Survey of predation by domestic cats
ID ROBERTSON
Division of Veterinary and Biomedical Sciences, Murdoch University, Murdoch, Western Australia 6150
Objectives To calculate the proportion of house cats
which were observed by their owners to have caught prey and
to describe the characteristics of these cats.
Design and procedure A telephone questionnaire was
administered to a randomly selected population of 458 cat
owners in metropolitan Perth. Specific questions were asked
about demographic characteristics, habits and diets of the
cats, and whether the owners had observed their cats catch
prey in the 12 month period preceding the survey.
Results The owners of 36% of 644 cats had observed
their cats with prey in the 12 month period preceding the
survey. Cats which spent more time outside, were neutered,
cross-bred, originated from households with only one or two
cats or were not fed meat were significantly more likely to be
observed to predate. The body condition and diet (other than
feeding meat) of cats did not influence the reported frequency
of predation.
Conclusion Although cats are only one factor involved in
the reduction in the numbers and diversity of Australian
wildlife, restriction of the outside activities of cats is likely to
diminish predation, particularly in areas close to native vege-
tation.
Aust Vet J 1998;76:551-554
Key words: Cat, predation, fauna.
A
lmost 50% of all recent extinctions of mammals
throughout the world have occurred on the Australian
continent.1 In contrast, less than 1% of the original
bird, reptile and amphibian species of Australia are now
extinct.2 Extinction and reductions of range of native fauna in
Australia can be attributed to a variety of factors including
predation from feral animals, competition and habitat degrada-
tion from rabbits and introduced livestock, clearing of vegeta-
tion, changes in fire regimes and human persecution. 2
Predation, competition for resources and transmission of disease
by feral cats have been implicated in the regional extinction of
several species of native mammals,2,3 and the decline in native
fauna, particularly rodents and small marsupials, has paralleled
the increase in the feral cat population. 4 In contrast, on
Christmas Island, it has been proposed that cats play a benefi-
cial role in the control of introduced rodents, and have not
necessarily had a deleterious effect on the native species of that
island.5
The effect of feral cats on native fauna has been well docu-
mented.1-4 However the effect of predation by house cats is
difficult to quantify and their effect on native fauna remains
unresolved. Barratt6 reported that although introduced rodents
are common prey of house cats in Canberra, house cats can have
a significant effect on locally abundant native fauna. The aims
of the current study were to determine the proportion of house
Aust Vet J Vol 76, No 8, August 1998
cats that were observed by their owners to have caught prey, and
to describe the characteristics of these cats.
Materials and methods
A random telephone survey of households in the Perth
metropolitan region was conducted to interview the owners of
cats. A random number generator was used to select the page,
column and position of a telephone number in the current tele-
phone directory. The telephone number corresponding to this
position was then selected. Only those numbers which appeared
to be private residences were included in the survey. If no
answer was obtained after three calls a replacement number was
selected using the same procedure. Records were kept of the
number of pet owners, households without pets, people who
declined to be involved in the survey, businesses, non-English
speakers, disconnected lines and fax machines contacted. The
non-response rate was calculated by dividing the number of
refusals by the total number of households interviewed.
The potential for misclassification bias was limited by having
a fixed questionnaire with ordered and specific procedures to be
followed. Specific data were collected about the age, sex, breed
and neutering status of cats, their diet and number of times fed
daily, weight, body condition, the number of cats present in
each household, the amount of time cats spent outdoors during
night and day time in winter and in summer, and whether the
owners had observed their cats with prey in the 12 month
period immediately preceding the survey. Information was not
collected on the species of prey.
Comparisons between groups for discrete data were initially
made using univariate analyses of odds ratios and their 95%
confidence intervals, and the chi-square test for independence.
Continuous data were initially analysed using the analysis of
variance. Logistic regression was then used to quantify the asso-
ciation between reported predation and each variable after
adjusting for the other variables. Only variables significant at
P ≤ 0.25 were considered eligible for inclusion in the multiple
logistic regression analysis.7,8 Dummy variables were generated
for any categorical variable with more than two levels. Backward
elimination was used to determine which factors could be
dropped from the multivariable model. The level of significance
for a factor to remain in the final model was set at 10%. The
goodness of fit of the model was assessed with the Hosmer-
Lemeshow statistic.9
Cats were used as the unit of analysis in this study. To account
for possible cluster effects in households containing several cats,
a within-household correlation coefficient was calculated and
used in an adjusted chi-square test.10,11 Statistical comparisons
were performed using Statistix for Windows (Analytical soft-
ware, Tallahassee, Florida) and Excel 97 (Microsoft). Pearson’s
correlation coefficients were calculated to determine correlation
between variables in the model. In Statistix a variable is auto-
matically dropped from the model if that variable is too highly
correlated with other independent variables.
