Biological Invasions

, Volume 13, Issue 3, pp 581–603

The diet of feral cats on islands: a review and a call for more studies

Authors

    • Mediterranean Institute for Ecology and Palaeoecology (UMR CNRS/IRD)Aix-Marseille University (Université P. Cezanne)
    • Ecology Systematic and Evolution, UMR CNRS 8079, Paris Sud University
  • F. M. Medina
    • Consejería de Medio AmbienteCabildo Insular de La Palma
  • E. Vidal
    • Mediterranean Institute for Ecology and Palaeoecology (UMR CNRS/IRD)Aix-Marseille University (Université P. Cezanne)
  • M. Nogales
    • Island Ecology and Evolution Research Group (IPNA-CSIC)
  • B. Tershy
    • Ecology & Evolutionary Biology DepartmentUniversity of California, Santa Cruz
  • E. Zavaleta
    • Environmental Studies DepartmentUniversity of California
  • C. J. Donlan
    • Advanced Conservation Strategies
    • Department of Ecology and Evolutionary BiologyCornell University
  • B. Keitt
    • Island Conservation, Long Marine LaboratoryUniversity of California
  • M. Le Corre
    • Lab ECOMAR, Université de La Réunion
  • S. V. Horwath
    • Island Conservation, Long Marine LaboratoryUniversity of California
Original Paper

DOI: 10.1007/s10530-010-9851-3

Cite this article as:
Bonnaud, E., Medina, F.M., Vidal, E. et al. Biol Invasions (2011) 13: 581. doi:10.1007/s10530-010-9851-3

Abstract

Cats are among the most successful and damaging invaders on islands and a significant driver of extinction and endangerment. Better understanding of their ecology can improve effective management actions such as eradication. We reviewed 72 studies of insular feral cat diet from 40 islands worldwide. Cats fed on a wide range of species from large birds and medium sized mammals to small insects with at least 248 species consumed (27 mammals, 113 birds, 34 reptiles, 3 amphibians, 2 fish and 69 invertebrates). Three mammals, 29 birds and 3 reptiles recorded in the diet of cats are listed as threatened by the IUCN. However, a few species of introduced mammals were the most frequent prey, and on almost all islands mammals and birds contributed most of the daily food intake. Latitude was positively correlated with the predation of rabbits and negatively with the predation of reptiles and invertebrates. Distance from landmass was positively correlated with predation on birds and negatively correlated with the predation of reptiles. The broad range of taxa consumed by feral cats on islands suggests that they have the potential to impact almost any native species, even the smallest ones under several grams, that lack behavioral, morphological or life history adaptations to mammalian predators. Insular feral cat’s reliance on introduced mammals, which evolved with cat predation, suggests that on many islands, populations of native species have already been reduced.

Keywords

Domestic catFelis catusFeeding behaviourFood webIsland ecosystemConservation

Introduction

Invasive species, particularly mammals, have caused population declines, local extinctions and global extinctions of many island endemic species making them leading contributors to biodiversity loss (Diamond 1989; Fritts and Rodda 1998; Balmford 1996; Stattersfield and Capper 2000; Atkinson 2001; Dowding and Murphy 2001; Courchamp et al. 2003; Long 2003; Aguirre-Muñoz et al. 2008). Cats Felis catus (Driscoll et al. 2007) were first introduced to islands in the Mediterranean in 9000 BP (Vigne et al. 2004; Driscoll et al. 2007), and have since been introduced to islands worldwide from the sub Antarctic to the sub Arctic including the most arid and mesic islands (Courchamp et al. 2003). Cats are successful invaders on islands because they can survive without access to fresh water, have high fecundity, are highly adaptable to new environments, and have generalist predatory behaviours that allow them to feed on the most available prey species (Van Aarde 1986; Konecny 1987; Say et al. 2002; Fitzgerald 1988; Fitzgerald and Turner 2000; Pearre and Maass 1998; Stattersfield and Capper 2000).

