Home range

Home range influences how widely the cats travel which in turn influences their contacts with other animals as well as humans. Traditionally, home range was considered to be the total space used by an animal (Powell & Mitchel 2012).The description of the use of space by an animal has been expanded to include how much time the animal spent in a given space. A more complex view of the home range has been proposed to be more constructive (Powell & Mitchel 2012; Recio & Seddon 2013). This view supposes that the behavior of having a home range is the result of the interplay between the environment and the animals’ understanding of that environment (Recio & Seddon 2013) in a cognitive map. The authors state that mammals plan their movements and have cognitive maps relative to where in the environment they are, what sorts of selection pressures exist, nutritional status, and other factors. Discussions of home range are complicated by the many different methods used to estimate them and the validity and utility of each (Powell & Mitchel 2012).

A clear understanding of where the cat spends time, why the cat is there, and what resources are available and used within the cat’s range can inform plans to decrease the numbers of free-roaming cats. Identifying patterns of activity by time and place can help with optimizing trapping or monitoring protocols. Knowing sources of food such as dumpsters could allow for managing those sources in such a way as to decrease the numbers of cats in the area. This understanding of the use of space also has implications for studying and understanding predation as well as cat’s social interactions (Recio & Seddon 2013).

One recent study evaluated both the use of space and home range but also the distribution of food sources (in this New Zealand study the food source was rabbits) (Recio & Seddon 2013). The authors found that home range size was a function of both the cats’ sex (males had larger home ranges) and season. In this setting, with only naturally distributed prey, cats shared certain high use locations as well as the rabbit hunting areas but showed no dramatic social interactions. How often cats spent time near each other also varied by season and by sex.

There are many published studies on cats’ home ranges (Liberg et al. 2000). In general, the higher the density of cats, the smaller the home range size of the individual cats. Female home ranges have been reported to be 0.27 ha (about 52 m²) in highly urbanized Jerusalem up to 170 ha (about 1300 m²) in the Australian bush where cats must find and catch naturally distributed prey. Male home ranges are generally about three times larger than females in similar settings. Male home ranges tend to vary by season and by size of the males. During the non-mating season, food is likely the most important force behind range size. During the mating season, male ranges overlap extensively and are larger than non-mating season ranges. Adult male cats do tend to disperse, and assuming that food and shelter are available, they will be limited in their density by the availability of females. There can be considerable variation even in the same study for home ranges (Horn et al. 2011; Wierzbowska et al. 2012). Owned cats have smaller ranges centered on their house, regardless of their sexes, than unowned or unfed cats (Schmidt et al. 2007; Horn et al. 2011).

Density and social interactions

Generally, cat density is thought to be dependent on resource availability, primarily food and shelter (Liberg et al. 2000). Very high densities (>100 cats/km²) were found in urban areas with great availability of garbage or pet food. Cats tended to have intermediate densities (2–100 cats/km²) where they were fed on farms or where there were natural clusters of prey species. Lowest densities (<5 cats/km²) were found in very rural populations where they had to hunt dispersed prey species such as rabbits or rodents.

Group living is possible when there are plentiful and consistent food sources including garbage or prey attracted to human refuse. When female cats are living in groups, their home ranges have very little overlap with other female groups (Macdonald et al. 2000). Female cat lineages appear to be the primary structure for naturally forming cat colonies. Large colonies may have multiple female families. Females do assist each other during birth by cleaning the new born kittens as well as group-mothering kittens by nursing and caring for other queen’s litters (Crowell-Davis 2005). This has survival value since the kittens will spend less time alone and have additional nursing and grooming sources. At higher densities of female cats, large numbers of males can aggregate, exacerbating nuisance and noise issues.

Since cat density may be very high in some locations, cats must have behaviors that allow them to adapt to living together at different densities (Bradshaw et al. 2012). The interactions of cats living in groups are not random; there are specific social interactions and favoritism among the cats. Age, sex, and blood relationships are important factors that influence how and why cats interacted with each other. Cats do appear to choose to live in some groups, they form attachments to specific members of the group, and they show affiliative behaviors. Therefore, the older and sometimes still popular view of cats as solitary and asocial is only true in some situations (Crowell-Davis et al. 2004). Cats who are preferred associates may spend quite a lot of time together during the day and this could have implications for trapping cats. Cats have been shown also to rest together, even in very hot, humid weather when close proximity to keep warm would not explain their behavior.

Social rolling, rubbing, and grooming were originally considered to be sexually linked behaviors (Bradshaw et al. 2012). However, rolling may be used during social nonsexual interactions of cats (and humans). Rubbing and grooming have also been broadened to a greeting display. Rubbing against another cat has been reported to be more common for feral cats after reuniting following hunting (Crowell-Davis 2005). Tail up occurs in all species of felids during urine spraying. However, domestic cats, including free-roaming and unsocialized cats, exhibit tail up in conjunction with affiliative behaviors especially social rubbing and during walking around (Cafazzo & Natoli 2009). Tail up has been hypothesized to be an affiliative behavior in cats associated with domestication and group living (Bradshaw et al. 2012; Brown & Bradshaw 2013).

