Citation
Instraspecific and interspecific communication on prairie dog colonies

Material Information

Title:
Instraspecific and interspecific communication on prairie dog colonies black tailed prairie dogs and burrowing owls
Creator:
Bryan, Rebecca Dyckes
Publication Date:
Language:
English
Physical Description:
1 electronic file : illustrations (some color), maps ;

Subjects

Subjects / Keywords:
Animal communication ( lcsh )
Black-tailed prairie dog -- Vocalization ( lcsh )
Burrowing owl ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Review:
ABSTRACT: This thesis is a compilation of research from two different yet related projects. The common thread in these projects is alarm calling and communication on black-tailed prairie dog (Cynomys ludovicianus) colonies. I considered both intra-specific communication: how black-tailed prairie dogs communicate with each other, and inter-specific communication: how another species, the Western Burrowing Owl (Athene cunicularia), benefits from prairie dog communication. The first chapter consists of background information pertinent to communication and alarm calling, and introduces both study species. The second and third chapters are formatted as stand-alone manuscripts reporting the results of my field studies. I have also included appendices of additional data analysis information that, while important, for various reasons was not included in the prepared manuscripts.
Bibliography:
Includes bibliographical references.
Statement of Responsibility:
by Rebecca Dyckes Bryan.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
856152459 ( OCLC )
ocn856152459

Auraria Membership

Aggregations:
Auraria Library
University of Colorado Denver

Downloads

This item has the following downloads:


Full Text
INTRASPECIFIC AND INTERSPECIFIC COMMUNICATION
ON PRAIRIE DOG COLONIES
BLACK TAILED PRAIRIE DOGS AND BURROWING OWLS
by
Rebecca Dyckes Bryan
B.A., University of Colorado Denver, 1997
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Science
Biology
2012


This thesis for the Masters degree by
Rebecca D. Bryan
has been approved for the
Department of Integrative Biology
by
Michael Wunder, Advisor
Michael Greene
Laurel Hartley
Date
April 26, 2012


Bryan, Rebecca D. (M.S., Biology)
Interspecific and intraspecific communication on prairie dog colonies: black-tailed
prairie dogs and burrowing owls.
Thesis directed by Assistant Professor Michael Wunder.
This thesis is a compilation of research from two different yet related projects. The
common thread in these projects is alarm calling and communication on black-tailed
prairie dog (Cynomys ludovicianus) colonies. I considered both intra-specific
communication: how black-tailed prarie dogs communicate with each other, and inter-
specific communication: how another species, the Western Burrowing Owl {Athene
cunicularia), benefits from prairie dog communication. The first chapter consists of
background information pertinent to communication and alarm calling, and introduces
both study species. The second and third chapters are formatted as stand-alone
manuscripts reporting the results of my field studies. I have also included appendices of
additional data analysis information that, while important, for various reasons was not
included in the prepared manuscripts.
The form and content of this abstract are approved. I recommend its publication.
Approved: Michael Wunder
m


DEDICATION
I dedicate this work to my husband and best friend, George.


ACKNOWLEDGMENTS
It is impossible to acknowledge every single person who contributed to the
completion of this project, but I do want to name a few very important individuals
without whom I would never have gotten this far. First and foremost I want to say thanks
to my parents, Martha and Doug, who have always known what I am capable of
achieving even when I cant see it. And to the best big sister ever, Melissa: thanks for
giving me much needed teaching and learning advice. Huge thanks to George for playing
the role of supportive husband in every way; I could never have done this without you.
Also, I want to thank my committee members: Mike Greene, for reinforcing my love of
animal behavior (its so cool!); Laurel Hartley, for encouraging me, offering ideas and
advice, and most importantly, friendship; and last, but certainly not least, Mike Wunder,
for introducing me to this project idea and letting me run with it even when I sometimes
ran into dead ends. As much as I struggled with it at the time, I will always appreciate
that you gave me the freedom to find those dead ends for myself, and that you were there
to help me find my way back out.
I would also like to acknowledge everyone in the Ecology Group who helped me
brainstorm ideas and let me vent when necessary. Susan Spaulding at BCPOS and A.E.
Nash at Colorado Reptile Humane Society made the first years studies possible; the
snake sheds made all the difference! Thanks to the Audubon Society of Greater Denver
for funding, the CPER (especially Mary Ashby, who went above and beyond to help me
find owl nests there) and the National Forest Service for enabling me to do research on
the Pawnee. Also, Marina Corcetti and Zara Hickman: your help in the field was
invaluable and I could not have accomplished half as much without you. Thank you.
v


TABLE OF CONTENTS
CHAPTER
I. INTRODUCTION.........................................................1
Prairie Dogs......................................................1
Burrowing Owls....................................................4
Alarm Calling and Eavesdropping...................................6
Figures and Tables...............................................10
II. CHARACTERISTICS OF BLACK-TAILED PRAIRIE DOG (Cynomys
ludovicianus) ALARM CALLS IN RESPONSE TO SIMULATED PREDATOR
THREATS.................................................................13
Abstract.........................................................13
Background.......................................................14
Materials and Methods............................................15
Study Area....................................................16
Field Methods.................................................16
Data Analysis.................................................18
Results..........................................................18
Duration......................................................19
Frequency.....................................................19
IQR bandwidth.................................................20
Average Power.................................................20
Discussion.......................................................20
Figures and Tables...............................................25
III. BURROWING OWLS EAVESDROP ON ALARM CALLS OF BLACK TAILED
PRAIRIE DOGS............................................................28
Abstract.........................................................28
vi


Introduction
29
Methods............................................................31
Field Methods..................................................33
Data Analysis..................................................34
Analysis I Latency to Response............................34
Analysis II Intensity of Response.........................35
Results............................................................35
Latency to Response Times......................................36
Intensity of Response..........................................36
Discussion.........................................................37
Figures and Tables.................................................41
APPENDIX I: PRAIRIE DOG ALARM CALL ANALYSIS..............................45
Results from Balloon Trials........................................45
APPENDIX II: BURROWING OWL RESPONSE ANALYSIS.............................47
Playback Trial Locations...........................................47
REFERENCES...............................................................50
vii


LIST OF TABLES
Table
I. 1 Cross-taxa heterospecific eavesdropping examples............................12
II. 1 Structural attributes of alarm spectrograms................................27
III. 1 Ethogram of Scored Burrowing Owl Behaviors................................44
III.2 Median Times to Alert Response.............................................44
AI. 1 Structural attributes of alarm spectrograms................................46
AIL 1 Order of sound cue presentation effect on time to alert behavior...........48
AIL2 Pair-wise Wilcoxon test, median time to alert behavior......................48
AIL 3 Order of sound cue presentation effect on response intensity...............49
AII.4 Chi-squared comparison of response strength by stimulus....................49
viii


LIST OF FIGURES
Figure
1.1 Historic Range of Prairie Dogs...............................................10
1.2 Historic Range of Burrowing Owls.............................................11
11.1 Map of Prairie Dog Study Area, Boulder County, CO...........................25
11.2 Spectrogram measurements....................................................26
11.3 Spectrogram of Snake Responses..............................................27
III. 1 Map of Burrowing Owl Study Area, Pawnee National Grasslands, CO...........41
III. 2 Latency to Response results...............................................42
III. 3 Intensity of Response Results.............................................43
IX


CHAPTER
I. INTRODUCTION
This project examines both interspecific and intraspecific communication that
takes place on black-tailed prairie dog (Cynomys ludovicianus) colonies. Black tailed
prairie dogs emit alarm calls in the presence of threats, and these alarms are specific to
the type or immediacy of the perceived threat (Smith et al. 1977, Hoogland 1995).
Relatives and distant kin are alerted to danger and are therefore more likely to be able to
avoid it. This enhances the fitness of the alarm caller, since its offspring and relatives are
more likely to avoid the threat (Hoogland 1995). I hypothesize that these alarm calls
serve an unintended purpose as well, alerting another species living on the colony to
threats. Western Burrowing owls (Athene cunicularia) have high nesting success on
black-tailed prairie dog colonies. Although there are multiple hypotheses about why this
is the case, (Desmond et al. 2000, Beradelli et al. 2010) one unexplored idea is that
burrowing owls attend and respond to the alarm calls black-tailed prairie dogs emit. This
would enhance the birds survival on colonies and offer a further explanation of increased
nesting success.
Prairie Dogs
Black-tailed prairie dogs (hereafter, prairie dogs) are the most common of the 5
prairie dog species found in North America, which include Gunnisons prairie dogs
(Cynomys gunnisoni), white-tailed prairie dogs (Cynomys leucurus), Mexican prairie
dogs (Cynomys mexicanus), and Utah prairie dogs (Cynomys parvidens). The black-tailed
prairie dog is an herbivorous rodent of the sciurid family. Adult weights average about 1
1


kg, and full grown black-tailed prairie dogs stand 30-35 cm high. They are native to the
western Great Plains of North America, and their range extends from southern Canada
through Montana, the Dakotas, Wyoming, Colorado, Utah, Arizona, and Texas (Figure
1.1).
Prairie dogs are diurnal fossorial mammals that live in colonies. Average colony
area is 20 to 60 hectares, and colonies can contain from 25 to upwards of one million
individuals (Bishop and Culbertson 1976). Prairie dogs live in territorial family groups
known as coteries within the colonies. Coteries normally consist of one male, 2 to 4
breeding females, and juvenile offspring (Hoogland 1995). Unlike other rodent species,
prairie dogs breed only once a year. Females give birth to litters of 1 to 7 pups in the
spring between the months of March and June, and care of pups is shared among the
female coterie members (Hoogland 1995). Young emerge from the burrows for the first
time in mid- to late- May. Predators of adult and juvenile prairie dogs include American
badgers (Taxidea taxus), swift foxes ( Vulpes velox), coyotes (Canis latrans), black-footed
ferrets (Mustela nigripes), golden eagles (Aquila chrysaetos), and red-tailed hawks
(Buteo jamaicensis). Bullsnakes (Pituophis catenifer) and prairie rattlesnakes (Crotalus
viridis) will predate newborn and juvenile prairie dogs (Halpin 1983).
The geographic range of the black-tailed prairie dog has decreased dramatically
over the past 200 years. Fossil records indicate that prairie dogs had a relatively stable,
large range from 40,000 years ago until the early 19th century (Lomolino and Smith 2001,
Forrest 2005). Since then their geographic range has shrunk considerably: the historic
occupied habitat of the black-tailed prairie dog is estimated at 35 to 40 million hectares,
but is currently estimated to be less than 1 million hectares (USFWS 2011). This is a
2


direct result of habitat conversion for agriculture, human land development and
consequent eradication efforts (Miller et al. 1994), and the introduction of plague
(Yersiniapestis) (Lomolino and Smith 2001, Antolin 2002). The decline of active black
tailed prairie dog colonies has not only affected this species, but several others as well,
including small mammals and birds such as mountain plover (Charadrius montanus) and
burrowing owl (Athene cunicularia) (Miller et al. 1994, Wuerthner 1997, McDonald et al.
2004, Knopf 2008, Shipley et al. 2008, Cully et al. 2010).
Prairie dogs can be considered a keystone species (sensu Power 1996): their
influence on plant and animal communities is disproportionately large compared to their
relative abundance (Kotliar et al. 1999, Kotliar 2000, but see Stapp 1998). Prairie dogs
also serve the function of ecosystem engineers (sensu Jones et al. 1994) because they
modify their environment in ways that influence resource availability to other organisms.
Prairie dog colonies have lower vegetation height, lower standing biomass, and more bare
ground than the surrounding grassland (Whicker and Dettling 1988). Prairie dog colonies
provide a large prey base for carnivores and offer shelter and nesting habitat for various
species of birds and small mammals (Barko 1997, Wuerthner 1997, Cully et al. 2010). In
short- and mixed-grass prairies of North America, prairie dog towns are associated with
higher bird species richness and abundance than un-colonized areas (Agnew et al. 1986,
Barko et al. 1999). Prairie dog grazing on colony sites supplies better visibility of insect
prey for insectivorous bird species (Barko 1997, Schneider et al. 2006), better access to
seed for granivorous birds (Agnew et al. 1986), better detection of predators for ground-
nesting birds (Knopf 2006), and the prairie dogs themselves serve as a food source for
raptors (Barko et al. 1999).
3


Burrowing Owls
Western Burrowing Owls {Athene cunicularia) are small ground-dwelling raptors.
Burrowing owl nesting habitat consists primarily of open treeless areas on plains and
prairies with short vegetation (McDonald et al. 2004). They often nest in prairie dog
burrows in active prairie dog colonies on the Great Plains. The burrowing owl breeding
range extends to the middle of Alberta and Saskatchewan in the north, and their winter
range extends south to Honduras. They can be found in Florida and Mexico year-round
(Figure 1.2). A burrowing owl home range is around 2.5km2 (Poulin et al. 2011).
Burrowing owls begin nesting as early as March, but in northern latitudes nesting may
not begin until May. Only female burrowing owls incubate the eggs; males of the mated
pair hunt and forage for both adults and young hatchlings. Burrowing owl young are
altricial and completely dependent on parental care until they are 2 months old (Poulin et
al. 2011). Juvenile survival is around 57%, and mortality rates for these birds are highest
when they first leave the nest before they can fly and when they first become
independent from adult owls (Davies and Restani 2006).
Burrowing owl predators include badgers, long-tailed weasels {Mustela frenata\
skunks (Conepatus spp., Spilogale spp.), coyotes, and domestic dogs {Cants familiaris)
and cats (Felts domesticus). Other raptors (hawks, owls, and falcons) also hunt burrowing
owls. Burrowing owl eggs and chicks can be predated by prairie rattlesnakes and
bullsnakes. During the nesting season, burrowing owls will dive or fly at mammalian
predators near the nest. To avoid avian predators, burrowing owls will escape into their
burrows (Poulin et al. 2011). Young burrowing owls will respond to a nest disturbance
with a rattlesnake-like buzzing vocalization which is thought to deter predators (Owings
4


et al. 2002). If a nest is predated, burrowing owls will disperse and re-nest in a different
burrow, as long as it is early enough in the breeding season (Catlin and Rosenberg 2008).
Burrowing owls will return to the same nesting site year after year if they are successful
there (McDonald et al. 2004). Within the nesting site, owls may make use of more than
one burrow; nearby burrows can act as escape holes or alternate nest sites if the original
site becomes unusable (Winchell 1994), and young disperse to the nearby burrows after
their first 2 to 3 weeks of life (Poulin et al. 2011).
Burrowing owls show a preference for prairie dog burrows on active colonies. A
study on National Grasslands on the Great Plains showed owls present on 69% of active
prairie dog colonies, but they were present on only 11% of inactive colonies (Sidle et al.
2001). Burrowing owls are crepuscular, nocturnal, or diurnal depending on the time of
year. During breeding season, they are most active at dawn and dusk, but are alert and
active during the day as well (Poulin et al. 2011).
In Colorado, burrowing owls are found more commonly on than off prairie dog
colonies (Smith and Lomolino 2004, Plumpton and Lutz 1993). Burrowing owls have
shown an overall declining trend in population in recent years (Desmond et al. 2000), and
the Colorado Division of Wildlife (2011) lists burrowing owls as a state threatened
species. Because of their close association on with prairie dogs, the decline in prairie dog
geographic range may have large effect on how well owls survive. Conservation efforts
for burrowing owls are closely tied with those for the black-tailed prairie dog.
5


Alarm Calling and Eavesdropping
In many species living in social groups, individuals rely on collective vigilance to
increase their foraging time while simultaneously decreasing their risk of predation. One
or a few group members monitor their surroundings for potential threats while others
forage (Bell et al., 2009). Scanning the surroundings is synchronized between group
members (Pays et al. 2007) such that each group member has an opportunity to forage
without having to maintain strict individualized vigilance. Vigilant animals emit alarms
when they detect a risk. Alarm signals can be visual, auditory, or chemical depending on
the species. For species using audible alarm calls, the alarm signals vary in specificity;
some calls are urgency-based only, and others encode information about the particular
type of threat (review: Macedonia and Evans 1993; Furrer and Manser 2009). A
collective-vigilance and alarm-calling system minimizes the energy required by any
single individual to avoid predators (Beuchamp and Ruxton 2007).
Prairie dogs are one such species that uses a sophisticated auditory alarm system
to warn other colony members of potential danger. In the presence of perceived danger,
prairie dogs emit an anti-predator bark (Hoogland 1995). This anti-predator bark varies in
rate and in acoustic structure depending on the predator threat and, possibly, the mode of
approach (Hoogland 1995, Fredrickson and Slobodchikoff 2007). Barks vary in
frequency, duration, and harshness depending on the situation (Smith et al. 1977). Single
barks are uncommon, while continuous bouts of chirping can go on for several minutes at
a time. Black-tailed prairie dogs also display a jump-yip behavior which can occur in
territorial disputes and chases, in response to low risk stimuli, or in the absence of any
apparent stimulus at all (Smith et al. 1976). Senkiw (2007) hypothesized that the jump-
6


yip may also serve to provide information about vigilance among group members,
alerting foraging prairie dogs in the vicinity to the callers vigilance status.
Hollen and Radford (2009) define three important aspects of animal alarm calling
behavior as follows: 1) call production, where a call having a specific acoustic structure
is created, 2) call usage, the context in which the call is used, and 3) call responses, the
reactions of others to the call. Normally all three of these aspects apply within a single
species. For example, if a prairie dog (1) produces an alarm call (2) in the presence of an
aerial predator, (3) other prairie dogs will react with the appropriate avoidance behavior,
whether they have themselves seen the predator or not (Kiriazis and Slobodchikoff 2006).
However, there are cases where aspects (1) and (2) will lead to a response in a different
species. This is known as heterospecific eavesdropping.
Heterospecific eavesdropping has been studied extensively in several bird species
(Magrath et al. 2007, 2009), and evidence of eavesdropping on another species alarm
calls has been noted in non-avian species as well. Evidence of social and non-social
animals eavesdropping on avian alarm calls has been documented in several species of
mammals and some reptiles (Table 1.1). Banded mongooses (Mungos mungo) increase
their vigilance in response to plover (Vanellus spp.) alarms (Mueller and Manser 2008),
and red squirrels (Sciurus vulgaris), increase vigilance and escape responses when they
eavesdrop on alarm calls given by Eurasian jays (Garralus glandarius) (Randier 2005).
Schmidt et al. (2008) demonstrated that eastern chipmunks (Tamias striatus) eavesdrop
on tufted titmouse (Baeolophus bicolor) alarms, and show a stronger anti-predator
response to that indirect cue than they show to a direct predator cue. Rasa (1983) showed
that dwarf mongooses (Helogale undulata rufula) forage with two species of hornbill
7


