Content uploaded by Benjamin E. Lenth
Author content
All content in this area was uploaded by Benjamin E. Lenth on Apr 10, 2014
Content may be subject to copyright.
The Effects of Dogs on Wildlife Communities
Benjamin Lenth1*, Mark Brennan2, and Richard L. Knight3
February, 2006
Final research report submitted to:
City of Boulder Open Space and Mountain Parks
1 Graduate Degree Program in Ecology, Colorado State University
2 Boulder County Parks and Open Space
3 Department of Forest, Rangeland, and Watershed Stewardship, Colorado State University
*Corresponding author: ben.lenth@gmail.com
Abstract
Domestic dogs (Canis familiaris) are frequent visitors to open space areas, but
little is known about their ecological impacts. We studied the effects of dogs on wildlife
by comparing the activity levels of wildlife in areas that prohibit dogs, with areas that
allow dogs off-leash under “voice and sight” control. To measure wildlife activity both on
trail and up to 200 m off-trail, we used four methods: pellet surveys, scented tracking
plates, remote triggered cameras, and on-trail scat surveys. Additionally, in prairie dog
(Cyonomys ludocivianus) colonies we measured the distances of prairie dog burrows to the
nearest trail, and compared the density of prairie dog burrows between areas with and
without dogs. The presence of dogs along recreational trails correlated with altered
patterns of habitat utilization by several wildlife species. Mule deer (Odocoileus
hemionus) activity was significantly lower in proximity to trails in areas that allow dogs,
and this effect extended at least 100 m off-trail. Small mammals, including squirrels
(Sciurus spp.), rabbits (Sylviagus spp.), chipmunks (Eutamias spp.), and mice (Peromyscus
spp., Reithrodontomys spp., Onychomys spp., Zapus spp.), also exhibited reduced levels of
activity in proximity to trails in areas with dogs, and this effect extended at least 50 m off-
trail. Furthermore, the density of prairie dog burrows was lower within 25 m of trails in
areas that allow dogs. The presence of dogs also affected carnivore activity, although in
varying ways. Red fox (Vulpes vulpes) detections were higher in areas that allowed dogs,
and bobcat (Felis rufus) detections were lower. These findings have implications for the
management of natural areas regarding dog policies, particularly those that allow dogs off-
leash.
2
Introduction
Domestic dogs (Canis familiaris) are ubiquitous in American society, yet we know
relatively little about their ecology or interspecific interactions with wildlife. Numbering
approximately 400 million worldwide, with 61.6 million in America alone, domestic dogs
far outnumber all other canids combined (Coppinger and Coppinger 2001, U.S. Pet
Ownership and Demographics Sourcebook 2002). The vast majority of dogs in the U.S.
are owned as pets, though some range freely, and some are feral.
Outdoor recreation in North America is growing rapidly in popularity, and this
trend is expected to continue in the coming decades (Flather and Cordell 1995), with a
variety of impacts on wildlife (for a review see Knight and Gutzwiller 1995). Dogs often
accompany recreationists, necessitating policies that manage the perceived impacts of dogs
on protected lands. Often, leash laws restrict dogs to trails and limit their interaction with
other dogs, people, livestock and wildlife. Little scientific data exists to justify these
policies in terms of ecological impacts. When dogs accompany recreationists on trails,
their activity is usually concentrated along a relatively narrow corridor within a natural
setting, providing a unique opportunity to examine their effects on wildlife.
Increasing numbers of dogs in natural areas could have varied and complex
ecological effects, potentially influencing community dynamics in myriad ways including
indirect effects that could cascade down through trophic levels (Kay 1998). Dog’s closest
relative, the gray wolf (Canis lupus) has demonstrated disproportionate ecological impacts
that can cascade through ecosystems, affecting plants, animals, and ecological processes
(Smith et al. 2003). In addition to direct ecological effects through predation, increased
activity by wolves in Yellowstone has caused elk (Cervus elaphus) to avoid willow
thickets, where the limited visibility increases their susceptibility to ambush. Here, the
mere possibility of wolf predation can change patterns of habitat utilization by elk and
other ungulates, allowing willows to regain vitality and support a host of other species
(Ripple and Beschta 2004). Dogs are inefficient hunters, and could not regulate
ecosystems with the efficacy of wolves (Serpell 1995). However, dogs are avid chasers,
and through chasing could displace wildlife from their habitats, particularly when certain
species, such as deer, perceive dogs as predators and avoid areas where they could be
chased. Ungulates and herbivores are thus susceptible to disturbance by dogs, additive to
3
predation pressures already present by native carnivores such as mountain lions (Felis
concolor) and coyotes (Canis latrans).
Dogs and native carnivores rarely have opportunities to interact directly, but may
interact indirectly through scent marking (Bekoff 1979). Indirect interactions could either
displace or attract carnivores depending upon the perceived competition or threats dogs
offer. Carnivores determine and identify territories primarily through olfactory cues, and
recognize and avoid areas scent-marked by other individuals or packs (Gorman and
Towbridge 1989, Pal 2003). Recreational trails with abundant dog scent could appear to
carnivores to be linear dog territories, necessitating increased caution and possibly
deterring their activity. However, the interaction could have the opposite effect on certain
carnivore species. Carnivore activity is sometimes elevated along portions of territories,
such as boundaries, where they encounter novel stimuli and invest considerable time in
territorial surveillance and maintenance (Allen et al. 1999). Along a busy trail that
consistently provides novel scents from different dogs, carnivores could increase their
activity investigating and marking. Such opportunities may arise particularly often due to
carnivores’ propensity to travel along trails, which often present the easiest route in rough
terrain (Kevin Crooks, pers. comm.). Thus, as dogs could potentially attract or repel native
carnivores, the indirect effects of dogs on carnivore activity is difficult to predict and is
likely species-specific. The repercussions of carnivore disturbance can be
disproportionately important to the structure and function of ecosystems (Estes 1996).
In this study, we investigated the indirect effects of the presence of dogs on wildlife
activity. We wanted to know if the presence of dogs in natural areas influenced the
activity of wildlife, and if so, how far the effects of dogs extended from their location. To
do this, in 2004 and 2005 we compared the activity levels of mammals on two open spaces
in Boulder County that prohibit dogs, with two areas that allow dogs off-leash under
“voice and sight control”. We selected trails with roughly equal levels of recreational use,
and randomly located transects along these trails. To create indices of wildlife activity for
comparison across these two policies, we used four methods: track plates, pellet plots, scat
transects, and remote-triggered cameras. Dog activity was also sampled with these
methods to ascertain their spatial distributions on open space. Camera stations and scat
transects were performed on the trails only, and track plates and pellet plots were
4
performed within three distance categories perpendicular from the trail: 0-5 m, 50-100 m
(randomly selected), and 150-200 m (100 m beyond the second plot). Before describing
our study, we provide a comprehensive summary of the existing literature on dog ecology.