Territoriality may be defined as actively defending the home range against other cats (at least of the same sex) so that there are not overlapping ranges (Liberg et al. 2000). Among the lay audience, territoriality has been considered to be preventing other cats from joining their group. In both contexts, there appears to be some variability. One explanation is that a large enough concentrated and stable food source leads to adult female cats living in groups and to their defending that resource from nonrelated cats (Macdonald et al. 2000; Crowell-Davis 2005). However, there are clearly data that support the influx of new cats into locations with food sources (Castillo & Clarke 2003). And many pet cats grow to be very comfortable with new housemate cats. Whether the lack of relatedness of the original cats decreases their defense of territory or if the plentiful food supply makes defending territory energetically less worthwhile is unknown. Some reports indicated that males are more likely to be accepted by females than other, nonrelated females (Liberg et al. 2000). This makes sense in that a new female could be adding a whole new lineage of offspring, while a male would only be adding himself.

The presence and type of social hierarchies that groups of cats form have been debated (Crowell-Davis 2005). Dominance may be defined as priority access to resources including food (Bonanni et al. 2007) or as deferring to another cat based on past interactions (Crowell-Davis 2005). However, there are conflicting views about the nature and extent of a hierarchical social structure based on dominant or subordinate group living cats. Hierarchies in larger groups of cats appear to be only partially linear (Crowell-Davis 2005). Cats signal deference and avoid confrontation in many ways as far as we can determine: by walking around other cats, waiting for another cat to pass or time sharing (different cats use a highly desirable location during different times), retreating, and avoiding eye contact. These interactions are often very subtle and result in little physical conflict.

Higher ranking cats based on success in aggressive encounters may control access to food but not always. An isolated and undisturbed colony of 10–17 cats in Rome was studied for access to food (Bonanni et al. 2007). Each cat was individually identifiable by coat color and hair length. Dyadic interactions based on aggression were examined to determine who was dominant. Within 1 m of the food, there was a linear hierarchy topped by a female, and further, females were dominant over all of the males at some point. Aggressive behavior included threats (striking, biting, threatening body postures, pointing staring, and baring of canines), chasing, vocal duels, and real duels. Submissive behavior included crouching with ears flattened, avoiding, retreating, and fleeing. Away from food, male cats tended to be dominant. While juveniles less than 12 months of age were at the bottom of the aggression hierarchy, at 4–6 months of age they were often the first to eat. The authors suggest that there was social tolerance of these juveniles, possibly due to fewer total numbers of cats. They also propose that female cats value food more than males due to reproductive stresses even during less reproductively active the time of year they are less active reproductively, possibly because the hierarchy, once established is maintained. The social tolerance to kittens and of males to females around food may be a product of domestication and population density and the greater sociability and behavioral flexibility required. Alternatively, this tolerance could be due to the wide variety of ecological conditions in which cats live.

Another explanation for the conflicting views of cat social structure is that groups of cats are likely quite heterogeneous in their sources: one mother and her offspring, groups of mothers (related or not) and their offspring, groups of females with well-known and well-liked males, intact cats brought together due to food sources who have found themselves in the area, altered cats from various sources, and mixtures of all of these. It seems plausible that some of the lack of clarity and the reported differences in cats’ behavior with each other is due to these differences in group membership. In addition, there are clearly differences in individual cats and their interest and skill in interacting with other cats (Mendl & Harcourt 2000). I believe that a detailed understanding of the social hierarchy or lack thereof for community cats relative to food and shelter is not critical since cats are clearly able to survive and often thrive in whatever group settings they inhabit.

The significance of dominance may be more important to reproductive interactions. First, there are some behavioral characteristics such as boldness and friendliness to humans that are paternal in origin and could influence the ability to socialize cats to humans. Second, if there is a hierarchy or some method to determine who successfully impregnates the female cats, this has implications for population control strategies. Definitions of “dominance” among male cats relative to breeding and access to females are varied. In some cases, dominant males are just more successful at breeding than subordinates (Liberg et al. 2000). Based on a few studies, these dominant males are typically the larger, heavier, and mature cats.

In describing “dominance,” Natoli and De Vito (1991) defined it as winning the encounter with another male. However, only threats given, threats received, and cheek rubbing were associated with being “dominant.” Location, urine spraying, vocalization, leaving, and fleeing were not significantly associated with cats who won their interactions. Males who were considered to be “dominant” by this definition did not interfere with other males’ copulating nor did they have increased copulation success. Based on the data, the most “dominant” cat appeared to be the one involved in the most agonistic encounters without this behavior providing any obvious benefits. In another study, similar criteria were used and here the larger and dominant males were more likely to access females (Natoli et al. 2007). However, aggression was seen only rarely and many lower ranking males bred females.

Other factors clearly come into play in determining breeding success including the males’ relationship with the local females. Some males spent much more time with the females than others but both groups had similar frequencies of copulation (Natoli & De Vito 1991). Some authors suggest that there may be two groups of males based on their social interactions with the females: those who spend much of their time with a specific group of females and those who do not (Liberg et al. 2000; Crowell-Davis 2005). This could influence breeding success since some female cats do show some choices in who they allow to breed with them. There are advantages to cultivating specific females including possible greater mating success with familiar females, less energy expended roaming and seeking mates, and participation in protecting their kittens which improves the chances of their genes surviving. Advantages to roaming more might include optimizing total mating success and success as a larger male (Liberg et al. 2000). While unrelated male cats have been documented to kill kittens, males can be helpful in caring for and playing with kittens, particularly if they have been breeding females in that colony.