(Tockus deckeni and T. flavirostris), and benefit from eavesdropping on those birds
alarm calls. This relationship is so important that the mongooses often will not begin to
forage without hornbills present, since their presence significantly enhances mongoose
survival (Rasa 1983). Foraging activity decreases and vigilance increases in Gunthers
dik-dik antelope (Madoqua guentheri) when they hear go-away bird (Corythiaxoides
leucogaster) alarm calls (Lea et al. 2008). Two species of iguanas, the Galapagos marine
iguana (Amblyrhynchus cristatus) and the Madagascan spiny tailed iguana (Opiums
cuvieri cuvieri) eavesdrop on closely associated birds: Galapagos mocking birds
(Nisomimusparvulus) and paradise flycatchers (Terpsiphone mutata), respectively.
Galapagos marine iguanas respond to mocking bird alarm calls (but not song) with
increased vigilance behavior (Vitousek et al. 2007), and spiny tailed iguanas will also
increase their vigilance in response to flycatcher alarms, but not their songs (Ito and Mori
2010).
In most documented eavesdropping cases a bird species eavesdrops on another
bird species, a mammal species eavesdrops on another mammal species, or a mammal
species eavesdrops on a bird species. In an exception to this, Rainey et al. (2004) found
evidence of a bird species that eavesdrops on a mammalian alarm call. Yellow casqued
hornbills (Ceratogymna data) eavesdrop on alarm calls that Diana monkeys
(Cercopithecus diana) produce. The hornbills do not respond to every alarm vocalization
Diana monkeys make, however. Diana monkeys have different alarm calls in response to
different predators, and the hornbills are able to differentiate between those alarms,
adding to the usefulness of the eavesdropping behavior. Hornbills only use energy to
respond to alarm calls for predators that they share with the monkeys, and do not waste
8


energy responding to alarms for other, unshared predators (Rainey et al. 2004). In all
eavesdropping cases, the eavesdropping individuals are able to enhance their own fitness
by paying attention to other species alarms and taking anti-predator measures, usually in
the form of increased vigilance.
If burrowing owls living on prairie dog colonies take advantage of a similar
eavesdropping behavior, it might add to the explanation of their increased nesting success
on prairie dog colonies. Not only do they benefit from ready-made burrows, shortened
grass, higher prey availability, and a possible predator dilution effect provided by the
prairie dogs (Plumpton and Lutz 1993, Desmond et al. 2000, Orth and Kennedy 2001),
they also have a built-in look-out system alerting them to possible threats. Eavesdropping
on prairie dog alarm calls could be very valuable to burrowing owls.
9


Figures and Tables
Figure 1.1 Historic Range of Prairie Dogs.
Range of the five prairie dog species in the United States. Black tailed prairie dogs
('Cynomys ludovicianus) have the broadest range, stretching from Canada to Northern
Mexico. Map downloaded from:
http://www.defenders.org/images/programs_policy/Imperiled_Species/prairie_dog/pd_his
toric_range.jpg
10


Burrowing Owl Range
0
1,120
1,680 2,240
m--------- i Miles
Burrowing Owl Range
season
breeding
winter
year-round
Burrowing Owl Range data from
httpy/bna birds Cornell edu/bna/speaes/061/
World Map from
ESRI. Redlands. CA
Projection North America Equidistant Conic
Figure 1.2 Historic Range of Burrowing Owls.
Breeding and wintering ranges of the Western Burrowing Owl {Athene cunicularia)
in the United States.
11


Table 1.1 Cross-taxa heterospecific eavesdropping examples.
A selection of cross-taxa eavesdroppers that use avian alarm calls to their advantage.
Avian alarm caller Responding heterospecific Eavesdroppers response Reference
Hombills (Tockus deckeni and T. Dwarf mongoose (Helogale undulata Decreases own vigilance behavior in presence of hombills and attends to Rasa 1983
flavirostris) rufula) hombill alarms while foraging
Eurasian jay (Garralus glandarius) Red squirrel (Sciurus vulgaris) Increases vigilance and escape responses when jays alarm Randier 2005
Galapagos mocking bird (Nisomimus parvulus) Galapagos marine iguana (Amblyrhynchus cristatus) Increases vigilance in response to alarms, but not in response to song Vitousek et al. 2007
White-bellied go-away bird (Corythiaxoides leucogaster) Gunthers dik-dik antelope (Madoqua guentheri) Decreases foraging activity, increases vigilance in response to alarm calls Lea et al. 2008
Plovers (Vanellus spp.) Banded mongoose (Mungos mungo) Increases vigilance in response to alarm, but does not appear to respond differently when alarm Mueller and Manser 2008
urgency varies
Tufted titmouse (Baeolophus bicolor) Eastern chipmunk (Tamias striatus) Stronger anti-predator response to the indirect cue than to direct predator cues Schmidt et al. 2008
Paradise flycatcher (Terpsiphone mutata) Madagascan spiny tailed iguana (iOplurus cuvieri Increases vigilance in response to alarm calls Ito and Mori 2010
cuvieri)
12


CHAPTER
II. CHARACTERISTICS OF BLACK-TAILED PRAIRIE DOG
(Cynomys ludovicianus) ALARM CALLS IN RESPONSE TO
SIMULATED PREDATOR THREATS
Abstract
I presented potentially threatening stimuli to a colony of black-tailed prairie dogs
{Cynomys ludovicanus) in Boulder County, Colorado and recorded the subsequent
vocalized alarm calls. I measured characteristics of prairie dog vocalizations in response
to three stimuli: a model snake, a domestic dog, and a human on foot. For each trial, I
analyzed the following characteristics of the spectrogram of the first distinct bark in
response to the stimulus: 1) the duration of the bark (time in seconds), 2) the central
frequency, 3) low frequency, and 4) high frequency (Hz) of the vocalization, 5) the first
and third quartile frequencies (Hz), and 6) the average power (db) of the vocalization. I
compared these structural characteristics of the alarm calls across stimulus type and
found that there were significant differences in bark duration, IQR bandwidth, and
frequency change between the human stimulus and the model snake stimulus. The greater
duration and higher frequency of the snake alarm bark may indicate that prairie dogs
perceived the simulated snake as a more imminent, localized, and urgent threat and that
the human may or may not have been perceived as a threat. I was unable to draw
conclusions about more specific semantic encoding contained in the alarm from these
measurements.
13


Background
In many species living in social groups, individuals rely on collective vigilance to
increase their foraging time while simultaneously decreasing their risk of predation. One
or a few group members monitor their surroundings for potential threats while others
forage (Bell et al., 2009). Scanning the surroundings is synchronized between group
members (Pays et al., 2007) such that each group member has an opportunity to forage
without having to maintain strict individualized vigilance. Vigilant animals may emit
audible alarm calls when they detect a threat. Auditory alarm signals vary in specificity;
some calls are urgency-based only, and others encode information about the particular
type of threat (review: Macedonia and Evans, 1993; Furrer and Manser, 2009). Alarm-
calling systems like these minimize the energy required by any single individual to avoid
predators (Beuchamp and Ruxton, 2007). Several studies of alarm calling social animals
have shown that if an animal produces a specific alarm call in the presence of a given
predator or a given threat level, other group members will react with the appropriate
avoidance behavior whether they have themselves seen the predator or not (Seyfarth and
Cheney, 1984, 2003; Kiriazis and Slobodchikoff, 2006; Furrer and Manser, 2009).
Gunnisons prairie dogs (Cynomys gunnisoni) use alarm calls with referential
specificity. The calls differ in structure based on predator and mode of approach (Ackers
and Slobdchikoff, 1999). For instance, the alarm call for a hawk is structurally distinct
from the alarm call for a coyote, and prairie dogs take different evasive measures based
on these call differences. There is also evidence that Gunnisons prairie dogs can
recognize individuals of the same predator species and alarm based on that individuals
threat level (Ackers and Slobodchikoff, 1999).
14


Another species of prairie dog, the black-tailed prairie dog (Cynomys
ludovicianus) also uses a sophisticated alarm system to warn other colony members of
potential danger. In the presence of perceived danger, black-tailed prairie dogs emit an
anti-predator alarm bark (Hoogland, 1995): repeated chirps about 0.1s duration. The anti-
predator bark may vary in rate and in sound structure depending on the predator threat
and, possibly, the mode of approach (Smith etal., 1977; Hoogland, 1995; Fredrickson
and Slobodchikoff, 2007). Black-tailed prairie dogs also display a jump-yip behavior
which can occur in territorial disputes and chases, in response to low risk stimuli, or in
the absence of any apparent stimulus at all (Smith et al., 1976). Halpin (1983) and
Loughry (1987) noted that jump-yips occurred most commonly in the presence of snakes,
and Owens and Loughry (1985) posit that the jump-yip is a reaction to a less immediate
threat while an alarm bark indicates more urgency.
Fredrickson and Slobodchikoff (2007) showed that black-tailed prairie dogs might
also encode information about predator threat in their alarm calls. Alarm calls were
significantly different in structure when a human previously seen with a weapon,
indicating a potential high predatory threat, approached and when a human associated
with food, indicating a low predatory threat, approached.
Based on the studies of Gunnisons prairie dog vocabulary and observations of
black-tailed prairie dog vocalization behaviors, I tested the hypothesis that presenting
black-tailed prairie dogs with different potentially threatening stimuli would elicit alarm
calls that were measurably different in structure. I used a domestic dog (Canis lupus
familiaris), a simulated snake, and a human as stimuli.
Materials and Methods
15


Study Area Experimental trials took place on Carolyn Holmberg Preserve at
Rock Creek Farm in southeast Boulder County, Colorado (Figure II. 1). This public open
space property covers approximately 1100 acres, and is a popular recreation area for
hikers, runners, anglers, and bicyclists. The property is surrounded by a shopping mall,
major highway, railroad, and neighborhood subdivision. The area was primarily native
short-grass prairie before its conversion to rangeland, and there are some parcels of
uncultivated land containing primarily native vegetation, such as buffalo grass (Buchloe
dactyloides), blue grama (Bouteloua gracilis), and prickly pear cactus (Opuntia
polyacantha). All trials conducted took place on rangeland which included, along with
native grasses, non-native grasses and forbs such as cheatgrass (Bromus tectorum) and
field bindweed (Convolvulus arvensis). All of the trial locations were within one half
mile of wetlands in the preserve. Common plant species in the wetland areas are plains
cottonwood (Populus sargentii) and coyote willow (Salix exigua). The habitat is also
home to prairie dog predators including coyotes (Canis latrans), red-tailed hawks (Buteo
jamaicencis), bull snakes (.Pituophis catenifer), and western rattlesnakes (Crotalus
viridis). Grazing cattle were rotated around the study site. I did not make recordings on
colonies when cattle were present.
Field Methods I presented stimuli to prairie dogs daily between 0500 h and 1100
h from 25 May to 29 July 2010. I digitally recorded all audible responses to each stimulus
presented. Stimuli included 1) a domestic dog (German shepherd mix) walking into the
colony, 2) a rubber snake pulled through the colony on fishing line, and 3) a human
walking into the colony at 3-4 kph, toward a single focal prairie dog. I also presented
prairie dogs with a helium-filled Mylar balloon to simulate an airborne threat (Appendix
16


I). However, the presentation of the balloon added confounding variables, particularly
noise, that likely affected response behaviors, so analysis of those trials has been omitted
from results.
To ensure that the prairie dogs were not associating me with the presented stimuli,
I remained hidden in a blind throughout all trials, with the exception of the human trial.
Individual animals were not marked, but I rotated the placement of blinds over nine
different areas of the colony to decrease the probability of recording responses from the
same individuals repeatedly. Each of the nine testing sites was at least 100m from any of
the others, and I estimate that at least 20 different animals contributed to my samples. At
each testing location, I gave prairie dogs in the area two to three days to acclimatize to
the presence of the blind prior to running any trials. I only ran domestic dog trials on days
that I arrived before prairie dogs had begun their morning foraging. I kept the dog on a
15-foot lead anchored behind the blind during all presentations. The dog was very
accustomed to prairie dogs, and did not vocalize or attempt to chase the prairie dogs at
any time. For snake trials, I hid the snake at least 20 meters away from the blind under
debris on-colony so that it would not be visible until I began the trial and started pulling it
from behind the blind. To simulate a more natural smell, I also wrapped the snake in
sheds from a bull snake. I alternated the presentation of two different observers (RDB
and a technician) as the human stimulus. Because there were no systematic differences in
the alarm calls in response to the two different people, I considered response alarm calls
to both humans as one group.
In all trials, I waited until one or more prairie dogs were visible in the trial area
before beginning. I made mp3 audio recordings using a MarantzPMD661 digital
17


recorder (Mahwah, NJ) and AudioTechnica shotgun microphone (Stow, OH), which
was set up to protrude just outside of the blind, aimed at the prairie dog colony. I started
recording 5 seconds before presenting a stimulus, and stopped recording when the
stimulus was no longer visible to the prairie dogs or when they stopped reacting. I
recorded calls from prairie dogs that were 15-60 meters from the blind. The microphone
did not clearly pick up prairie dog vocalizations from farther than 60 meters away. Trial
recordings lasted a minimum of one minute. At the end of each day, I downloaded the
recordings into a computer running RavenPro 1.4 sound analysis software (Ithaca, NY).
Data Analysis For each alarm call recorded, I determined alarm rate by
calculating the number of barks in a randomly selected 10 second interval during the
stimulus presentation. I compared the averages of alarm rate across trials. For consistency
in measuring acoustical structure, I used the first audible bark in response to each
presented stimulus for analysis. I measured the duration of the bark (time in seconds), the
central, low, and high frequencies (Hz) of the vocalization, the first and third quartile
frequencies (Hz), the IQR bandwidth (difference between the first and third quartile
frequencies) (Hz), and the average power of the vocalization (db; Figure II.2). From these
measurements, I also calculated the change in frequency: the difference between the
highest frequency and lowest frequency harmonic in the vocalization.
I used R version 2.11.1 (2010: The R foundation for statistical computing) to
run one-way ANOVA tests on each of the above measurements by stimulus. I used post-
hoc multiple comparisons analysis using Tukeys HSD to determine significant
differences between measurements by stimulus.
Results
18


I was able to take measurements from a total of 33 snake trails, 34 human trials,
and 15 dog trials. If sonograms were unclear or had more than one prairie dog bark
overlapping, I discarded them from my measurements. Because I could only run dog
trials on days that I arrived before the prairie dogs emerged, I was unable to run as many
dog trials. Other stimuli allowed me to run trials at different blinds on the same day, so I
did not have this constraint.
Prairie dogs alarmed in response to each of the presented stimuli. Responses
immediately followed the appearance of each stimulus. There were no significant
differences among trials in the average number of barks per random 10 second interval;
prairie dogs alarmed at an average rate of 16.2 5.6 barks/ten seconds, with dog predator
responses showing the highest amount of variance.
Duration One-way ANOVA showed a significant effect of stimulus type on the
length of the first bark in response to the stimuli (F2,79 = 7.0, P = 0.0014). Specifically,
post-hoc multiple comparisons showed that the response to the snake stimulus was longer
in duration than that for the human stimulus (P = 0.013), but the average response
durations for snake and human stimuli were not significantly different from that recorded
for the dog stimulus (Table II. 1).
Frequency Central, low, and high frequency measurements alone did not show
significant differences based on stimulus type presented. However, there was a
significant effect of stimulus type on the average change in frequency for each alarm call
(F2,79 = 3.3, P = 0.039). Post-hoc multiple comparisons showed that snake alarms were
associated with a higher average change in frequency than human alarms (P = 0.04), but
neither differed significantly from average changes for dog alarms (Table II. 1).
19