Previous literature on dog ecology
Studies of domestic dog ecology include both free-ranging and pet dogs in the
United States and overseas, and focus on direct interactions with wildlife and the
transmission of diseases. The earliest article expresses concerns with free-ranging dogs in
the southeastern U.S. and their capacity to harass and kill deer, spread rabies, and eat small
mammals and birds (Giles 1960). Later field studies in the same region demonstrated that
dogs chased deer (Sweeney et al. 1971), sometimes leading to deer mortality through direct
attacks or exhaustion (Corbett et al. 1971). Other researchers have found similar results
with pet dogs in Virginia (Gavitt et al. 1974), Indiana (Olson 1974), Arkansas (Gipson et
al. 1977) and Idaho (Lowry and McCarthur 1978). Concerns over such attacks led Denny
(1974) to conduct a survey of state wildlife agencies, agricultural agencies, and zoos,
regarding the costs of uncontrolled dogs. The responses were guesses rather than
systematic evaluations, but reported unusually high impacts (e.g., 5000 annual deer
moralities caused by dogs in Kansas). Gentry (1983) used these data are used as the
primary source for the sensationalist book When Dogs Run Wild. Other reports questioned
the severity of the impacts of dogs on deer populations (Perry and Giles 1971). However,
in Florida, the endangered Key deer (Odocoileus virginianus clavium) is particularly
susceptible to attacks by free-ranging dogs due to its small size (U.S. Fish and Wildlife
Service 1999). Also in the southeastern U.S., feral dogs have been observed to form
packs, behave aggressively, and eat small mammals, garbage, and vegetative material
(Scott 1973). Prey can include the endangered gopher tortoise (Gopherus polyphemus) in
Alabama (Causey and Cude 1978). Dogs are also resilient scavengers in urban areas as
Beck (1973) documented in a study of stray dogs in Baltimore, Maryland.
Overseas and in certain areas of the U.S., dogs that live in rural villages are
commonly free-ranging but derive much of their nutritional needs from people’s leftovers
and waste. In the southwestern U.S., numerous feral and abandoned dogs clustered their
activity and home ranges at garbage dumps (Daniels and Bekoff 1989). When villages
5
border protected areas, dogs are known to roam inside the borders of such areas, competing
with scavengers and carnivores in Zimbabwe (Butler and Du Toit 2002) and wolves (Canis
lupus) in Italy (Boitani 1983). Also within protected areas, dogs are known to eat wombats
(Vombatus ursinus) and wallabies (Macropus rufogriseus and Wallabia bicolor) in
Australia (Tiggs et al. 1984), marine iguanas (Amblyrhrynchus cristatus) in the Galapagos
Islands (Kruuk and Snell 1981, Barnett and Rudd 1983), wild turkeys (Meleagris
gallopavo) in the United States (Miller and Leopold 1992), capybaras (Hydrochoerus
hydrochoerus) in Venezuela (McDonald 1981), and the endangered golden langur
(Trachypithicus geei) in India and Bhutan (Medhi et al. 2004). In a nature reserve in
Isarel, dogs are known to chase Nubian ibexes (Cpara ibex nubiana) and rock hyraxes
(Procaviacapensis syriaca) (Brickner 2002). Dog home ranges average 57.8 ha in Italy
(Boitani and Ciucci, 1995), and can be over 900 ha for transient dogs in protected areas in
southeast Australia (Meek 1999).
Dog interactions with other carnivores are highly variable and sometimes fatal.
Anecdotes of dogs both attacking and being attacked by other carnivores are abundant.
Documentation includes wolves attacking dogs in Finland (Kojola et al. 2004) and Alaska
(Gipson 1983), dogs killing coyotes (Canis latrans) (Kamler et al. 2003), and coyotes, lynx
(Lynx canadensis), and leopards (Pantherinae panthera) killing dogs (Palomares and Karo
1999). In one bizarre report, a pack of black squirrels (Spermophilus dauricas) killed a
stray dog in Russia’s Maritime Territory (BBC News 2005). In Alaska, feral dogs
vigorously scent-marked spots previously marked by wolves and coyotes (Gipson 1983).
Dogs are also known to interbreed with coyotes and wolves (Mengel 1971).
For wildlife populations, the greatest consequences from interactions with dogs
may come from the role dogs play as a vector for the transmission of disease. In a review,
Sime (1999) notes that dogs are a potential vector for canine distemper, rabies, parvovirus,
plague, giardia, and muscle cysts. Dogs have been implicated in transmitting rabies to two
species of jackals (Canis spp.) in Zimbabwe (Rhodes et al. 1997), canine distemper to lions
(Panthera leo) in Tanzania (Cleveland et al. 2000) and African wild dogs (Lycaon pictus)
in Kenya (Alexander and Appel 1994), and rabies, canine distemper and parvovirus to the
highly endangered Ethiopian wolf (Canis simensis) (Laurenson et al. 1998). In Ethiopia,
dogs also interbreed with wolves, diluting their genetic stock and further imperiling their
6
survival (Laurenson et al. 1997). In Zimbabwe, Butler et al. (2004) note that free-ranging
dogs are the most common carnivore in a communal land bordering a reserve, where they
are often preyed upon by native carnivores, creating positive conditions for disease
transmission. Recently, threatened gray wolves (Canis lupus) in Yellowstone National
Park have suffered an epidemic of parvovirus or canine distemper, believed to have been
transmitted by dogs (Smith 2006, Doug Smith, pers. comm.).
Examinations of dog ecology along recreational trails are most relevant to this
study. MacArthur et al. (1979, 1982) found that dogs increased the flushing distance and
heart rates of bighorn sheep (Ovis canadensis) in Canada. In Great Britain, dogs increased
the sensitivity of breeding golden plovers (Pluvialis apricaria) in proximity to trails
(Yalden and Yalden 1989). Along trails in Switzerland, dogs increased the flushing
distances for marmots (Marmota marmota), particularly when on a very long leash
(Mainini et al. 1993). In California, the presence of dogs along trails correlated with lower
detections of raptors and egrets (Abraham 2001). Three previous studies have been
performed in Boulder County: Bekoff and Meaney (1997) found very little interaction
between dogs and wildlife; Bekoff and Ickes (1999) found that dogs increased prairie dog
(Cynomys ludovicianus) vigilance and often chased them; and Miller et al. (2001) found
mule deer (Odocoileus hemionus), vesper sparrows (Pooecetes gramineus), western
meadowlarks (Sturnella neglecta), and American robins (Turdus migratorius) showed
elevated sensitivity and flushing distances when dogs accompanied hikers, particularly
when off-trail.
Methods
Study site selection
The City and County of Boulder, Colorado, have protected over 45,810 ha
(113,200 acres) of open space lands, with over 320 km of recreational trails. Most of the
lands managed by the City of Boulder Open Space and Mountain Parks (OSMP) allow
dogs to be off-leash under “voice and sight control”. A portion of trails require leashes,
and some prohibit dogs all together. Open spaces managed by Boulder County Parks and
Open Space (BCPOS) require dogs to be on-leash on most trails, and prohibit dogs from
some trails. In northern Boulder County, BCPOS manages Heil Valley Ranch (1,993 ha)
7
and Hall Ranch (1,297 ha), both of which prohibit dogs and which were chosen as study
sites. For comparison, we then chose two study sites that allow dogs off-leash under
“voice and sight control”. Based on GIS analysis and the recommendations of OSMP and
BCPOS staff, study sites were selected to match the following criteria as closely as
possible.
1. Ecological characteristics: Elevation ranging between 1,615 and 2,590 m in Ponderosa
pine (Pinus ponderosa) and Douglas fir (Pseudotsuga mensiesii) forests, with meadows
and slopes less than 40º on all aspects.
2. Equivalent visitation: Trails were chosen using best estimates and expert opinions of
levels of recreational visitation to each trail. OSMP trails were roughly matched to
visitation levels for Heil Valley Ranch (48,890 annual visits in 2003) and Hall Ranch
Ranch (74,112 visits) (BCPOS staff). Previous visitation estimates for OSMP trails
were ranked on a scale of 1 to 5, with 1 as “very low” and 5 as “very high”. Study
trails ranked 3 or 4 on this scale (Steve Armstead, pers. comm.). This assumption was
tested empirically in the course of the study.