Cats are one of the most widespread invasive predator on islands causing strong negative impacts on native wildlife (Fitzgerald 1988; Macdonald and Thom 2001) and are responsible, at least in part, for 8% of global bird, mammal and reptile extinctions and a significant threat to almost 10% of critically endangered birds, mammals and reptiles (Medina et al. In Review). Eradicating cats from islands can protect native species from the threat of extinction (Keitt and Tershy 2003; Nogales et al. 2004; Ratcliffe et al. 2009) and research on the ecology of insular feral cats can improve the efficacy and prioritization of cat eradications, particularly when (i) complex biotic interactions occurred on multi-invaded islands and (ii) highly vulnerable native prey are present (Fitzgerald 1988; Fitzgerald and Turner 2000; Macdonald and Thom 2001).

Here, we reviewed all available published studies of insular feral cat diet to determine: (1) what general patterns define the diet of feral cats on islands, (2) how does biogeography influence the diet of feral cats, (3) what are the taxonomic and size diversity of cat prey species, and (4) how many threatened species were found in the diet of feral cats on islands?

Materials and methods

Data compilation

We compiled data from published literature using electronic databases (Webspire, Web of Knowledge, Ovid SP, Inist, Blackwell Publishing, Science Direct) using the key words: cat(s), feral cat(s), Felis catus, island(s), diet, predation, and native species. Data compilation was conducted through July 2008. References cited in the resulting papers were examined for additional sources. The diet studies we analyzed used three types of samples: scats, gut contents and stomach contents. Some studies utilized more than one of these sample types. For the analyses in this study we did not differentiate between the types of samples because several studies found only minor differences between sample types (Lozano et al. 2006; Souza and Bager 2008), but see Jones et al. (2003). When studies used more than one type of sample, we analyzed each sample type as a separate study (e.g. Berruti 1986; Fitzgerald et al. 1991; Martínez-Gómez and Jacobsen 2004). Australia was excluded from our review because it is traditionally defined as a continent and cat diet in Australia has been reviewed elsewhere (Dickman 1996; Burbidge and Morris 2002). Lastly, we selected only data from feral cats, feeding on wild prey (Artois et al. 2002) and only incorporated data from papers giving detailed information about cat diet.

The following data were extracted from each paper: geography (latitude, distance from landmass, elevation and area) of each study island, study date, type of sample analyzed (guts, stomachs and scats), prey sample size and characteristics of sampling sessions (i.e. number of sessions, periods of the year).

Global patterns of feral cat diet

To describe global patterns of feral cat diet, when one study described cat diet on the same island but with two different sample types (i.e. scats and stomachs), we eliminated the less representative, i.e. the one with the smallest sample size, to avoid redundancy. We calculated the percentage of the 66 studies where the main prey categories (i.e. mammals, birds, reptiles, fishes, invertebrates) were represented by at least one species. Then we constructed a global diet pattern for the main prey categories using data from papers that gave feral cat diet in frequencies of occurrences (% FO) for all taxa (65 studies).

Biogeographical patterns of feral cat diet

For the following analyses, we only selected the most representative (the largest sample size and the most number of seasons covered) study per island, avoiding those based on excessively small sample size and which described diet with insufficient accuracy (less than 30 samples). We tested for correlations between the frequency of occurrence (%FO) of the six main prey categories (rabbits, rats, mice, birds, reptiles and invertebrates such as dependent variables) and five biogeographical variables (latitude, distance from nearest continental landmass, area, elevation, area*elevation). The dependent variables were arcsine transformed and the quantitative descriptive variables were log transformed for normality (Sokal and Rohlf 1995). Variables not normally distributed after transformation were analyzed using Spearman rank tests.

We also performed a PCA analysis, using the %FO as dependent variable, to obtain a spatial ordination of the different main prey categories and islands included in this review. For this, we applied a varimax rotation in order to identify the contributions of the different prey items. In all analyses, all data concerning %FO were arcsin transformed to better meet the assumption of normality (Tabachnick and Fidell 1996; Fowler et al. 1998).