In a study of several colonies of cats on an island in Japan, they found that females were more selective about whom they allowed to copulate with them than with whom they allowed to mount them (Ishida et al. 2001). This difference could partially explain some of the disparities reported in the literature about female cat preferences. This study also found that weight, age, and length of estrus influenced frequency of accepted copulation by females. They reported that females were less likely to breed with close kin than with distant kin. Heavier males were somewhat more likely to successfully breed the females, but familiarity with the males by the females was also a driving force behind successful mating (Yamane et al. 1996). When the paternity of the kittens was evaluated using polymorphic microsatellite DNA, more than half of the kittens in a given group were found to be sired by males from outside the group (Yamane 1998). This DNA evidence brought into question the ability of observation only to determine copulation success. In contrast, on a sub-Antarctic island, where cats live at a very low cat density of about 1.5 cats/km² and are primarily independent of humans, hunting for their food, litters of kittens did not show multiple paternity (Say et al. 2002). Only one male sired each litter when 13 litters (from 9 queens) were genotyped. This would suggest that the location constraints (islands), population density, and home ranges could also be involved in breeding success.

Understanding cat populations

Cats have spread throughout the world as pets and concomitantly unowned free-roaming cats have become a concern (Macdonald et al. 2000). In some locations, free-roaming cats were deliberately or accidentally introduced while in other locations, these populations grew from the pet population. Cat population size and growth is a function of the cats’ reproductive capacity, their ability to survive, and their ability to migrate into (immigration) and out of (emigration) the population of interest (Slater & Weiss 2014). I will provide a brief overview of some typical data on these vital rates (fertility, mortality, immigration, and emigration). Fertility rates quantify the ability of female cats to produce kittens and are often expressed as litters per year. Free-roaming cat females produce an average of 1.4 litters per year, with a median of 3 kittens per litter (range 1–6 per litter) (Nutter et al. 2004b). A report from seven trap, neuter and return (TNR) surgery programs from around the USA reported March and April as the peak months for pregnant cats with continued pregnancies seen in substantial numbers through the summer (Wallace & Levy 2006). Reproductive capabilities in community cat populations are especially critical since low sterilization rates of 2% have been reported.

Mortality rates are the frequency of cats dying during a given time period and are often expressed as deaths per year or the percentage of cats dying in a year. Mortality rates for adult and juveniles have been reported in the literature in several regions. For adults, mortality is generally less varied and has been reliably reported between 22% and 45% (Warner 1985). For kittens, a much greater range of mortality has been reported from 20% (Gunther et al. 2011) to 75% (for kittens up to 6 month of age) (Nutter et al. 2004b).

Migration in and out of populations or colonies is less well studied and varies widely. One example is a study in urban Israel which identified four urban feeding groups, two of which underwent TNR (Gunther et al. 2011). In this study, resident cats were those who were seen in at least 4 of the 30 observation periods during the year. Immigrants appeared in the colony after the initial 5–7 weeks of observations and were seen for more than 4 observations and emigrants were cats seen in the initial 5–7 weeks but not again for the rest of the year. In the smallest colony (12 cats initially), there were 3 cats who were residents (25%) and 9 who left the colony (75%). In one of the larger colonies, there were 16 resident cats (31%), 23 immigrants (44%), and 11 emigrants (25%). In Rome, immigration in 103 cat colonies across a 10 year period was about 21% of the total population size (Natoli et al. 2006). Another study, where female cats were hysterectomized each year for 3 years, found a decreasing trend for immigration from 54% in the first year down to 15% in the third year (Mendes-de-Almeida et al. 2006). In North Carolina, six colonies were studied for up to 7 years (Nutter 2005). Number of immigrants per colony ranged from 1 to 14. In total, there were 9 immigrant females and 27 males; all females and 5 of the males were assumed to be abandoned. With such variability, it is difficult to make predictions and generalizations. In addition, because of the high variability, a large sample size is ideally required so that statistically significant and biologically appropriate conclusions can be drawn.

There are also external factors that influence population size by their impact upon the vital rates of fertility, mortality, immigration, and emigration. While all of the factors below may influence mammalian species, the precise extent of their effect on cats is undetermined. Carrying capacity is the maximum population size that can be supported by a specific environment/location (Gotelli 2001). Carrying capacity in cats is probably determined by food, shelter, and space availability. Increasing cat density beyond the food supply could influence both fertility and survival. However, I believe that kitten mortality increases and decreases the population size when carrying capacity is exceeded. Disease outbreaks also decrease population size through decreasing survival. Catastrophes like hurricanes or floods also can have dramatic influences on cat numbers. Geography has strong influences on climate and on the type of catastrophes that are possible and also anecdotally on cat population size and growth rates. Generally, northern USA and Canada will probably have higher mortality rates due to limited resources and extremely cold winters. Cats do survive in these areas; however, population growth is likely to be slower and in extreme winters, cats probably survive based on the availability of shelter, much of it man-made.

It is not simple to measure the vital rates of fertility, mortality, immigration, and emigration in community cat populations. However, it is somewhat less complex to decide if the total population size is increasing or decreasing. The methods to quantify cat populations have been extrapolated from the wildlife monitoring literature (Boone & Slater 2013). The precise choice of how to measure the numbers of cats depends on what is going to be done with the data. In some situations, individual colonies of cats may be a useful unit of analysis. Data to collect include the original size of the colony, when and how many kittens were born, and how many cats have joined or left or died. Careful tracking of colonies can provide very useful data if the caregivers can keep and share basic records. The ideal solution to recruit caretakers to collect data remains elusive but monthly reminders or incentives have been suggested.

