DOMESTIC CAT (Felis catus ) PREDATION OF BIRDS IN AN URBAN ENVIRONMENT
Carol A. Fiore and Karen Brown Sullivan
To quantify the effects of urban domestic cat predation on birds in Wichita, Kansas, a city of approximately 300,000
residents, we collected birds killed by study cats, analyzed fecal material, and tracked cats using radio collars to estimate
mean numbers of birds killed per cat. A random survey and information from local veterinarians were used to calculate pet
cat density. The results indicate that the average urban cat in Wichita kills 4.2 birds per year. Additionally, we found that
the majority of cats (83&(Troglodytes aedon ) appear to be at increased risk (9&
Given the current climate of increasing concern over declining songbird populations (Franzreb and Phillips 1995), the need
for better documentation of the impact of domestic cats on birds is warranted. Much of the information on the hunting
abilities of domestic cats (Felis catus), an introduced predator in the United States, is anecdotal; few studies have been
done to assess the impact that free-roaming domestic house cats have on native bird populations. The negative impact of
cats on island birds has been well documented (Fitzgerald and Karl 1979; Ludwig 1994; Turner and Bateson 1988;
Heidemann and Vauk 1970). Since the fauna on most islands did not evolve in the presence of mammals, many insular
birds are ill-equipped to deal with the predatory domestic cat.
Several different approaches have been taken to determine the relative contribution of birds to the diet of domestic cats.
Hubbs (1951) examined stomach contents of feral cats and found birds in 74 of 184 stomachs examined; when present,
bird remains comprised 25.2&
Churcher and Lawton (1987, 1989) enlisted the help of all but one of the cat owners in the small English village of
Bedfordshire, England, and for one year all prey items observed by owners were collected and turned over to the
researchers. Fecal material was not analyzed nor were cats tracked and observed by the researchers. Their estimate of 20
million birds per year killed in England by house cats was a very conservative one because it did not account for kills not
brought to the owners. The proportion of unobserved kills may be significant. For example, a researcher in southern Illinois
speculated that his three house cats, which he followed for six years, brought home only about 50 Eberhard 1954;
Fitzgerald 1988; Mitchell and Beck 1992).
Patronek (1998), in a critique of published studies focused on cat predation, pointed out several common methodological
problems. For example, extrapolation of data from the England study (Churcher and Lawton 1987, 1989) to other
ecological areas with different habitats and bird species is problematic. Coleman and Temple (1993) may have
overestimated cat density populations by polling residents when kittens were likely to be most abundant. The use of small
non-random data sets of cats, as in Mitchell and Beck (1992), and a lack of physical evidence when polls are taken (e.g.,
Potter 1991) were all noted. Patronek also claims that cats hunt birds during the day and tend to bring prey home.
A lack of studies exists on urban cats in general, and the methodology used in much of the research to date is limited in its
application. The purpose of our study was to collect information necessary to estimate reliably the impact of domestic cats
on birds in an urban environment. To date, no other research has attempted to estimate average bird kills per cat by
combining information from bird collection, scat analysis, and radio tracking of cats. Additionally, we estimated pet cat
density in the city and collected information regarding public perceptions and opinions about cat predation and regulation of
cats, an undertaking not attempted by other researchers.
MATERIALS AND METHODS
The Study Site.&
The Study Cats. a few numbers were unused.
Determination of Average Bird Kills per Cat.&
The second methodology used to calculate average number of bird kills per cat was to examine fecal material of the study
cats for the presence of bird parts, primarily feathers. Persuading volunteers to turn over scat was difficult for several
reasons. More than half of the volunteers did not own a litter box and some volunteers were reluctant to participate in this
particular activity. A total of 14 households collected scat at least once during the study.
