantial quantities of non-food material, which might become detrimental to their health. In addition to anthropogenic food items, the faeces of raccoons from urban sites include a variety of non-food items (e.g. plastic, rubber bands) that probably came from raided bins (Hoffmann & Gottschang, 1977). Even though coyotes (Gehrt, 2007) and stone martens (Eskreys-Wójcik & Wierzbowska, 2007) are not noted as bin raiders, 2% of Chicago coyotes’ scats have evidence of human refuse, for example, fast food wrappers, pieces of rubber, sweet wrappers, plastic, string and aluminum foil (Morey, Gese & Gehrt, 2007), and 17% of stone marten scats from urban areas contained rubber and plastic, etc. (Eskreys-Wójcik & Wierzbowska, 2007).

Crops (fruit, vegetables and grain)

Fruit is of major seasonal importance to badgers, making up 48–61% of the diet (stomach contents and faeces) of Bristol badgers (Harris, 1984), and persimmons are found in 100% of autumn-collected Japanese badger scats in urban Tokyo. Stone martens also rely heavily on fruit (present in 43% of scats, Baghli, Engel & Verhagen, 2002; Lanszki, 2003). Even species such as coyotes and foxes may use fruit as a significant food source. Fruit is present in 23% of Chicago coyote scats (Morey et al., 2007), and 43% of urban Washington State coyote scats (Quinn, 1997a). Lewis et al. (1993) reported seeds of >44 plant genera (from >28 plant families) present in 73% of the scats of red foxes from Orange County, California (with seasonal differences: greater occurrence in autumn). Contesse et al. (2004) recorded wild fruit in the stomachs of 23% of urban Zürich red foxes examined, and cultivated fruit and crops in 49%. These carnivore species may therefore significantly benefit from orchards and market gardens around towns and cities, and the badger, for example, appears to take less scavenged food at the time of year that seasonal fruit is available (Harris, 1984).

Synanthropic prey – rodents and birds

The rich resource of rodents in urban areas is likely to have encouraged the first cats into close association with humans, as discussed earlier. Rodents and birds (especially synanthropic species, e.g. sparrows, pigeons) are also a major food source for a number of other carnivore species, most notably coyotes, red foxes, stone martens and badgers. Rodents are present in 42% of Chicago coyote scats (Morey et al., 2007) and 26% of Zürich red fox stomachs (where they make up 11% of total stomach content; Contesse et al., 2004). Rodent remains are present in 14.3% of Tokyo Japanese badger scats in spring (Kaneko et al., 2006).

Although they only make up 5% of stomach volume, bird prey were present in 24% of Zürich red fox stomachs (Contesse et al., 2004). In California, bird remains are present in 70% of fox scats in built-up areas including extensive amounts of duck and passerine remains, with egg shell in 5% of scats (Lewis et al., 1993). 6.2% of badger samples collected in Bristol (Harris, 1984), but 29% of urban Tokyo Japanese badger scats (Kaneko et al., 2006) include bird remains during spring (when birds are breeding).

In urban Luxembourg, stone martens prey principally on synanthropic birds and mammals (Herr, 2008). Lanszki (2003) compared stone marten scats from a small village and surrounding agricultural area in Hungary. Stone martens from both areas relied heavily on fruit (cultivated fruit for village animals, more wild fruit in rural animals), while village martens included a high proportion of birds (e.g. house sparrows) in their diet (20% for village compared with 11% for rural animals) but fewer mammals (13% for village compared with 35% for rural animals).

Pets, livestock and road-kill

Urban carnivores may also make use of domestic animals as prey. For example, three studies report that between 1 and 13% of the diet of urban coyotes is made up of cats (MacCracken, 1982; Quinn, 1997b; Morey et al., 2007). Urban areas may also provide food for scavenging in that the numbers of road kills around towns and cities is likely to be higher than it is for rural areas. For example, in a park surrounded by urbanization in Ohio, US, coyotes eat a primarily ‘natural’ diet of small to large mammals, but they also take advantage of the many white-tailed deer road-kill carcasses (Cepak, 2004), a resource that would normally be rare. Pets and livestock (including hens, cats, dogs, rabbits and cattle) make up 4.5% of the gut volume of Zürich red foxes (Contesse et al., 2004) and a small proportion of the diet of Californian red foxes (Lewis et al., 1993). In Europe, grey wolf and brown bear numbers have been increasing, which is likely to result in increased interaction with humans at the rural-urban interface: because of the decline in the number of wild game, they have begun to prey on domestic horses, cattle and dogs (and will also take fruit and vegetables, e.g. potatoes) (Pasitschniak-Arts, 1993; Cosmosmith, 2011).

Deliberate feeding

In many urban areas across developed countries, households may regularly put out food for urban carnivores such as badgers and even foxes. Roper (2010) reported that 29% of householders surveyed in Brighton deliberately provided food for foxes, badgers and other mammals, and over half of these households were providing food every night. Lewis et al. (1993) reported an individual person regularly feeding red foxes within a Californian urban park, providing an average (±sd) of 7.12 ± 0.23 kg day⁻¹ of beef, chicken, turkey and fish (measured over a 48-day period) to the ∼40 foxes present in the park (∼0.177 kg per fox per day). Even if the food is not left deliberately, many wild carnivores will regularly take dog or cat food left accessible. For example, in Zürich, when pet food was present in a fox stomach, it made up the majority of the stomach contents (Contesse et al., 2004). With the high energy content of anthropogenic food, one or two households leaving out food may have a significant effect on the foraging behaviour of these animals.