This process resulted in the selection of OSMP land east of the Flatirons along the Mesa
Trail corridor from Skunk Canyon south. This area was divided in half to create two sites:
Skunk Canyon south to the Big Bluestem trail, including Shanahan Mesa (~1203 ha), and
from Big Bluestem south, including Doudy Draw and forested portions of Flatirons Vista
(~927 ha). Areas in the study sites that were within 300 m from roads, houses, and other
developments were excluded from sampling. Different types of recreationists—hikers,
mountain bikers, and equestrians—were assumed to have similar impacts to wildlife. This
assumption is based on documentation that these recreational groups cause similar flushing
responses for large herbivores in Antelope Island State Park, Utah (Taylor and Knight
2003).
Field methods
At each site, sampling locations were located randomly along trails, and spaced a
minimum of 500 m apart. Pellet plots and track plates were performed along transects that
ran perpendicular to the trail on one side, away from other trails and steep slopes.
Transects consisted of three plots or track plates within three distance categories
8
perpendicular to the trail: 0-5 m, 50-100 m (randomly selected), and 150-200 m (100 m
beyond the 2nd distance). Cameras and scat surveys were performed on-trail only. For all
field methods, sampling effort was equivalent across policies.
1. Pellet plots: Activity and habitat utilization of herbivores (deer and rabbits) were
measured using pellet-group counts in 100 m2 circular plots (Bennett et al. 1940, Neff
1968, Collins 1981). One set of plots was cleared summer 2004 and re-checked 12
months later; a second set of plots was cleared early summer 2005 and re-checked three
months later.
2. Track plates: Mammal activity was determined by the use of scented track plates,
which consisted of a 1 m2 aluminum plate, sprayed with talc isolated in ethanol, with a
scented carnivore lure (Carmen’s Pro’s Choice and Canine Call, Sterling Trap and Fur,
Sterling, Iowa) in a sponge in the middle (Linheart and Knowlton 1975, Zielinski et al.
1996, Belant 2003, Sargeant et al. 2003). While the primary targets of these plates
were mid-sized carnivores, track plates were also used in the detection of small
mammals (Drennan et al. 1998, Glennon et al. 2002). Plates were checked daily for
three rain-free nights. Each day, tracks were photographed and identified using field
guides (Halfpenny 2001), plates were cleaned and re-sprayed with talc, and lure was
reapplied. The three nights were not considered independent and were therefore
collapsed into a single data point. Not all small mammal prints were identifiable to
species, so plausible species were listed.
3. Remote-triggered cameras: TrailMaster® brand remote-triggered cameras were used to
monitor a variety of mammal species with unambiguous identifications (Kucera and
Barrett 1993, Cutler and Swann 1999, but see Rice 1995). In addition to triggering
pictures, Trailmaster units also record the timing of “events”, which occur when the
infrared trigger is broken—usually because a person, dog or animal passes by. In 2004
we used five cameras paired with track plates, both on and off-trail, but with
disappointing results. In 2005 we restricted camera placement to recreational trails
only and had 10-11 cameras set up at a time at two sites simultaneously. Cameras were
set to be active for two weeks (mean=13.5 nights), at a height of .25-.75 m to detect
medium-sized carnivores. Vegetation was cleared from the beam path to avoid false
events. Camera locations were baited using a mixture of commercial carnivore lures,
9
and were spaced a minimum of 500 linear m apart. Cameras were revisited every 2-5
days to check on their function, replace film and batteries, and reapply the lure. Due to
high numbers of visitors on trails during the day, cameras were only active from
approximately 8:00 PM to 8:00 AM. Infrared trigger sensitivity was set between 2 to
4, with a delay of 3 minutes to avoid repeated photos of the same animal. Cameras
were hidden using camouflage materials, and were cable-locked to trees with small
signs explaining their purpose.
4. On-trail scat surveys: Because circular pellet plots were not detecting enough carnivore
scat, and as we frequently observed carnivore scat along trails, we initiated on-trail scat
surveys. Scat surveys were conducted continuously from July 2004 through October
2005 and were conducted on a sub-set of study trails. Trails were walked and scat was
identified using field guides (Halfpenny 2001) and cleared from the trail, in two-week
intervals (mean=17 days). Pictures and samples were taken for further analysis. To
test the accuracy of the scat identifications, the DNA from 50 scat samples was
identified to species (Paxinos et al. 1997, Mills et al. 2000).
5. Prairie dog study: We identified seven recreational trails in Boulder County that occur
adjacent to prairie dog colonies, each with a relatively “unbounded” geography that did
not limit the movement of prairie dogs or the location of their burrows (Johnson and
Collinge 2004). Three of these trails were in areas that did not allow dogs, and four
were in areas that allow dogs off-leash under “voice and sight control”. We selected
segments of these trails where prairie dog burrows existed continuously from the trail
up to at least 200 m from the trail on one side. We identified active burrows by
observation of prairie dog activity, including fresh scat, evidence of digging, tracks,
clear burrow openings, and prairie dogs themselves (Powell et al. 1994). With a laser
rangefinder, we measured the perpendicular distance from the trail to each active
burrow within this trail segment, up to 200 m from the trail.
Statistical analyses
We established α = 0.1 a priori for all analyses to minimize the possibility of Type
II errors (Holling and Allen 2002). All effects and interactions included in models were
10
selected a priori as relevant to the biology of the response organisms. In all cases we
eliminated effects to include only those significant at α = 0.1.
1. Pellet plots: Data were converted to a density of pellet piles per hectare sampled, and
were square-root transformed to normalize variance. For both rabbits and mule deer, a
mixed model analysis of covariance (ANCOVA) was performed using PROC MIXED
in Statistical Analysis Software (SAS, 2005). The response was modeled to include the
fixed effects of dog policy, site, date, and distance from trail (three distance
categories), a continuous covariate of an estimation of hourly visitation to that segment
of trail, a random effect of transect location, and relevant interactions between these
effects. When the overall F-test was significant for the effect of distance categories or
the interaction of distance categories and dog policy, pairwise comparisons between
these categories were made with the least-significant-difference method (Ott &
Longnecker 2001).
2. Track plates: The three nights of track plate data were combined into one data point
with whole number counts of each species detected; species not detected were counted
as “zeros” in the dataset. These count data were then converted to an index of
detection frequency (I) by dividing the count of detections per species (X) by the
number of trap nights (n): I=X/n. This dataset was then analyzed using a mixed model
ANCOVA for binomial data using PROC GLIMMIX (SAS, 2005). The model effects
and their associated F-tests were the same as in the pellet plot analysis described above.
Response variables included domestic dogs and a combination of small mammals
including squirrels, rabbits, chipmunks, mice and voles.
3a. Remote-triggered camera event data: Trailmaster events are non-specific recordings of
the breaking of the infrared beam, recorded as number of events per hour. Events were
presumed to bear a consistent relationship with visitation along a trail, and the majority
of events were assumed to be human and pet trail visitors. Trailmaster cameras
frequently malfunctioned (maxed out events, transmitter out of alignment, batteries
died, etc.), so the event dataset had variable gaps throughout the two weeks they were
active at each location. We did not adjust for double counts (when single visitors
passed monitors more than once) or for multiple counts (when visitors passed more
than one monitor on a single outing). Because visitation peaked on weekends, we
11
normalized the dataset to give equal weight to weekday and weekend visitation using
the following adjustment: 1) all incomplete days were dropped from the dataset; 2)
separate hourly visitation averages were computed for weekdays and weekends; 3) data
were recombined to normalize weekday and weekend hourly visitation using the
formula: Visitation (normalized) = (5/7)*weekday visitation + (2/7)*weekend
visitation. Normalized hourly visitation data for each trail segment was combined by
dog policy and by site. Tests of equivalent visitation across dog policies were
conducted using a Sattherwaite t-test, which allows for unequal variance between the
samples (PROC TTEST in SAS).