551
Results and discussion Table 1. Prevalence of owner-reported predation, odds ratios and their
95% confidence intervals, and significance obtained from the screening
During the survey 2195 households were successfully inter- of factors with predation by univariate analysis.
viewed, of which 458 (20.9%) owned a total of 644 cats
(average 1.4 cats per cat-owning household). One hundred and Factor Prevalence of Association with P
thirty-one (5.6%) people declined to be involved in the survey. predation reported predation odds ratio
by owners (%) (95% CI)a
A further 313 households could not be contacted after three
attempts on consecutive days, 170 telephone numbers were Female cats 33.0 1.0
Male cats 39.2 1.2 (0.9, 1.7) 0.25
either disconnected or connected to fax machines, 65 people
were unable to speak English, and 44 numbers were businesses Entire cats 23.7 1.0
Neutered cats 37.9 2.2 (1.3, 3.8) 0.0019
and therefore excluded from the survey.
Thirty-six percent of cats were reported to have caught prey Purebred cat 24.5 1.0
Crossbred cat 38.0 2.4 (1.5, 4.0) 0.0004
in the 12 month period preceding the interview, 44% of cats
were reported not to have caught prey and 20% of owners were Households with > 2 cats 25.2 1.0
Households with ≤ 2 cats 38.1 2.1 (1.3, 3.3) 0.003
unsure of the predatory habits of their pets. This reported
degree of predation is likely to be an underestimate of the real Inside in summer nights 28.2 1.0
situation, as 70% of cats spent most (> 50% of the time) of Outside in summer nights 45.2 2.6 (1.8, 3.7) < 0.0001

their summer days outside, compared with 54% during winter Inside in summer days 23.4 1.0
days, 45% during summer nights and 33% during winter Outside in summer days 41.2 2.9 (2.0, 4.4) < 0.0001

nights. Not surprisingly, cats that spent more than 50% of their Inside in winter nights 30.9 1.0
day or night time outside (in either season) were reported more Outside in winter nights 46.2 2.3 (1.6, 3.3) < 0.0001

often to predate (P < 0.0001) than cats which spent more time Inside in winter days 26.8 1.0
inside (Table 1). However after multivariate analysis only cats Outside in winter days 43.6 2.8 (1.9, 4.0) < 0.0001
which were outside at night time in summer or in the day time Normal body condition 34.3 1.0
in winter were significantly more likely to be observed to Fat body condition 41.0 1.3 (0.9, 2.0) 0.18
predate (Table 2). These times correspond with that when more Fed more than once a day 31.4 1.0
cats were outside (70 and 54%, respectively). Fennel12 reported Fed only once a day 36.7 1.0 (0.6, 1.6) 0.86
that domesticated cats were generally more active, and were Fed dry pet food 34.4 1.0
active for longer periods during the seasons of spring and Not fed dry pet food 36.0 1.1 (0.6, 2.0) 0.79
summer than in autumn and winter. These seasons correspond Not fed canned pet food 33.3 1.0
with the breeding cycle of many Australian native animals, as Fed canned pet food 36.3 1.2 (0.8, 2.0) 0.38
well as introduced species. Although cats have historically been
Fed meat 33.9 1.0
considered nocturnal animals, Turner and Meister13 reported Not fed meat 37.8 1.3 (0.9, 1.8) 0.18
that house cats are as active throughout the day as night, and
Fed offal 32.3 1.0
have shifted their activity, including predation, into the daylight Not fed offal 36.1 1.9 (0.8, 4.3) 0.12
hours in contrast to their wild ancestors. This observation could
Not fed ‘other food’b 32.4 1.0
mean that, in the current investigation, up to 70% of cats have Fed ‘other food’ 37.6 1.3 (0.9, 1.8) 0.23
significant opportunities to catch prey in Perth. A similar degree
Fed supplements 23.9 1.0
of predation was reported by Paton14 (62% of cats were Not fed supplements 36.8 1.9 (0.9, 4.0) 0.08
observed to catch birds, 60% mammals and 20% reptiles) in a
Cats receiving snacks 36.1 1.0
study involving a questionnaire administered to cat owners in Cats not receiving snacks 36.2 1.1 (0.8, 1.7) 0.55
metropolitan Adelaide.