Food security in urban areas

One of the greatest advantages of anthropogenic food sources may be that they are more reliable compared with natural food sources. For example, urban coyotes show a seasonal pattern in some dietary foods (e.g. fruit) but also eat refuse (as do those in more rural areas if they can access it) (Quinn, 1997a), which is less likely to be seasonally affected. Similarly, although red foxes are eclectic feeders and can easily adapt to variation in food types available (Reynolds & Tapper, 1995), seasonal variation of London fox diet appears to be less pronounced than in rural foxes (Harris, 1981b). Even so, some seasonal variation in diet has still been demonstrated for certain urban red fox populations (Oxford: Doncaster et al., 1990, e.g. Zürich: Contesse et al., 2004).

In rural areas of Britain and Ireland, the most favoured badger habitats are broad-leaf woodlands and meadows (Feore & Montgomery, 1999) that provide them with access to large numbers of earthworms (Kruuk, 1978, 1989). However, in an urban environment, badgers seem to avoid open grasslands (lawns, playing fields, etc.) within their home ranges (supporting the contention that they are opportunistic generalists rather than earthworm specialists; Roper, 1994). Instead, urban badgers expand their diet range to include more anthropogenic food sources (e.g. refuse and garden crops) to the extent that earthworms are seasonally only a minor dietary component (Harris, 1984; Huck et al., 2008b).

What kills urban carnivores?

Review of the literature indicates many anecdotal statements (but few records) regarding causes of mortality in urban carnivores. Causes of mortality can also be dynamic, with principal causes shifting over time, making it difficult to carry out direct comparison between urban and rural environments. For example, the principal cause of mortality in red foxes in the US has shifted away from hunting due to an over 10-fold fall in fur prices making hunting less profitable (Gosselink et al., 2007). Additionally, avoidance of an encroaching competitor/predator (the coyote) has resulted in increased road mortality in red foxes because they are utilizing habitat that brings them closer to human habitation (Gosselink et al., 2007). Waves of disease have also resulted in significant mortality in carnivores. In dense urban populations, where individuals live in closer proximity to each other, it is intuitive that the likelihood of an infectious disease spreading may be increased (but see also White, Harris & Smith, 1995, who predicted that heterogeneity of urban habitats meant lower frequency of contact between rabies infected and uninfected British foxes than in rural populations of a similar density).

We summarized 29 studies that included cause of death statistics for red fox, coyote, badgers, bobcats and raccoon to investigate whether the causes of death differed between urban and rural areas (Fig. 2). We identified the absolute numbers of animals where cause of death was identified as due to motor vehicles (‘cars’ or ‘road-kill’), hunting/euthanasia, toxicity, predation/aggression, disease, starvation/emaciation and unknown/other.

Figure 2. Causes of mortality in rural and urban populations of badgers Meles meles, red foxes Vulpes vulpes, coyotes Canis latrans, raccoons Procyon lotor, bobcats Lynx rufus and skunks Mephitis mephitis. Data have been extracted from various sources and percentages have been calculated from the total absolute number of animal deaths across all studies inclusive: coyote (Andelt & Mahan, 1980; Atkinson & Shackleton, 1991; Holzman, Conroy & Davidson, 1992; Chamberlain & Leopold, 2001; Way et al., 2001; Grinder & Krausman, 2001a; Tigas, Vuren & Sauvajot, 2002; Riley et al., 2003; Van Deelen & Gosselink, 2006; Berger & Gese, 2007; Gosselink et al., 2007; Schrecengost et al., 2009; Gehrt, 2011); red fox (Harris & Smith, 1987; Lewis et al., 1993; Gosselink et al., 2007; Soulsbury et al., 2007); skunks (Gehrt, 2005); badgers (Cheeseman, Wilesmith & Stuart, 1989; Kowalczyk et al., 2003); and raccoons (Mech, Barnes & Tester, 1968; Glueck, Clark & Andrews, 1988; Clark et al., 1989; Hasbrouck, Clark & Andrews, 1992; Riley et al., 1998; Chamberlain et al., 1999; Gehrt & Fritzell, 1999; Prange et al., 2003; Gehrt & Prange, 2007; Urbanek, Nielsen & Wilson, 2009). *Data for 1636 red foxes in Bristol, UK, were assumed to include data collected previously by the same researchers (1978–80, n = 87 Harris 1981a, 1977–86, n = 564, Harris & Trewhella 1988).

Road accident has been listed as a major cause of mortality in carnivores, killing a large proportion of badgers (57%), red foxes (40%), coyotes (31%), bobcats (38%) and skunks (30%), with little difference evident between urban and rural habitats where these data are available (Fig. 2). Road death is likely to be biased towards individuals that disperse further, for example, males and juveniles (Baker et al., 2007). Of the 151 recorded deaths of black bears in urban environments (over a 10-year period), all were due to humans, and 89 of 151 (59%) were killed by vehicles (Beckmann & Lackey, 2008). In urban areas, deaths exceeded recruitment meaning urban areas were sinks for this species (Beckmann & Lackey, 2008). Notably, an estimated 50 000 badgers are believed to die on British roads each year (Harris et al., 1992, 1995), which equates to 49% of all adult and post-emergence cub fatalities. We could not find published mortality statistics specifically for urban badgers for comparison. Road accident is a major cause of death in urban raccoons (31%), but less so for rural animals (8%). Roads can act as barriers to dispersing wildlife (e.g. pumas Beier, 1995; bobcats and coyotes; Riley et al., 2003), although this can be mitigated by culverts and underpasses (Grilo, Bissonette & Santos-Reis, 2008; Harris et al., 2010a), while Bristol red foxes change their activity patterns, avoiding roads prior to midnight when traffic volume is higher (Baker et al., 2007).

Hunting and destruction (i.e. euthanasia) are the next most common causes of death among carnivores (Fig. 2). Hunting/destruction is the major cause of death for raccoons (61% of mortalities), especially for rural raccoons (64%), while almost a quarter (22%) of rural coyotes die at the hands of hunters (compared with 9% in urban environments). Thirty-eight per cent of deaths of red foxes in European cities are due to animals being destroyed. Similar figures exist for urban areas in the US (35%), but in rural US, hunting is a minor cause of death in foxes (9%), where predation (38%) and death on roads (40%) are the major causes of mortality.