Diet diversity and species threatened by feral cats

Levin’s index (trophic niche breadth) was applied to calculate diet diversity. This index reaches maximum values when all diet items are consumed in equal proportions, the minimum value indicating more specialized diets. The normalized Levin’s index (hereafter Levin’s-std) was used to avoid variation caused by different samples sizes (Krebs 1989). We also tested for correlations between the Levin’s-std index and the five biogeographic variables mentioned, and we performed Spearman correlation tests when variables were not normally distributed, even after transformation. Then, mean weights (from the literature) of vertebrates species preyed on by feral cats were listed. This information was not available for invertebrates.

Lastly, we compared the prey species from our review to those listed in a review of feral cat impacts on insular endangered species (Medina et al. In Review) with an emphasis on prey species that were listed by the IUCN 2008 Red List of Threatened Species in one of the five most threatened categories of the (Extinct, Extinct in the Wild, Critically Endangered, Endangered and Vulnerable).

Results

Data compilation and global patterns of feral cat diet

We reviewed 58 papers with data on the diet of feral cats on islands. Those papers described results from 72 diet studies (Appendix 1) conducted on 40 different islands (Fig. 1). Studies of feral cat diets on islands were not evenly distributed with relatively few studies in the tropics and no studies in important biogeographic regions including the Caribbean, Indonesia, Japan and French Polynesia (Fig. 1). Most of the studies were of diet found in scat and covered only one or two seasons with a single sample session (18/56) (Table 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs10530-010-9851-3/MediaObjects/10530_2010_9851_Fig1_HTML.gif
Fig. 1

Location of the diet studies of feral cats carried out on islands worldwide

Table 1

Sampling description of the diet studies of feral cats: type of samples, sample size; seasons covered by sampling sessions and their number

Samplings

Number of studies

Sample type

 Scats

46

 Guts

13

 Stomachs

13

Sample size

 0-9

3

 10-49

24

 50-99

12

 100-499

27

 >500

6

Studies covered

 1 season

18

 2 seasons

17

 3 seasons

6

 All year

24

 Not mentioned

7

Number of samplings

 1

19

 2

12

 3

4

 Several (collected occasionally)

19

 Several (planned)

8

 Not mentioned

10

Mammal and bird species were found in nearly all diet studies we reviewed, followed by reptiles and invertebrates (Fig. 2a). For the subset of studies that recorded frequency of occurrence, mammals were most frequent (Fig. 2b).
https://static-content.springer.com/image/art%3A10.1007%2Fs10530-010-9851-3/MediaObjects/10530_2010_9851_Fig2_HTML.gif
Fig. 2

a The percentage of diet studies in which the main prey categories were found. The number of studies is given above each bar. b The frequencies of occurrences of main prey categories in each diet study. The number of studies with frequencies of occurrences is given above each bar. The error bars represent standard deviation

Biogeographical patterns of feral cat diet

Latitude was positively correlated with %FO of rabbits and negatively with %FO of reptiles and invertebrates in feral cat diet. Distance from landmass was positively correlated with %FO of birds and negatively with %FO of reptiles (Table 2).
Table 2

Spearman Rank tests describing the frequencies of occurrence of rats, rabbits, mice, birds, reptiles and invertebrates in the diet of feral cats related to biogeographic variables

 