Two different methods were used in analyzing the fecal material. In the first, four volunteers separately bagged each day’s
scat for five consecutive days each month. Each day was treated as a separate analysis. After approximately 12 hours,
most of the contents of a cat’s stomach have been passed (Hubbs 1951). For this reason, a feather(s) in each day’s scat
was conservatively counted as one kill. The data indicate that this approach is valid since we rarely found feathers on
consecutive days. In the second type of collection, 10 households bagged the entire contents of the litter box, in which case
the entire collection was treated as a single analysis. The procedure used to check for feathers in scat was similar to that
used by Day (1966), Liberg (1982, 1984), and Fitzgerald and Karl (1979). Scat from each separate collection was
soaked in water in glass beakers for at least 24 hours. The contents of each beaker was rinsed in a 1.4 mm wire mesh sieve
which was held over an opaque plastic container. The water in the container and the material in the sieve were then
screened for feathers. Questionable items were removed with tweezers and deposited in a separate viewing container for
examination under a dissecting microscope. When feathers were found they were placed in vials, marked, and information
recorded. No attempt was made to identify the species of bird nor to assess the proportion of bird parts to other prey
items. The objective was solely to ascertain whether the cat had ingested a bird for which the owner had not collected
remains. Scat data were used in two different ways. When feathers were found during examination of scat, an additional kill
was recorded for that cat’s kill record if there had been no remains collected at that time. If more than one cat was using
the litter box, kills were alternated between cats (starting with the lowest numbered cat), but only one cat in the household
was credited with a kill. Only one kill was counted per analysis regardless of the quantity of feathers recovered from the
scat. Since the scat data set was based on a small number of households, a second method was developed to quantify the
data. The following formula was used for each cat:
total number of times feathers were found (without owner knowledge)
total number of analyses
An average of all values was used to arrive at the percentage of time a cat could be expected to have ingested a bird with
no owner knowledge. This value was used to adjust the total number of birds killed per cat per year.
The third method used was to track cats using radio collars. Although tracking provided important supplemental information
about cat activities and ranges, cat/owner relationships, and owner knowledge of cat location and habits, the information
gained was qualitative. The equipment consisted of a TRX-3S receiver, a large directional antennae, and an SOC-2270
transmitter collar (Wildlife Materials, Carbondale, Illinois). Binoculars were used to watch the cat once it was located. The
waterproof transmitter was mounted on a 1.59 cm wide nylon collar. The antennae was hidden between the fabric on the
collar, and the signal had a range of about 0.62 - 0.93 km. Preliminary observations suggested that the collar did not affect
behavior. Habituation was achieved by fitting the collar to the cat several days before tracking started.
The majority of the cats were not available for tracking due to owner or neighbor objections, physical barriers (e.g., privacy
fences), and trespassing issues. Although tracking proved difficult in an urban setting, we observed eight cats during daylight
hours for a total of 57 hours.
Cat Density Estimation.&
The second approach utilized to determine cat density used the results of a random telephone survey. Written by Dr. Ellie
Shore of Wichita State University (WSU), with a grant from the Edith J.Goode Residuary Trust, this Pet Ownership Survey
was conducted in January 1999 by the Elliot School of Communication at WSU and was aimed at understanding pet
ownership trends. For this study, only data from city residents (N = 622) were used and relevant responses included those
concerning observations of feral and stray cats, current vaccines of pet cats, and questions regarding feelings about cat
predation and regulation. These data allowed calculation of two independent estimates of pet cat density. In the first, the
total number of cats vaccinated by veterinarians was divided by the percentage of cats that survey respondents claimed to
have been vaccinated (e.g., if 500 cats received vaccinations and respondents indicated that 50&
Bird Kills.& (Table 1). Due to the poor condition of some collected remains, it was not possible to identify 17 of the birds,
some of which may have represented additional species. The most birds collected were European Starlings (14& (Table 1)
than those that nest and feed above ground level. Cats in this study preyed on 7.3& (Table 2); inclusion of scat data brings
the total to 23 birds.
Our data indicated that 72 the top predator cat was declawed. The owner of the top predator cat lived on an older small
suburban lot on the north edge of the city, and a review of the birds killed by this cat suggest the animal had a large range.
The greatest number of bird collections occurred during the months of May and June, with secondary peaks during the
months of April and July (Figure 1). Bird kills were counted only when actual physical remains were recovered (or feathers
found in scat). Verbal reports from volunteers were never counted as kills; the 113 birds represent actual remains collected
from volunteers (scat analysis added additional birds to this total). Taking into account the actual number of days each
volunteer participated in the study and the number of bird kills recorded, we calculated a mean of 3.44 &
Out of 215 separate scat analyses, each of which could have composed several beakers of fecal material, feathers were
found a total of 28 times. In only one instance, however, did the owner know that a bird had been killed and/or consumed.
When scat data and bird collection data were combined, the number of kills recorded for cat 13 totaled 17 birds (all but
three were credited with scat information), which puts this cat alongside cat 18 for the highest number of kills in the study
(Table 2). A mean value of the percentage of time a cat could be expected to ingest a bird with no owner knowledge was
calculated to be 21&
The Pet Ownership Survey indicates a total of 35.2 research from this study, however, indicates that cat owners often give
inaccurate information about activities of their pets, and disappearances of supposedly indoor cats, especially at night, are
common. An indoor cat as defined in this study meant a cat that never goes outdoors. People who interpreted this question
to mean that because the cat was inside &
Results of Survey Concerning Cat Regulation. over 50&
Any study that utilizes human volunteers, especially for such a long period of time, is prone to problems. This study was no
exception; the many sources of bias all tended to underestimate the numbers of birds killed (i.e., missed birds, missing cats,
absence and relocation of owners and schedule conflicts, remains thrown away, lack of active participation). Volunteers
reported bird escapes and although the extent of the injuries was unknown, it is likely that most of these birds later died.