Pollutants (e.g. motor oil and antifreeze) and poisons (particularly anticoagulant rodenticides, directly poisoning the animal or where the carnivore takes poisoned rodents) are likely to be a significant cause of mortality in urban carnivores. However, our literature review indicated that toxicity is listed as a cause of death for only a few urban and rural coyotes (Riley et al., 2003; Van Deelen & Gosselink, 2006) (Fig. 2) and for kit foxes (Cypher, 2010). The lack of reports may be related to difficulty in ascertaining poisoning as a cause of death, particularly when carcasses are located some time after death. Organochlorines (Dip et al., 2003) and lead (Dip et al., 2001) are found in higher concentrations in urban than rural red foxes in Zürich. Organochlorine levels increase in adult male foxes but not vixens, which appear to pass the compounds to their offspring through lactation.

Protection from predators is likely to play an important role in selection of urban habitats. Predation or aggression is responsible for the death of only 10% of urban coyotes compared with 25% for rural populations, where they conflict with wolves (Fig. 2) and it has been suggested that the massive increase in coyote numbers over recent decades is likely due to reduction in the numbers of grey wolf across North America (Gese & Bekoff, 2004). In turn, coyotes tend to avoid landscapes with extensive human presence, and their conflict with red foxes means that foxes end up being relegated to areas with relatively more intense human activity (e.g. roads, farmsteads) (Gosselink et al., 2007). An estimated 38% of red foxes in rural US die due to predation/aggression, largely due to conflict with coyotes, compared with only 12% in urban US (Gosselink et al., 2007). In the UK, the absence of a natural predator for the fox results in less predation. Nevertheless, even in urban UK (London and Bristol), a high proportion of red foxes die due to wounds incurred during aggression, principally from stray dogs or conspecifics (Harris & Smith, 1987; Soulsbury et al., 2007). Recent control of stray dog numbers, however, has reduced the incidence of aggression as a cause of death (S. Harris, pers. comm. 2010).

Disease has been recorded as the major cause of mortality for urban raccoons, accounting for an average of 50% of deaths in urban areas compared with only 19% of rural raccoons (Fig. 2). High levels of sarcoptic mange have been recorded in urban red foxes in Britain, causing population crashes (Baker et al., 2000; Soulsbury et al., 2007) and to a lesser extent in urban coyotes (Grinder & Krausman, 2001a). Disease outbreaks may be important factors affecting populations of other carnivore species; however, we note that not all authors indicate a breakdown for disease that would allow comparison – for example, disease and starvation/emaciation are often not distinguished.

Consequences of increased food and water, protection from predation

As a consequence of increased food and water availability in urban habitats, coupled with protection from predators, growth rate, body condition, survival and population densities of carnivores are predicted to be favoured.

Accelerated growth rates and reduced weight loss over winter

The presence of abundant, high-energy, non-seasonal food sources in urban areas may have a significant effect on the growth of carnivore species. Yom-Tov (2003) examined museum specimens collected from Israel over 60 years (from 1945 to 2005), a time span when human population in the country increased approximately eightfold, resulting in a significant increase in anthropogenic food sources (Yom-Tov, 2003). He recorded that, over this time, species that do not use anthropogenic food (the caracal Caracal caracal and jungle cat Felis chaus) did not significantly change in mass or size; however, wolves, golden jackals Canis aureus and striped hyaenas, which all feed from garbage dumps and make use of livestock carcasses, increased in body mass. The larger species appeared to be more capable of exploiting the extra food provided by humans (Yom-Tov, 2003). A similar pattern of size increase in skull measurements was also recorded for badger and red fox populations in Denmark from 1862 to 2000, which again could be related to altered human agriculture and therefore food sources (Yom-Tov, Yom-Tov & Baagøe, 2003).

Starvation due to substantial weight loss over winter is a significant cause of death in skunks, but urban skunks fare better over winter than their rural counterparts (Rosatte et al., 2010). Similarly, urban raccoons exhibit better physical condition than rural ones, possibly due to anthropogenic food (Rosatte, Power & Macinnes, 1991). Black bears in urbanized Nevada average 30% heavier than bears in rural areas due to a diet heavily supplemented by garbage (Beckmann & Lackey, 2008). Urban kit foxes demonstrate greater body mass compared with non-urban individuals (especially for juveniles) and also demonstrate different haematological characteristics (Cypher, 2010). Urban Eurasian badgers can be heavier than nearby rural badgers, presumably due to the availability of anthropogenic food (Roper 2010 and references therein). More research in this area is needed.

Increased reproduction and survival in urban habitats

Increased survivorship has been recorded for a number of urban carnivore species ( Table 1). Opossums are recognized as bin-raiders par excellence (Clark, 1994), and their reliance on anthropogenic sources of food is such that, in areas where one would expect their range to have been limited by the winter cold and lack of natural food, they are, in fact, well-established (Kanda, 2005). Similarly, in addition to their better physical condition, urban raccoons demonstrate increased survival and higher annual recruitment compared with rural animals and subsequently increased site fidelity and higher population densities (Hoffmann & Gottschang, 1977; Rosatte et al., 1991; Riley, Hadidian & Manski, 1998; Smith & Engeman, 2002; Prange et al., 2003).