Larger cat populations than individual colonies are somewhat more complicated to assess. There are two main approaches: trends to see if the population is growing or declining (relative estimates) or counts to see how many cats are estimated to be present (absolute estimates). In general, if relative estimates or general trends are adequate, then more rapid methods, performed in a consistent manner across a long enough time period should provide information about changes in cat population sizes across time. These changes may be due to active programs in an area (TNR or changes in animal control policies) or to natural fluctuations in cat populations. A commitment of years will be needed for useful trend data due to the inherent variability in cat populations. These rapid surveys are less precise than intensive methods but are also less costly and can provide a broad-based estimate that can be used as a baseline. Rapid surveys for cats could involve walking specific areas and recording the number of cats seen according to a specific detailed plan. In more rural settings, spot light counts of cats at night have been used to determine relative abundance (Edwards et al. 2000).

If an actual and relatively precise estimate of cat numbers is required, then a combination of rapid and intensive techniques will be needed (Boone & Slater 2013). The rapid surveys would provide the broad strokes, identifying where cats are commonly found. In contrast, the intensive surveys would focus on a smaller number of locations and collect much more detailed information about the individual cats. Methods for intensive surveys often involve identifying individual cats repeatedly over time using photographs or descriptions. Intensive surveys may also provide welfare and behavioral information on the cats. While many animal shelters may not have the staff to perform these surveys, trained volunteers could be used. Another potential source of assistance is college or universities. The colleges may have courses where fieldwork is required and students in those courses could assist with the design and initial survey work. More sophisticated analyses would be needed of the intensive survey methods to calculate reproduction, survival, migration, behavior, and home ranges, but these additional analyses are possible. In addition, an intensive survey of cat numbers can be used as a check to see how accurate the observers appear to be in performing the rapid methods in the same location.

Cat behavior and influence on health

The health status of the cats has intrinsic importance due to concerns about animal welfare and may be strongly influenced by cat behavior. Additionally, some cat diseases are important due to their potential to be transmitted to wildlife or humans. Factors that might influence disease frequency are cat density, interactive behavior (including sexual behaviors), nutrition, human management practices such as vaccination or deworming, geography (some diseases are more common in some parts of the county or some climates), etc.

Another issue to consider is how the free-roaming unowned cat population disease frequency differs from that of the free-roaming owned pet population since their behaviors with humans and possibly each other can vary. One study in France found quite different disease prevalence for owned and unowned cats (Hellard et al. 2011). In many situations, free-roaming cats are considered to be the source of the disease problem. Yet, owned cats allowed outside who do leave their yards and mingle with other cats are also a real source. Only a few studies have made those direct comparisons (Nutter et al. 2004a; Hellard et al. 2011).

This section will focus on the commonly discussed diseases of rabies, feline immunodeficiency virus (FIV), and feline leukemia virus (FeLV) in the context of human and cat behavior. There is also the potential for cat diseases to be transmitted to other species. Some transmission from domestic cats to wild cats or other wild species has been documented. FeLV has been reported rarely in wild felines (Foley et al. 2013). However, FIV appears to be host species specific and endemic in many felids (Pecon-Slattery et al. 2008). The actual source cats for these viruses are difficult to determine. Was this an owned cat allowed to roam or was this an unowned cat? Was the disease passed from one wild felid to another? Answers to these questions are as yet unknown. There has also been a little research on how cat behavior may influence disease in general. Regardless, decreasing the numbers of community cats will decrease disease transmission opportunities.

The cat’s ability to live in groups at various densities could influence disease occurrence and transmission. The density of cats is a function of resources and has been examined relative to parasite load and disease prevalence. One mainland, high density cat population in Lyon, France, and a low density cat population on an island were studied (Fromont et al. 2001). In Lyon, previously reported viral prevalence was 14% for FIV and 5% for FeLV. On the island, all 104 live and 46 culled cats were negative for viruses.

In the USA, the rabies reservoirs are wildlife and the cat is a spill over species with 92% of positive animals tested being wildlife species (Dyer et al. 2013). Among the domestic species, cats are the most commonly diagnosed; however, the frequency of cats testing positive for rabies is decreasing. Only 1.1% of cats submitted for testing in 2012 were positive. Only 19% had a vaccination status recorded. Rabies is of primary importance due to human exposures and subsequent postexposure treatments. Better awareness of the signs of rabies and prevention in owned cats could decrease human health risks. Rabies should be in the differential diagnosis of any aggressive behavior or sudden change in behavior (Frymus et al. 2009) and can occur even in young kittens (Bretous et al. 2008).

It is important to consider that while rabies is a reportable disease for all states, information about the animal tested varies widely from state to state. This makes it impossible to determine if the cats were considered to be owned or not, and vaccination status is often missing or unknown. Risk of contact with different populations such as feral, owned, or socialized but unowned is likely different since truly feral cats, unless they have the furious form of rabies, will avoid human contact. Advocating for and providing free or low cost rabies vaccination for all cats as well as teaching about rabies prevention is the best protection against rabies.

From a control perspective, rabies vaccinations in cats are extremely effective (Frymus et al. 2009; Jas et al. 2012). In addition, cats are less susceptible to rabies than dogs and their susceptibility decreases with age. Cats also tend to mount a better immune response than dogs with 97% developing antibodies after vaccination (Frymus et al. 2009). National or local laws often guide vaccination application; however, a single rabies vaccination induces long-lasting immunity, defined as a neutralization titer above 0.5 IU/ml. These results imply that even a single rabies vaccination may produce protection against the disease in cats.