Number of samples

Statistical values

Explanatory variables

Latitude

Distance from landmass

Area

Elevation

Area * Elevation

Dependent variables

FO of rats

34

R

−0.2284

0.1501

0.0569

0.0598

0.0761

P

ns

ns

ns

ns

ns

FO of rabbits

34

R

0.4104

−0.3229

0.1781

0.1939

0.2285

P

*

ns

ns

ns

ns

FO of mice

34

R

0.0920

−0.1947

−0.0996

0.1878

0.0386

P

ns

ns

ns

ns

ns

FO of birds

34

R

−0.0995

0.4626

−0.0891

−0.0717

−0.0836

P

ns

**

ns

ns

ns

FO of reptiles

34

R

−0.4532

−0.4216

0.2983

0.1071

0.1866

P

**

*

ns

ns

ns

FO of invertebrates

34

R

−0.4164

0.1116

−0.0381

0.1652

0.0585

P

*

ns

ns

ns

ns

P < 0.05, ** P < 0.01, *** P < 0.001

The first three components of the PCA analysis explained 75.6% (PCI = 35.9%, PCII = 21.2%, and PCIII = 18.5%) of the whole variance. We represented these three main PCA components and the islands in order to understand the spatial ordination (Fig. 3a, b). The first two components separated those islands where rabbits were important in cat diet (component I with negative values), grouping sub-Antarctic and subtropical islands (e.g. Canaries) (Fig. 3a) and where rats were important (component I, with positive values) like Galapagos, Port-Cros island, Japanese islands and some islands north of New Zealand. The component II segregated those islands where the consumption of birds (with negative values) was an important component of the diet (e.g. Herekopare, Marion, or Jarvis), than the rest of prey items. With respect to the components I and III (Fig. 3b), the first separated well those islands with diet characterized by the presence of rabbits (negative values) from those in which rats were the main component (positive values) was relevant (e.g. tropical and subtropical archipelagos such as Aldabra, Galapagos or Canaries). Table 3 showed the factor loadings of the main components of the diet. The factor I indicated that while rabbits are highly preyed on by cats, birds show a scarce presence in the diet. The factor 2 and 3 were related to a negative relationship between the consumption of reptiles vs. mice and birds, respectively.
https://static-content.springer.com/image/art%3A10.1007%2Fs10530-010-9851-3/MediaObjects/10530_2010_9851_Fig3_HTML.gif
Fig. 3

Plot of the location of the main %FO of the main prey consumed by feral cats on islands. Key to island legend: ALD Aldabra, ALE Alegranza, AMS Amsterdam, CAB Cabrera, CHR Christmas, DAS Dassen, ELH El Hierro, FUE Fuerteventura, GRA Gran Canaria, DOG Great Dog, HAM Haha-jima, HER Herekopare, IRI Iriomote, ISA Isabela, JAR Jarvis, JUA Juan de Nova, KER Kerguelen, LAG La Gomera, LAP La Palma, LIT Little Barrier, MAC Macquarie, MAR Marion, HAW Hawaii (the big island), NEW New Island, NZN New Zealand North, NZS New Zealand South, POR Port-Cros, RAO Raoul, REU Réunion, SCL San Clemente, STC Santa Cruz, SOC Socorro, STE Stewart, TFE Tenerife

Table 3

Principal component analysis of the correlation matrix between percentages of %FO (arc-sin transformed) for the main prey items included in the diet of feral cats (Felis catus) on the different islands studied

Main prey items

Factor loadings

 

F1

F2

F3

Rabbits

−0.70

0.49

0.06

Rats

0.60

−0.48

0.34

Mice

−0.39

−0.63

0.43

Birds

0.64

−0.03

−0.66

Reptiles

0.42

0.60

0.56

Invertebrates

0.76

0.20

0.25

Factor loadings >0.5 and <-0.5 are indicated in boldface type

Diet diversity and species threatened by feral cats

The studies we reviewed identified 248 prey species in the diet of insular feral cats: 27 mammals (12 natives and 17 non-natives), 113 birds (38% of which were seabirds), 34 reptiles, 3 amphibians, 2 fishes and 69 invertebrates, mostly insects (Appendix 2). The Levin’s-std index was only calculated for 21 studies where prey number was given (Table 4). This index was positively correlated with distance from landmass, elevation and area*elevation (Table 5). The size of feral cat prey ranged from several grams to more than 2 kg; lagomorphs and seabirds (mainly Procellariidae) constituted the largest species, while small rodents (mice) and reptiles (lizards and geckos) were the smallest (Fig. 4).
Table 4

Diet studies of feral cats expressed in prey number. In this study sample, Levin’s index was calculated

Studies

Islands

Latitude

Distance from landmass (km)a

Area (km2)

Elevation (m)

Prey number

Number of prey categories

Levin’s- std

Jones (1977)

Macquarie

54.3

1990.0

128.0

240.0

60

7

0.387

Karl and Best (1982)