Table 1. As a consequence of abundant aseasonal food resources and protection from predation, many carnivore species demonstrate increased survival in urban habitats compared with their rural counterparts, consequently reaching higher population densities and exhibiting smaller home range sizes.? indicates that comparative data for urban and rural locations do not appear in the literature

	Survival	Density	Territory size	Reference
Coyote	Higher	Higher	Larger, smaller	Fedriani et al., 2001; Riley et al., 2003; Atwood et al., 2004; Gehrt, 2007, and references therein, Gehrt & Riley, 2010
Red fox	Higher	Higher	Not different, smaller	Lyoyd, 1980; Harris, 1981a; Harris & Rayner, 1986a; Adkins & Stott, 1998; Marks & Bloomfield, 1999a; Heydon, Reynolds & Short, 2000, data cited by Contesse et al., 2004; Webbon et al., 2004; Iossa et al., 2010
Eastern striped skunks	?	Equal or higher	Smaller, not different	Verts, 1967; Bixler & Gittleman, 2000; Gehrt, 2004; Prange & Gehrt, 2004; Rosatte et al., 2010
Stone martens	?	Higher	Not different	Herr, 2008; Herr et al., 2009a
Eurasian badger	?	Equal or higher	Larger, smaller	Harris, 1982; Cresswell & Harris, 1988a; Cheeseman et al., 1988b; Feore & Montgomery, 1999; Kowalczyk et al., 2000; Davison et al., 2009; Huck et al., 2008a
Raccoon	Higher	Higher	Smaller	Hoffmann & Gottschang, 1977; Riley et al., 1998; Smith & Engeman, 2002; Prange et al., 2003; Prange, Gehrt & Wiggers, 2004; Hadidian et al., 2010
Opossums	Higher	Equal or higher	?	Prange & Gehrt, 2004; Kanda, 2005

Although most coyotes in urban Chicago die before reaching their second year (Gehrt, 2011), urban coyote populations nevertheless show higher survival compared with rural studies, where coyotes are exposed to wolf predation, as well as hunting and trapping by humans (Gehrt, 2007 and references therein). Female black bears in urban areas of Nevada give birth much earlier (between 4 and 5 years of age, some as early as 2–3 years of age) than rural bears (7–8 years; Beckmann & Lackey, 2008). Urban black bear survival was, however, so much lower that this higher fecundity does not translate to higher recruitment and urban areas act as sinks. The evidence for reproductive rate and survival in red foxes seems to be mixed: even if urban animals do exhibit higher reproductive rates, this may, however, be countered by lower survivorship (e.g. Harris, 1977; Doncaster & Macdonald, 1991).

In their taxonomic review of urban carnivores, Iossa et al. (2010) indicated that although juvenile and adult survivorship for urban carnivore species tends to be higher than for their rural counterparts, the pattern is not statistically significant across taxa (n = 4 species for juvenile survivorship and n = 8 species for adult survivorship).

Increased population density

Carnivore species that are able to exploit additional food resources are likely to exhibit higher population densities in urban compared with rural environments. For example, coyotes, red foxes, eastern striped skunks, stone martens, badgers, raccoons and opossums, all may reach higher densities in cities compared with rural areas (Table 1) (Iossa et al., 2010). Carnivores may reach extremely high densities in urban areas. For example, Fedriani, Fuller & Sauvajot (2001) reported densities of 3 coyotes km⁻² in urban southern California, which is approximately seven times higher than that in rural locations. The highest badger density may be 33 adults km⁻² recorded for Brighton, UK (Huck et al., 2008a). The highest density recorded for raccoons is an astonishing 333 individuals km⁻² (estimated for an urban park in Fort Lauderdale, Florida), which is ∼4 to ∼400 times the density recorded for rural populations (Riley et al., 1998; Smith & Engeman, 2002).

Although 87% of the total British red fox population may be located in rural areas (Webbon, Baker & Harris, 2004), foxes may reach much higher densities in urban than rural locations. In Bristol, red fox densities of up to 37 individuals km⁻² have been recorded (Baker et al., 2001), while 16 individuals km⁻² were recorded for Melbourne, Australia (White et al., 2006). Town councils and local mammal groups across Great Britain record either stability in numbers or more often an increase in urban fox populations, with more cities in the south-east and coast areas hosting foxes (Wilkinson & Smith, 2001, and references therein). Foxes are also now present in cities that earlier models of population growth (Harris & Smith, 1987) predicted would not host foxes. Although the most common reason given for perceived increases in urban fox numbers is increased food availability, Harris (1981b) found that, at least in some urban areas, waste food formed a small part of a fox's diet, suggesting that other variables are involved.

In contrast with the species listed earlier, there seem to be conflicting data for opossums. Prange & Gehrt (2004) suggested that opossum densities are not increased in urban areas, with opossums being relatively more common in rural than urban parts of north-eastern Illinois, US. Kanda et al. (2006), however, reported that road-killed opossums in Massachusetts, US, are more common in areas of low forest cover and more human development, and the authors considered them urban generalists. Similarly, striped skunks can be regarded as generalists par excellence, being found in nearly all habitats across North America (Verts, 1967). Densities, however, do not generally seem to differ between urban and rural locations (Gehrt, 2004; Prange & Gehrt, 2004), suggesting either an inability to make extensive use of anthropogenic resources as successfully as other carnivore species or some other constraints.

Social systems and territorial behaviour

Greater resource availability and increased population density for urban carnivores are likely to determine their social behaviour. The corollary of having more animals resident in urban areas is that either the individuals have smaller exclusive territories or that their home ranges overlap with more individuals, implying considerable changes in sociality and behaviour. Creel & Macdonald (1995) discussed five selective pressures that appear to influence sociality in carnivores (Table 2). Examining the potential action of these factors in the urban environment suggests that resource dispersion and dispersal costs are likely to have the greatest influence on carnivore sociality, and we predict reduced territoriality or aggression for urban carnivores, reduced home range area for individuals, increased group sizes, greater dispersal of individuals from their natal sites and altered mating systems (Table 2). Reviewing the literature suggests that there is evidence to support these predictions of social plasticity (e.g. for foxes and coyotes), although we need more direct comparisons between rural and urban using standard methods to make general conclusions regarding these aspects of carnivore biology.