TNR programs should always include rabies vaccination with a 3 or 4 year vaccine in areas where rabies is present. By stabilizing the population through sterilization, TNR results in a population of cats that has been vaccinated against rabies at least once. Removal of cats is sometimes intuitively (or legally) performed to control rabies in cats. A recent review article by public health and bird protection officials argued that because TNR does not decrease populations, it is not an effective method for reducing public health concerns (Roebling et al. 2013). The authors recommended removal of stray or unwanted cats. However, removal programs almost never catch all of the cats in the area and do not typically alter the habitat enough to prevent new cats from migrating in and existing cats from reproducing to fill the niche. Consequently, removal results in a population of young, unvaccinated, breeding cats which does nothing to prevent the transmission of rabies. Based on the substantial research on controlling rabies in dogs, there is no evidence that removal of dogs in countries where they are the reservoir species has ever had a substantial impact on population density or spread of rabies, likely due to the high population turnover (World Health Organization 2013). This finding also supports stabilizing the population and mass vaccination of cats as the primary methods to control rabies spillover from wildlife and protect human health.

Cat behavior is important in the spread and control of FeLV and FIV. FeLV is an infection spread by prolonged close contact and from infected mother to her kittens (Levy & Crawford 2004). This would tend to make this a disease of both sexes, possibly in clusters or colonies. There could be a genetic component as well. FIV is a disease spread by fighting and biting and is more commonly reported in males. TNR reduces the spread of both of these diseases by preventing litters of kittens from being born to FeLV mothers and decreasing fighting of males by neutering. Testing all cats for these diseases in community cat populations may not be the best use of resources if the colonies are monitored. If a cat is positive for FeLV, will the cat be euthanized even if apparently healthy? If the cat is feral, adoption is not an option. If a cat is positive for FIV, euthanasia is probably not appropriate since that cat will likely live a long life. In a set of calculations where all the resources are put into sterilization instead of testing, fewer FeLV positive cats were left than if a smaller number were tested and sterilized (Levy & Crawford 2004). Cats with illnesses that are associated with FeLV or FIV may be considered for testing if euthanasia is an option for positive cats. Cats placed for adoption should always be tested according to AAFP guidelines.

FIV and FeLV are a concern in free-roaming cats because of their transmission to other cats including owned pets and wild felines. The welfare of infected cats is also a common consideration. FeLV is the much more serious disease with more clinical problems much earlier in the course of infection than FIV (Hartmann 2012). Median survival of pet cats with FeLV is 2–3 years. However, in contrast to FeLV, most FIV-infected cats live many years without health problems. In pet cats seen in veterinary clinics, cats infected with FIV had similar survival to uninfected cats with a median life expectance of about 15 years (Liem et al. 2013).

One study explicitly linked cat behavior and the prevalence of FIV in two Roman and one French cat colonies (Natoli et al. 2005). Two of the three colonies had male cats with FIV with prevalence of 29% and 19%, and all but one infected male was among the more aggressive cats based on encounters with other cats. These findings fit with the disease transmission of biting.

The prevalence of these two diseases has been reported to be somewhat varied. This may be due to the choice about which cats to test. Cats who have a high suspicion of FIV, such as adult intact males, or cats with wounds, illness, or dental disease, would yield a higher prevalence of FIV. Cats from a colony with known FeLV cats are also somewhat more likely to test positive to FeLV. The number of cats sampled and the number of positive cats can also have implications for the interpretation of prevalence. In one study in Ottawa, Canada, the feral cats from a single colony of about 40 cats had 20 cats tested (Little 2005). The prevalence of FIV was 5% (one male cat). I calculated the 95% confidence interval as 0.1–25%. Therefore, the prevalence could be as low as 0.1% and as high as 25%. This same study of stray cats in Ottawa selected another group of cats for testing based on the cats being intact males or ill and all came from a high cat-density area without previous TNR (Little 2005). In this situation, a higher prevalence of FIV would be expected and was indeed found at 23%; all were males. Cats brought to a veterinary clinic were also tested for a variety of reasons including general screening of healthy cats. The prevalence of FIV in this group was 5.9% and all were males.

In the USA, other studies have sampled a large number of cats, all of them entering a TNR surgery program (Lee et al. 2002). In Raleigh, NC, the prevalence of FIV was 2% (1–4%, 95% confidence interval) and FeLV 5% (4–7%). In Gainesville, FL, the prevalence of FIV was 4% (3–6%) and for FeLV 4% (3–5%). Here the 95% confidence intervals are much smaller and the estimates much more precise and useful. Males were about four times more likely to be positive for FIV than females in both locations.

One study in Mauna Kea, Hawaii, found higher prevalence: 9% for FIV and 16% for FeLV (Danner et al. 2010). FIV was found only in adult males. These prevalence differences could be due to genetics, density, or selection of cats for testing. Predicting the prevalence is impossible without testing. If a high prevalence of these diseases is thought to be likely, testing at least 100 random cats is needed to give an estimate of prevalence ±10%. A careful plan for positive cats and serious conversations with caretakers are also necessary.

One study compared FeLV and FIV prevalence in cats trapped for TNR with local community pet cats (Nutter et al. 2004a). Pet cats had similar FIV and FeLV antigen or higher FeLV antibody prevalence compared to the TNR cats. This implies that the feral cat population in this area was not a more important source of disease than the pet cat population.