Stewart

47.3

1792.0

1746.0

979.0

737

5

0.451

Apps (1983) (/year)

Dassen

33.3

8.5

2.2

19.0

138

6

0.128

Fitzgerald and Veitch (1985)

Herekopare

46.5

1792.0

0.3

90.0

90

9

0.157

Van Rensburg (1985)

Marion

46.5

1725.0

316.0

1230.0

294

10

0.208

Santana et al. (1986)

Gran Canaria

27.6

196.5

1560.0

1949.0

332

23

0.246

Fitzgerald et al. (1991)

Raoul

29.2

2784.5

31.8

516.0

103

9

0.243

Nogales et al. (1992)

Alegranza

29.4

168.5

10.2

289.0

199

5

0.312

Medina and Nogales (1993)

Tenerife

28.2

288.0

2034.0

3717.0

477

10

0.294

Nogales and Medina (1996)

La Gomera

28.1

335.0

372.0

1487.0

57

8

0.440

García-Márquez et al. (1999)

El Hierro

27.5

383.0

278.0

1501.0

326

8

0.421

Smucker et al. (2000)

HawaT

19.5

3705.0

117.0

4205.0

290

3

0.651

Espinosa-Gayosso et al. (2006)

San Jose

28.0

20.0

194.0

133.0

159

5

0.088

Medina et al. (2006)

La Palma

28.4

416.5

706.0

2400.0

967

14

0.282

Port-Cros

43.0

8.5

6.4

119.0

667

14

0.115

Bonnaud et al. (2007)

Peck et al. (2008)

Juan de Nova

17.0

296.5

4.4

10.0

228

12

0.316

Phillips et al. (2007)

San Clemente

32.6

100.0

56.8

599.0

14246

13

0.185

Matias and Catry (2008)

New Island

51.4

401.5

1.97

100

842

14

0.335

Medina et al. (2008)

Fuerteventura

28.2

96.5

1660

807

614

27

0.210

Kawakami and Mashiko (2008)

Haha-jima

26.4

142.09

1968.0

20.2

210

16

0.230

Faulquier et al. (2009)

La Reunion

21.1

677.0

2512.0

3069.0

336

7

0.519

aAustralia was considered as the landmass for New Zealand islands and its surrounding islands

Table 5

Results of the Spearman Rank test describing the Levin’s index of the diet of feral cats related to biogeographical variables

  

Number of samples

Statistical values

Explanatory variables

Latitude

Distance from landmass

Area

Elevation

Area * Elevation

Dependent variables

Levin’s-std

21

R

−0.2532

0.6186

0.3013

0.4961

0.4390

P

ns

**

ns

*

*

P < 0.05, ** P < 0.01, *** P < 0.001

https://static-content.springer.com/image/art%3A10.1007%2Fs10530-010-9851-3/MediaObjects/10530_2010_9851_Fig4_HTML.gif
Fig. 4

The weight pattern of vertebrates preyed on by feral cats on islands: a mammals, b birds, c reptiles

We identified 36 IUCN-listed threatened species (3 mammals, 29 birds, 3 reptiles and 1 amphibian (not represented in the figure) that were preyed upon by cats (Fig. 5).
https://static-content.springer.com/image/art%3A10.1007%2Fs10530-010-9851-3/MediaObjects/10530_2010_9851_Fig5_HTML.gif
Fig. 5

A comparison between endangered species listed in diet studies (this review) with those listed in the review of the impact of feral cats on globally threatened insular species. Legend: EX Extinct, EW Extinct in the wild, CR Critically endangered, EN Endangered, VU Vulnerable and NT Near threatened

Discussion

Data compilation

We compiled 58 scientific papers, describing 72 studies on the diet of feral cats, in different archipelagos worldwide. Few data were available on feral cat diet from islands in the Caribbean, Indonesia or French Polynesia, three biodiversity hotspots with large numbers of threatened endemic species (Brooks and Smith 2001; Long 2003). Surprisingly, there were also only a few studies from the Mediterranean Basin, despite the presence of cats on some islands since 9000 BP (Vigne et al. 2004), and on Japanese islands. Studies in under-sampled regions will undoubtedly add new species of prey items to the nearly 250 reported here.