Table 2. Prediction table describing how the factors influencing sociality in carnivores (sensu Creel & Macdonald, 1995) are influenced by living in the urban environment

	Mechanism for action	Prediction for urban environment
Resource dispersion	Abundant prey, or rich or highly variable resource patches result in conspecifics being more tolerated	Reduced competition would result in reduced aggression, overlapping home ranges and the formation of larger groups.
Dispersal costs	High numbers of mates, plenty of suitable habitat and reduced intra-specific competition	Reduced competition for resources would encourage dispersal.
Resource acquisition	Groups needed to acquire and retain resources, e.g. through hunting	Largely negated by the abundance of ‘easily captured’ food
Defence against predators	Groups less vulnerable than solitary animals	Largely negated by the absence of natural predators
Reproductive advantages	Groups better at feeding and protecting young than are solitary animals	Largely negated by abundant food resources year round

Generally, red foxes appear to have smaller home ranges and shorter dispersal distances under higher population densities (Macdonald, 1980; Adkins & Stott, 1998). However, red foxes are so behaviourally plastic that it is often difficult to demonstrate any overt territorial and social behaviour (Cavallini, 1996). This plasticity is demonstrated upon release from population pressures. For example, an outbreak of sarcoptic mange in Bristol foxes caused a population crash that resulted in the remaining foxes increasing their home range size, even though food availability did not change (Baker et al., 2000). Additionally, in Oxford and Toronto, Canada, suburban populations have more stable territories than foxes closer into the cities (Doncaster & Macdonald, 1991; Adkins & Stott, 1998). Diet and home range were not different between red foxes in the two areas in Oxford, and the shifting territories were likely to be due to a higher turnover of the fox population in the more disturbed city centre (Doncaster & Macdonald, 1991). Changes in the distribution of food has a rapid effect on social structure: Macdonald et al. (1999) found that otherwise sparsely distributed red foxes in Saudi Arabia centred their activities around often ephemeral but important food resources such as camel carcasses and shifting human camps and were tolerant of the presence of other foxes. Most major cities in Switzerland support red fox populations, most likely due to the anthropogenic food supplies available; as rural foxes are shy, Contesse et al. (2004) suggested that this colonization must have entailed ‘behavioural ontogenetic adaptations’.

Like red foxes, raccoons appear to have a plastic social system. Generally thought to be solitary and asocial, there is some evidence that loose groups of males maintain territories that overlap with those of solitary females (Chamberlain & Leopold, 2002). Territoriality collapses in urban areas with concentrated sources of food, where raccoons can reach extraordinary densities (Smith & Engeman, 2002).

Badger society is based around their setts (Kruuk, 1989) and badger distribution in urban areas seems to be partly dependent on suitable areas for digging setts (Huck et al., 2008a). In urban areas, distribution of suitable soil (with appropriate drainage) is patchy, and it has been noted that zones of intermediate human population density are apparently favoured (Huck et al., 2008a; Davison et al., 2008). Badger setts in some urban UK sites are smaller than nearby rural setts, possibly indicating their more recent provenance and therefore an active colonisation process (Davison et al., 2008). Bristol and Brighton (UK) urban badgers demonstrate less territorial behaviour (e.g. no scent marking of boundaries) and higher rates of dispersal than rural populations (Harris, 1982; Cheeseman et al., 1988a; Cresswell & Harris, 1988b; Davison et al., 2009), but while Bristol badgers had larger but more overlapping home ranges, the Brighton badgers had small non-contiguous territories typical of low density rural populations (Davison et al., 2009). The Brighton population had extremely high population density, however (Huck et al., 2008a), with much dispersal between groups, suggesting that differences between Bristol and Brighton are less to do with badger population density than the nature of the urban environment itself (Roper, 2010). The presence of solitary animals in urban environments, not obviously affiliated with any particular group, reflects the diverse range of social organization in badgers (Kowalczyk, Bunevich & Jedrzejewska, 2000).

In general, coyote densities are higher for urban compared with rural areas. Urban coyotes demonstrate both smaller (Andelt & Mahan, 1980; Atwood et al., 2004; Gehrt, 2007, and references therein) and larger (Riley et al., 2003) home ranges than their rural counterparts. Home range size may largely be driven by resources available rather than population densities as coyotes avoid the most built-up areas, preferring wooded/shrubby areas to open areas (Quinn, 1997b) and need areas of ‘natural’ cover (vegetation) within their urban territories, which influences dispersal patterns (Grinder & Krausman, 2001b). Water, which may limit coyote distribution and density in deserts (Gese & Bekoff, 2004), is not likely to be a limiting factor in urban areas. Gehrt and colleagues (Gehrt & Prange, 2007; Gehrt, Anchor & White, 2009; Gehrt, 2011) refer to urban coyotes forming packs and suggest that, although coyotes prefer to hunt alone, they form packs to defend territories, with roughly half of all urban coyotes living in territorial packs that consist of five to six adults and their pups that were born that year.

This pattern of altered territories does not, however, hold true for all carnivore species in an urban area. Despite reaching moderately higher population densities in urban compared with rural locations, stone martens show no significant differences in territory size between the two habitats, even though the territories of urban martens fell entirely within built-up areas (Herr, Schley & Roper, 2009a).

Which species make the best urban adapter/exploiter?

Understanding the biology of urban adapters and exploiters may enable us to explain their role in cities and also allow predictions regarding their future and the abilities of other carnivores to establish within urban areas. It is possible that some ‘non-adapted’ species (i.e. ‘urban avoiders’, sensu McKinney, 2006) may adapt to the urban environment sometime in the future or even come to exploit human resources in and around cities. What, though, are the features that make some species better than others at becoming urban dwellers?