A large study provides some baseline information on FIV antibody and FeLV antigen seroprevalence in the USA. The study enrolled 345 veterinary clinics and 145 animal shelters in the USA and Canada (Levy et al. 2006). Over 18,000 cats were tested during this time following the guidelines and recommendations from the American Association of Feline Practitioners. However, the ultimate decision about testing was made by the owners. Prevalence for cats tested at the animal shelters was lower for both FIV (1.5%) and FeLV (1.7%) than for veterinary clinics (3.1 and 2.9%, respectively). In models that accounted for type of clinic, region, type of animal (indoor pet, relinquished pet at shelter, stay or feral at shelter, and outdoor access pet), age, and health status (sick or healthy), adult cats had higher prevalence for FeLV and FIV than juveniles. Intact cats of either sex were at risk for FeLV and FIV. For FIV, risk was higher than indoor pets for relinquished pets and stray cats in the shelter and higher still for shelter feral or outdoor access pets. For FeLV, only outdoor access pets had higher prevalence than all other groups. Different regions showed different prevalence for each virus. Despite the fact that the cats and clinics were not randomly sampled, this study is the best published to date to examine the regional distributions of these viruses and risk factors in a comprehensive fashion and provides useful information on baseline prevalence of these viruses.

Another recent study examined the prevalence of FIV and FeLV using samples submitted for testing with Idexx Laboratories (Chhetri et al. 2013). The authors used an approach where the ratio of positive FIV to positive FeLV was examined geographically to see if there were clusters where one virus was more or less common than the other. The eastern and southern parts of the USA had more FIV relative to FeLV while the west had the reverse pattern. Different clades or strains of FIV are found in different geographic regions and could be partly responsible for the patterns. In addition, factors relating to how cats are kept, when and why cats are tested, and age distributions could also be involved in the geographic patterns. These data could be helpful understanding the likelihood of cats in a given location being exposed to one or the other of these viruses.

Measuring predation

There are several common ways to measure predation by cats. To understand the implications of studies using different measurement methods, a general understanding of the methods’ strengths and limitations is useful. Overall, any of these methods may be used to compare predation but they will have different biases and may not be directly comparable (Fitzgerald & Turner 2000). Gut samples require that cats be euthanized or found dead and contents of the digestive system be analyzed for the species of animal present. Results of gut samples are reported as the percentage of occurrences of a specific species. This type of analysis overestimates small prey animals and underestimates large prey on a volume basis. Fecal samples are analyzed and summarized similarly to gut samples but are not linked to a specific cat and can be collected without killing the cats.

Counting what prey animals were brought home by cats has the limitation of only being applicable to owned cats with compliant owners. This method may underestimate prey since it excludes prey that might have been killed and/or eaten and left in the field (Fitzgerald & Turner 2000). However, this method may also overestimate prey if the cat brings home items that the cat did not kill. In addition, these studies usually require volunteers and the owners of cats who are known to be good hunters and could be more eager to volunteer (or vise versa). Counts are commonly summarized by mean prey by species. However, due to the wide variability of hunting interest and skill (and preference about bringing prey home), the mean estimate for a group of cats is typically much higher than the more representative median estimate. These studies also tend to downplay the number of cats who bring nothing home at all. One study in Great Britain reported that older and fatter cats brought home fewer prey (Woods et al. 2003). The age and body condition of cats in these studies are often not examined.

A method using motion-activated video cameras on owned cats’ collars recently reported the number of prey successfully captured as well as the percentage eaten, brought home, or left where captured (Loyd et al. 2013). This study found that 44% of cats (24/55) actually hunted wildlife, only 30% were successful hunters (16/55), and the majority of predation was during the warm season. The 16 cats who successfully captured prey ate 28% of their catches, returned 23% to their houses, and left 49% at the capture site. Increasing age was associated with decreasing frequency of prey captured. This study raises additional questions about the accuracy of counting what the pet cat brought home. Additional research would be needed to see if unowned cats or cats in other locations show similar foraging patterns.

A final method is identification in the field of dead prey species and determination of what animal killed or partially ate the prey. This method appears to be used primarily on oceanic islands (Bradshaw et al. 2012). Limitations are that determining cause of death may be difficult. This is due to some similarity in teeth marks between predators, some prey being eaten almost entirely or prey being removed.

The human side of cats and predation

Human beliefs and behaviors about cats have substantially influenced the concerns and potential responses to cat predation. Predation of individual wild animals is a commonly expressed concern. There are several arguments as to why cats should not be allowed to prey on wildlife. Philosophically, some individuals believe that it is objectionable that cats should be permitted to kill individual wild animals (Barrows 2004; Tantillo 2006). These individuals argue that since cats are non-native predators subsidized by human feeding and care, they should not be allowed to hunt native species. The value of the cats’ need to hunt should not outweigh the value of the hunted animal. Therefore, humans should keep the domestic cat under control by confining them indoors or to a yard or enclosure. Alternatively, some individuals may have different values about cats compared to other species and may prefer to let a native rodent species go extinct rather than lethally control feral cats (Gorman & Levy 2004).

Another frequently expressed belief about cats is that since cats are an introduced species, they should be removed. This belief is based on an underlying assumption that we should protect native species from introduced ones. This is clearly not always the case since we routinely kill native predators to protect livestock. A corollary of this belief is that removing these non-native cats will “fix” the ecological problem. Ecosystems are much more complex than this and usually cats are not the only introduced species or habitat alteration. Typically rats, mice, rabbits, birds, or other predators have also been introduced with the cats. In addition, human habitation has changed water flow, fire cycles, and vegetation. Livestock also changes the landscape in substantial ways. So this assumption that removing cats will return the ecosystem to “normal” is fundamentally flawed.

A recent and balanced review of the cat on pacific islands recognizes the interplay of human perspectives and conflicting priorities (Duffy & Capese 2012). The author points out differences in what is considered to be humane treatment of cats, in what is the goal of cat control, and in the value of an introduced species like the cat over the native species. Eradication is potentially possible on islands or the mainland where there is no immigration from the owned population. However, this requires a high rate of removal, no sources of new cats, removal of all cats, and the resources available to do this.