Such tests of equivalent visitation refer to humans, not dogs, but visitors to OSMP
trails have been estimated to be accompanied by at least one dog 30% of the time
(Mertz 2002). Thus, we expected that a proportion of events were caused by dogs,
additive to the events triggered by humans. To test only visitation rates by humans
across policies, we adjusted the event data for OSMP sites by assuming that 30% of
visitors were indeed accompanied by dogs, and so the events should equal 130% of
human-triggered events. Thus, the proportion of events excluding dogs to total events
was 100% / 130% = 0.769; we multiplied this number by all OSMP visitation estimates
to gain an adjusted hourly visitation estimate, and re-tested the equivalence of
visitation using a t-test (PROC TTEST in SAS).
3b. Remote triggered camera photo data: Similar to the track plate data, an index (I) of
activity for photos was calculated by dividing the number of photos (X) of any species
by the number of nights (n) the Trailmaster cameras were active: I=X/n. This dataset
was square-root transformed to normalize the variance, then analyzed using a mixed
model ANCOVA with the fixed effects of dog policy and site, the continuous covariate
of visitation, and the random effect of transect location (PROC MIXED in SAS).
When the overall F-test for site was significant, pairwise comparisons between sites
were performed. Only data from 2005 was used in this analysis. Also, we
hypothesized that dog activity was correlated with average visitation levels, so we
tested this by fitting a linear regression between these variables (PROC REG in SAS).
4. On-trail scat surveys: For each survey, trails were broken into 1-km segments, and data
were converted to a density of scats per km, for each species detected. Data were
12
square-root transformed to stabilize the variance. DNA results indicated variable
precision with scat identification, so all native carnivore scat was combined, excluding
black bears, which were not confused with other scats. Data for each trail segment
were averaged per month, and these data were then modeled using a mixed model
repeated-effects analysis of variance (ANOVA), with interactions. We expected that
adjacent months and adjacent kilometer segments would be correlated, so we
considered a first-order autoregression matrix (AR-1) on both these variables to model
our assumption that the data was both temporally and spatially autocorrelated (Ott and
Longnecker 2001). We tested the appropriateness of the AR(1) covariance structure
with asymptotic Wald Z-tests on the covariance parameter estimates within each model
and eliminated the autoregressive covariance structure when it did not have significant
influence on the model (SAS 2004). We initially considered the fixed effects of dog
policy, site, month, the presence of a trailhead, and relevant 2- and 3-way interactions
of these effects, plus kilometer segment location as a random effect and month as a
repeated measures variable, and dropped insignificant effects. Visitation estimates
were not available for these trail segments. The final model was used for overall F-
tests on each effect.
5. Prairie dog burrow location: Distance from each prairie dog burrow to the trail was
pooled by dog policies. To normalize for the variable widths of the trail segments
running through the prairie dog colonies that were measured, the raw number of
burrows extending fixed distances away from the trail were divided by the width of the
trail segment to create a density of burrows per meter of trail: density = # burrows /
meters of trail. This density was calculated for the complete dataset, extending up to
200 m from the trail, and for truncations of the dataset focusing progressively closer to
the trail: 100 m, 50 m, 25 m, and 10 m. Within each set distance from the trail, we
compared the mean burrow densities across dog policies using Sattherwaite t-tests to
deal with unequal variance (PROC TTEST in SAS).
13
Results
Visitor use
A t-test on all normalized event data indicated overall visitation between policies to
be significantly higher on OSMP lands where dogs are allowed (df=55.8, t=2.13,
p=0.0375). However, a second t-test on adjusted visitation data (presuming 30%
accompaniment by dogs) indicated that overall visitation did not differ between policies
(df=60, t=0.57, p=0.5738), thus meeting the assumption of equal visitor use across
policies. Despite meeting this assumption, visitation estimates for individual trail
segments were used as covariates in all analyses except on-trail scat surveys and prairie
dog burrow densities.
Pellet plots
Pellet plot data indicated that the activity of mule deer and rabbits were both lower
along trails in areas that allow dogs and that the strength of this effect was influenced by
the distance from the trails. For both 12-month and 3-month summer pellet plots, dog
policy and distance from the trails were significant predictors of deer activity (Fig. 1). In
areas that allow dogs, deer were significantly less active within 0-5 m of trails than 50-100
m of trails (year: F44=-3.28, p=0.002; summer: F28=-2.25 p=0.0328), and within 50-100 m
than 150-200 m from trails (year: F44=-3.97, p=0.0003; summer: F28=-4.07 p=0.0003).
However, even in areas that prohibit dogs, deer activity was lower within 5 m of trails than
50-100 m from trails (year: F44=-5.38, p<0.0001; summer: F28=-3.45 p=0.0018), but deer
activity did not differ between 50-100 m and 150-200 m from trails (year: F44=1.49,
p=0.1435; summer: F28=-0.22 p=0.8247). Comparing between dog policies, where dogs
are allowed deer were significantly less active at two distances from the trails: within 5 m
(year: F44=1.89, p=0.0657; summer: F28=1.73 p=0.0938) and within 50-100 m (year:
F44=3.56, p=0.0009; summer: F28=2.93 p=0.0067). Within 150-200 m of trails deer
activity did not differ by dog policy (year: F44=-0.80, p=0.4268; summer: F28=-0.90
p=0.3736).
Rabbit activity was also significantly lower along trails in areas that allow dogs, but
trends varied between the one-year and three-month plots (Fig. 2). For the one year plots,
dog policy (F22=4.93, p=0.0369) and distance from trail (F44=2.89, p=0.0665) were
14
significant predictors of rabbit activity, but over the summer only dog policy was
significant (F12=9.56, p=0.0093). In areas that allow dogs, rabbit activity was similarly
low within 5 m of trails and within 50-100 m of trails (year: F44=-0.89, p=0.3779) but was
significantly higher at 150-200 m of trails (year: F44=-2.10, p=0.0419).
0
500
1000
1500
2000
2500
3000
3500
4000
.
a
c c
b
c
b
0
2000
4000
6000
8000
10000
12000
0-5 m
50-100 m
150-200 m
0-5 m
50-100 m
150-200 m
-------Dogs off-leash------- -------No dogs-------
a
b
b
c
c
c
Fig. 1. 12-month and 3-month deer pellet densities by dog policy and distance from trail.
Different letters above columns indicate significant differences (p<0.1) between 1)
adjacent distance categories within policies, and 2) the same distance category between
policies, based on a square root transformation of the data presented.
Pellet
p
iles/ ha ±1 s.e.m.
3 months
12 months
15
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
a
a
a
b
b b
0
1000
2000
3000
4000
5000
0-5 m
50-100 m
150-200 m
0-5 m
50-100 m
150-200 m
-------Dogs off-leash------- -------No dogs-------
a
a
b
b
b
b
Fig. 2. 12-month and 3-month rabbit pellet densities by dog policy and distance from trail.
Different letters above columns indicate significant differences (p<0.1) between 1)
adjacent distance categories within policies, and 2) the same distance category between
policies, based on a square root transformation of the data presented.
In areas that prohibit dogs, distance from the trail did not influence rabbit activity.
Comparing across dog policies, areas that allow dogs had lower rabbit activity over the
course of a year at two distance categories from trails: within 5 m (year: F44=2.49,
p=0.0168), and within 50-100 m of trails (year: F44=2.32, p=0.0246), but not within 150-
Pellet
p
iles/ ha ±1 s.e.m.
3 months
12 months
16
200 m of trails (year: F44=0.73, p=0.4697). This trend was not present with the summer
plots.