IQR bandwidth The effect of stimulus type on average IQR bandwidth was also
significant (F2;79 = 4.4, P = 0.015). Post-hoc multiple comparisons showed that alarms in
response to the snake stimulus had a smaller IQR bandwidth than those in response to the
human stimulus (P = 0.029). Dog alarm barks were not significantly different from either
stimulus (Table II. 1).
Average Power There was a significant effect of stimulus type on the average
power of alarm barks (F2;79 = 7.2, P = 0.013). Specifically, post-hoc multiple
comparisons of average power showed that calls in response to the snake stimulus were
higher power than those in response to the human stimulus (P = 0.0012). Responses did
not differ significantly between human and dog or snake and dog (Table II. 1).
Discussion
My results support the hypothesis that black-tailed prairie dog alarm calls differ in
acoustical structure based on the stimulus presented. These results support previous
research on Gunnisons prairie dogs (Slobodchikoff el al., 1991; Ackers and
Slobodchikoff, 1999) and black-tailed prairie dogs (Smith et al., 1977; Halpin, 1983;
Owens and Loughry, 1985; Fredrickson and Slobodchikoff, 2007) suggesting that alarm
calls vary in structure according to predator threat.
Alarm calls in response to the snake stimulus showed a higher maximum power
(db) than alarm calls in response the human stimulus. Because I do not know exact
distances to the prairie dogs recorded in each trial, it is unclear if the calls were higher in
power simply due to the distance to the alarming prairie dog or if the increased decibel
level was reflective of the stimulus. I could not control for distance of prairie dogs that
responded, but because my recordings were at random distances across all samples, it
20


seems unlikely that any distance-based bias influenced the results. In fact, the
consistently random nature of the sampling distance offers further support to the
statistically different results being a result of differences in stimulus rather than in
distance.
Change in frequency indicates the difference between the highest frequency and
lowest frequency harmonic in the vocalization, and IQR bandwidth indicates the
difference between first and third quartile frequencies (Figure II.2). Both of these
measurements were significantly different in the snake response than the human
response: the difference between the lowest and highest harmonic was larger in the snake
than in the human, but the difference in the third and first quartile frequencies was
smaller. This combination of factors would result in a different sound timbre.
In the presence of the snake stimulus, I recorded two distinct types of
vocalization: a normal alarm bark and a jump-yip vocalization (Figure II.3). Based on
previous observations, the latter was likely a snake-specific alarm call. The sonograms
shape and the behavior accompanying the vocalization are both similar in structure to
those described by Smith et al. (1976), Owings and Owings (1979), and Loughry (1987).
However, I observed this call on only 3 of 30 snake presentations. For data analysis, I
used the alarm bark that coincided with the snake presentation and not the jump-yip.
The acoustical differences I measured may convey information about the
immediacy of the threat. Longer vocalizations at a higher calling rate may indicate more
urgent contexts (review: Taylor and Reby, 2010). Alarming prairie dogs in the vicinity of
the model snake emitted significantly longer barks in response to the snake stimulus than
to the human, but the rate of calling did not vary significantly. It is possible that the threat
21


of the snake was more immediate than that of the approaching human, resulting in a more
potentially dangerous situation and a longer vocalization. The preserve where trials took
place is a very heavily used recreation area in an outdoors-oriented community. Prairie
dogs encounter humans and dogs in this area on a daily, and at times an hourly basis.
Snakes are common, but are likely considered more of a threat in this area than are dogs
or humans.
The snake-specific jump-yip call occurred on three occasions during trials. The
spectrogram associated with this call had a triangular shape distinctly different from the
parallel lines on the spectrograms associated with more general alarm barks in response
to the model snake (Figure II.3). A sound that rises and falls equally in a chevron shape
may indicate ambivalence on the part of the vocalizing animal (Morton, 1977): the
animal has conflicting motivations as to whether it should stay or flee, resulting in the
animal doing a combination of both of those behaviors. This is manifested in rapid
movements and conspicuous displays known as mobbing (Lord etal., 2009). Mobbing is
a common anti-predator behavior seen in many birds and mammals in response to a
predator or an unknown stimulus. An individual giving a mobbing call and/or displaying
a mobbing behavioral pattern signals others in the area to join. Whether or not others
join, the behavior pattern reflects the ambivalence noted above. The chevron shape of the
snake-specific jump-yip which accompanied both tentative stalking and aggressive
attacks on the model snake supports this hypothesis, although stalking and mobbing of
the snake occurred without the jump-yip as well; on 18 such occasions, prairie dogs
directly approached the snake stimulus. On four of these occasions they attacked and tore
pieces off of the model snake. Prairie dogs retreated from the model snake only twice,
22


and one of those times the retreating animal was a juvenile. I did not observe more than
one adult prairie dog approaching and attacking the model snake at any given time, but
did observe other adult prairie dogs in the vicinity observing the model and the attacking
prairie dog from nearby.
The observation of more alarm behaviors than jump-yip behaviors in response to
the snake stimulus contrasts with results from previous studies where the jump-yip
response to snakes was more commonly recorded (Halpin, 1983; Loughry, 1987).
Perhaps this has to do with the sudden appearance of the snake in the colony when pulled
out of its hiding place. Macedonia and Evans (1993) noted that some alarm-calling
mammals use different alarm calls to differentiate high threat from lower threat predators,
but will use a high-threat predator call if danger from a normally low-threat predator is
immediate. If a jump-yip indicates a lower-threat predator, but that lower-threat predator
appears quite suddenly and/or unexpectedly, a more urgent alarm may be used instead. It
is also possible that the snake model was not recognized as a snake, and the prairie dogs
were alarming in response to a novel threat.
Ackers and Slobodchikoff (1999) showed that Gunnisons prairie dogs are able to
encode referential specificity in their alarm calls. Through discriminate function analysis,
they showed that alarm calls for a model skunk, a model coyote, and an oval control
showed qualitative and quantitative differences, and were consistent for an individual
prairie dog. This suggests that random variation in sonogram characteristics among
prairie dogs is less relevant than systematic variation in sonogram characteristics
resulting from encoded semantic information. Using a similar analysis method,
Fredrickson and Slobodchikoff (2007) showed that black-tailed prairie dogs might also
23


encode semantic information about predator type in their alarm calls. Because I did not
mark individuals, I could not be sure that I presented different stimuli to the same animal
at a trial site, and so I did not use discriminate function analysis to incorporate all
measured parameters in statistical analysis. However, I was still able to quantify
differences in alarm calls based on the stimulus type. It is therefore quite likely that the
prairie dogs recorded in this study are also encoding qualitative information about
potential threats in their alarm calls.
Because black-tailed prairie dog alarm calls serve to alert relatives and distant kin
of danger (Hoogland, 1995), the more specific the call is to a threat, the more likely it is
that those relatives and distant kin will react appropriately. Results of this study suggest
that black-tailed prairie dog alarm calls contain information about threat urgency.
However, although urgency information may be contained in the call, the existence of the
information does not necessarily make it informative to other prairie dogs (Seyfarth and
Cheney, 2003), and I did not specifically observe prairie dog responses to these calls. The
mobbing behavior witnessed in response to the snake model indicates a possibility that
the alarm calls convey information to other prairie dogs in the area, but without more
specific playback experiments, I do not know if these behaviors were a result of visual or
auditory cues.
24


Figures and Tables
Figure II.1 Map of Prairie Dog Study Area, Boulder County, CO.
All prairie dog response trials took place at the Carolyn Holmberg Preserve at Rock
Creek Park, an Open Space property in Boulder County, Colorado. Inset: blind locations
marked with stars. Blinds and trials were rotated around these nine sites over the two
month period.
25


U
i:-
High frequency
I *
6-
5-Central frequency
Low frequency

Third quartile frequency

IQR bandwidth
First quartile frequency
Figure II.2 Spectrogram measurements.
Spectrogram of two alarm barks in response to the domestic dog. Measurements of
(a) central, low, and high frequencies (in Hertz), (b) duration (Atime), (c) the first and
third quartile frequencies, and (d) IQR bandwidth. Measurements were made using
RavenPro1.4 software.
26


i h n
12-
a
b
i
Figure II.3 Spectrogram of Snake Responses.
Spectrogram of (a) normal bark in response to the snake stimulus, (b) jump-yip
recorded in presence of snake stimulus. The normal bark is much shorter in duration, with
less harmonic variation, and the jump-yip vocalization has a distinctive chevron shape
seen in animal vocalizations that are commonly associated with mobbing behaviors.
Table II. 1 Structural attributes of alarm spectrograms.
Mean standard error for the first alarm bark in response to a predator stimulus.
Stimuli that share the same superscript letters are not significantly different (P > 0.05).
Different superscript letters indicate significant differences in the means (P < 0.05).
mean A time (ms) mean A frequency (Hz)
dog (n= 15) 0.074 0.0035ab 6314.1 221.58ab
snake (n = 33) 0.076 0.0029a 7126.5 188.81b
human (n = 34) 0.063 0.0022b 6267.1 302.83a
mean IQR bandwidth (Hz) mean average power (db)
dog (n= 15) 1986.8 282.24ab 64.93 2 47ab
snake (n = 33) 1247.6 178.99a 59.03 1.14b
human (n = 34) 1889.9 162.71b 66.66 1.67a
27


CHAPTER
III. BURROWING OWLS EAVESDROP ON ALARM CALLS OF
BLACK TAILED PRAIRIE DOGS
Abstract
I investigated whether Western Burrowing Owls {Athene cunicularia) nesting on
black-tailed prairie dog {Cynomys ludovicianus) colonies responded to playbacks of
prairie dog alarm calls. Western burrowing owls are closely associated with black-tailed
prairie dogs in Colorado and neighboring states on the Great Plains of the United States.
Owls have high nesting success on prairie dog colonies: they have pre-made nest-sites
and good foraging habitat, and prairie dogs may act as an alternate prey source for
predators. Owls nesting on prairie dog colonies may eavesdrop on the prairie dogs as
well: owls attending and responding to the alarm calls that the prairie dogs emit may
enhance their survival and nesting success on colonies. I performed playback experiments
with three different treatments: a prairie dog alarm call, a biological control (cattle
mooing), and a non-biological control (an airplane engine). Owls reacted more quickly to
the prairie dog alarm than to the biological control. They increased their vigilance
behaviors significantly in response to playbacks of the alarm, but did not show an
increased reaction to either the biological or the non-biological control. My results
suggest that owls nesting on prairie dog colonies eavesdrop on, and increase their
vigilance in response to, prairie dog alarm calls.
Key words: Burrowing Owl {Athene cunicularia), eavesdropping, black-tailed prairie dog
{Cynomys ludovicianus), alarm calling, vigilance
28


Introduction
When an individual animal gathers information about the environment it might
produce a cue (a vocal or chemical signal) in response to that information for the benefit
of other group members or kin. When the cue is produced, there is the potential for
unintentional recipients as well (Goodale et al. 2010). Animals sharing a common habitat
and common predators will sometimes take advantage of each others vigilance by
eavesdropping on their neighbors, particularly when the information comes in the form of
a vocal alarm call. Heterospecific eavesdropping, when members of a different species
acquire and benefit from the cue another species produces, has been studied extensively
in birds. Magrath et al. (2007) and Fallow et al. (2011) found that different bird species
with similar predators are able to cue in to each others alarm calls when they hear them.
Mobbing behaviors in many bird species are a result of responses to heterospecific alarms
(Chu 2001, Johnson et al. 2003, Nocera and Ratcliffe 2010).
Evidence of social and non-social animals eavesdropping on avian alarm calls has
been documented in mammalian and reptilian species as well. Mammalian examples
include red squirrels (Sciurus vulgaris), dik-dik antelope (Madoqua guentheri), banded
mongooses (Mungos mungo), and Eastern chipmunks (Tamias striatus) (Randier 2005,
Lea et al. 2008, Mueller and Manser 2008, Schmidt 2008), and reptilian examples include
Galapagos marine iguanas {Amblyrhynchus cristatus) and Madagascan spiny tailed
iguanas (Opiums cuvieri cuvieri) (Vitousek et al. 2007, Ito and Mori 2010). Rainey et al.
(2004) found evidence of a more unusual eavesdropper: a bird which eavesdrops on a
non-avian alarm call. Yellow casqued hornbills (Ceratogymna elata) eavesdrop on alarm
calls of Diana monkeys (Cercopithecus diana). In this particular case of eavesdropping,
29


the birds differentiate between alarm calls they hear: they only respond to alarm calls for
predators that they share with the monkeys, and not to alarm calls for other, unshared
predators (Rainey et al. 2004). In all cases, eavesdropping individuals are able to enhance
their own fitness by paying attention to other species alarms and taking anti-predator
measures, usually in the form of increased vigilance.
Western burrowing owls (.Athene cunicularia) are listed as a species of special
concern in several states in the Great Plains, including Montana, Oklahoma, Utah, and
Wyoming (Poulin et al. 2011), and are listed as a state-threatened species in Colorado
(Colorado Division of Wildlife 2011). Western burrowing owls (hereafter, burrowing
owls or owls) are closely associated with black-tailed prairie dogs (Cynomys
ludovicianus) in these areas, and the decline in black-tailed prairie dogs (hereafter, prairie
dogs) due to management and sylvatic plague has been correlated with declines in
burrowing owl populations (Desmond et al. 2000). Burrowing owls have high nesting
success on prairie dog colonies, and are found more often on than off prairie dog colonies
in the Western Great Plains (Holroyd et al. 2001). Burrowing owls commonly nest on
prairie dog colonies with high burrow densities (Plumpton and Lutz 1993), and show a
preference for active prairie dog colonies over abandoned colonies (Restani et al. 2001,
Sidle et al. 2001). Multiple hypotheses have been put forth to explain this association
between owls and prairie dog colonies. The disturbances prairie dogs create provide pre-
made nest sites, and the short grass provides good foraging habitat (Poulin et al. 2011).
Also, having prairie dogs nearby may cause a dilution effect if prairie dogs act as an
alternative prey source for predators (Desmond et al. 2000). Desmond et al. (2000) and
Beradelli et al. (2010) suggested that burrowing owls nesting on black-tailed prairie dog
30


colonies may eavesdrop on the prairie dogs as well: owls could attend and respond to the
alarm calls that the prairie dogs emit. This would enhance the birds survival on colonies
and offer a further explanation of increased nesting success.
Burrowing owls use vocal alarm signals to alert each other to dangers, and those
alerted (young or a nearby member of a pair) take appropriate actions to avoid the danger
(Coulombe 1971). Burrowing owls sharing a habitat with another species that uses an
alarm calling behavior (eg. prairie dogs or ground squirrels) may increase their vigilance
when they hear an alarm from that species. Black-tailed prairie dogs emit alarm calls in
the presence of predatory threats to alert relatives and distant kin of danger (Hoogland
1995). Burrowing owls, especially eggs and chicks, share predators with prairie dogs:
American badgers (Taxidea taxus), coyotes (Canis latrans), red-tailed hawks (Buteo
jamaicensis), and possibly prairie rattlesnakes (Crotalus viridis) and bullsnakes
(Pituophis catenifer) predate both prairie dog and burrowing owl young (Halpin 1983,
Coulombe 1971). Burrowing owl adults that eavesdrop on alarming prairie dogs can
potentially learn when predators are approaching. If they increase their vigilance so they
can respond appropriately to protect their young, they may enhance their fitness. In this
study I tested whether burrowing owls respond to alarm calls that prairie dogs emit.
Methods
With permission from the Central Plains Experimental Range (CPER) and the
National Forest Service, I conducted playback experiments on nesting owls on 14
different prairie dog colonies on the CPER and the Pawnee National Grasslands (PNG),
located in Weld County, Colorado (Figure III. 1). Experiments took place June through
July, 2011. The Pawnee National Grasslands is a short-grass steppe ecosystem. The
31


average summer temperature is 22 C, and annual precipitation patterns can vary greatly.
The average precipitation is 32cm, most commonly occurring between April and
September. Precipitation manifests primarily as soaking rains in spring and as heavy
localized thunderstorms in summer (SGS-LTER 2007). All experiments took place on
prairie dog colonies on active ranchland, where vegetation included buffalo grass
(Buchloe dactyloides), blue grama (Bouteloua gracilis), prickly pear cactus (Opuntia
Polyacantha), scarlet globemallow (,Sphaeralcea coccinea), saltbush (Atriplex
canescens), and field bindweed (Convolvulus arvensis).
I conducted audio playbacks to owls nesting at 20 different locations on 14 active
prairie dog colonies; several colonies had more than one nesting owl pair present, and
when that was the case, I rotated trials between the active nests. I played 5 minute
recordings of a randomized sequence of sounds to each bird: a prairie dog alarm call (the
experimental treatment), a small airplane engine (a non-biological control treatment),
cattle mooing (a biological control treatment), and ambient noise: intermittent birdsong,
insect chatter, and other typical prairie sounds. I collected digital recordings of all sounds
using a MarantzPMD661 digital recorder (Mahwah, NJ) and AudioTechnica shotgun
microphone (Stow, OH). With the exception of the plane sound, which was collected
near the Rocky Mountain Metropolitan Airport in Broomfield, Colorado, I recorded all
sound stimuli on the PNG or the CPER. I recorded a herd of mooing cattle on open range
on the PNG. I recorded prairie dog alarm call responses to a simulated snake predator (a
toy snake pulled through the colony on a fishing line) on the CPER. I edited and
combined the sounds using RavenPro 1.4 sound analysis software (Ithaca, NY). To
ensure that the order the birds heard the sounds did not affect their responses, I created
32