Impact of cats on wildlife

Despite the large body of work on cats and predation, relatively little research has been done that contemporaneously examines cat numbers, cat predation, and prey populations size and success (Tantillo 2006). This lack of a clear connection between predation by cats and response by prey species leads to continued debate about the impact of cats on prey populations. In many situations then, we do not know if predation is compensatory or additive.

Compensatory predation replaces other forms of mortality for the prey species. This is what is usually thought to happen with wild predators and prey in a balanced ecosystem. Predators are a normal part of the mortality of the prey species. For example, foxes eat rabbits and help keep rabbit numbers in check and avoid environmental damage. If rabbit numbers begin to decline, fox numbers would also decline, and the balance between the prey and predator is reestablished.

Additive predation by cats would mean that cats are incrementally increasing the total amount of predation on a population. Prey populations would not be able to compensate and would decline. If there are other sources of food, cat populations would not decrease. Instead cats would switch to hunting other species or other sources of food. It is difficult to design studies to assess the true impact of cats on prey species. And because there have been some instances where cats appear to be a primary cause of species extinction on an island, many wildlife biologists assume that cat predation is additive. This leads to studies that are extrapolated inappropriately or cited without critical thought regarding their actual implications.

One article on predator removal and water bird success begins by stating: “the efficacy of long-term predator removal in urbanized areas is poorly understood” (Meckstroth & Miles 2005). In this specific situation in San Francisco Bay, CA, striped skunks were the primary predator with other species present including cats. The authors reported that the current level of predator removal did not result in better success for the studied bird species producing more offspring. The authors indicated that the lack of predator decline in the face of predator removal could be due to several possibilities: (i) other sources of food that could encourage predators to migrate into space vacated by the removed predators; (ii) a compensatory density dependent increase in survival or reproduction by predators; or (iii) insufficient numbers of predators removed. Yet cats are routinely blamed for bird declines without clear evidence.

A study in suburban areas around Washington, DC, concluded that cats were a primary reason for failure of Grey Catbird nesting success and poor post-fledging survival (Balogh et al. 2011). They reported varied bird success rates between the three studied locations. However, cats were identified as being responsible for 47% of known predation deaths. What required close reading of the article to determine was that 43% of predation events could not be assigned to a predator species. The authors acknowledged that they could not determine if cat predation was additive or compensatory. Therefore, the actual effect of this predation on the catbirds as a species was unknown.

One study might suggest that cat predation is compensatory. The authors used song bird spleens as a measure of immune competence (Moller & Erritzoe 2000). They found that cat-killed birds had smaller spleens. There were otherwise no differences between the cat-killed birds and birds killed by other means (mostly cars and windows) and brought in by the public. The characteristics examined were sex, age, month of death, body weight or condition, liver mass, and wing or tarsus length. These results suggest that cats killed birds with reduced immunocompetence that would likely have had reduced viability.

One study of cat predation in England did include a crude estimate of prey population. The authors estimated cat density by surveying residents and then estimated cat predation by recording what the cats brought home (Baker et al. 2008). Finally, they compared these data to bird densities and productivity. They also evaluated the body condition of birds killed by cats compared to those killed by hitting a window. They found that about 60% of cats studied did not bring any prey home, and wood mice were the most common species for cats who did bring home prey (53% of prey brought home). Individual prey animal per cat per year varied in the four studied locations from 1.5 to 12. Cat-killed birds had poorer body condition than those killed by a window strike. This suggested that cats could be killing the weaker members of the population and providing a compensatory rather than additive form of mortality. However, the authors concluded that cats could have had an additive impact on a few species of local birds. This conclusion is not completely supported by the data as they did not study long-term prey population changes to measure cat impact. Instead they examined breeding success, extrapolated predation to all cats including those who brought nothing home, and defined impact relative to numbers of fledged birds.

In another study of outdoor-owned cats, the authors found no relationship between the number of cats and small mammal abundance or foraging. This study included 11 radio-collared owned cats from 8 households bordering a suburban nature preserve (Kays & DeWan 2004). Most of the cats spent their outside time in the neighbor’s yard or the nearby forest edge (within 10 m). The authors speculated that the lack of impact on prey species was due to cold weather limiting cat activity in the winter and animals who prey on cats keeping cats out of the nature preserve. They also suggested that the prey species in that area may have been more resistant to cat predation than species in other areas. These authors seemed to work hard to justify the fact that the cats did not impact local prey species.

Extinction of species by cats is an often repeated accusation. In some instances, authors are concerned about localized extinctions of a specific species. In other cases, particularly on islands, extinction may be of all or most of a species. Ground-nesting birds are particularly at risk. Ecosystems without ground living mammals have species at particularly high risk because the native species did not evolve with a predator like a cat (or fox or rat). Much of the literature on extinction is retrospective and looks at available historical data. This limits our ability to account for factors apart from the presence of cats, such as other predator species, native and introduced plant life, and alterations from human habitation.

One article estimated that conservation actions between 1994 and 2004 prevented the extinction of 16 bird species with 10 of them on islands (Butchart et al. 2006). However, the primary threats were multifactorial and included habitat loss and degradation (88% of the bird species were threatened by this change), invasive species (50% of bird species), and exploitation by humans (38% of bird species). Successful conservation efforts were reported to require multiple approaches including habitat protection and management (75% of threatened bird species), control of invasive species (50%), and captive breeding and release programs (33%). In all cases where cats were introduced, a species of rat had also been introduced. In most situations, multiple introduced species were present and controlled. Substantial motivation, resources, and efforts were needed to prevent these extinctions, and no one solution, such as eradication of cats, was sufficient.