Track plates
Track plate data indicated that in areas that allow dogs off-leash, dogs traveled up
to 85 m from trails, but the vast majority of dog detections were within 5 m of trails. Small
mammal tracks were not identifiable to all species, but included rabbits (Sylvilagus spp.),
squirrels (Sciurus spp.), chipmunks (Eutamias spp.), ground squirrels (Spermophilus spp.),
mice (Peromyscus spp., Reithrodontomys spp., Onychomys spp., Zapus spp., Mus spp.),
voles (Clethrionomys spp., Phenacomys spp., Microtus spp.), and rats (Rattus spp.,
Neotoma spp.) (Fitzgerald et al. 1994). Together, the activity of these species was
inversely correlated with the presence of dogs (Fig. 3). Within 5 m of trails, small
mammal activity was significantly lower in areas that allowed dogs than in areas that
prohibit dogs (T94=3.36, p=0.0011). Within areas that allow dogs, small mammals were
less active within 5 m of trails than 50 m or further from trails (T94=-3.63, p=0.0005).
Remote-triggered cameras
Dog policy was a significant predictor of activity levels for a variety of species,
including dogs, red foxes (Vulpes vulpes), bobcats (Lynx rufus), and all native carnivores
combined (Fig. 4). Other carnivores detected included grey foxes (Urocyon
cinereoargenteus), black bears (Ursus americanus), mountain lions (Felis concolor),
striped skunks (Mephitis mephitis), coyotes (Canis latrans) and raccoons (Procyon locor).
As expected, dogs were photographed significantly more on trails that allow dogs than
trails that prohibit dogs (F1,57 =77.19, p<0.0001). Dog activity was also somewhat
correlated with visitation (R2=0.218, F1,29=1.82, p=0.008). On trails that allow dogs,
activity was also higher for native carnivores (F1,60 =16.09, p=0.0002), including native
canids (F1,57 =18.06, p<0.001) and especially red foxes (F1,57 =13.97, p=0.0004). Red
foxes constituted 49% of all native carnivore photos; 58% when excluding black bears.
Dog activity was inversely correlated with bobcat activity (F1,57 =5.89, p=0.0184). Unlike
the results from our pellet data, mule deer activity on trails was not different between
policies (F1,60=0.16, p=0.6929).
17
Track plate detection frequencies
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0-5 m 50-100 m 150-200 m 0-5 m 50-100 m 150-200 m
--------------Dogs-------------- --------------No Dogs--------------
Detection frequency ±1 s.e.m
Dogs
Small Ma mma ls
a
b
b
b
b
b
Fig. 3. Track plate detection frequencies for dogs and small mammals. Different letters
above columns indicate significant differences between distance categories and policies for
small mammals (p<0.1) based on a square root transformation of the data presented.
On-trail scat surveys
On-trail scat surveys also showed dog policy to be a strong predictor of dog activity
(F1, 34.5=61.32, p<0.0001) (Fig. 5). Similar to the photo data, these trails also had higher
levels of native carnivore activity (F1, 38.2=19.01, p<0.0001). Dog policy was not the sole
significant predictor of scat density, however. For both dogs and native carnivores,
seasonal variation was considerable, and the month of the survey influenced scat density
(dogs: F11,230=2.18, p=0.0161; carnivores: F11,231=3.17, p=0.0005), as did the interaction
of month with dog policy (dogs: F10,215=3.85, p<0.0001; carnivores: F10,209=3.68,
p=0.0001). The presence of a trailhead within a kilometer segment increased dog activity
(F1,24.5=15.12, p=0.0004) and decreased carnivore activity (F1,42.2=5.95, p<0.019) (Fig.
6). For dogs, the strength of the trailhead effect depended on dog policy (F1,34.5=14.26,
p=0.0006) and also on the month (F21,207=1.86, p=0.0149). The lower density of carnivore
scat near trailheads was also dependent upon month (F11,231=1.77, p=0.0603). For
18
carnivores, significant variation in activity levels also existed between the four study sites
(F2, 27.9=9.51, p=0.0004).
For both dogs and native carnivores, modeling the temporal autocorrelation
between adjacent months using an AR(1) autocovariance structure significantly reduced
the variance in the final model (dogs: Z= 3.03, P=0.0024; carnivores: Z= -1.65, P=0.0993).
Dogs
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Dogs No d ogs
a
b
Native carnivores
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
Dogs No dogs
a
b
b
Red Fox
0.00
0.05
0.10
0.15
0.20
0.25
Dogs N o do gs
b
Bobcats
0.00
0.01
0.02
0.03
0.04
0.05
0.06
Dogs No dogs
a
b
Fig. 4. Photo detection frequencies (photos / night) of four groups of animals by dog
policy. Different letters above columns indicate significant differences (p<0.1) based on a
square root transformation of the data presented.
Prairie dog burrow densities
Dog policy did not significantly predict prairie dog burrow density within 200 m
(t2.3=1.14, p=0.3574), 100 m (t3.65=1.07, p=0.3487), or 50 m of the trail (t3.17 =1.72,
p=0.1796). However, in areas where dogs are prohibited there were significantly higher
Detection frequencies, ±1 s.e.m.
Dog policy
a
a
19
densities of prairie dog burrows within 25 m (t4.59=3.78, p=0.0151) and 10 m of trails
(t4.65=3.86, p=0.0136).
0
0.5
1
1.5
2
Dogs No Dogs
Scats / km trail ±1 s.e.m.
Dog s
Native Carnivores
Fig. 5. On-trail scat densities by dog policy for dogs and combined native carnivores,
including coyotes, red foxes, grey foxes, bobcats and mountain lions. Different letters
above columns indicate significant differences (p<0.1) based on a square root
transformation of the data presented. Capital letters represent statistical comparisons for
native carnivores, and lower case letters refer to comparisons for dogs.
Discussion
The presence and spatial use patterns of dogs in open spaces of Boulder County
correlated with altered patterns of wildlife activity. As expected, dog activity was focused
along recreational trails, but was not restricted to the trail itself. In areas both with and
without dogs, the density of mule deer pellets indicated that deer were less active close to
trails. In areas that prohibit dogs, deer activity was decreased within 50 m of trails; this
represents the baseline recreational disturbance on mule deer by hikers, equestrians and
mountain bikers. Where dogs are allowed the disturbance was greater: deer activity was
decreased within 100 m of trails. Photo data indicated that trails are not boundaries for
deer movement, as frequencies of deer photos on trails did not differ between dog policies.
But pellet densities are a more sensitive measure of deer activity and habitat utilization
because they indicate areas where deer are able to pause, rest, and bed down—areas where
deer are not disturbed, but relaxed.
a
b
A
B
20
0
0.5
1
1.5
2
2.5
3
3.5
trailhead uptrail trailhead uptrail
-------Dogs------- -------No Dogs-------
Scats / km trail ±1 s.e.m.
Dogs
Native Carn ivo res
Fig. 6. On-trail scat densities by dog policy, within 1 km of trailheads or uptrail, for dogs
and combined native carnivores, including coyotes, red foxes, grey foxes, bobcats and
mountain lions. Different letters above columns indicate significant differences (p<0.1)
based on a square root transformation of the data presented. Capital letters represent
statistical comparisons for native carnivores, and lower case letters refer to comparisons
for dogs. No trailhead effect was found in areas that prohibit dogs.
These findings are consistent with the behavioral response of mule deer to dogs.
Miller et al. (1999), also in Boulder County, observed that mule deer were more sensitive
and flushed at greater distances from pedestrians when they were accompanied by a dog,
and that the behavioral response of deer was greatest when pedestrians with dogs were off-
trail. Dogs off-leash frequently travel off-trail—up to 85 m in this study. Wildlife that are
sensitive to recreational disturbance are generally most sensitive to unpredictable spatial
patterns of disturbance (MacArthur et al. 1982, Knight and Cole 1995), whereas
predictable spatial activities, such as activity restricted to trails, allows wildlife to habituate
to the disturbance (Whittaker and Knight 1999). The spatial behavior of off-leash dogs is
unpredictable, and when dogs wander off-trail they are likely to elicit flushing responses
from deer, even if they do not give chase (Miller et al. 2001). Decreased detections of deer
pellets within 100 m of trails in areas that allow dogs indicate these behavioral responses
translate to altered patterns of habitat utilization, and make the trail corridor less suitable
for mule deer. The same is true for small mammals including squirrels, rabbits, chipmunks
a
b
A
B
21
and mice, though these species’ activity resumes normal levels within 50 m from trails in
areas that allow dogs.