Received a kitten diet 32.3 1.0
Telephone surveys are an inexpensive means of rapidly gath-
Didn’t receive a kitten diet 33.0 1.4 (0.7, 2.5) 0.32
ering data. However biases can influence the quality of the
Fed a diet for health reasons 25.0 1.0
information collected. In this survey the small number of
Not fed a specific diet 36.4 2.0 (0.8, 5.0) 0.12
people who declined to be interviewed (5.6%) helped minimise
a95% confidence intervals
sampling bias, although households without telephones, those bFood excluding proprietary pet food, meat or offal
with unlisted numbers and those that did not answer their tele-
phones on three consecutive days were obviously not included.
This survey relied on observations of owners to quantify and
identify characteristics of house cats which predated. surveys6 are based on the assumption that prey brought home
Observation of predation by owners is primarily influenced by and observed by the owner is representative of the cats’ total
the time spent with their pet and their pets’ frequency of preda- catch. Turner and Meister13 found that mammals were either
tion. The frequency of predation is influenced by the amount of eaten near the place of capture or were carried home, while
time the cat is outside, the agility of the cat, and whether deter- invertebrates and smaller animals were usually devoured at the
rents, such as bell collars, are used. Furthermore, recent adverse place of capture. This may account for the large numbers of
publicity on the role of the cat in the demise of certain species mammals and small number of reptiles detected by owners in
of native fauna may have influenced owners’ responses. These the study of Barratt.6 Similar results may have lead to a degree
factors may lead to an underestimation of the true frequency of of under-reporting of predation by owners in the current study.
predation by cats in Perth. Although it is difficult to make inferences concerning the actual
Most studies investigating the predatory habits of feral cats effect of cats on populations from dietary studies, Martin and
have identified prey in gut contents,15,16 however this procedure colleagues16 considered that cats could have a significant effect
is not appropriate for house cats. The findings of this and other on some native animal populations.

552 Aust Vet J Vol 76, No 8, August 1998

Table 2. Characteristics of cats found to be significantly associated with
predation in the logistic regression model.
Factor Association with predation
odds ratio (95% CI)a
Entire cats 1.0
Neutered cats 2.4 (1.4, 4.1)
Purebred 1.0
Crossbred cat 1.9 (1.1, 3.2)
Households with > 2 cats 1.0
Households with ≤ 2 cats 2.0 (1.2, 3.4)
Fed meat 1.0
Not fed meat 1.6 (1.1, 2.3)
Inside in summer nights 1.0
Outside in summer nights 2.1 (1.4, 3.1)
Inside in winter days 1.0
Outside in winter days 2.7 (1.8, 4.0)
a95% confidence intervals
After adjusting for the influence of household,10,11 cats origi-
nating from households with only one or two cats were found
to predate more frequently than cats from households with
more than two cats (adjusted χ2 = 6.44, P < 0.05). All variables
significant at P ≤ 0.25 (Table 1) were included in the initial
multivariable model. Neutering, number of cats in a household,
breed, not eating meat, time spent outside in summer nights
and winter days were significant factors in the final model
(Table 2). The Hosmer-Lemeshow statistic was 4.79 (P = 0.78)
and therefore the model was judged to fit the observed data
well.
Crossbred cats were more likely to be observed by owners to
predate than purebred cats (adjusted OR 1.9, CI 1.1, 3.2; P <
0.05 – see Table 2). Although the owners of purebred cats were
more likely to keep their cats inside, it was apparent after multi-
variable analysis that the difference between purebred and cross-
bred cats was not solely the result of their time spent outside.
There were no significant differences reported in predation
between male and female cats, although desexed cats were
reported to predate more frequently than entire cats (adjusted
OR 2.4, CI 1.4, 4.1; P < 0.005). Turner and Meister13 observed
that domestic cats with kittens were more likely to predate than
were other cats. However in the current study there were not
enough lactating queens to confirm this observation. Although
the average age of cats which were observed to predate was
slightly older (5.34 years) than cats which didn’t predate (5.17
years), the difference was not significant (P = 0.65).
Significantly, there were no statistical associations between
observed predation and a cat’s body condition, weight, number
of daily feeds, or whether snacks or treats were given by the
owner throughout the day. These findings agree with those
reported by others13,17 who considered that capturing and
killing of prey by domestic cats were independent of hunger.
Turner and Meister13 reported that although feeding of
domestic cats reduced the amount of time spent hunting, it did
not eliminate hunting. They found that female house cats on
farms in Switzerland spent about half as much time hunting as
did feral female cats. Interestingly, in the current study cats not
fed meat were more likely to be observed to predate than cats
Aust Vet J Vol 76, No 8, August 1998
fed meat (adjusted OR 1.6, CI 1.1, 2.3; P < 0.05 – Table 2). No
other dietary factors were found to influence predation (Table 1).