Variation in predation impact

All geographic and ecological locations are not the same, particularly with regard to the risk to prey species from cats. I have already made some comments about the differences between isolated islands and mainland ecosystems and the consequences of predation. However, there is some debate as to whether fragmented mainland habitats are similar to true island habitats. One author proposed that insular island species have been selected for and are adapted to the unique physical landscape of their island (Walter 2004). This would make them especially vulnerable to changes like introduced animals. On the other hand, species on the mainland continents, even with currently fragmented habitats, were evolutionarily exposed to a much more complex, variable, and dynamic landscape with more climatic changes. This complexity also is visible in more species richness and increased diversity of predators and diseases. These fundamental differences influence the species’ ability to adjust to changes. In addition, the presence of humans widely influences species evolution as well as current behavior and environment.

The proposal that the evolutionary history of island and mainland species is fundamentally different would support the conclusion that habitat fragmentation does not have the same impact on continental species as habitat changes do on island species. This does not mean that habitat fragmentation has no influence on continental species but that it is not strictly analogous to the island situation. Fragmentation can result in different types and distributions of species, both prey and predator. For example, coyotes, bobcats, mountain lions, and spotted skunks were all found more often in larger fragment areas and less often as the fragments became smaller in coastal San Diego County (Crooks 2002). Cats showed the opposite pattern and were found more often in smaller fragments. Cats were also more common in more isolated fragments (those farther away from other fragments) and on the urban edge than the native predators. These findings support that habitat size can influence predators but that the precise changes depend on the species and location.

The patterns seen in fragmented habitat for cats could be due to the close proximity of homes (the cats in this study could be owned or unowned), or due to the absence of larger predators who eat cats like coyotes (Crooks 2002). An alternative reason could be a phenomenon called mesopredator release. Mesopredators are mid-sized predators like cats, raccoons, and skunks and are subordinate to larger “top” predators like coyotes, mountain lions, and bobcats. When the large predators leave or are removed, the mesopredators are “released” and their numbers may increase, potentially increasing their impact on prey species (Courchamp & Sugihara 1999). There is still some debate about whether this release is real or not. However, I believe that some of the contradictions in findings are due to the variability between locations and that mesopredator release may only occur in specific types of situations.

There are also varied densities of the different species in different geographic areas. For example, urban natural areas may have higher population densities of some bird species than nearby natural spaces in more rural locations. A variety of reasons have been suggested for this phenomenon included possible decreases in predation (Sorace 2002). However, one study examining this hypothesis reported that predators such as kestrels, nocturnal raptors, crows, rats, foxes, cats, and dogs were at higher densities in urban parks, in the very same location where bird and pest density was higher.

The previously referenced research illustrates just a fraction of the variability in the ecosystem found in different locations. This makes it impossible to apply regional data or models to diverse areas without careful examination of similarities and differences. Furthermore, ecosystems, even on islands, are quite complex and the relationships among the different species may not be well understood. Concerns about mesopredator release and other unintended consequences must be taken into account (Duffy & Capese 2012). Unintended consequences were seen on Macquarie Island where successful cat eradication resulted in a boom in rabbit population and dramatic habitat damage (Bergstrom et al. 2009). Birds’ behavior may influence their survival on islands with cats but the details are complex and dependent on several additional factors (Pontier et al. 2010). One author traces the impact of humans, landscape change, and biodiversity loss on Socorro in the Mexican Pacific (Walter 2004). In 1869 sheep were introduced. No documented bird population changes were seen until the Mexican Navy built a settlement there in the 1950s. An elf owl became extinct for unknown reasons after that. The Socorro dove became extinct by 1972 due to human persecution, possible feral cat predation, and many landscape changes. This example is one of several that has illustrated the complexity of the situation even on an isolated island.

Decreasing hunting by cats

Relatively little work has been done on ways to specifically decrease free-roaming cat hunting activities. Feeding cats has been suggested to reduce their urge to hunt. There are no data to support or deny this claim. However, cats who are poorly fed are more susceptible to disease and have higher kitten mortality (Fitzgerald & Turner 2000). A lack of a focal food source also tends to increase range size meaning that cats will spend less time in locations where their rodent control skills are desired, potentially have more environmental impact by having a larger range and be more difficult to count or monitor (Liberg et al. 2000). Another approach that has been used successfully is to exclude cats and other predators from sensitive areas using fencing (Young et al. 2013).

The use of bells on cat collars used to be considered ineffective. However, recent research in New Zealand has found that wearing a bell on the collars of owned cats who go outside and hunt does decrease the predation of birds and rodents by 50% or more (Gordon et al. 2010). This study had the same cats wearing and not wearing the bell alternately, in random order, for 6 weeks at a time. Bell wearing did not decrease predation of rats, lizards, and insects but did decrease mice and bird predation. Cats 6 years old and older caught more rats than younger cats. A similar study in England on the cats that first wore and then did not wear bells in random order for 4 weeks at a time found a decrease in total prey of almost 50% with the bells on (Ruxton et al. 2002). Another study in England also randomly assigned the order of belled and non-belled collars to 68 cats for 4 weeks at a time (Nelson et al. 2005). An additional treatment group was a collar with a sonic device. The belled collar decreased prey brought home by about 33% compared to a plain collar. The sonic device performed similarly to the belled collar. These results support that a bell on a hunting cat’s collar may reduce hunting success of cats who are allowed outside.