Similar altered patterns of habitat utilization were found in prairie dog colonies
adjacent to recreational trails. Bekoff and Ickes (1999) demonstrated prairie dogs behave
more vigilantly in the presence of dogs, even when they are not chased. Our results
indicate that this behavioral response can translate into altered patterns of burrow
placement, which is less frequent within 25 m of trails in areas that allow dogs than in
areas that prohibit dogs.
On-trail scat surveys and remote-triggered photos indicated several relationships
between the activity of dogs and carnivores. Along trails that allow dogs, native carnivore
activity—predominantly red foxes—was higher than in areas that prohibit dogs. Dog scat
densities were highest and native carnivore scats were lowest within a kilometer of
trailheads. The opposite was true up-trail, where dog scat densities were lower, and native
carnivore scat densities, as well as photo detection frequencies, were higher. Trailheads
did not influence carnivore activity where dogs were prohibited. Several observations may
explain these trends. Most dogs defecate very soon after arriving at a trail, and many
visitors do not walk dogs much beyond the trailhead. However, the dogs that are allowed
to continue along the course of the trail continue to urinate and scent-mark areas previously
marked by other dogs and carnivores, creating what may seem to native carnivores to be a
dog territory. Foxes and other carnivores, especially canids, may be attracted to these
areas to patrol, maintain, and possibly defend their territories (Henry 1977). We may
speculate that up-trail, elevated detections of red foxes (and other carnivores) where dogs
are allowed hints at increased vigilance and territorial maintenance along trails. As this
trend does not hold within a kilometer of trailheads, perhaps this elevated carnivore
response only occurs up to a certain threshold of dog and human activity that is exceeded
near trailheads, where dog scat is abundant and carnivore scat is rare. However, these
speculations cannot be confirmed through our methods or through any index of activity
(Anderson 2001).
Very low frequencies of bobcat photos on trails in areas that allow dogs indicate
that bobcats may also be sensitive to the presence of dogs, and may avoid areas they
frequent. Bobcats are secretive and elusive from humans, though they are considered
22
resilient to broad-scale land use change (Woolf and Hubert 1998). In Colorado, coyotes
are an important natural predator of bobcats (Fitzgerald et al. 1994). Bobcats may perceive
dogs as canid threats similar to coyotes, so it is possible that the presence of dogs may
deter them from otherwise suitable habitat.
Several alternative explanations of the data exist, and warrant further study. First,
this study was designed to analyze spatial patterns of wildlife activity, but the influence of
recreationists, including dogs, could lead instead to alterations in temporal activity
patterns. Such is the case in southern California, where high levels of recreational
visitation correlate with altered temporal patterns of wildlife activity (George and Crooks,
in press). Thus, the spatial activity patterns we studied may not illustrate the complete
picture of how wildlife manages to co-exist with high levels of human recreation and dogs
in a limited area.
Second, landscape-scale geographic differences exist between our no-dog sites
(BCPOS in northern Boulder County) and dog sites (OSMP in southern Boulder County).
OSMP sites are in closer proximity to an urban area (Boulder), which could allow
subsidization of human-commensal species such as red foxes and black bears, both of
which were detected more frequently on these sites. Also, the location of trails within the
landscape may differ. In particular, the trails on Heil Valley Ranch (no dogs) are limited to
one plateau with little habitat diversity and lower levels of wildlife activity than other parts
of the same open space (Given 2001).
Conclusion
We studied the indirect interactions of dogs with wildlife and documented that
trails in areas that allow dogs have a wider area of influence on mule deer and small
mammals, including prairie dogs, rabbits, squirrels and mice. This area of influence
surrounding trails is potentially unsuitable habitat for these species. In these areas, visitors
are less likely to view these wildlife species. Carnivores, such as red foxes and bobcats,
may also alter their behavior in these areas.
23
Management Recommendations
Policies regulating types of recreational use on trails should be considered within a
regional context, including trails managed by Boulder County Parks and Open Space, the
City of Boulder Open Space and Mountain Parks, the Arapahoe-Roosevelt National Forest,
Rocky Mountain National Park, Eldorado Canyon State Park, and other public and private
lands. The types of recreational uses allowed on these trails influence the rates of
recreational visitation and their associated ecological impacts. Trails that allow dogs or
mountain bikes lead to higher recreational use, and these types of recreation may not
always be entirely compatible (BCPOS 2004). Within such a regional context, trails that
are kept dog-free could protect against the demonstrated ecological impacts that dogs have
on wildlife communities and could facilitate wildlife viewing opportunities for trail users.
Recommendations for further research
To further explore the effects of dogs on wildlife communities, we recommend
before-after control-impact studies with any new trails that are created. Particularly strong
inference could be gained through a crossover study design where trail policies are
switched temporarily from dogs to no-dogs, and vice-versa, with target species monitored
as desired. Also, studies that can estimate population sizes, such as through a mark-
recapture study using DNA or photos, would allow determination of relative detection
frequency on trails for individual animals. Then it would be possible to know if carnivores
spend relatively more time on trails more when dogs are present, which would indicate
elevated investment towards territorial maintenance and defense.
Acknowledgements
We are grateful to many individuals for their assistance with this study. For advice
on study design, we would like to thank Kevin Crooks, Cameron Ghalambor, Mark
Gershman, Gary White, William Andelt, and Buffy Hastings. Colin Talbert assisted with
GIS, and Marty Colòn taught us the art of tracking. Field work was performed tirelessly
by Art Roberts, Kelly Matheson, Renée Culver, Sam Clapper, Tim Sichmeller, Heather
Hill, Laria Hill, Justin Severyn, Colin Talbert, Buffy Hastings, Dave Hoerath and John
Justus. Trailmaster cameras were generously provided by Robert Skorkowsky, of Hahn’s
24
Peak/Bears Ear Ranger District, Medicine Bow National Forest, and Roland Wostl, of the
Colorado Department of Transportation, via Carron Meaney. Phil Chapman provided
assistance with statistical analysis. Wildlife scat DNA analysis was generously performed
by Emily Ruell, in the laboratory of Kevin Crooks. This study would not have been
possible without funding from Boulder County Parks and Open Space, the City of Boulder
Open Space and Mountain Parks, and the Boulder County Nature Association.
Literature cited
Abraham, K. 2001. Interactions between dogs and wildlife in parks on the Berkeley
Marina. Unpublished Report, submitted to Berkeley Parks and Recreation. Online
at: http://ist-socrates.berkeley.edu/~es196/projects/2001final/Abraham.pdf
Alexander, K.A. and M.J. Appel. 1994. African wild dogs (Lycaon pictus) endangered by a
canine distemper epizootic among domestic dogs near the Masai Mara National
Reserve, Kenya. Journal of Wildlife Diseases 30:481-485.
Allen, J. J., M. Bekoff, and R. L. Crabtree. 1999. An observational study of coyote (Canis
latrans) scent-marking and territoriality in Yellowstone National Park. Ethology
105:289-302.
Anderson, D. R. 2001. The need to get the basics right in wildlife field studies. Wildlife
Society Bulletin 29:1294-1297.
Barnett, B.D. and R.L. Rudd. 1983. Feral dogs of the Galapagos Islands: impact and
control. International Journal on Studies on Animal Problems 4:44-58.