Scats were the most frequent type of sample analyzed (Mukherjee et al. 2004). A minimum size of ~100 scats is considered necessary to identify principal prey remains occurring in 5% of scats and ~100 samples is also required when comparing diets to distinguish moderate effects over time or between areas (Trites and Joy 2005). Less than half of the 72 studies were based on more than 100 samples, thus most diet studies reviewed here probably missed some prey items (Table 1a). Moreover, most diet studies were conducted with fewer than two sampling sessions and during just part of the year (Table 1b). This effort is likely to miss rare species, which suggests that the 36 of IUCN listed threatened species we found may be a conservative estimate.

Global patterns of feral cat diet

Feral cats feed on a wide range of prey from large birds to small insects (Langham 1990; Tidemann et al. 1994; Peck et al. 2008), yet mammals and birds were present in practically all studies. Introduced mammals (mainly rabbits, rats and mice) were the main prey and birds, reptiles and invertebrates were often consumed as secondary prey. Birds were important on islands that harbour important colonies of seabirds and, in these cases, they constitute the main source of biomass in the diet of feral cats (Jones and Barmuta 1998; Bonnaud et al. 2007; Nogales and Medina 2009, Peck et al. 2008, Faulquier et al. 2009). This observation is enhanced by cats eating bird eggs, which are undetectable in the gut, but amount to a significant loss of birdlife in certain areas (Rauzon 1983). Reptiles and invertebrates were not systematically recorded in feral cat diets but they were frequently eaten when present, particularly on tropical and subtropical islands (Konecny 1983; Nogales et al. 1990, Seabrook 1990). However, due to their small size, reptiles rarely represented a significant percentage of prey biomass (Nogales et al. 1988; Medina and Nogales 1993; Casañas-Acosta et al. 1999) but cat predation can have a high impact on them.

Large differences in the diet of cats between islands indicated considerable trophic plasticity, with feral cats eating the most available items on each island (Arnaud et al. 1993; Martínez-Gómez and Jacobsen 2004). This same pattern was also seen between habitats and seasons within an island (Clevenger 1995; Nogales and Medina 1996; Medina et al. 2006).

This ability to seasonally adjust diet preference on an island has significant ramifications for native species, especially seabirds. Both rabbits and rats were regular items in the diet of insular feral cats (Dilks 1979; Fitzgerald et al. 1991; Nogales et al. 1992; Fitzgerald and Turner 2000; Medina et al. 2006). However, during months when breeding seabirds are present on an island they can become the preferred prey item (Rauzon 1983, Bloomer and Bester 1990, Keitt et al. 2002, Peck et al. 2008). The presence of the alternative prey items, rabbits or rats, means the feral cat population can sustain itself at a larger population throughout the year and thus cat impacts on native species are likely to be exacerbated via hyperpredation (Taylor 1979; Jones 2002; Keitt et al. 2002; Oro et al. 2004). This can also facilitate the complete eradication of native seabirds because the alternative prey source allows the feral cat population to sustain itself even as the seabird resource moves to a zero population (Courchamp et al. 1999; Roemer et al. 2002). Conversely, feral cats can eradicate populations of rodents when seabirds are absent during severe climatic conditions (Rauzon et al. In Review 2011).

Biogeographical patterns of feral cat diet

The increase in rabbit consumption with latitude can be explained by their presence on most sub-Antarctic islands where this lagomorph has become a staple prey inside a low prey diversity environment. The increase of birds in the diet of cats on more geographically isolated islands probably reflected the relative absence of native mammals on these insular areas (Vitousek et al. 1995; Whittaker and Fernández-Palacios 2007). These results reinforced the opportunistic feeding behavior of feral cats. As expected by the global distribution of reptiles, higher latitudes support fewer reptiles (Schall and Pianka 1978), and pointed out by Fitzgerald (1988), the presence of reptiles in insular feral cat diet decreased with increasing latitude. The same trend was recorded for invertebrates suggesting that cats eat a wide range of prey as soon as they are available.