three different treatment orders: Order 1 prairie dog alarm (A), cattle mooing
(biological control B), airplane (non-biological control N); Order 2 B, N, A; and
Order 3 N, A, B. I started every trial with two min of ambient noise, and played
stimulus sounds at a consistent volume and for the same amount of time across each trial.
Each stimulus sound lasted 30 seconds, and the stimulus sounds were separated from
each other by 1 min of ambient noise. I video recorded all trials using a JVC Everio
HD camcorder (Wayne, NJ). The time frame over which I ran trials coincided with the
owls nesting and chick-rearing stages. I did not know if this would affect adult vigilance
patterns, so I divided all trials into two categories: early (when eggs have not yet
hatched or have hatched but chicks have not yet surfaced) and late (when nestlings
have surfaced and are beginning to disperse to nearby burrows). I then tested for
differences in the adults reactions to the stimulus cues based on nest stage.
Field Methods For each trial, I placed a SME-AFS portable field speaker
(Elmont, NY) within 10 m of an active nest burrow. Owls usually flushed or ducked into
the nest or a nearby burrow upon my arrival. Before beginning any playbacks, I waited in
a blind for the bird to return to the vicinity of the speaker. For most trials the owl was
within 10 to 20 m of the speaker. If the owl was more than 30 m away from the speaker, I
did not score the trial in data analysis. The blind was 25 to 50 m from the nest burrow
and/or focal owl for all trials. I used a pop-up hunting blind when it was not possible to
use the truck as a blind. The truck was preferable because the owls are accustomed to
seeing ranchers vehicles on the prairie, so I used it as much as possible. I tested for
differences in responses based on blind type when running analyses. Birds were
unmarked, so it was not always possible to determine which experiments were repeated
33


on the same bird on different days. I rotated nest sites visited daily to ensure that I did not
subject the same owl to the same tests consecutively.
Data Analysis I scored video files using JWatcher1.0 software (Blumstein and
Daniel 2007) using the ethogram in Table III. 1. If the owl moved beyond 30 m of the
speaker during the playback, or ducked underground and did not re-emerge, I included
only those parts of the trial with the bird within 30 m of the speaker in the analysis. If
there were prairie dogs vocalizing in the background of the video, I did not include the
video in analyses.
For analysis, I classified responses as either alert or relaxed. Alert responses
included vigilance (head turns), bobbing, locomotion, vocalizing, or stretching tall or
squatting low while looking. Relaxed responses included preening, yawning, or sleeping
(Table III. 1). For the purposes of this study, I categorized foraging or looking down at
the ground as relaxed behaviors, because the owls vigilance was directed toward prey
and not toward external threats like predators. I calculated 1) latency to first alert
response to presented sound cues (how quickly the owl reacted to the sound cue), and 2)
the intensity of the response (how actively the owl responded) as determined by the
number of alert responses within 10 sec of the presented cue. I considered each alert
response mutually exclusive. I used R version 2.11.1 (2010: The R foundation for
statistical computing) for all statistical analyses.
Analysis I Latency to Response. I measured how quickly the owls responded
to each sound stimulus, and compared these measurements by sound cue. To find the
owls response times in milliseconds, I subtracted the start time of a given sound cue
from the start time of the first alert response after that sound began playing. Because
34


results were not normally distributed, my data did not meet assumptions for parametric
statistical tests. I used a Kruskal-Wallis rank-sum test to determine whether there were
differences in median time to response in each of the sound stimulus categories: alarm,
biological control, and non-biological control. In order to determine which categories
showed significant differences from one another, I ran pair-wise Wilcoxon tests with
Bonferroni adjustment.
Analysis II Intensity of Response. Here I examined the rate of response to
sound cues: i.e. how many alert responses the owl showed per unit time after hearing the
sound cue. I categorized the owls responses to the sound cues using behaviors within the
first 10 sec during the playback of any given sound cue. I counted the number of alert
responses during that 10 sec time frame and assigned the responses to one of three
categories. Category 1 was defined as a low response rate: 0-2 alert behavioral
responses; Category 2 was a moderate response rate: 3-4 alert behavioral responses;
and Category 3 was a high response rate: 5 or more alert behavioral responses.
I administered a chi-squared analysis on these data with a Fishers exact test given
that one of the categories (Non-biological control, category 3) had less than 5 data points.
Due to differences in response based on blind type (Fishers Exact test P< 0.01), I
analyzed truck and pop-up blind data separately.
Results
I conducted a total of 93 playback experiment trials on adult owls over the two-
month period (truck blind n = 51, pop-up blind n = 42). I collected useful data from 58 of
those trials (truck blind n = 38, pop-up blind n = 20), which included 16 different nesting
pairs on 14 prairie dog colonies (Appendix II). Thirty-five of the 93 trials were censored
35


for various reasons, including excessive background noise: high winds; interruptions such
as vehicles, prairie dogs, or cattle; owls leaving the experimental trial area; or too much
distance between the video camera and the bird for accurate behavioral scoring.
Latency to Response Times There were no systematic differences between the
owls responses by sound presentation order by nest stage (2-way ANOVA stimulus and
phenology: F2,151 = 0.62, P = 0.54), or by blind (2-way ANOVA stimulus and blind: F2,
151 = 0.32, P = 0.72), so I pooled all data in subsequent analyses. Owls showed shorter
median response times to the prairie dog alarm cue and the non-biological control cue as
compared to the biological control sound (Pairwise Wilcoxon test with Bonferroni
adjustment: Alarm and Non-biological cue /J=0,045, Alarm and Biological control cue
/J=0.028). There was no significant difference in the median response time to the non-
biological control sound cue and the prairie dog alarm sound cue (P = 1.0) (Table III.2,
Figure III. 2).
Intensity of Response There were no significant differences in response rate
category (1 = low, 2 = moderate, 3 = high) based on the order the stimuli were presented
(Fishers Exact test: Alarm order: P = 0.31, Biological Control order: P = 0.20, Non-
biological control order: P = 0.32) or nesting stage (Fishers Exact test P = 0.30). When I
conducted Chi-squared analysis by blind-type, I found fewer category 3 responses (n=3)
than expected with the pop-up blind (Fishers Exact test P < 0.01). Because of this
discrepancy, I analyzed truck data and pop-up blind data separately.
In trials run from the truck (n = 38), owls showed more category 3 responses and
fewer category 1 responses than expected when presented with the prairie dog alarm cue.
When presented with the biological control sound cue, they showed fewer category 3
36


responses and more category 1 responses than expected. Owls showed expected
responses to the non-biological sound cue (Fishers Exact test P = 0.024; Figure IIF3).
In trials run from the pop-up (n = 20), there were no significant differences in observed
and expected values for all sound cues (Fishers Exact test P = 0.15). However, I
recorded only three category 3 responses in trials run from the pop-up blind, and all of
those were in response to the prairie dog alarm cue. This low number of responses makes
statistical analysis difficult.
Discussion
Animals can indirectly receive information about their environment by observing
information exchanges between other animals. The eavesdropping animal can obtain this
information from visual, auditory, or chemical cues given by the donor. Eavesdroppers
get information about traits such as size or strength of the donor, or they can gain
information about resources and predation threats (Sih et al. 2009). For these reasons,
eavesdropping can be useful in mate finding and competition, and is especially valuable
in predator evasion.
Animals that use an anti-predator alarm calling system are an important indirect
information source for eavesdropping individuals in their environments. When animals
emit alarm calls, the calls may be urgency based (Furrer and Manser 2009), they may
have predator specific information encoded (Seyfarth and Cheney 2003), and/or they may
induce group members to mob a potential threat (Morton 1977). Animals that share a
habitat, have similar predators, and are similar in body size to the alarm caller are likely
to pay attention to those alarms because doing so can increase their fitness (Templeton
and Greene 2007). Some eavesdroppers are able to accurately interpret the information
37


the alarms encode (Seyfarth and Cheney 1984,2003; Rainey 2004; Hetrick and Sieving
2012), and some eavesdroppers show more anti-predator behaviors in response to a
heterospecific alarm call than they do to direct evidence of a predator (Schmidt et al.
2008). Eavesdropping behavior is common between species of birds and between species
of mammals, and cross-taxa eavesdropping is effective in many systems as well.
My study shows an example of a cross-taxa eavesdropping network that likely
enhances the fitness of the eavesdropper by alerting them to potential threats in their
environments. I found that burrowing owls living on prairie dog colonies increased their
vigilance behaviors in response to prairie dog alarm calls: the owls reacted more quickly
to the prairie dog alarm sound cue than to a non-alarm biological sound cue (cattle
mooing), and continued to show increased vigilance behaviors after their initial response.
Upon hearing the prairie dog alarm call, the owls had fewer category 1 (low rate)
responses than expected based on chance alone; they showed significantly more category
2 (moderate rate) or category 3 (high rate) response rates than expected. However, the
owls did not flee in response to the alarms and very rarely (n = 2) vocalized in the 10
seconds following the prairie dog alarm call. This potentially indicates that the owls were
increasing their vigilance, but needed first-hand information before they would act on
those cues.
Although the owls reacted equally quickly to the non-biological sound cue, they
did not continue to show increased vigilance. The non-biological sound cue was a sudden
loud airplane engine noise that did not particularly resemble any sounds that the owls
were accustomed to hearing at the study site, with the possible exception of occasional
passing trucks. It appears that this sudden noise startled the owls into a quick alert
38


response, but since the noise came from a single direction and was not accompanied by
any physical apparition, it was likely not seen as an immediate threat. The owls did not
flee nor increase their vigilance over the following 10 second interval.
The owls nesting on this study site are very accustomed to cattle, encountering
them on a nearly daily basis. This might explain the birds response (or lack thereof) to
the biological control sound cue. Cattle moo to communicate with each other, and
although some cow vocalizations may be in response to threats (Watts and Stookey
2000), cattle do not share predators with burrowing owls, and the presence of cattle on
burrowing owl and prairie dog habitat is relatively recent from an evolutionary
perspective. Owls and prairie dogs have shared habitat for thousands of years, whereas
organized cattle ranching on the Great Plains did not begin until the late 1860s (ERS
2000), so even if eavesdropping made sense in this case, it might not have had time to
evolve. Also, it may not be necessary for owls to become or to remain highly vigilant in
the presence of cattle: it is possible that with the cattle nearby there is a reduced risk to
the owls of predation as well as reduced foraging capabilities, both of which could result
in lower vigilance over all.
Owls in this area have a very close association with prairie dogs; the owls live in
abandoned burrows on active colonies and occasionally interact with foraging or cover-
seeking prairie dogs that have gotten close to their nests by flying at them or vocalizing
toward them (personal observation). Foraging black-tailed prairie dogs are generally
quiet and intermittently vigilant. Black-tailed prairie dogs do not rely on a single sentinel
keeping watch for the entire group; instead all individuals regularly survey their
surroundings, and when a threat is detected, the individual detecting the threat will alert


others in the group with a bark alarm (Hoogland 1995). Like prairie dogs in the alarming
animals vicinity, owls in the area also increase their vigilance behaviors in response to
the alarm in an attempt to ascertain the threat.
My data suggest that owls nesting on prairie dog colonies likely gain some benefit
from eavesdropping on prairie dog alarm calls. Owls increasing their vigilance in
response to a prairie dog alarm may be more aware of threats in their environment as a
result. However, whether the owls are reacting to the urgency of the alarm or semantic
information contained in the alarm is still unknown and deserves further exploration. In
many eavesdropping species, eavesdroppers show some ability to differentiate alarm calls
from non-alarm vocalizations (Vitousek et al. 2007, Goodale et al. 2010, Ito and Mori
2010), as well as predator-specific alarm calls from each other (Rainey 2004).
Eavesdroppers able to differentiate would have the advantage of saving energy required
for increased vigilance by not responding to calls that do not indicate a threat to the
eavesdropper. Rainey (2004) demonstrated that yellow casqued hornbills are able to
distinguish between predator-specific calls given by Diana monkeys, and only show
increased anti-predator behaviors when they hear the alarm for a shared predator. Like
Diana monkeys, prairie dogs also emit predator-specific calls which differ in response to
predator type and mode of approach (Fredrickson and Slobodchikoff 2007). Burrowing
owls may possess the ability to differentiate between types of alarm calls. In this study
we only looked at reactions to the snake predator alarm. Further studies of owl reactions
to specific alarm and non-alarm calls could show whether owls are able to decipher the
information coded in prairie dog alarms.
40


Figures and Tables
Figure IIL1 Map of Burrowing Owl Study Area, Pawnee National Grasslands, CO.
Trials took place on the Pawnee National Grasslands and Central Plains Experimental
Range in Northeastern Colorado. I conducted a total of 93 playback experimental trials
on adult owls over the two-month period of June and July 2011.1 was able to collect
useful data from 58 of those trials, which included 16 different nesting pairs.
Approximate locations of those nesting pairs are indicated by stars.
41


' ii i SI so Prairie Dog Alarm 1 (experimental treatment)
> u C Lf) 0) rH S W .'r o m J 1 Airplane Engine 1 (non-biological control)
m rH in o - Cattle Mooing L (biological control) M IL . / /
1 1 1 1 1 / / '
0 10000 20000 30000 40000 120000 130000
Time in milliseconds
Figure III.2 Latency to Response results.
Latency to response times showed a skewed distribution. Non-parametric analysis of
data showed owls had shorter median response times to the prairie dog alarm cue and the
non-biological control cue as compared to the biological control sound (Pairwise
Wilcoxon test with Bonferroni adjustment: Alarm vs Non-biological cue P=0.045, Alarm
vs Biological control cue P=0.028). There was no significant difference in the median
response time to the non-biological control sound cue and the prairie dog alarm sound
cue. Responses to the biological sound cue showed more variance in the length of time
from sound cue to first alert response.
42


o
ID
O

CD
O
CO
o
fc
_Q
E
d
c
o
OJ
o
o
category3
category2
categoryl
Alarm
Biological Control Non-biological control
trial type
Figure III.3 Intensity of Response Results.
I presented owls with three different sound stimuli: the experimental treatment:
(Alarm) a prairie dog barking in response to a snake predator, a biological control:
cattle mooing, and a non-biological control: a low-flying airplane engine. All responses
are categorized as 1- low (0-2 vigilant behaviors in the first 10 seconds of playback), 2-
moderate (3-4 vigilant behaviors in the first 10 seconds of playback), or 3- high rate (5 or
more vigilant behaviors in the first 10 seconds of playback). Owls showed more moderate
and high rate responses and fewer low rate responses to the prairie dog alarm call than to
the biological or non-biological controls. They showed more low-rate responses to the
biological control. Fishers P=0.024.
43


Table III.1 Ethogram of Scored Burrowing Owl Behaviors.
Ethogram of behaviors scored in J-Watcher 1.0. Each of these behaviors was scored
individually, and then classified as alert or relaxed for data analysis.
Alert behaviors DESCRIPTION Relaxed behaviors DESCRIPTION
vigilance standing or perching and looking in any direction. Each time bird re-focuses (turns head) is a new vigilance bout. sleeping eyes closed or heavily lidded
stretching up tall preening actively preening feathers or stretching wings
squatting low yawning
locomotion walking, running, or flying foraging looking at the ground or following insects
vocalizing cooing or alarming
bobbing full body or head bob
Table III.2 Median Times to Alert Response.
I measured the latency from stimulus presentation to alert behavioral response in
three categories, Alarm (experimental treatment), Biological Control (cattle mooing), and
Non-biological control (airplane engine). The median time to reaction was faster for the
Alarm and the Non-biological control than it was for the Biological control. Different
superscript letters indicate significant differences in medians.
MEDIAN TIME TO ALERT BEHAVIOR
Sound MEDIAN TIME (MS) (RANGE)
Prairie dog Alarma 1299.5 (216-32912)
Biological Control13 2620.0 (250-125824)
Non-biological Control21 748.0 (279-35417)
Kruskal-Wallace % 8.697
T*-value 0.012
44


DATA APPENDIX
APPENDIX I: PRAIRIE DOG ALARM CALL ANALYSIS
Results from Balloon Trials
On 13 of the 17 bird balloon presentations, prairie dogs ran to nearby burrows. They
proceeded underground into burrows on three of those occasions. On the other 10
occasions, they remained at the lip of a burrow. At least one prairie dog responded
vocally to the bird stimulus in 15 of the 17 trials. The number of barks in the subsequent
alarm calls ranged from 1-250 during a 3-minute period. Bird balloon alarm calls were
significantly different in duration and change in frequency than those emitted in the
presence of the snake stimulus, (duration: F3; 95 = 6.5, P=0.0005; delta frequency: F3; 95 =
3.4, P=0.02). Post-hoc comparisons using Tukeys HSD showed bird balloon alarms
were on average significantly shorter in duration (P=0.01) and higher in delta frequency
(P=0.04) than snake alarms. Values did not differ significantly from any of the other calls
or in any of the other measurement categories. We cannot be sure whether the responses
elicited by the bird stimulus were in response to the shape of the balloon, or in response
to the loud noise made by the Mylar material in the wind.
45