B.B.C. News—British Broadcasting Corporation. 2005. Russian squirrel pack kills dog.
December 1. Online at: http://news.bbc.co.uk/2/hi/europe/4489792.stm.
Beck, A.M. 1973. The ecology of stray dogs. York Press, Baltimore, MD.
Bekoff, M. 1979. Scent-marking by free-ranging domestic dogs—olfactory and visual
components. Biology of Behavior 4:123-139.
Bekoff, M. and R. Ickes. 1999. Behavioral interactions and conflict among domestic dogs,
black-tailed prairie dogs, and people in Boulder, Colorado. Anthrozoos 12:105-110.
Bekoff, M. and C. A. Meaney. 1997. Interactions among dogs, people, and the
environment in Boulder, Colorado: A case study. Anthrozoos 10:23-31.
Bennett, L.J., P.F. English and R. McCain. 1940. A study of deer populations by use of
pellet-group counts. Journal of Wildlife Management 4:399-403.
Boitani, L. 1983. Wolf and dog competition in Italy. Acta Zoologica Fennica 174:259-264.
Boitani, L. and P. Ciucci. 1995. Comparative social ecology of feral dogs and wolves.
Ethology, Ecology and Evolution 7:49-72.
Boulder County Parks and Open Space. 2004. Recreation conflict at six Boulder County
Parks and Open Space Properties: a baseline study. September 10.
Brickner, I. 2002. The impact of domestic dogs (Canis familiaris) on wildlife welfare and
conservation: a literature review. With a situational summary from Israel.
Unpublished Report. Available online at:
http://www.tau.ac.il/lifesci/zoology/members/yom-tov/inbal/dogs.pdf
25
Butler, J. R. A. and J. T. du Toit. 2002. Diet of free-ranging domestic dogs (Canis
familiaris) in rural Zimbabwe: implications for wild scavengers on the periphery of
wildlife reserves. Animal Conservation 5: 29-37.
Butler, J.R.A., J.T. du Toit and J.A. Bingham. 2004. Free-ranging domestic dogs (Canis
familiaris) as predators and prey in rural Zimbabwe: threats of competition and
disease to large wild carnivores. Biological Conservation 115:369-378.
Causey, M.K. and C.A. Cude. 1978. Feral dog predation of the Gopher Tortoise in
Southeast Alabama. Herpetological Review 9:94-95.
Cleaveland, S., M.G.J. Appel, W.S.K. Chalmers, C. Chillingworth, M. Kaare and C. Dye.
2000. Serological and demographic evidence for domestic dogs as a source of
canine distemper virus infection for Serengeti wildlife. Veterinary Microbiology
72:217-227.
Collins, W.B. 1981. Habitat preferences of mule deer as rated by pellet-group distributions.
Journal of Wildlife Management 45: 969-972.
Coppinger, R. and L. Coppinger. 2001. Dogs: a startling new understanding of canine
origin, behavior, and evolution. Scribner, New York.
Corbett, R. L., R. L. Marchinton and C. E. Hill. 1971. Preliminary study of the effects of
dogs on radio-equipped deer in a mountainous habitat. Proceedings of the
Southeastern Association of Game and Fish Commmissioners. 25:67-71.
Cutler, T.L. and D.E. Swann. 1999. Using remote photography in wildlife ecology: a
review. Wildlife Society Bulletin 27:571-581.
Daniels T.J. and M. Bekoff. 1989. Spatial and temporal resource use by feral and
abandoned dogs. Ethology 81:300-312.
Denny, R.N. 1974. The impact of uncontrolled dogs on wildlife and livestock. The Texas
Journal of Agriculture and Nature. 39:257-291.
Drennan, J.E., P. Beier and N.L Dodd. 1998. Use of track stations to index abundance of
sciurids. Journal of Mammalogy 79:352-359.
Estes, J. A. 1996. Predators and ecosystem management. Wildlife Society Bulletin 24:390-
396.
Fitzgerald, J. P., C. A. Meaney, and D. M. Armstrong. 1994. Mammals of Colorado.
University of Colorado Press, Niwot, CO.
Flather, C. H. and H. K. Cordell. 1995. Outdoor recreation: historical and anticipated
trends. In R.L. Knight and K.G. Gutzwiller, editors. Wildlife and recreationists:
coexistence through research and management. Island Press, Washington, D.C.
Gavitt, J.D., R.L. Downing and B.S. McGinnes. 1974. Effects of dogs on deer reproduction
in Virginia. Proceedings of the annual conference of the Southeastern Association
of Fish and Wildlife Agencies 28:532-539.
Gentry, C. 1983. When dogs run wild: the sociology of feral dogs and wildlife. McFarland
& Company, Inc., Jefferson, N.C.
George, S. and K Crooks. In press. The Effects of Recreation on Large Mammals in an
Urban Nature Reserve. Biological Conservation.
Gill, J.A., K. Norris and W.J. Sutherland. 2001. Why behavioral responses may not reflect
the population consequences of human disturbance. Biological Conservation
97:265-268.
Gipson, P.S. and J.A. Sealander. 1977. Ecological relationships of white-tailed deer and
dogs in Arkansas. Pages 3-16 in Philips, R.C. and C. Jonkel, eds. Proceedings of
26
the 1975 Predator Symposium. Bulletin of Montana Forest Conservation
Experimental Station. Missoula: University Press.
Gipson, P.S. 1983. Evaluation and control implications of behavior of feral dogs in interior
Alaska. Pages 285-294 in D.E. Kaukemen, ed., Vertebrate pest control and
management materials: fourth symposium. ASTM STP 817. American society for
testing and materials: Philadelphia, Pennsylvania.
Given, B. 2001. Utilization of scent-station surveys to determine the response of predators
to recreation on Heil Ranch. Unpublished Report to Boulder County Parks and
Open Space, Small Grant Program.
Glennon, M.J., W.F. Porter and C.L. Demers. 2002. An alternative field technique for
estimating diversity of small-mammal populations. Journal of Mammalogy
83:734-742.
Gorman, M. L. and B. J. Trowbridge. 1989. The role of odor in the social lives of
carnivores. Pages 57-88 in J. L. Gittleman, ed. Carnivore behavior, ecology and
evolution. Cornell University Press, Ithaca, New York.
Halfpenny, J.C. 2001. Scats and tracks of the Rocky Mountains, second edition. Globe
Pequot Press, Guilford, CT.
Henry, J.D. 1977. Use of urine marking in scavenging behavior of red foxes, Vulpes
vulpes. Behaviour 61:82-106.
Holling, C.S. and C.R. Allen. 2002. Adaptive inference for distinguishing credible from
incredible patterns in nature. Ecosystems 5:319-328.
Johnson, W.C. and S.K. Collinge. 2005. Landscape effects on black-tailed prairie dog
colonies. Biological Conservation 115:487-497.
Kamler, J.F., W. Ballard and P.S. Gipson. 2003. Occurrence of feral dogs (Canis lupus
familiaris) in Northwest Texas: an observation. The Texas Journal of Agriculture
and Natural Resources 16:75-77.
Kay, C. E. 1998. Are ecosystems structured from top-down or bottom-up: a new look at an
old debate. Wildlife Society Bulletin 26:484-498.
Knight, R.L. and D. Cole. 1995. Wildlife responses to recreationists. Pages 51-69 in R.L.
Knight and D. Cole, ed. Wildlife and recreationists: coexistence through
management and research. Island Press, Washington D.C.
Knight, R. L. and K. J. Gutzwiller. 1995. Wildlife and recreationists: coexistence through
management and research. Island Press, Washington, D.C..