Geographic location

Although we have reviewed literature from all continents across the globe, we note that there is a bias towards ‘western’ societies in terms of the reporting rate for urban carnivores: we found few or no reports of urban dwellers other than anecdotal information outside of Europe, North America, Japan and Australia. We suggest that this bias may, firstly, reflect differences in human population densities. Higher human population density results in an increased proportion of ‘urbanized’ land and reduced availability of undeveloped landscape, pressurizing or enticing animals to use urban habitat. Secondly, the nature of cities will influence whether carnivores reside there. Compared with densely populated city centres, suburbs and towns support greater natural resources and therefore provide more opportunity for urban carnivores. Thirdly, Iossa et al. (2010) pointed out that there is a high prevalence of populations of feral and stray dogs in developing countries, which might limit the presence of carnivore species (e.g. Vanak & Gompper, 2009; Vanak, Thaker & Gompper, 2009). Finally, across the globe, people will respond differently to carnivores entering urban environments, which may contribute to differences in reporting ratio. In India, culturally based tolerance towards carnivores allows many small carnivores and even leopards Panthera pardus, wolves, sloth bears Melursus ursinus and striped hyaenas to persist among high human population densities, albeit in agricultural landscapes (Karanth & Chellam, 2009). In south China, large and small carnivore species have been extirpated or greatly reduced in numbers; ironically, it is in mostly highly urbanized Hong Kong, with strong legal protection, where surviving species can be most easily encountered (Lau, Fellowes & Chan, 2010). Outside of anecdotal information, we could find no reports of carnivores living in African cities, despite a vast array of carnivore species on the continent. This may reflect the nature of urbanization or the nature of predator guilds in Africa: large expanses of adjacent rural or undeveloped habitat may provide sufficient alternative resources, while human self-preservation or protection of livestock may preclude the establishment of some carnivore species close to urban areas.

The role of taxon

All major terrestrial carnivore families have representatives that show some degree of association with human settlement (Fig. 3a). There appears to be no taxonomic restriction in terms of an ability to exploit urban environments. The major restrictions may therefore be in terms of body size and dietary flexibility.

Figure 3. There are no clear phylogenetic determinants of whether a species will show some adaptation to anthropogenically disturbed habitat (a), although body mass (b) does appear to be relevant: most eutherian Carnivora species that are known to associate with towns and cities are in the range of 1–30 kg. Data do not include non-terrestrial carnivores or vegetarian diet specialists (e.g. red panda, Ailurus fulgens; Ailuridae).

Body size

Body size plays an important part in determining whether a carnivore species uses the urban environment. The proportion of species that utilize human habitat – from villages through to cities – is not spread evenly across the range of eutherian terrestrial carnivore body masses (Fig. 3; χ²₆ = 12.60, P = 0.05). Both small and large carnivores are under-represented in the urban environment.

Body size is important in terms of how a species is able to deal with the habitat fragmentation implicit with urban environments. Larger body size is a benefit in human-fragmented agricultural landscapes if it aids the animals’ ability to move in and out of the fragment matrix (e.g. coyotes), but body size should not be too large that viable populations cannot survive in small habitat fragments (Gehring & Swihart, 2003). Crooks (2002) reported that of 11 predator species in southern California, the four largest (puma, coyote, bobcat and American badger, Taxidea taxus) and two smallest (western spotted skunk Spilogale gracilis and long-tailed weasel Mustela frenata) species were most sensitive to fragmentation of natural habitat. The medium-sized species (raccoon, gray fox, cat, opossum and striped skunk) were the best at adapting to fragmented and anthropogenically modified habitats. Gehring & Swihart (2003) found a similar result for eight carnivore species at Indiana, US (coyote, red fox, gray fox, raccoon, striped skunk, opossum, cat and long-tailed weasel).

In addition to compromised mobility, small carnivores are also likely to conflict with domestic cats and dogs. For example, Harris (1981a) reported that 15% of red fox cubs were killed by animals; in most cases, these were known to have been stray dogs. The British cat population (total ∼9 million cats) killed an estimated 92 million prey items over a period of 5 months (from April to August), of which 57 million were mammals (Woods, Macdonald & Harris, 2003). Although only 0.1% of this mammal prey could be identified as other carnivores, 9 million cats is 20 times the population of weasels Mustela nivalis and stoats M. erminea and 38 times the population of red foxes in Britain (Woods et al., 2003), implying the possibility of intense competition.

Despite their size, some large carnivores have managed to maintain an uneasy truce at some urban interfaces by moving in and out of the urban matrix, for example, brown bears (Swenson et al., 2000; Kaczensky et al., 2003; Rauer, Kaczensky & Knauer, 2003), black bears (Witmer & Whittaker, 2001; Beckmann & Berger, 2003; Beckmann & Lackey, 2008) and spotted hyaenas (Patterson et al., 2004; Kolowski & Holekamp, 2006). Although they are also active killers of live prey, these species scavenge, making use of the rich resources available around cities. Wolves can also come into surprisingly close contact with humans in rural (Bangs & Shivik, 2001; Musiani et al., 2003; Wydeven et al., 2004) and urban (Promberger et al., 1998) areas.