Table AI.1 Structural attributes of alarm spectrograms.
Mean standard error for the first alarm bark in response to a predator stimulus.
Stimuli that share the same superscript letters are not significantly different (P > 0.05).
Different superscript letters indicate significant differences in the means (P < 0.05).
mean A time (ms) mean A frequency (Hz)
dog (n =15) 0.074 0.0035ab 6314.1 221.58ab
snake (n = 33) 0.076 0.0029a 7126.5 188.81b
human (n = 34) 0.063 0.0022b 6267.1 302.83ab
bird (n =17) 0.063 0.0027b 6008.1 233.04a
mean IQR bandwidth (Hz) mean average power (db)
dog (n= 15) 1986.8 282.24ab 64.93 2 47ab
snake (n = 33) 1247.6 178.99b 59.03 1.14b
human (n = 34) 1889.9 162.71a 66.66 1.67a
bird (n= 17) 1256.5 200.05ab 60.60 2.20 ab
mean high frequency mean central frequency
dog (n =15) 7967.3 230.57a 4054.0 156.8a
snake (n = 33) 8572.2 195.82a 3659.3 89.02a
human (n = 34) 7912.0 298.06a 3668.2 145.4a
bird (n= 17) 7559.13 246.20a 3992.5 143.4a
mean low frequency mean T 5t quartile frequency
dog (n= 15) 1653.3 104.7a 3353.4 177.0a
snake {n = 33) 1481.2 51.3 5a 3053.8 131.0a
human (n = 34) 1644.9 79.87a 2898.1 150.7a
bird (n= 17) 1551.1 35.55a 3414.9 203.2a
mean 3 rd quartile freqency
dog (n = 15) 4880.9 267.2a
snake (n = 33) 4301.4 144.7a
human (n = 34) 4788.0 171.2a
bird (n = 17) 4761.4 218.6a
46


APPENDIX II: BURROWING OWL RESPONSE ANALYSIS
Playback Trial Locations
I ran playback trials to 16 different nesting pairs of owls and their chicks on the
Pawnee National Grasslands and Central Plains Experimental Range. GPS coordinates
and colony names of nests are listed below in alphabetical order:
Central Plains Experimental Range:
N40 48.001 W104 42.293
N40 48.115 W104 41.426
N40 48.059 W104 41.278
N40 48.074 W104 46.255
N40 49.032 W104 45.564
Pawnee National Grasslands:
Carroll:
N40 42.850 W104 36.884
East Stoneham:
N40 38.732 W103 34.882
N40 38.934 W103 34.914
Halter:
N40 47.911 W104 35.031
N40 48.138 W104 35.062
Keota:
N40 38.770 W104 08.036
Reno:
N40 44.506 W104 26.993
N40 44.461 W104 26.895
Tappy:
N40 49.546 W103 38.664
N40 49.475 W103 38.611
West Stoneham:
N40 38.447 W103 44.122
47


Table AII.1 Order of sound cue presentation effect on time to alert behavior.
The order stimuli were presented did not affect response times. Owls showed similar
response times to each sound stimulus regardless of where it occurred in the order.
Median time to alert behavior (ms) (range)
Prairie dog Alarm Biological Control Non-biological Control
Sound order
A-B-N 1129.0 3216 735
(216-32912) (344-42489) (461-8948)
B-N-A 1597.5 2369 876
(296-12874) (250-125824) (279-35417)
N-A-B 1190.5 1865 748
(221-19432) (293-30328) (310-11713)
K-W chi squared 1.2569 2.2518 0.0271
p value 0.5334 0.3244 0.9866
Table AII.2 Pair-wise Wilcoxon test, median time to alert behavior.
Bonnferroni-adjusted pairwise comparisons of median times to alert behavior show
that the response to the prairie dog alarm was significantly shorter than the response to
the biological control, but not the non-biological control.
Sound
Biological Control
Non-biological Control
Prairie dog Alarm
P = 0.028
P = 1
Biological Control
P = 0.045
48


Table AII.3 Order of sound cue presentation effect on response intensity.
The order in which stimuli were presented did not affect the intensity of response.
Observed values of intensity did not differ from expected intensity values regardless of
where the sound cue occurred in the order.
observed/expected (components)
stimulus A-B-N N-A-B B-N-A
Alarm: category 1 10/7.30(1.00) 3/5.08 (0.85) 7/7.62 (0.05)
Fisher's Exact test category 2 6/9.85 (1.51) 9/6.85 (0.67) 12/10.28 (0.29)
p-value = 0.3066 category 3 7/5.84 (0.23) 4/4.06 (0.00) 5/6.10(0.20)
Biological control: category 1 14/13.61 (0.01) 10/89.72 (0.01) 11/11.67 (0.04)
Fisher's Exact test category 2 6/5.83 (0.00) 2/4.17(1.13) 7/5.00 (0.80)
p-value = 0.19562 category 3 1/1.56(0.20) 3/1.11 (3.21) 0/1.33 (1.33)
Non-biological category 1 9/9.3 (0.01) 11/7.85 (1.26) 6/8.83 (0.91)
Control: Fisher's Exact test category 2 7/6.81 (0.01) 3/5.74(1.30) 9/6.45 (1.01)
p-value = 0.3287 category 3 3/2.87 (0.01) 2/2.42 (0.07) 3/2.71 (0.03)
Table AII.4 Chi-squared comparison of response strength by stimulus.
Owls showed fewer mild responses and more moderate and strong responses than
expected in response to the prairie dog alarm call. The showed more mild responses than
expected in response to the biological control sound (mooing cattle). Responses were as
expected for the non-biological control sound (airplane engine). We ran a Fishers Exact
test because there were fewer than 5 responses in the non-biological control category 3.
Category observed/expected (components)
Alarm Biological Control Non-biological Control
1- minimal 13/20 (2.82) 24/17 (2.70) 18/17.19 (0.04)
2- moderate 16/12 (1.33) 7/10 (1.8) 9/10 (0.10)
3- strong 13/9 (1.40) 4/7.8 (1.86) 8/7.8 (0.00)
Fisher's Exact test p-value = 0.02449
49


REFERENCES
Ackers, S. H. and C. N. Slobodchikoff. 1999. Communication of stimulus size and
shape in alarm calls of Gunnisons prairie dogs, Cynomys gunnisoni. Ethology,
105:149-162.
Agnew, W., D.W. Uresk, and R.M. Hansen. 1986. Flora and fauna associated with
prairie dog colonies and adjacent ungrazed mixed-grass prairie in western South
Dakota. Journal of Range Management 39(2): 135-139.
Antolin, M.F., P. Gober, D. Luce, D.E. Biggins, W.E. Van Pelt, D.B. Seery, M.
Lockhart, and M. Ball. 2002. The influence of sylvatic plague on North
American wildlife at the landscape level, with special emphasis on black-footed
ferret and prairie dog conservation. Transactions of the 67th North American
Wildlife and Natural Resources Conference 67:104-127.
Barko, V.A. 1997. History of policies concerning the black-tailed prairie dog: a review.
Proceedings of the Okahoma Academy of Sciences 77:27-33.
Barko, V.A., J.H. Shaw, D.M. Leslie, Jr. 1999. Birds associated with black tailed
prairie dog colonies in southern shortgrass prairie. The Southwestern Naturalist
44(4):484-489.
Bell, M. B. V., A. N. Radford, R. Rose, H. M. Wade, and A. R. Ridley. 2009. The
value of constant surveillance in a risky environment. Proceedings of the Royal
Society Biological Sciences, 276: 2997-3005.
Beradelli, D., M.J. Desmond, and L. Murray. 2010. Reproductive success of
burrowing owls in urban and grassland habitats in southern New Mexico. The
Wilson Journal of Ornithology 122(1): 51-59.
Beuchamp, G. and G. D. Ruxton. 2007. False alarms and the evolution of antipredator
vigilance. Animal Behavior, 74:1199-1206.
Bishop, N.G. and J.L. Culbertson. 1976. Decline of prairie dog towns in southwestern
North Dakota. Journal of Range Management 29:217-220.
Blumstein, D.T., C. Evans, and J.C. Daniel. 2006. J-watcher 1.0. See
http://www.jwatcher.ucla.edu/
Catlin, D.H. andD.K. Rosenberg. 2008. Breeding dispersal and nesting behavior of
burrowing owls following experimental nest predation. American Midland
Naturalist 159:1-7.
50


Chu, M. 2001. Heterospecific responses to scream calls and vocal mimicry by
Phainopeplas (Phainopepla nitens) in distress. Behaviour 138(6): 775-787.
Colorado Division of Wildlife. 2011. [Internet], Wildlife species profiles: burrowing
owl. Colorado Department of Natural Resources, Denver, CO U.S.A.
http://wildlife.state.co.us/WildlifeSpecies/Profiles/Birds/Pages/BurrowingOwl.asp
x. (last accessed 24 January 2012).
COULOMBE, H.N. 1971. Behavior and population ecology of the burrowing owl, Speotyto
cunicularia, in the Imperial Valley of California. The Condor 73(2): 162-176.
Cully, J.F., S.K. Collinge, R.E. VanNimwegen, C. Ray, W.C. Johnson, B.
Thiagarajan, Conlin, andB.E. Holmes. 2010. Spatial variation in keystone
effects: small mammal diversity associated with black-tailed prairie dog colonies.
Ecography 33(4): 667-677.
Davies, J.M. andM. Restani. 2006. Survival and movements of juvenile burrowing
owls during the postfledging period. The Condor 108(2):282-291.
Desmond, M. J., J. A. Savidge, and K.M. Eskridge. 2000. Correlations between
burrowing owl and black-tailed prairie dog declines: a 7-year analysis. The
Journal of Wildlife Management 64(4): 1067-1075.
Economic Research Service (ERS). 2000. [Internet] A history of American
agriculture, 1607-2000. (ERS-POST-12.) Washington, DC. U.S.A.
http://www.agclassroom.org/gan/timeline/index.htm (last accessed 1 March
2012).
Fallow, P.M., J.L. Gardner, and R.D. Magrath. 2011. Sound familiar? Acoustic
similarity provokes responses to unfamiliar Heterospecific alarm calls. Behavioral
Ecology 22(2):401-410.
Forrest, S. 2005. Getting the story right: a response to Vermier and collegues.
Bioscience 55(6): 526-530.
Fredrickson, J.K. and C.N. Slobodchikoff. 2007. Referential specificity in the alarm
calls of the black-tailed prairie dog. Ethology Ecology & Evolution 19: 87-99.
FurrerR.D. andM.B. Manser. 2009. The evolution of urgency-based and functionally
referential alarm calls in ground-dwelling species. American Naturalist 173(3):
400-410.
Goodale, E., G. Beauchamp, R.D. Magrath, J.C. Nieh, and G.D. Ruxton. 2010.
Interspecific information transfer influences animal community structure. Trends
in Ecology and Evolution 25(6): 354-361.
51


Halpin, Z.T. 1983. Naturally occurring encounters between black-tailed prairie dogs
(Cynomys ludovicianus) and snakes. American Midland Naturalist 109(1): 50-54.
Hetrick, S.A. andK.E. Sieving. 2012. Antipredator calls of tufted titmice and
interspecific transfer of encoded threat information. Behavioral Ecology 23(1):
83-92.
Hollen L.I. and A.N. Radford. 2009. The development of alarm call behavior in
mammals and birds. Animal Behavior 78:791-800.
Holroyd, G.L., R. Rodriguez-Estrella, and S.R. Sheffield. 2001. Conservation of
the burrowing owl in western North America: issues, challenges, and
recommendations. Journal of Raptor Research 35(4): 399-407.
Hoogland J.L. 1995. The black-tailed prairie dog: social life of a burrowing mammal.
University of Chicago Press, Chicago.
Ito, R. and A. Mori. 2010. Vigilance against predators induced by eavesdropping on
heterospecific alarm calls in a non-vocal lizard Opiums cuvieri cuvieri
(ReptiliaTguana). Proceedings of the Royal Society Biological Sciences
277(1685): 1275-1280.
Jachowski, D.S., J.J. Millspaugh, D.E. Biggins, T.M. Livieri, and M.R. Matchett.
2008. Implications of black-tailed prairie dog spatial dynamics to black-footed
ferrets. Natural Areas Journal 28(1): 14-25
Johnson, F.R., E.J. McNaughton, C.D. Shelley, andD.T. Bleimstein. 2003.
Mechanisms of Heterospecific recognition in avian mobbing calls. Australian
Journal of Zoology 51(6) 577-585.
Jones, C.G., J.H. Lawton, and M. Shachak. 1994. Organisms as ecosystem engineers.
Oikos 69:373-386.
Kiriazis, J. and C. N. Slobodchikoff. 2006. Perceptual specificity in the alarm calls of
Gunnisons prairie dogs. Behavioral Processes, 73:29-35.
Knopf, F.L. 2008. [Internet], Mountain plover studies Pawnee National Grassland 1985-
2007. Littleton(CO): The Audubon Society of Greater Denver; 2009. 74p.
Available from:
http://www.denveraudubon.com/documents/Knopf%202008%20PNG%20Plover
%20Studies.pdf (Last accessed 21 April 2011).
Kotliar, N.B., B.W. Baker, A.D. Whicker, and G. Plumb. 1999. A critical review of
assumptions about the prairie dog as a keystone species. Environmental
Management 24:177-192.
52


Kotliar, N.B. 2000. Application of the new keystone-species concept to prairie dogs:
how well does it work? Conservation Biology 14:1715-1721.
Lantz, S.J., C.J. Conway, and S.H. Anderson. 2007. Multiscale habitat selection by
burrowing owls in black-tailed prairie dog colonies. Journal of Wildlife
Management 71(8):2664-2672.
Lea, A.J., J.P. Barrera, L.M. Tom, andD.T. Blemstein. 2008. Heterospecific
eavesdropping in a nonsocial species. Behavioral Ecology 19(5): 1041-1046.
Lomolino, M.V. and G. Smith. 2001. Dynamic biogeography of prairie dog (Cynomys
ludovicianus) towns near the edge of their range. Journal of Mammalogy 82(4):
937-945.
Lord, K., M. Feinstein, andR. Coppinger. 2009. Barking and Mobbing. Behavioral
Processes, 81:358-368.
Loughry, W. J. 1987. The dynamics of snake harassment by black-tailed prairie dogs.
Behaviour 103:27-48.
Macedonia, J.M and C.S. Evans. 1993. Variation among mammalian alarm call
systems and the problem of meaning in animal signals. Ethology 93:177-197.
Magrath, R.D., B.J. Pitcher, and J.L. Gardner. 2007. A mutual understanding?
Interspecific responses by birds to each others aerial alarm calls. Behavioral
Ecology 18(5):944-951.
-----. 2009. An avian eavesdropping network: alarm signal reliability and heterospecific
response. Behavioral Ecology 20(4): 745-752.
Martin, D.J. 1973. Selected aspects of burrowing owl ecology and behavior. The
Condor 75(4):446-456.
McDonald, D., N.M. Korfanta, and S.J. Lantz. 2004. [Internet], The Burrowing Owl
(Athene cunicularia)'. a technical conservation assessment. USDA Forest Service,
Rocky Mountain Region. Available from:
http://www.fs.fed.us/r2/projects/scp/assessments/burrowingowl.pdf (last accessed
10 Oct 2010).
Miller, B., G.C. Ceballos, and R. Reading. 1994. The prairie dog and biotic diversity.
Conservation Biology 8(3):677-681.
Morton, E.S. 1977. On the occurrence and significance of motivation-structural rules in
some bird and mammal sounds. The American Naturalist 111:855-869.
53