Several birds were retrieved alive from cats, but all the birds died subsequently as a result of the injuries. Many wildlife
rehabilitators cite poor survival rates for cat-caught birds (Dowling et al. 1994; Ludwig 1994; Potter 1991).
Although birds are of only secondary importance in the diet of domestic cats, small mammals such as mice and young
rabbits being the primary prey taken (George 1974; Churcher and Lawton 1987, 1989), our scat and bird collection data
demonstrate that most domestic cats kill birds (83 Adamec 1976; Morris 1986). In fact, our scat analysis suggests that a
far greater number of birds are consumed than was previously thought. We questioned volunteers about feather cat toys
and presence of pet birds in the home; neither was deemed to be a source of error. Additionally, a few of the homes
owned indoor cats not included in the study and scat analysis may have included droppings from these cats. It is possible
samples from these litter boxes may have failed to include any droppings from study cats.
Turner and Bateson (1988) remark that today’s cats are not particularly nocturnal, but our research suggested otherwise.
Goldsmith et al. (1991) tracked cats primarily at night when they were most active. Calhoon and Haspel (1989) found stray
and feral cats avoided most humans by seeking shelter during the day and emerging only late at night. Main activity times for
cats appear to be bimodial with peaks occurring around midnight and before sunrise (Tabor 1983; Haspel and Calhoon
1993). Patronek (1998) has criticized cat predation studies for their lack of direct observation of bird kills by cats but,
given the secretive nature of cats, their solitary hunting and reluctance to tolerate observers, a lack of direct documentation
is not surprising. Our tracking data suggest that cats are extremely difficult to watch and even more difficult to observe in
acts of predation. This does not suggest a lack of predation; scat analysis and bird collection data have clearly shown
Our estimate of pet cat density is likely to be low for several reasons. The cat rabies data supplied by area veterinarians
were from 1997 and the population of Wichita, especially in the west part of the city, has increased over the past two
years, likely leading to a greater number of pets. Calculated density was based on data from the Pet Ownership Survey
which indicated 87.8&
The long-term ecological implications of cat predation on birds in Wichita are far from clear. More investigation is called
for, especially with regards to such sensitive birds as the Dickcissel. Certainly there are many threats to birds besides cat
predation; habitat loss on wintering and nesting grounds may be the primary factor in the decline of many songbirds. Cats
may not be one of the primary causes of avian mortality, but they do kill birds, and in some areas may well prove to be a
cause for serious concern. The latest estimate of the number of pet cats in the U.S. is from the 1999-2000 American Pet
Products Manufacturers Association’s Pet Owner Survey which estimates there are 64 million pet cats in the U.S. It seems
reasonable to assume that on the basis of the pet cat population alone in this country, that if each cat killed 4.2 birds per
year as did the average cat in this study, this would result in the death of at least 269 million birds per year due to predation
by pet cats alone. Further assuming that half of these cats never go outside (e.g., the phone survey in this study indicated
The common misconception that declawing cats prevents bird killing has been proven false by this and other studies. Also,
results from several studies clearly show that bells do not prevent cats from taking prey (e.g., Potter 1991). The cat owner
volunteers in this study were asked upon its completion if they would keep their cats inside as a result of the study, and they
were told that the greatest threat to birds in Wichita occurs during the months of May and June. Of the 26 volunteers who
answered this question, 73&
We thank Jerry Davis for his special inspiration, Donal O’Brien for financial support, Bob Gress, Max Thompson, and
Gene Young for help in identifying the birds, and Linda Winter and Ron Jurek for their wonderful advice and many
documents and letters. A special thanks to Ellie Shore who designed a wonderful survey and allowed us to incorporate our
questions, and to Dr. Jon Piper for his words of encouragement and for reading the manuscript. Free cat products, which
were used to compensate volunteers for their efforts, were provided by A&M Products, Wal-Mart of West Wichita, and
PetsMart. Without the support of Wichita State University and our 27 cat owners, this project would not have been
possible. We would also like to thank our families, who have been incredibly supportive of our work.
No part of this article may be reproduced without the authors'