The PCA analysis showed that on most islands cats had primary prey like rabbits, rats, mice or birds and this preference depended on the availability and phenology of prey in the different islands. Some general trends indicated that (i) rats were the main prey especially where rabbits were absent, (ii) rabbit predation could mitigate bird predation and (iii) reptiles and invertebrate consumption was higher in tropical and subtropical islands probably due to the higher abundance of these prey.

Diet diversity and species threatened by feral cats

The 179 different vertebrate prey species we found in the diets of feral cats on islands confirmed the wide niche breadth of this efficient generalist. Introduced rodents and lagomorphs, large and small birds and small reptiles appeared more regularly in island cat diet. Data on invertebrate prey species diversity was much less consistently reported, however, cats preyed upon a large range of invertebrate prey taxa from small hymenoptera to large crustaceans. Insects were the invertebrates most frequently preyed on by feral cats (Fitzgerald and Karl 1979; Kirkpatrick and Rauzon 1986, Konecny 1987; Medina and García 2007) and most insect orders were recorded, with Orthoptera, Coleoptera and Lepidoptera most commonly recorded.

As predicted by island biogeography theory (Vitousek et al. 1995; Whittaker and Fernández-Palacios 2007), diet diversity increased with elevation and area*elevation. Surprisingly, however, diet diversity also increased with distance from nearest landmass. Perhaps this is due to confounding variables of islands size and distance from shore, because (i) more oceanic islands were more highly invaded, or (ii) more evolutionarily naïve prey species are present on oceanic islands, or (iii) more eradications were done and save island species.

Feral cats have contributed to over 8% of all bird, mammal and reptile extinctions and to the declines of almost 10% of critically endangered birds, mammals and reptiles (Medina et al. In Review). However, we found relatively few globally threatened vertebrates in the diet of feral cats (Fig. 4). Globally threatened species were underrepresented in our study because there was no study of cat diet on islands within their range and because most studies we reviewed were not designed to detect rare prey species (Blackburn et al. 2004). Threats to birds and other native species when they are the primary prey are obvious; however, there is no clear relationship between frequency of prey in the diet and impact of the predator on that prey species. Moreover, the effects of predation are usually inversely dependent on prey density when native species are not the primary prey (Sinclair et al. 1998; Cuthbert 2002). Although none of the studies in our analysis accounted for time since cat introduction, it is likely that the most vulnerable native prey were reduced in number or extirpated long before diet data were collected.

Conclusions

This study confirms that feral cats on islands are extremely opportunistic, generalist predators capable of consuming most available vertebrate species. This combined with their ability to survive without drinking water and their large thermal tolerance (Bradshaw 1992) explains their tremendous geographic range. Our metadata analysis also suggests that cats can negatively impact a large percentage of the native vertebrates on any island. Future studies of cat diet on islands should be, (1) conducted so as to sample during the presence of “temporary” prey such as seabirds, (2), represented by a sufficient number of samples (> 100), (3) published by indicating the results at least for the main prey categories not only in frequency of occurrence but also in prey frequencies, biomass and index of relative abundance, and (4) preferentially dedicated to under-sampled and under-studied insular regions (e.g. Caribbean, Indonesia, Japan and French Polynesia).

Acknowledgments

This contribution is dedicated to all persons who have provided information on the diet of feral cat in all islands worldwide. We are very grateful to F. Torre, P. Campagne and K. Bourgeois for helping with statistical analyses and special thanks for S. Fadda. We would like to thanks M. Fitzgerald and two anonymous referees for their helpful comments. This work has received support from the European Union and the DIREN PACA via a Life Nature project (ref. LIFE03NAT/F000105), the French National Research Agency (ALIENS project), the MEDAD (Ecotropic programme), the project CGL-2004-0161 BOS co-financed by the Spanish Ministry of Science and Education and the European Union, and a CR PACA PhD fellowship to EB.

Copyright information

© Springer Science+Business Media B.V. 2010