Mueller, C.A., andM.B. Manser. 2008. The information banded mongooses extract
from Heterospecific alarms. Animal Behaviour 75(3):897-904.
Nocera, J.J. andL.M. Ratcliffe. 2010. Migrant and resident birds adjust antipredator
behavior in response to social information accuracy. Behavioral Ecology 21:121-
128.
Orth, P.B. andP.L. Kennedy. 2001. Do land-use patterns influence nest-site selection
by burrowing owls (Athene cunicularia hypugaea) in northeastern Colorado?
Canadian Journal of Zoology 79: 1038-1045.
Owings, D.H. and W.J. Loughry. 1985. Variation in snake-elicited jump-yipping by
black tailed prairie dogs: ontogeny and snake specificity. Zeitschriftfuer
Tierpsychologie, 70:177-200.
Owings, D. H. and S. C. Owings. 1979. Snake-directed behavior by black-tailed prairie
dogs (Cynomys ludovicianus). Zeitschrift fuer Tierpsychologie, 49:35-54.
Owings, D.H., M.P. Rowe, and A. S. Rundus. 2002. The rattling sound of rattlesnakes
(Crotalus viridis) as a communicative resource for ground squirrels
{Spermophilus beecheyi) and burrowing owls {Athene cunicularia). Journal of
Comparative Psychology 116(2): 197-205.
Pays, O., P. C. Renaud, P. Loisel, M. Petit, J. Gerard, and P. J. Jarman. 2007. Prey
synchronize their vigilant behavior with other group members. Proceedings of the
Royal Society Biological Sciences, 274:1287-1291.
Plumpton, D.L. and S.R. Lutz. 1993. Influence of vehicular traffic on time budgets of
nesting burrowing owls. The Journal of Wildlife Management 57(3): 612-616.
Poulin, R. L. D. Todd, E.A. Haug, B.A. Millsap, andM.S. Martell. 2011. [Internet],
Burrowing Owl {Athene cunicularia), The Birds of North America Online (A.
Poole, Ed.). Cornell Lab of Ornithology, Ithaca, NY, U.S.A.
http://bna.birds.cornell.edu/bna/species/061 (last accessed 24 January 2012).
Power, M.E., D. Tilman, J.A. Estes, B.A. Menge, W.J. Bond, L.S. Mills, G. Daily J.
C. Castilla, J. Lubchenco, and R.T. Paine. 1996. Challenges in the quest for
keystones. BioScience 46:609-620
Rainey, H, J. K. Zuberbuehler, andP.J.B. Slater. 2004. Hornbills can distinguish
between primate alarm calls. Proceedings of the Royal Society Biological
Sciences 271(1540): 755-759.
Randler, C. 2005. Red squirrels {Sciurus vulgaris) respond to alarm calls of Eurasian
jays {Garrulus glandarius). Ethology 112:411-416.
54


Rasa, O.A.E. 1983. Dwarf mongoose and hornbill mutualism in the Taru Desert, Kenya.
Behavioral Ecology and Sociobiology 12:181-190.
Restani, M., L.R. Rau, andD.L. Flath. 2001. Nesting ecology of burrowing owls
occupying black-tailed prairie dog towns in southeastern Montana. Journal of
Raptor Research 35(4):296-303.
Schmidt, K.A., E. Lee, R.S. Ostfeld, and K. Sieving. 2008. Eastern chipmunks
increase their perception of predation risk in response to titmouse alarm calls.
Behavioral Ecology 19(4): 759-763.
Schneider, S.C., M.B. Wunder, andF.L. Knopf. 2006. Relationship between shrubs
and foods in mountain plover habitat in Park County, Colorado. The Southwestern
Naturalist 51 (2): 197-202.
SENKIW, R.W. 2007. The jump-yip display, vigilance, and foraging behavior of the black-
tailed prairie dog (Cynomys ludovicianus). M.Sc. thesis, University of Manitoba,
Winnipeg.
Seyfarth, R.M. andD.L. Cheney. 1984. The acoustic features of vervet monkey
grunts. Journal of the Acoustic Society of America 75(5): 1623-1628.
-----. 2003. Signalers and receivers in animal communication. Annual Review of
Psychology 54:145-173.
SGS-LTER. 2007. [Internet], About the shortgrass steppe, Colorado State University, Ft.
Collins CO, U.S.A. http://sgs.cnr.colostate.edu/about_location.aspx (last accessed
24 January 2012).
Shipley, B.K., R.P. Reading, and B.J. Miller. 2008. Capture rates of reptiles and
amphibians on black-tailed prairie dog (Cynomys ludovicianus) colonies and on
uncolonized prairie in Colorado. Western North American Naturalist 68(2):245-
248.
Sidle, J.G., M. Ball, T. Byer, J.J. Chynoweth, G. Foli, R. Hodorff, G. Moravek, R.
Peterson, andD.N. Svingen. 2001. Occurrence of Burrowing Owls in Black-
Tailed Prairie Dog colonies on Great Plains National Grasslands. Journal of
Raptor Research 35:316-321.
SiH, A., S.F. Hanser, and K. A. McHugh. 2009. Social network theory: new insights and
issues for behavioral ecologists. Behavioral Ecology and Sociobiology 63: 975-
988.
Slobodchikoff, C. N., J. Kiriazis, C. Fischer, and E. Creef. 1991. Semantic
information distinguishing individual predators in the alarm calls of Gunnison's
prairie dogs. Animal Behavior, 42:713-719.
55


Smith, G. andM.V. Lomolino. 2004. Black-tailed prairie dogs and the structure of
avian communities on the shortgrass plains. Oecologia 138(4):592-602.
Smith, J. W., S. L. Smith, J. G. Devilla, and E. C. Oppenheimer. 1976. The Jump-yip
display of the black-tailed prairie dog Cynomys ludovicianus. Animal Behavior,
24:609-621.
Smith, J. W., S. L. Smith, E. C. Oppenheimer, and J. G. Devilla. 1977. Vocalizations of
the black-tailed prairie dog (Cynomys ludovicianus). Animal Behavior, 25:152-
164.
Stapp, P. 1998. A reevalutaion of the role of prairie dogs in Great Plains grasslands.
Conservation Biology 12:1253-1259.
Taylor, A. M. and D. Reby. 2010. The contribution of source-filter theory to mammal
vocal communication research. Journal of Zoology, 280:221-236.
Templeton, C.N. andE. Greene. 2007. Nuthatches eavesdrop on variations in
heterospecific chickadee mobbing alarm calls. Proceedings of the National
Academy of Sciences 104(13): 5479-5482.
Trefry, S.A. andD.S. Hik. 2009. Eavesdropping on the neighbourhood: collared pika
{Ochotona collaris) responses to playback calls of conspecifics and
heterospecifics. Ethology 115: 928-938.
United States Fish and Wildlife Service: black-tailed prairie dog. [Internet],
Mountain-prairie region. Lakewood, CO, USA. updated 2011 May 19. Available
from: http://www.fws.gov/mountain-prairie/species/mammals/btprairiedog/ (last
accessed 24 Feb 2012).
Vitousek, M. N., J.S. Adelman, N.C. Gregory, and J.J.H. St Clair. 2007.
Heterospecific alarm call recognition in a non-vocal reptile. Biology Letters
3:632-634.
Watts, J.M. and J.M. Stookey. 2000. Vocal behavior in cattle: the animals
commentary on its biological processes and welfare. Applied Animal Behaviour
Science 67:15-33.
Whicker, A.D. and J.K. Detling. 1988. Ecological consequences of prairie dog
disturbances. Bioscience 38:778-785.
WlNCHELL, C.S. 1994. Natural History of burrowing owls. Proceedings of the Sixteenth
Vertebrate Pest Conference, Lincoln, NE. 83-86.
56


WUERTHNER, G. 1997. The black-tailed prairie dog: headed for extinction? Journal of
Range Management 50(5): 459-466.
57


Full Text

PAGE 1

INTRASPECIFIC AND INTERSPECIFIC COMMUNICATION ON PRAIRIE DOG COLONIES BLACK TAILED PRAIRIE DOGS AND BURROWING OWLS by Rebecca Dyckes Bryan B.A. University of Colorado Denver 1997 A thesis submitted to the Faculty of the Graduate School of the Un iversity of Colorado in partial fulfillment of the requirements for the degree of Master of Science Biology 2012

PAGE 2

ii This thesis for the degree by Rebecca D. Bryan has been approved for the Department of Integrative Biology by Michael Wunder Advisor Michael Greene Laurel Hartley Date ____ April 26, 2012 _______

PAGE 3

iii Bryan, Rebecca D. ( M.S ., Biology ) Interspecific and intraspecific communication on prairie dog colonies: black tailed prairie dogs and burrowing owls. Thesis directed by Assistant Professor Michael Wunder This thesis is a compilation of research from two different yet related projects. The common thread in these projects is alarm calling and communication on black tailed prairie dog ( Cynomys ludovicianus ) colonies I considered b oth intra specific communication: how black tailed prarie dogs communicate with each other, and inter specific communication: how another species, the Western Burrowing Owl ( Athene cunicularia ), benefits from prairie dog communication. The first chapter co nsists of background information pertinent to communication and alarm calling, and introduces both study species. The second and third chapters are formatted as stand alone manuscripts reporting the results of my field studies I have also included append i c es of additional data analysis information that, while important, for various reasons was not included in the prepared manuscripts. The form and content of this abstract are approved. I recommend its publication. Approved: Michael Wunder

PAGE 4

iv DEDICATION I dedicate this work to my husband and best friend, George.

PAGE 5

v ACKNOWLEDGMENTS I t is impossible to acknowledge every single person who contributed to the completion of this project, but I do want to name a few very important individuals without whom I wo uld never have gotten this far. First and foremost I want to say thanks to my parents, Martha and Doug, who have always known what I am capable of achieving even when I see it. And to the best big sister ever, Melissa: thanks for giving me much neede d teaching and lea rning advice Huge thanks to George for playing the role of supportive husband in every way ; I could never have done this without you. Also, I want to thank my committee members : Mike Greene for reinforcing my love of animal behavior (it ; Laurel Hartley, for encouraging me, offering ideas and advice, and most importantly, friendship; and last, but certainly not least, Mike Wunder, for introducing me to this project idea and letting me run with it even when I sometimes ran in to dead ends. As much as I struggled with it at the time, I will always appreciate that you gave me the freedom to find those dead ends for myself, and that you were there to help me find my way back out. I would also like to acknowledge everyone in the Ec ology Group who helped me brainstorm ideas and let me vent when necessary. Susan Spaulding at BCPOS and A E snake sheds made all the difference! Thanks to the Audubon Soci ety of Greater Denver for funding, the CPER (especially Mary Ashby, who went above and beyond to help me find owl nests there) and the National Forest Service for enabling me to do research on the Pawnee. Also, Marina Corcetti and Zara Hickman : your help i n the field was invaluable and I could not have accomplished half as much without you. Thank you.

PAGE 6

vi TABLE OF CONTENTS CHAPTER I. INTRODUCTION ................................ ................................ ................................ .... 1 Prairie Dogs ................................ ................................ ................................ ......... 1 Burrowing Owls ................................ ................................ ................................ ... 4 Alarm Calling and Eavesdropping ................................ ................................ ........ 6 Figures and Tables ................................ ................................ ............................. 10 II. CHARACTERISTICS OF BLACK TAILED PRAIRIE DOG ( Cynomys ludovicianus) ALARM CALLS IN RESPONSE TO SIMULATED PREDATOR THREATS ................................ ................................ ................................ ..................... 13 Abstract ................................ ................................ ................................ ............. 13 Background ................................ ................................ ................................ ........ 14 Materials and Methods ................................ ................................ ....................... 15 Study Area ................................ ................................ ................................ ... 16 Field Methods ................................ ................................ .............................. 16 Data Analysis ................................ ................................ ............................... 18 Results ................................ ................................ ................................ ............... 18 Duration ................................ ................................ ................................ ....... 19 Frequency ................................ ................................ ................................ .... 19 IQR bandwidth ................................ ................................ ............................. 20 Average Power ................................ ................................ ............................. 20 Discussion ................................ ................................ ................................ .......... 20 Figures and Tables ................................ ................................ ............................. 25 III. BURROWING OWLS EAVESDROP ON ALARM CALLS OF BLACK TAILED PRAIRIE DOGS ................................ ................................ ................................ ............ 28 Abstract ................................ ................................ ................................ ............. 28

PAGE 7

vii Introduction ................................ ................................ ................................ ....... 29 Methods ................................ ................................ ................................ ............. 31 Field Methods ................................ ................................ .............................. 33 Data Analysi s ................................ ................................ ............................... 34 Analysis I Latency to Response. ................................ .......................... 34 Analysis II Intensity of Response. ................................ ........................ 35 Results ................................ ................................ ................................ ............... 35 Latency to Response Times ................................ ................................ .......... 36 Intensity of Response ................................ ................................ ................... 36 Discussion ................................ ................................ ................................ .......... 37 Figures and Tables ................................ ................................ ............................. 41 APPENDIX I: PRAIRIE DOG ALARM CALL ANALYSIS ................................ ......... 45 Results from Balloon Trials ................................ ................................ ................ 45 APPENDIX II: BURROWING OWL RESPONSE ANALYSIS ................................ .... 47 Playback Tr ial Locations ................................ ................................ .................... 47 REFERENCES ................................ ................................ ................................ .............. 50

PAGE 8

viii LIST OF TABLES T able I.1 Cross taxa heterospecific eavesdropping e xamples. ................................ ................. 12 II.1 Structural attributes of alarm spectrograms. ................................ ............................ 27 III.1 Ethogram of Scored Burrowing Owl Behaviors. ................................ .................... 44 III.2 Median Times to Alert Response. ................................ ................................ ......... 44 AI.1 Structural attributes of alarm spectrograms. ................................ .......................... 46 AII.1 Order of sound cue presentation effect on time to alert behavior. ......................... 48 AII.2 Pair wise Wilcoxon test, median time to alert behavior. ................................ ....... 48 AII.3 Order of sound cue presentation effect on response intensity. .............................. 49 AII.4 Chi squared comparison of response strength by stimulus. ................................ .. 49

PAGE 9

ix LIST OF FIGURES Figure I.1 Historic Range of Prairie Dogs. ................................ ................................ ............... 10 I.2 Historic Range of Burrowing Owls. ................................ ................................ ......... 11 II.1 Map of Prairie Dog Study Area, Boulder County, CO. ................................ ........... 25 II.2 Spectrogram measurements. ................................ ................................ ................... 26 II.3 Spectrogram of Snake Responses. ................................ ................................ .......... 27 III.1 Map of Burrowing Owl Study Area, Pawnee National Grasslands, CO. ................ 41 III.2 Latency to Response results. ................................ ................................ ................. 42 III.3 Intensity of Response Results. ................................ ................................ ............... 43

PAGE 10

1 CHAPTER I. INTRODUCTION This project examines b oth i nterspecific and intraspecific communication that takes place on black tailed prairie dog ( Cynomys ludovicianus ) colonies Black tailed prairie dogs emit alarm calls in the presence of threats, and these alarms are specific to the type or i mmediacy of the perceived threat ( Smith et al. 1977 Hoogland 1995) Relatives and distant kin are alerted to danger and are therefore more likely to be able to avoid it This enhances the fitness of the alarm caller, since its offspring and relatives are more likely to avoid the threat (Hoogland 1995) I hypothesize that these alarm calls serve an unintended purpose as well, alerting an other species living on the colony to threats. Western Burrowing owls ( Athene cunicularia ) have high nesting success on bl ack tailed prairie dog colonies. Although there are multiple hypotheses about why this is the case (Desmond et al. 2000, Beradelli et al. 2010) one unexplored idea is that burrowing owls attend and respond to the alarm calls black tailed prairie dogs emit This nesting success. Prairie Dogs Black tailed prairie dogs (hereafter, prairie dogs) are the most common of the 5 prairie dog species found in North America, whi ( Cynomys gunnisoni ) white tailed prairie dogs ( Cynomys leucurus ) Mexican prairie dogs ( Cynomys mexican us ) and Utah prairie dogs ( Cynomys parvidens ) The black tailed prairie dog is an herbivorous rodent of the sciurid family. Adult weights average about 1

PAGE 11

2 kg, and full grown black tailed prairie dogs stand 30 35 cm high. They are native to the western Great Plains of North America, and their range extends from southern Canada through Montana, the Dakotas, Wyoming, Colora do, Ut ah, Arizona, and Texas (Figure I.1) P rairie dogs are diurnal fossorial mammals that live in colonies A verage colony area is 20 to 60 hectares, and colonies can contain from 25 to upwards of one million individuals (Bishop and Culbertson 1976). P rai rie dogs live in territorial family groups known as coteries within the colonies. Coteries normally consist of one male, 2 to 4 breeding females, and juvenile offspring (Hoogland 199 5). Unlike other rodent species, prairie dogs breed only once a year. Fema les give birth to litte rs of 1 to 7 pups in the spring between the months of March and June and care of pups is shared among the female coterie members (Hoogland 1995) Young emerge from the burrows for the first time in mid to late May. Predators of ad ult and juvenile prairie dogs include American badgers ( Taxidea taxus ), swift foxes ( Vulpes velox ), coyotes ( Canis latrans ), black footed ferrets ( Mustela nigripes ), golden eagles ( Aquila chrysaetos ), and red tailed hawks ( Buteo jamaicensis ). Bullsnakes ( P ituophis catenifer ) and prairie rattlesnakes ( Crotalus viridis ) will predate newborn and juvenile prairie dogs (Halpin 1983). The geographic range of the black tailed prairie dog has decreased dramatically over the past 200 years. Fossil records indicate t hat prairie dogs had a relatively stable, large range from 40,000 years ago until the early 19 th century (Lomolino and Smith 2001 Forrest 2005 ). Since then their geographic range has shrunk considerably : the historic occupied habitat of the black tailed p rairie dog is estimated at 35 to 40 million hectar es, but is currently estimated to be less than 1 million hectares (USFWS 2011). This is a

PAGE 12

3 direct result of habitat conversion for agriculture, human land development and consequent eradication efforts (Mill er et al. 1994), and the introduction of plague ( Yersinia pestis ) (Lomolino and Smith 2001 Antolin 2002 ). The decline of active black tailed prairie dog colonies has not only affected this species, but several others as well including small mammals and b irds such as mountain plover ( Charadrius montanus ) and burrowing owl ( Athene cunicularia ) (Miller et al. 1994, Wuerthner 1997, McDonald et al. 2004, Knopf 2008, Shipley et al. 2008, Cully et al. 2010) P rairie dogs can be considered a keystone species (se nsu Power 1996) : their influence on plant and animal communities is disproportionately large compared to their relative abundance (Kotliar et al. 1999, Kotliar 2000, but see Stapp 1998). Prairie dogs also serve the function of ecosystem engineers (sensu Jo nes et al. 1994) because they modify their environment in ways that influence resource availability to other organisms. Prairie dog colonies have lower vegetation height, lower standing biomass, and more bare ground than the surrounding grassland (W h icker and Dettling 1988). Prairie dog colonies provide a large prey base for carnivores and offer shelter and nesting habitat for various species of birds and small mammals (Barko 1997, Wuerthner 1997, Cully et al. 2010) In short and mixed grass prairies of N orth America, prairie dog towns are associated with higher bird species richness and abundance than un colonized areas (Agnew et al. 1986 Barko et al. 1999). Prairie dog grazing on colony sites supplies better visibility of insect prey for insectivorous b ird species (Barko 1997, Schneider et al 2006), better access to seed for granivorous birds (Agnew et al. 1986), better detection of predators for ground nesting birds (Knopf 2006), and the prairie dogs themselves serve as a food source for raptors (Barko et al. 1999).