Kojola, I., S. Ronkainen, A. Hakala, S. Heikkinen and S. Kokko. 2004. Interactions
between wolves (Canis lupus) and dogs (C. familiaris) in Finland. Wildlife Biology
10:101-105.
Kruuk, H. and H. Snell. 1981. Prey selection by feral dogs from a population of marine
iguanas (Amblyrhynchus cristatus). Journal of Applied Ecology 18:197-204.
Kucera, T.E. and R.H. Barrett. 1993. The TrailMaster® camera system for detecting
wildlife. Wildlife Society Bulletin 21:505-508.
Laurenson, K., F. Shiferaw and C. Sillero-Zubiri. 1997. Disease, domestic dogs and the
Ethiopian wolf: the current situation. Pages 32-42 in C. Sillero-Zubiri, D.W.
MacDonald and the IUCN/SSC Canid Specialist Group. The Ethiopian wolf—
status survey and conservation action plan. IUCN, Gland, Switzerland.
27
Laurenson, K., C. Sillero-Zubiri, H. Thompson, F. Shiferaw, S. Thirgood and J. Malcolm.
1998. Disease as a threat to endangered species: Ethiopian wolves, domestic dogs
and canine pathogens. Animal Conservation 1:273-280.
Lowry, D. A. and K. L. McArthur. 1978. Domestic dogs as predators on deer. Wildlife
Society Bulletin 6:38-39.
MacArthur, R.A., R.H. Johnston and V. Geist. 1979. Factors influencing heart-rate in free-
ranging bighorn sheep—physiological approach to the study of wildlife
harassment. Canadian Journal of Zoology 57:1020-2021.
MacArthur, R.A., V. Geist and R.H. Johnston. 1982. Cardiac and behavioral responses of
mountain sheep to human disturbance. Journal of Wildlife Management 46:351-
358.
Mainini, B., P. Neuhaus and P. Ingold. 1993. Behavior of marmots (Marmota marmota)
under the influence of different hiking activities. Biological Conservation 64:161-
164.
McDonald, D.W. 1981. Dwlindling resources and the social behavior of capybaras
(Hydrochoerus hydrochaeris). Journal of Zoology 194:371-391.
Medhi, R., D. Chetry, P.C. Bhattacharjee and B.N. Patiri. 2004. Status of Trachypithecus
geei in a rubber plantation in Western Assam, India. International Journal of
Primatology 25:1331-1337.
Meek, P.D. 1999. The movement, roaming behaviour and home range of free-roaming
domestic dogs, Canis lupus familiaris, in coast New South Wales. Wildlife
Research 26:847-855.
Mengel, R.M. 1971. A study of dog-coyote hybrids and implications concerning
hybridization in Canis. Journal of Mammalogy 52:316-336.
Mertz, S. 2002. Compliance with Leave No Trace frontcountry principles—a preliminary
examination of visitor behavior. Report for City of Boulder Open Space and
Mountain Parks.
Miller, J.E. and B.D. Leopold. 1992. Population influences: predators. Pages 119-128 in
J.G. Dickson, ed. The wild turkey: biology and management. Stackpole books,
Mechanicsburg, PA.
Miller, S.G., R.L. Knight and C.K. Miller. 2001. Wildlife response to pedestrians and
dogs. Wildlife Society Bulletin 29:124-132.
Mills, L.S., K.L. Pilgrim, M.K. Schwartz and K. McKelvey. 2000. Identifying lynx and
other North American felids based on MtDNA analysis. Conservation genetics
1:285-288.
Neff, D.L. 1968. The pellet-group count technique for big game trend, census, and
distribution: a review. Journal of Wildlife Management 32:597-614.
Olson, J.C. 1974. Movements of deer as influenced by dogs. Indiana Department of
Natural Resources, Job Progress Report Project W-26-12-5: 1-36.
Ott, R. L. and M. Longnecker. 2001. An introduction to statistical methods and data
analysis. Fifth edition. Duxbury, Pacific Grove, CA.
Pal, S. K. 2003. Urine marking by free-ranging dogs (Canis familiaris) in relation to sex,
season, place and posture. Applied Animal Behavior Science 80:45-59.
Palomares, F. and T.M. Caro. 1999. Interspecific killing among mammalian carnivores.
The American Naturalist 153:492-508.
28
Paxinos, E., C. McIntosh, K. Ralls and R. Fleischer. 1997. A noninvasive method for
distinguishing among canid species: amplification and enzyme restriction of DNA
from dung. Molecular ecology 6:483-486.
Perry, M.C. and R.H. Giles. 1971. Free running dogs. Virginia Wildlife 32:17-19.
Powell, K.L., Robel, R.J., Kemp, K.E., Nellis, M.D., 1994. Aboveground counts of black-
tailed prairie dogs—temporal nature and relationship to burrow entrance density.
Journal of Wildlife Management 58:351-355.
Rhodes, C.J., R.P.D. Atkinson, R.M. Anderson and D.W. Macdonald. 1997. Rabies in
Zimbabwe: reservoir dogs and the implications for disease control. Phil.
Transactions of the Royal Society of London B. 353:999-1010.
Rice, C.G. 1995. TrailMaster® camera system: the dark side. Wildlife Society Bulletin
23:110-113.
Ripple, W.J and R.L Beschta. 2004. Wolves and the ecology of fear: Can predation risk
structure ecosystems? Bioscience 54:755-766.
Sargeant, G.A., D.H. Johnson and W.E. Berg. 2003. Sampling designs for carnivore scent-
station surveys. Journal of Wildlife Management 67:289-298.
SAS Institute. 1999. SAS/STAT user’s guide. Version 8.0. SAS Institute, Cary, North
Carolina.
Scott, M.D. and K. Causey. 1973. Ecology of feral dogs in Alabama. Journal of Wildlife
Management 37:253-265.
Serpell, J. 1995. The domestic dog: it’s evolution, behavior and interactions with people.
Cambridge: University Press.
Sime, C. A. 1999. Domestic dogs in wildlife habitats. Pages 8.1-8.17 in G. Joslin, and H.
Youmans, coordinators. Effects of recreation on Rocky Mountain wildlife: A
review for Montana. Committee on effects of recreation on wildlife, Montana
chapter of The Wildlife Society.
Smith, D.W., R.O. Peterson and D.B. Houston. 2003. Yellowstone after wolves.
Bioscience 53:330-340.
Smith, D.W. 2006. Interviewed by John Nielson. All Things Considered, National Public
Radio, 17 Jan. 2006. Online at:
www.npr.org/templates/story/story.php?storyId=5160919
Sweeney, J. R., R. L. Marchinton and J. M. Sweeney. 1971. Responses of radio-monitored
white-tailed deer chased by hunting dogs. Journal of Wildlife Management 35:707-
716.
Taylor, A. R. and R. L. Knight. 2003. Wildlife responses to recreation and associated
visitor perceptions. Ecological Applications 13:951-963.
Tiggs, B., H. Brunner and J.M. Cullen. 1984. The food of fox, dog and cat in
Croajingalong National Park, South-Eastern Victoria. Australian Wildlife Research
11:491-499.
U.S. Fish and Wildlife Service. 1999. South Florida multi-species recovery plan. Atlanta,
Georgia.
U.S. Pet Ownership and Demographics Sourcebook. 2002. American Veterinary Medical
Association, Schaumberg, IL.
Whittaker, D., and R. L. Knight. 1999. Understanding wildlife responses to humans.
Wildlife Society Bulletin 26:312–317.
29
Woolf, A. and G.F. Hubert, Jr.. 1998. Status and management of bobcats in the United
States over three decades: 1970s-1990s. Wildlife Society Bulletin 26:287-293.
Yalden D.W. and P.E. Yalden. 1990. The sensitivity of breeding golden plovers Pluvialis
apricaria to human intruders. Bird Study 36:49-55.