Although their size is an advantage in terms of accessing resources over a wide area, it can also make large carnivores a greater threat to humans and, clearly, human tolerance is a limiting factor for some species (Iossa et al., 2010). Most large (>20 kg, Carbone, Teacher & Rowcliffe, 2007) carnivores have given way before humans (Woodroffe, 2000; Cardillo et al., 2004), generally avoiding built-up areas. On average, felids (23.1 ± 39.7 kg, range 1.3–164 kg, n = 36 species) are larger than other carnivores (average 9.1 ± 22.8 kg, range 0.104–173, n = 173 species, t₂₀₇ = 2.90, P = 0.004; analysed from raw data presented by Meiri, Simberloff & Dayan, 2005); and their trend to hypercarnivory (> 70% meat in the diet) and propensity for killing rather than scavenging prey seems to preclude large felids from residing comfortably with humans. A greater proportion of the largest carnivores are felids, which include some of the most dangerous carnivores that have, or occasionally still do, live in close association with humans (e.g. lions Panthera leo and tigers Panthera tigris; Löe & Röskaft, 2004). In the Sundarbans region of Bangladesh, 392 people were killed by tigers between 1956 and 1970 (Hendrichs, 1975), and 79 people from villages close to the mangrove jungle were killed by tigers between 2002 and 2006 (Khan, 2009). Löe & Röskaft (2004) cited over 12 000 human deaths reported globally in the 20th century due to tigers (in the same period only 313 deaths from brown bears were recorded).

For carnivores, a body mass of 20 kg marks where a shift from small prey to large vertebrate prey occurs (Carbone et al., 2007). With the exception of the occasional coyote, all the well-established urban dwellers are well below this mass (average 4.60 ± 4.56, n = 11, min eastern spotted skunk: 0.34 kg, max coyote: 13.4 kg; Fig. 1). The coyote's success in urban environments appears to be due to their movements between urban and undeveloped areas, and switching between live prey and scavenging (Gehring & Swihart, 2003). Smaller (≤20 kg) carnivore species may be successful as urban dwellers due to release from competition with larger species (‘mesopredator release’, sensu Crooks & Soulé, 1999). Species with the most potential competitors (e.g. generalist diet species) may therefore have the greatest release from competition (Caro & Stoner, 2003) in urban zones.

Diet

Nearly all the well-established urban carnivores are generalists that are able to make use of carrion and human waste food (Fig. 1) (Crooks, 2002). The majority of these species are omnivorous, taking a wide range of diet items, including fruit, small mammals, invertebrates, lizards, and scavenged food (as discussed in the section: ‘What do they eat?’). McKinney (2006) terms these animals ‘edge’ species as they do well in the biodiverse and food-rich gardens and natural fragments that make up much of the urban landscape. Many carnivores that do not succeed in human-dominated landscapes (e.g. bobcats, American badgers, weasels and eastern spotted skunks) are hypercarnivore hunters of live prey or specialists (e.g. American badgers rely on digging out burrow-dwelling small mammals). For example, even when cohabiting with humans in farmland, the eastern spotted skunk relies on commensal rats and mice and takes no anthropogenic food (Crabb, 1941).

The most notable exception to this trend to omnivory is the domestic cat. While felids are adapted to hypercarnivory and can take prey as large as or larger than themselves (Kok & Nel, 2004), domestic cats may be exceptional in that, across multiple studies, they seem to subsist on prey averaging 1.1% their own mean body mass (Pearre & Maass, 1998), which is smaller than predicted based on their body mass (13%, Peters, 1983, 11%, Vézina, 1985) but larger than expected if they were considered specialist ‘small-prey eaters’ (Peters, 1983) or reliant on invertebrates (Vézina, 1985). Domestic cats appear to break the mould of ‘specialised’ felids and, like red foxes, are eclectic feeders that can adapt to local prey availability. We analysed data presented by Pearre & Maass (1998) and found that cats sampled from sites close to human habitation (farms, suburban and urban studies) take significantly smaller prey (23.2 ± 8.3 g; n = 16 studies) than cats in rural areas (72.6 ± 92.1 g, n = 28 studies). These data suggest that cats living close to human habitation modify their diet, which may explain how these hypercarnivores deal so well in anthropogenic environments.

Behavioural and biological flexibility

The ‘ideal’ urban carnivore should be highly adaptable in terms of diet, movement patterns and social behaviour (in the section: ‘How is the ecology of mammal carnivores influenced by urban living?’). However, there are some exceptions to this premise. For example, Herr et al. (2009a) found that stone martens in Luxembourg were almost entirely urban (their territories falling within the extent of the study towns), and their presence suggests that they successfully deal with the challenges of this environment. Their socio-spatial distribution, however, is almost exactly the same as recorded in non-urban habitats, and stone martens do not make much use of anthropogenic food sources (implying both social and dietary inflexibility). While stone martens are well-established urban carnivores, the congeneric pine marten Martes martes avoids human habitation (Baghli et al., 2002; Herr, 2008). This difference appears to be due to pine martens being less omnivorous than stone martens, and while pine martens are diurnal, the crepuscular stone marten is less susceptible to clashes with humans (Herr, 2008; Herr, Schley & Roper, 2009b).

Cardillo et al. (2004) demonstrated how biological features (e.g. geographic range, population density, reproductive rates and dietary requirements) explain 45% of variation in risk of extinction for carnivore species, or 80% when combined with high levels of exposure to human populations. Biological ‘inflexibility’ (small geographic ranges, low population density, low reproductive rates, need for large hunting areas or specific prey) in the face of increasing human populations and urbanization means potential extinction, while ‘flexible’ species (wide geographic range, potential high population density, high reproduction and generalist trophic niche) are more likely to adapt to increasing urbanization.

What are the consequences of carnivore presence in cities?

Although urban carnivores may be valued by large sectors of society (Baker & Harris, 2007) and even encouraged (e.g. through deliberate feeding section: ‘What do they eat?’), these animals can also clash with their human neighbours to a greater or lesser degree through disease transmission to humans and pets, damage to houses and gardens, general nuisance value (e.g. bin-raiding) or direct attack of humans or pets (Baker & Harris, 2007).