Others3,4,12,15 have concluded that cats are versatile predators
which can exploit a wide range of prey, and are able to readily
switch from one prey to another depending upon the avail-
ability and abundance of prey. Barratt6 reported that 64.9% of
all prey caught by house cats in Canberra were representatives of
10 mammalian species, 27% were birds (41 of the total 47
species were native) and only 6.7% were reptiles (7 species). The
introduced house mouse and black rat constituted 63% of all
prey, whilst the five native mammalian species caught composed
only 0.7% of all prey caught. In contrast Coman and Brunner15
reported that native mammals comprised 44% of the diet of
feral cats in a forest habitat in Victoria, whilst in agricultural
habitats no native mammals were predated. Martin and
colleagues16 demonstrated that small native mammals consti-
tuted a significant part of the diet of feral cats in pastoral areas
of Western Australia. It is apparent from these studies that prey
caught by both domesticated and feral cats is significantly influ-
enced by habitat. Although most animals caught by house cats
in Perth are likely to be introduced species, as corridors of
natural vegetation have been maintained in some areas of Perth,
it is possible the domestic cat can catch native species, with the
potential to hasten the regional extinction of these species.
Barratt6 concluded that in most suburban environments, locally
abundant species, which are predominantly introduced species,
comprised the primary prey of house cats. However, in less
disturbed environments, especially those adjoining new residen-
tial developments, predation by house cats may affect locally
abundant, but patchily distributed, populations of small native
mammals.
Although deterrents, such as bell collars, are at times recom-
mended to minimise a cat’s predatory forays, they are often not
successful.13,14 Imposing night curfews on cats could lessen
predation of native mammals which primarily are nocturnal in
nature. However, as a greater proportion of cats were outside
during the day in this study, night curfews alone are unlikely to
significantly reduce predation. The findings of this survey
suggest that crossbred, neutered cats not fed meat, which were
in a single or two cat household and which spend a significant
amount of time outside the house, are more likely to be
observed to predate. Significantly body condition did not influ-
ence reported predation. The major influence of domestic cats
on native fauna is likely to arise if these cats become feral. Feral
cats have been found in all regions of Australia, and they, in
conjunction with other introduced species and habitat destruc-
tion, have caused a significant decline in the diversity of
mammalian species present in some locations of Australia.1,2,18
This study found that a significant proportion of house cats in
Perth were observed to catch prey over a 12 month period.
However the direct role of the house cat in the predation and
extinction of native fauna remains unresolved. It is possible that
restriction of cats to only short periods of outdoor activity will
reduce predation of native fauna in locations where residential
areas and native habitats are in close proximity.
Acknowledgments
The author gratefully acknowledges the assistance of the final
year students of Murdoch University in 1994 and 1996 who
assisted with the interviews, and the cat owners who
contributed to the survey. The assistance of Professor Ian James
in interpretation of the analyses is gratefully acknowledged.
553
References 10. Donald A, Donner A. The analysis of variance adjustment to chi-square
tests when there is litter or herd correlation in surveys or field trials. Acta Vet
1. Short J, Smith AP. Mammal decline and recovery in Australia. J Mammal
1994;75:288-297.
Scand 1988;suppl 84:490-492.
11. Donner A. The comparison of proportions in the presence of litter effects.
2. Dickman CR, Pressey RL, Lim L, Parnaby HE. Mammals of particular conser-
vation concern in the western division of New South Wales. Biol Conserv
Prev Vet Med 1993;18:17-26.
12. Fennel C. Some demographic characteristics of the domestic cat population
1993;65:219-248.
in Great Britain with particular reference to feeding habits and the incidence of
3. Smith AP, Quin DG. Patterns and causes of extinction and decline in
the feline urological syndrome. J Small Anim Pract 1975;16:775-783.
Australian conilurine rodents [review]. Biol Conserv 1996;77:243-267.
13. Turner DC, Meister O. Hunting behaviour of the domestic cat. In: Turner
4. May SA, Norton TW. Influence of fragmentation and disturbance on the
DC, Bateson P, editors. The domestic cat: the biology of its behaviour.
potential impact of feral predators on native fauna in Australian forest ecosys-
Cambridge University Press, Cambridge, 1988:111-121.
tems. Wildl Res 1996;23:387-400.