PAGE 13

4 Burrowing Owls Western Burrowing Owl s ( Athene cunicularia ) are small ground dwelling raptors. Burrowing owl nesting habitat consists primarily of open treeless areas on plains and prairies with short vegetation (McDonald et al. 2004 ). They often ne st in prairie dog burrows in active prairie dog colonies on the Great Plains The burrowing owl breeding range extends to the middle of Alberta and Saskatchewan in the north, and their winter range extends south to Honduras. They can be found in Fl orida and Mexico year round ( F igure I. 2 ). A burrowing owl home range is around 2.5km 2 (Poulin et al. 2011). Burrowing owls begin nesting as early as March, but in northern latitudes nesting may not begin until May. Only female burrowing owls incubate the e ggs; males of the mated pair hunt and forage for both adults and young hatchlings. Burrowing owl young are altricial and completely dependent on parental care until they are 2 months old (Poulin et al. 2011). Juvenile survival is around 57%, and mortality rates for these birds are highest when they first leave the nest before they can fly and when they first become independent from adult owls (Davies and Restani 2006). Burrowing owl predators include badgers, long tailed weasels ( Mustela frenata ), skunk s ( Conepatus spp. Spilogale spp. ), coyotes, and domestic dogs ( Canis familiaris ) and cats ( Felis domesticus ). Other raptors (hawks, owls, and falcons) also hunt burrowing owls. Burrowing owl eggs and chicks can be predated by prairie rattlesnakes and bull snakes. During the nesting season, burrowing owls will dive or fly at mammalian predators near the nest. To avoid avian predators, burrowing owls will escape into their burrows (Poulin et al. 2011). Young burrowing owls will respond to a nest disturbance w ith a rattlesnake like buzzing vocalization which is thought to deter predators (Owings

PAGE 14

5 et al. 2002). If a nest is predated, burrowing owls will disperse and re nest in a different burrow, as long as it is early enough in the breeding season (Catlin and Ro senberg 2008). Burrowing owls will return to the same nesting site year after year if they are successful there (McDonald et al. 2004). Within the nesting site, owls may make use of more than one burrow; nearby burrows can act as escape holes or alternate nest sites if the original site becomes unusable (Winchell 1994), and young disperse to the nearby burrows after their first 2 to 3 weeks of life ( Poulin et al. 2011 ). Burrowing owls show a preference for prairie dog burrows on active colonies. A study on National Grasslands on the Great Plains showed owls present on 69% of active prairie dog colonies, but they were present on only 11% of inactive colonies (Sidle et al. 2001). Burrowing owls are crepuscular, nocturnal, or diurnal depending on the time of ye ar. During breeding season, they are most active at dawn and dusk, but are alert and active during the day as well (Poulin et al. 2011). In Colorado, burrowing owls are found more commonly on than off prairie dog colonies (Smith and Lomolino 2004 Plumpto n and Lutz 1993 ). Burrowing owls have shown an overall declining trend in population in recent ye ars (Desmond et al. 2000), and t he Colorado Division of Wildlife (20 11 ) lists b urrowing owls as a state threatened species. Because of their close association on with prairie dogs, t he decline in prairie dog geographic range may have large effect on how well owls survive C onservation efforts for burrowing owls are closely tied with those for the black tailed prairie dog.

PAGE 15

6 Alarm Calling and Eavesdropping In ma ny species living in social groups, individuals rely on collective vigilance to increase their foraging time while simultaneously decreasing their risk of predation. One or a few group members monitor their surroundings for potential threats while others f orage (Bell et al., 2009). Scanning the surroundings is synchronized between group members (Pays et al. 2007) such that each group member has an opportunity to forage without having to maintain strict individualized vigilance. Vigilant animals emit alarm s when they detect a risk. Alarm signals can be visual, auditory, or chemical depending on the species. For species using audible alarm calls, the a larm signals vary in specificity; some calls are urgency based only, and others encode information about the p articular type of threat (review: Macedonia and Evans 1993; Furrer and Manser 2009). A collective vigilance and a larm calling system minimizes the energy required by any single individual to avoid predators (Beuchamp and Ruxton 2007). P rairie dogs are one such species that use s a sophisticated auditory alarm system to warn other colony members of potential danger. In the presence of perceived danger, prairie dogs emit an anti predator bark (Hoogland 1995). This anti predator bark varies in rate and in acous tic structure depending on the predator threat and, possibly, the mode of approach (Hoogland 199 5 Fredrickson and Slobodchikoff 2007). Barks vary in frequency, duration, and harshness depending on the situation (Smith et al. 1977). Single barks are uncomm on, while continuous bouts of chirping can go on for several minutes at a time. Black territorial disputes and chases, in response to low risk stimuli, or in the absence of any appar ent stimulus at all (Smith et al. 1976). Senkiw (2007) hypothesized that the jump

PAGE 16

7 yip may also serve to provide information about vigilance among group members, Hollen and Ra dford (2009) define three important aspects of a nimal alarm calling behavior as follows: 1) call production, where a call having a specific acoustic structure is created, 2) call usage, the context in which the call is used, and 3) call responses the rea ctions of others to the call. Normally all three of these aspects apply within a single species. For example, if a prairie dog (1) produces an alarm call (2) in the presence of an aerial predator, (3) other prairie dogs will react with the appropriate avo idance behavior, whether they have themselves seen the predator or not (Kiriazis and Slobodchikoff 2006). However, there are cases where aspects (1) and (2) will lead to a response in a different species. This is known as heterospecific eavesdropping. Hete rospecific eavesdropping has been studied extensively in several bird species calls has been noted in n on avian species as well. Evidence of social and non social animals eavesdropping on avian alarm calls has been documented in several species of mammals and some reptiles (Table I.1) Banded mongooses ( Mungos mungo ) increase their vigilance in response to plover ( Vanellus spp. ) alarms (Mueller and Manser 2008), and red squ irrels ( Sciurus vulgaris ), increase vigilance and escape responses when they eavesdrop on alarm calls given by Eurasian jays ( Garralus glandarius ) (Randler 2005). Schmidt et al. (2008) demonstrated that eastern chipmunks ( Tamias striatus ) eavesdrop on tuft ed titmouse ( Baeolophus bicolor ) alarms, and show a stronger anti predator response to that indirect cue than they show to a direct predator cue. Rasa (1983) showed that dwarf mongooses ( Helogale undulata rufula ) forage with two species of hornbill

PAGE 17

8 ( Tocku s deckeni and T. flavirostris alarm calls. This relationship is so important that the mongooses often will not begin to forage without hornbills present, since their presence significantly enhances mongoose dik dik antelope ( Madoqua guentheri ) when they hear go away bird ( Corythiaxoides leucogaster ) alarm calls (Lea et al. 2008). Two species of iguanas, the Galapagos marine iguana ( Amblyrhynchus cristatus ) and the Madagascan spiny tailed iguana ( Oplurus cuvieri cuvieri ) eavesdrop on closely associated birds: Galapagos mocking birds ( Nisomimus parvulus ) and paradise flycatchers ( Terpsiphone mutata ), respectively. Galapagos ma rine iguanas respond to mocking bird alarm calls (but not song) with increased vigilance behavior (Vitousek et al. 2007), and spiny tailed iguanas will also increase their vigilance in response to flycatcher alarms, but not their songs (Ito and Mori 2010). In m ost documented eavesdropping cases a bird species eavesdrop s on a nother bird species, a mammal species eavesdrop s on a nother mammal species, or a mammal species eavesdrop s on a bird species. In an exception to this, Rainey et al. (2004) found eviden ce of a bird species that eavesdrop s on a mammalian alarm call. Yellow casqued hornbills ( Ceratogymna elata ) eavesdrop on alarm calls that Diana monkeys ( Cercopithecus diana ) produce T he h ornbills do not respond to every alarm vocalization Diana monkey s m ake however. Diana monkeys have different alarm calls in response to different predators, and the hornbills are able to differentiate between those alarms, adding to the usefulness of the eavesdropping behavior Hornbills only use energy to respond to ala rm calls for predators that they share with the monkeys and do not waste

PAGE 18

9 energy respond ing to alarms for other, unshared predators (Rainey et al. 2004). In all eavesdropping cases, the eavesdropping individuals are able to enhance their own fitness by pay predator measures, usually in the form of increased vigilance. If burrowing owls living on prairie dog colonies take advantage of a similar eavesdropping behavior, it might add to the explanation of th eir increased nesting success on prairie dog colonies. Not onl y do they benefit from ready made burrows, shortened grass, higher prey availability, and a possible predator dilution effect provided by the prairie dogs (Plumpton and Lutz 1993, Desmond et al. 2000, Orth and Kennedy 2001 ) they also have a built in look out system alerting them to possible threats. Eavesdropping on prairie dog alarm calls could be very valuable to burrowing owls.

PAGE 19

10 Figures and Tables Figure I 1 Historic Range of Prairie Dogs Range of the five prairie dog species in the United States. Black tailed prairie dogs ( Cynomys ludovicianus ) have the broadest range, stretching from Canada to Northern Mexico. Map downloaded from: http:// www.defenders.org/images/programs_policy/Imperiled_Species/prairie_dog/pd_his toric_range.jpg

PAGE 20

11 Figure I 2 Historic Range of Burrowing Owls Breeding and wintering range s of the Western Burrowing Owl ( Ath ene cunicularia ) in the United States.

PAGE 21

12 Table I 1 Cross taxa heterospecific eavesdropping examples A selection of cross taxa eavesdroppers that use avian alarm calls to their advantage. Avian alarm calle r Responding heterospecific esponse Reference Hornbills ( Tockus deckeni and T. flavirostris ) Dwarf mongoose ( Helogale undulata rufula ) Decreases own vigilance behavior in presence of hornbills and a ttends to hornbill alarms while foraging Rasa 1983 Eurasian jay ( Garralus glandarius ) Red squirrel ( Sciurus vulgaris ) Increases vigilance and escape responses when jays alarm Randler 2005 Galapagos mocking bird ( Nisomimus parvulus ) Galapagos marine iguana ( Amblyrhynchus cristatus ) Increase s vigilance in response to alarms, but not in response to song Vitousek et al. 2007 White bellied go away bird ( Corythiaxoides leucogaster ) dik antelope ( Madoqua guentheri ) Decreases foraging activity, increases vigilance in response to al arm calls Lea et al. 2008 Plovers ( Vanellus spp.) Banded mongoose ( Mungos mungo ) Increases vigilance in response to alarm, but does not appear to respond differently when alarm urgency varies Mueller and Manser 2008 Tufted titmouse ( Baeolophus bicolor ) Eastern chipmunk ( Tamias striatus ) Stronger anti predator response to the indirect cue than to direct predator cues Schmidt et al. 2008 Paradise flycatcher ( Terpsiphone mutata ) Madagascan spiny tailed iguana ( Oplurus cuvieri cuvieri ) Increases vigilance in response to alarm calls Ito and Mori 2010

PAGE 22

13 CHAPTER II. CHARACTERISTICS OF BLACK TAILED PRAIRIE DOG ( Cynomys ludovicianus) ALARM CALLS IN RESPONSE TO SIMULATED PREDATOR THREATS Abstract I presented potential ly threat ening stimuli to a colony of black tai led prairie dogs ( Cynomys ludovicanus ) in Boulder County, Colorado and recorded the subsequent vocalized alarm calls. I measured characteristics of prairie dog vocalizations in response to three stimuli: a model snake, a domestic dog, and a human on foot. For each trial, I analyzed the following characteristics of the spectrogram of the first distinct bark in response to the stimulus : 1) the duration of the bark (time in seconds), 2) the central frequency 3) low frequency and 4) high frequenc y (Hz) of the vocalization, 5) the first and third quartile frequencies (Hz), and 6) the average power (db) of the vocalization. I compared these structural characteristics of the a larm calls across stimulus type and found that there were significant differences in bar k duration, IQR bandwidth, and frequency change between the human stimulus and the model s nake stimulus. The greater duration and higher frequency of the snake alarm bark may indicate that prairie dogs perceived the simulated snake as a more imminent loca lized, and urgent threat and that the human may or may not have been perceived as a threat I was unable to draw conclusions about more specific semantic encoding contained in the alarm from these measurements.

PAGE 23

14 Background In many species living in social groups, individuals rely on collective vigilance to increase their foraging time while simultaneously decreasing their risk of predation. One or a few group members monitor their surroundings for potential threats while others forage (Bell et al 2009). S canning the surroundings is synchronized between group members (Pays et al. 2007) such that each group member has an opportunity to forage without having to maintain strict individualized vigilance. Vigilant animals may emit audible alarm calls when they detect a threat A uditory a larm signals vary in specificity; some calls are urgency based only, and others encode information about the particular type of threat (review: Macedonia and Evans 1993; Furrer and Manser 2009). Alarm calling systems like these minimize the energy required by any single individual to avoid predators (Beuchamp and Ruxton 2007). Several studies of alarm calling social anim als have shown that if an animal produces a specific alarm call in the presence of a given predator or a giv en threat level, other group members will react with the appropriate avoidance behavior whether they have themselves seen the predator or not (Seyfarth and Cheney 1984, 2003; Kiriazis and Slobodchikoff 2006; Furrer and Manser 2009). ogs ( Cynomys gunnisoni ) use alarm calls with referential specificity. The calls differ in structure based on predator and mode of approach (Ackers and Slobdchikoff 1999). For instance, the alarm call for a hawk is structurally distinct from the alarm call for a coyote, and prairie dogs take different evasive measures based threat level (A ckers and Slobodchikoff 1999).