Disease

The risk of zoonoses is a significant cause for concern. The public health issues of carnivore presence in cities have therefore been the focus of much research as well as the drive for extensive control measures. For example, the potential transmission of rabies, tuberculosis and parasites are potential dangers for humans, pets or livestock. Rabies control measures have seen significant numbers of carnivores killed (e.g. Tischendorf et al., 1998; Guerra et al., 2003; Bourhy et al., 2005) at substantial economic cost (Curtis & Hadidian, 2010). The greatest fear has been that rabies presence in established urban species is likely to increase the chance of transmission to pets or humans. Parasite transmission is also a significant risk. For example, raccoons carry a roundworm Baylisascaris procyonis, which causes no symptoms in the primary host but can be fatal to intermediate hosts (including humans) through visceral, neural or ocular larva migrans. As raccoons leave faeces in latrines in the open, risk of infection can be high for small children. Roussere et al. (2007) recorded that almost half of California residences surveyed had least one raccoon latrine containing B. procyonis eggs. Similarly, there is a high prevalence of Echinococcus multilocuralis in foxes in Zürich; this might be a source of infection for domestic carnivores and urban inhabitants (Stieger et al., 2002). Carnivores carry many other parasite diseases (see review by Soulsbury et al., 2010), which may have economic importance through transmission to domestic pets in urban environments.

Damage and nuisance

Carnivores may damage houses and gardens due to their diggings and residing in locations that may be problematic (e.g. roof spaces, where their movements are noisy and defecation or urination can cause damage) (e.g. Herr et al., 2010). Stone martens in Luxembourg climb into car engine compartments and, as part of territorial behaviour, destroy cables and rubber components and scent mark them (Herr et al., 2009b). In terms of general nuisance value, bin-raiding is a commonly reported problem with urban carnivores (Harris, 1984; Clark, 1994) (discussed in the section: ‘Refuse’). Digging activities may also cause damage; for example, badger setts can be extensive (e.g. have 80 entrance holes and 360 m of tunnels, Delahay et al., 2009, and references therein), and while badgers in Europe do not often use buildings, their excavations cause significant damage to roads, buildings and waterways (Delahay et al., 2009).

Pet mortality, nuisance and direct attacks

Larger carnivores using urban areas might also increase the chance of direct attacks upon humans and companion animals (e.g. Gehrt & Riley, 2010). Löe & Röskaft (2004) suggested that tiger attacks on humans are more likely when there is less natural prey (a situation typical of urban areas). Also, as some carnivores become used to human presence, they lose their fear, resulting in direct attacks. Non-threatening behaviour by humans and the presence of anthropogenic waste food may have contributed to the death of a geologist in Canada, allegedly due to grey wolves (Geist, 2007). Many people also seem to avoid acting aggressively when they encounter large carnivores in the hope this will prevent or stop an attack; at least for wolves (Geist, 2007) and for pumas (Beier, 1991) shouting and throwing objects is more effective. A recent account of a red fox attack on infant twins in London indicated that even screaming and lunging at the fox was not sufficient to scare it off (Anon, 2010). Rabid carnivores, particularly, act aggressively and this may increase their encounters with humans (Anon, 2008).

Arguably, the coyote may be the most directly dangerous carnivore to humans due to its reasonably large body size (10–16 kg), potential for hybridization with wolves in some part of its range (Curtis et al., 2007; Gehrt & Riley, 2010) and close association with urban areas. Urban coyotes show reduced fear of humans, even biting or acting aggressively towards them (Carrillo et al., 2007; Farrar, 2007; Schmidt & Timm, 2007; Shivik & Fagerstone, 2007). Potential hybridization with wolves may increase the incidence of this type of aggressive interaction.

The future of urban carnivores

As the human population grows and urban areas expand, it is likely that a growing number of animal species will come into contact with anthropogenically altered landscapes; the concomitant reduction in wilderness areas will make this inevitable. The availability of food and shelter resources within these landscapes will also entice species in. Beckmann & Lackey's (2008) study of black bears is a good example: bear numbers in urban areas of Nevada have increased more than three times the recorded historical baseline, and there has been a 10-fold increase in complaints about urban bears. The bears become fatter on anthropogenic food and breed younger, but mortality is so high that urban areas are sinks, particularly as urban black bears do not appear to be able to recolonize undeveloped areas. Consequently, bears are becoming concentrated around urban areas and rare in undeveloped areas. The pattern of increasing numbers of carnivore species present in towns and cities over recent decades (e.g. coyotes in the US: Gehrt, 2011; bears in the US: Beckmann & Lackey, 2008; and Europe: Quammen, 2003) may mark the future for the coexistence of carnivores with man. Understanding the biology of these animals is therefore going to become more important if we are to make the best of these unfolding circumstances towards the conservation of the carnivores as well as mitigating their potential impacts upon our lives.

We predict that on the outskirts of cities, more large species are likely to make use of urban resources (bears, wolves, possibly cougars and bobcats in America and Europe, and hyaenids in Africa and Asia). This may, however, be short-lived as cities become more intensively urban, the urban/wildland interface of suburbs becomes more blurred, and the extent of undeveloped land diminishes. Within cities themselves, if sufficient patches of vegetation remain, carnivores may continue to use urban habitats, as long as they are not outcompeted by established urban exploiters (e.g. cats, dogs) or destroyed through control measures due to disease concerns.

Although we predict the continuing increase in incidence of carnivores in urban areas, other authors have suggested that overall carnivore diversity is likely to decrease in the future due to human action. For example, although genets Genetta genetta are common and widely distributed, Cardillo et al. (2004) identify genets (along with several other viverrids) as likely to become endangered by 2030 due to their overlap in distribution with areas of high human population density. However, genets have been observed living alongside humans in urban habitats in Africa (PWB pers. obs.) and Europe (Larivière & Calzada, 2001 and references therein). The perseverance of carnivores such as genets in significantly anthropogenically disturbed habitats is likely to rely on physiological and behavioural adaptability of these charismatic animals.

Acknowledgements

Our thanks to Stephen Harris for his comments on red foxes in the UK and three referees for helpful suggestions. Thanks also to Rob Morley for sharing his observations.