14. Paton D. Loss of wildlife to domestic cats. In: Potter C, editor. The impact of
5. Tideman CR, Yorkston HD, Russack AJ. The diet of cats, Felis catus, on
Christmas Island, Indian Ocean. Wildl Res 1994;21:279-286.
cats on native wildlife. Proceedings of a workshop. Australian National Parks
and Wildlife Service, Canberra, 1991:64-69.
6. Barratt DG. Predation by house cats, Felis catus (L) in Canberra, Australia. 1.
15. Coman BJ, Brunner H. Food habits of the feral house cat in Victoria. J Wildl
Prey composition and preference. Wildl Res 1997;24:263-277.
7. Hosmer DW, Lemeshow S. Applied logistic regression. John Wiley & Sons,
Manage 1972;36:848-853.
16. Martin GR, Twigg LE, Robinson DJ. Comparison of the diet of feral cats
New York, 1989:82-134.
from rural and pastoral Western Australia. Wildl Res 1996;23:475-484.
8. Frankena K, Graat EAM. Multivariate analysis: logistic regression. In:
17. Fitzgerald BM. Diet of domestic cats and their impact on prey populations.
Application of quantitative methods in veterinary epidemiology. Noordhuizen
In: Turner DC, Bateson P, editors. The domestic cat: the biology of its
JPTM, Frankena K, van der Hoofd CM, Graat EAM. Wageningen Press,
Wageningen, 1997:137-178.
behaviour. Cambridge University Press, Cambridge, 1988:123-146.
18. Morton SR, Baynes A. Small mammal assemblages in arid Australia. Aust
9. Lemeshow S, Hosmer DW. The use of goodness-of-fit statistics in the devel-
opment of logistic regression models. Am J Epidemiol 1982;115:92-106.
Mammal 1985;8:159-169.
(Accepted for publication 5 June 1998)

OBITUARY
John Armstrong
ohn Armstrong passed away at his home near Mandurah, Western Australia, on 10 June 1998 after a long struggle
J with a brain tumour.
John’s veterinary career started when he was chosen as the first Veterinary Cadet to be appointed by the State
Government of Western Australia. Graduating with Honours from the University of Sydney Veterinary School in 1953, he
began as a pathologist at the Department of Agriculture Animal Health Laboratories. There under the direction of
eminent pathologist Dr HW Bennetts he developed an appreciation for research skills.
After marrying Gladys Cross, he went to Albany on the south coast, where as Government Veterinary Officer, he
applied his skills to the sheep and cattle industries. His next move was Kununurra, with overall responsibility for the East
and West Kimberley area where he established the structure that was to implement the final phases of the contagious
bovine pleuropneumonia eradication program in East Kimberley. He always maintained a special interest in the pastoral
industry of East and West Kimberley. This was to stand him in good stead later when the tuberculosis eradication
program commenced in these areas.
John was among Australian veterinarians sent to the UK to assist in the eradication of an outbreak of foot and mouth
disease. This gave him ‘hands on’ experience and an appreciation of the problems of eradicating an exotic disease.
From Kununurra John was transferred to Fremantle in 1967 as Veterinary Officer (Animal Quarantine).
His next appointment was Assistant CVO, administering field veterinary services until May 1977, then as Chief of the
Animal Health Division and Chief Veterinary Surgeon. Under his leadership a competent veterinary service was devel-
oped which included numerous veterinary officers trained under the veterinary cadetship scheme. His personal under-
standing of livestock industries and emphasis on cooperation and service was well respected within Western Australia.
As State veterinary representative on the Animal Health Committee, Chief Animal Quarantine Officers, National
Brucellosis and Tuberculosis Eradication Committees he was known for sound and reasoned judgement, especially in
times of crisis.
John served the AVA admirably and encouraged his staff to become involved in Association affairs. He presided over
the WA Division in 1969 and in his capacity as Chief, Animal Health Division was a member of the Veterinary Surgeon’s
Board, which he chaired over the six years almost up to his retirement in 1989. A Life Member of the AVA he was
awarded the Meritorious Service Medal in 1986.
After retirement from Government services John continued his interest in professional activities. He enjoyed the
company of local people in Fiji where he worked for two years with the Fijian Government. He was also associated with
the importation of ostriches as Government Veterinary Officer at Cocos Island, and as a private veterinarian.
John had the ability to encourage and cooperate. Those who knew him admired his conscientiousness and fairness.
These attributes contributed to continued friendships with his associates.
John maintained his love for the land and the sea. He enjoyed camping, sailing and a game of golf. Our sympathies
go to his wife Gladys and their children Grant, Ian and Sue.

Fred Wilkinson, Peter Lewis and Helen M Fairnie,

554 Aust Vet J Vol 76, No 8, August 1998