PAGE 24

15 Another species of prairie dog, the black tailed prairie dog ( Cynomys ludovicianus ) also use s a sophisticated alarm system to warn other colony members of potential danger. In the presence of perceived danger, black tailed p rairie dogs emit an anti predator alarm bark (Hoogland 1995) : repeated chirps about 0.1s duration The anti predator bark may vary in rate and in sound structure depending on the predator threat and, possibly, the mode of approach (Smith et al 1977 ; Hoo gland 1995 ; Fredrickson and Slobodchikoff 2007). Black which can occur in territorial disputes and chases, in response to low risk stimuli, or in the absence of any apparent stimulus at all (Smith et al 1976). Halpin (1983) and Loughry (1987) noted that jump yips occurred most commonly in the presence of snakes, and Owens and Loughry (1985) posit that the jump yip is a reaction to a less immediate threat while an alarm bark indicates more urgency. F redrickson and Slobodchikoff (2007) showed that black tailed prairie dogs might also encode information about predator threat in their alarm calls. Alarm calls were significantly different in structure when a human previously seen with a weapon, indicating a potential high predatory threat, approached and when a human associated with food, indicating a low predatory threat, approached. black tailed prairie dog vocalization behavio rs, I tested the hypothesis that presenting black tailed prairie dogs with different potentially threatening stimuli would elicit alarm calls that were measurably different in structure I used a domestic dog ( Canis lupus familiaris ), a simulated snake, an d a human as stimuli Materials and Methods

PAGE 25

16 Study Area Experimental trials took place on Carolyn Holmberg Preserve at Rock Creek Farm in southeast Boulder County, Colorado (Figure II.1) This public open space property covers approximately 1100 acres, and is a popular recreation area for hikers, runners, anglers, and bicyclists. The property is surrounded by a shopping mall, major highway, railroad, and neighborhood subdivision. The area was primarily native short grass prairie before its conversion to rang eland, and there are some parcels of uncultivated land containing primarily native vegetation, such as buffalo grass ( Buchloe dactyloides ), blue grama ( Bouteloua gracilis ), and prickly pear cactus ( Opuntia polyacantha ). All trials conducted took place on r angeland which included, along with native grasses, non native grasses and forbs such as cheatgrass ( Bromus tectorum ) and field bindweed ( Convolvulus arvensis ) All of the trial locations were within one half mile of wetlands in the preserve Common plant species in the wetland areas are plains cottonwood ( Populus sargentii ) and coyote willow ( Salix exigua ). The habitat is also home to prairie dog predators including coyotes ( Canis latrans ), red tailed hawks ( Buteo jamaicencis ), bull snakes ( Pituophis caten ifer ), and western rattlesnakes ( Crotalus viridis ). Grazing cattle were rotated around the study site. I did not make recordings on colonies when cattle were present Field Methods I presented stimuli to prairie dogs daily between 0500 h and 1100 h from 2 5 May to 29 July 2010. I digitally recorded all audible responses to each stimulus presented. Stimuli included 1) a domestic dog (German shepherd mix) walking into the colony, 2) a rubber snake pulled through the colony on fishing line, and 3) a human walk ing into the colony at 3 4 kph toward a single focal prairie dog I also presented prairie dogs with a helium filled Mylar balloon to simulate an airborne threat (Appendix

PAGE 26

17 I) However, the presentation of the balloon added confounding variables, particul arly noise, that likely affected response behaviors so analysis of those trials has been omitted from results. To ensure that the prairie dogs were not associating me with the presented stimuli, I remained hidden in a blind throughout all trials, with th e exception of the human trial. Individual animals were not marked, but I rotated the placement of blinds over nine different areas of the colony to decrease the probability of recording responses from the same individuals repeatedly Each of the nine test ing sites was at least 100m from any of the others, and I estimate that at least 20 different animals contributed to my samples. At each testing location I gave prairie dogs in the area two to three day s to acclimatize to the presence of the blind prior t o running any trials. I only ran domestic dog trials on days that I arrived before prairie dogs had begun their morning foraging. I kept the dog on a 15 foot lead anchored behind the blind during all presentations The dog was very accustomed to prairie do gs, and did not vocalize or attempt to chase the prairie dogs at any time For snake trials, I hid the snake at least 20 meters away from the blind under debris on colony so that it would not be visible until I began the trial and started pulling it from b ehind the blind. To simulate a more natural smell, I also wrapped the snake in sheds from a bull snake. I alternated the presentation of two different observers (RDB and a technician) as the human stimulus. Because there were no systematic differences in t he alarm calls in response to the two different people, I considered response alarm calls to both humans as one group In all t rials, I waited until one or more prairie dogs were visible in the trial area before beginning. I made mp3 audio recordings usin g a MarantzPMD661 digital

PAGE 27

18 recorder (Mahwah, NJ) and AudioTechnica shotgun microphone (Stow, OH) which was set up to protrude just outside of the blind, aimed at the prairie dog colony. I started recording 5 seconds before presenting a stimulus, and stop ped recording when the stimulus was no longer visible to the prairie dogs or when they stopped reacting. I recorded calls from prairie dogs that were 15 60 meters from the blind. The microphone did not clearly pick up prairie dog vocalizations from farther than 60 meters away. Trial recordings lasted a minimum of one minute At the end of each day, I downloaded the recordings into a computer running RavenPro 1.4 sound analysis software (Ithaca, NY) Data Analysis For each alarm call recorded, I determined alarm rate by calculating the number of barks in a random ly selected 10 second interval during the stimulus presentation. I compared the averages of alarm rate across trials. For consistency in measuring acoustical structure, I used the first audible bark in response to each presented stimulus for analysis. I measured the duration of the bark (time in seconds), the central, low, and high frequencies (Hz) of the vocalization, the first and third quartile frequencies (Hz), the IQR bandwidth (difference betwee n the first and third quartile frequencies) (Hz), and the average power of the vocalization (db; Fig ure II.2 ). From these measurements, I also calculated the change in frequency: the difference between the highest frequency and lowest frequency harmonic in the vocalization. I used R version 2.11.1 (2010: The R foundation for statistical computing) to run one way ANOVA tests on each of the above measurements by stimulus. I used post hoc multiple comparisons ant differences between measurements by stimulus Results

PAGE 28

19 I was able to take measurements from a total of 33 snake trails, 34 human trials, and 15 dog trials. If sonograms were unclear or had more than one prairie dog bark overlapping, I discarded them fro m my measurements. Because I could only run dog trials on days that I arrived before the prairie dogs emerged, I was unable to run as many dog trials. Other stimuli allowed me to run trials at different blinds on the same day, so I did not have this constr aint. Prairie dogs alarmed in response to each of the presented stimuli. Responses immediately followed the appearance of each stimulus. There were no significant differences among trials in the average number of barks per random 10 s econd interval; prair ie dogs alarmed at an average rate of 16.2 5.6 barks/ten s econds with dog predator responses showing the highest amount of variance Duration One way ANOVA showed a significant effect of stimulus type on the length of the first bark in response to the stimuli (F 2, 79 = 7.0, P = 0.0014). Specifically, post hoc multiple comparisons showed that the response to the snake stimulus was longer in duration than that for the human stimulus (P = 0.013), but the average response durations for snake and human stimu li were not significantly different from that recorded for the dog stimulus (Table II. 1). Frequency Central, low, and high frequency measurements alone did not show significant differences based on stimulus type presented. However, there was a significant effect of stimulus type on the average change in frequency for each alarm call (F 2, 79 = 3.3, P = 0.039). Post hoc multiple comparisons showed that snake alarms were associated with a higher average change in frequency than human alarms (P = 0.04), but nei ther differed significantly from average changes for dog alarms ( T able II. 1)

PAGE 29

20 IQR bandwidth The effect of stimulus type on average IQR bandwidth was also significant (F 2, 79 = 4.4, P = 0.015). Post hoc multiple comparisons showed that alarms in response t o the snake stimulus had a smaller IQR bandwidth than those in response to the human stimulus (P = 0.029). Dog alarm barks were not significantly different from either stimulus (Table II.1) Average Power There was a significant effect of stimulus type on the average power of alarm barks (F 2, 79 = 7.2, P = 0.013). Specifically, post hoc multiple comparisons of average power showed that calls in response to the snake stimulus were higher power than those in response to the human stimulus (P = 0.0012). Respon ses did not differ significantly between human and dog or snake and dog ( T able II. 1) Discussion My results support the hypothesis that black tailed prairie dog alarm calls differ in acoustical structure based on the stimulus presented. These results suppo rt previous et al 1991 ; Ackers and Slobodchikoff 1999) and black tailed prairie dogs (Smith et al. 1977 ; Halpin 1983 ; Owens and Loughry 1985 ; Fredrickson and Slobodchikoff 2007) suggesting that alar m calls vary in structure according to predator threat. Alarm calls in response to the snake stimulus showed a higher maximum power (db) than alarm calls in response the human stimulus. Because I do not know exact distances to the prairie dogs recorded in each trial, it is unclear if the calls were higher in power simply due to the distance to the alarming prairie dog or if the increased decibel level was reflective of the stimulus I could not control for distance of prairie dogs that responded, but becau se my recordings were at random distances across all samples, it

PAGE 30

21 seems unlikely that any distance based bias influenced the results In fact, the consistently random nature of the sampling distance offers further support to the statistically different resu lts being a result of differences in stimulus rather than in distance. Change in frequency indicates the difference between the highest frequency and lowest frequency harmonic in the vocalization, and IQR bandwidth indicates the difference between first an d th ird quartile frequencies (Fig ure II.2 ). Both of these measurements were significantly different in the snake response than the human response : the difference between the lowest and highest harmonic was larger in the snake than in the human, but the dif ference in the third and first quartile frequencies was smaller. This combination of factors would result in a different sound timbre. In the presence of the snake stimulus, I recorded two distinct types of vocalization: a normal alarm bark and a jump yip vocalization (Fig ure II.3 ). Based on previous observations the latter was likely a snake specific alarm call. shape and the behavior accompanying the vocalization are both similar in structure to those described by Smith et al. ( 1976 ), Owin gs and Owings ( 1979 ), and Loughry ( 1987 ). However, I observed this call on only 3 of 30 snake presentations For data analysis, I used the alarm bark that coincided with the snake presentation and not the jump yip. The acoustical differences I measured may convey information about the immediacy of the threat. Longer vocalizations at a higher calling rate may indicate more urgent contexts (review: Taylor and Reby 2010). Alarming prairie dogs in the vicinity of the model snake emitted significantly longer b arks in response to the snake stimulus than to the human, but the rate of calling did not vary significantly. It is possible that the threat

PAGE 31

22 of the snake was more immediate than that of the approaching human, resulting in a more potentially dangerous situa tion and a longer vocalization. The preserve where trials took place is a very heavily used recreation area in an outdoors oriented community. Prairie dogs encounter humans and dogs in this area on a daily, and at times an hourly basis. Snakes are common, but are likely considered more of a threat in this area than are dogs or humans. The snake specific jump yip call occurred on three occasions during trials. The spectrogram associated with this call had a triangular shape distinctly different from the para llel lines on the spectrograms associated with more general alarm barks in res ponse to the model snake (Fig ure II.3 ). A sound that rises and falls equally in a chevron shape may indicate ambivalence on the part of the vocalizing animal (Morton 1977) : the animal has conflicting motivations as to whether it should stay or flee, resulting in the animal doing a combination of both of those behaviors. This is manifested in rapid movements and conspicuous displays known as mobbing (Lord et al ., 2009). Mobbing i s a common anti predator behavior seen in many birds and mammals in response to a predator or an unknown stimulus. An individual giving a mobbing call and/or displaying a mobbing behavioral pattern signals others in the area to join. Whether or not others join, the behavior pattern reflects the ambivalence noted above. The chevron shape of the snake specific jump yip which accompanied both tentative stalking and aggressive attacks on the model snake supports this hypothesis, although stalking and mobbing of the snake occurred without the jump yip as well ; o n 18 such occasions, prairie dogs directly approached the snake stimulus. On four of these occasions they attacked and tore pieces off of the model snake. Prairie dogs retreated from the model snake only t wice,

PAGE 32

23 and one of those times the retreating animal was a juvenile. I did not observe more than one adult prairie dog approaching and attacking the model snake at any given time but did observe other adult prairie dogs in the vicinity observing the model a nd the attacking prairie dog from nearby The observation of more alarm behaviors than jump yip behaviors in response to the snake stimulus contrasts with results from previous studies where the jump yip response to snakes was more commonly recorded (Halpi n 1983 ; Loughry 1987). Perhaps this has to do with the sudden appearance of the snake in the colony when pulled out of its hiding place. Macedonia and Evans (1993) noted that some alarm calling mammals use different alarm calls to differentiate high thre at from lower threat predators, but will use a high threat predator call if danger from a normally low threat predator is immediate. If a jump yip indicates a lower threat predator, but that lower threat predator appears quite suddenly and/or unexpectedly, a more urgent alarm may be used instead. It is also possible that the snake model was not recognized as a snake, and the prairie dogs were alarming in response to a novel threat. to encode referential specificity in their alarm calls. Through discriminate function analysis, they showed that alarm calls for a model skunk, a model coyote, and an oval control showed qualitative and quantitative differences, and were consistent for an individual prairie dog. This suggests that random variation in sonogram characteristics among prairie dogs is less relevant than systematic variation in sonogram characteristics resulting from encoded semantic information. Using a similar analysis method, Fredrickson and Slobodchikoff (2007) showed that black tailed prairie dogs might also

PAGE 33

24 encode semantic information about predator type in their alarm calls. Because I did not mark individuals, I could not be sure that I presented different stimuli to the s ame animal at a trial site, and so I did not use discriminate function analysis to incorporate all measured parameters in statistical analysis. However, I was still able to quantify differences in alarm calls based on the stimulus type. It is therefore qu ite likely that the prairie dogs recorded in this study are also encoding qualitative information about potential threats in their alarm calls. Because black tailed prairie dog alarm calls serve to alert relatives and distant kin of danger (Hoogland 199 5), the more specific the call is to a threat, the more likely it is that those relatives and distant kin will react appropriately Results of this study suggest that black tailed prairie dog alarm calls contain information about threat urgency. However, a lthough urgency information may be contained in the call, the existence of the information does not necessarily make it informative to other prairie dogs (Seyfarth and Cheney 2003) and I did not specifically observe prairie dog responses to these calls The mobbing behavior witnessed in response to the snake model indicates a possibility that the alarm calls convey information to other prairie dogs in the area, but without more specific playback experiments, I do not know if these behaviors were a result of visual or auditory cues

PAGE 34

25 Figures and Tables Figure II 1 Map of Prairie Dog Study Area, Boulder County, CO All prairie dog response tr ia ls took place at the Carolyn Holmberg Preserve at Rock Creek Pa rk, an Open Space property in Boulder County, Colorado. Inset: blind locations marked with stars. Blinds and trials were rotated around these nine sites over the two month period.

PAGE 35

26 Figure II. 2 S pectr ogram measurements Spe ctrogram of two alarm barks in response to the domestic dog. Measurements of (a) central, low, and high frequencies (in Hertz the first and third quartile frequencies and (d) IQR bandwidth Measurements were made using RavenPro1.4 software.

PAGE 36

27 Figure II 3 Spectrogram of Snake Responses Spectrogra m of (a) normal bark in response to the snake stimulus, (b) jump yip recorded in presence of snake stimulus. The normal bark is much shorter in duration, with less harmonic variation and the jump yip vocalization has a distinctive chevron shape seen in an imal vocalizations that are commonly associated with mobbing behaviors Table II 1 Structural attributes of alarm s pectro grams Mean standard error for the first alarm bark in response to a predator stimu lus. Stimuli that share the same superscript letters are not significantly different ( P > 0.05). Different superscript letters indicate significant differences in the means ( P < 0.05). dog ( n = 15) 0.074 0.0035 a b 6314.1 221.58 ab snake ( n = 33) 0.076 0.0029 a 7126.5 188.81 b human ( n = 34) 0.063 0.0022 b 6267.1 302.83 a mean IQR bandwidth (Hz) mean average power (db) dog ( n = 15) 1986.8 282.24 ab 64.93 2.47 ab snake ( n = 33) 1247.6 178.99 a 59.03 1.14 b human ( n = 34) 1889.9 162.71 b 66.66 1.67 a

PAGE 37

28 CHAPTER III. BURROWING OWL S EAVESDROP ON ALARM CALLS OF BLACK TAILED PRAIRIE DOGS Abstract I investigated whether Western Burrowing Owls ( Athene cunicularia ) nesting on black tailed prairie dog ( Cynomys l udovicianus ) colonies responded to playbacks of prairie dog alarm calls. Western burrowing owls are closely associated with black tailed prairie dogs in Colorado and neighboring states on the Great Plains of the United States. Owls have high nesting succes s on prairie dog colonies: they have pre made nest sites and good foraging habitat, and prairie dogs may act as an alternate prey source for predators. Owls nesting on prairie dog colonies may eavesdrop on the prairie dogs as well: owls attending and respo nd ing to the alarm calls that the prairie dogs emit may enhanc e their survival and nesting success on colonies I performed playback experiments with three different treatments: a prairie dog alarm call, a biological control (cattle mooing), and a non biol ogical control (an airplane engine). Owls reacted more quickly to the prairie dog alarm than to the biological control. They increased their vigilance behaviors significantly in response to playbacks of the alarm, but did not s how an increased reaction to either the biological or the non biological control. My results suggest that owls nesting on prairie dog colonies eavesdrop on, and increase their vigilance in response to prairie dog alarm calls. Key words: Burrowing Owl ( Athene cunicularia ), eavesdropp ing, black tailed prairie dog ( Cynomys ludovicianus ), alarm calling, vigilance

PAGE 38

29 Introduction When an individual animal gathers information about the environment it might produce a cue (a vocal or chemical signal) in response to that information for the bene fit of other group members or kin When the cue is produced, there is the potential for unintentional recipients as well (Goodale et al. 2010). Animals sharing a common habitat eavesdropping on their neighbors, particularly when the information comes in the form of a vocal alarm call Heterospecific eavesdropping when members of a different species acquire and benefit from the cue another species produces has been studied ext ensively in birds Magrath et al. (2007) and Fallow et al. (2011) found that different bird species Mobbing behaviors in many bird species are a result of responses to heterospecific alarms (Chu 2001, Johnson et al. 2003, Nocera and Ratcliffe 2010). Evidence of social and non social animals eavesdropping on avian alarm calls has been documented in mammalian and reptilian species as well. M ammalian examples include re d squirrels ( Sciurus vulgaris ), dik dik antelope ( Madoqua guentheri ), banded mongooses ( Mungos mungo ), and Eastern chipmunks ( Tamias striatus ) (Randler 2005, Lea et al. 2008, Mueller and Manser 2008, Schmidt 2008) and reptilian examples include Galapagos marine iguanas ( Amblyrhynchus cristatus ) and Madagascan spiny tailed iguanas ( Oplurus cuvieri cuvieri ) (Vitousek et al. 2007, Ito and Mori 2 010) Rainey et al. (2004) found evidence of a mor e unusual eavesdropper: a bird which eavesdrops on a non avian ala rm call. Yellow casqued hornbills ( Ceratogymna elata ) eavesdrop on alarm calls of Diana monkeys ( Cercopithecus diana ). In this particular case of eavesdropping,

PAGE 39

30 the birds differentiate between alarm calls they hear: they only respond to alarm calls for pre dators that they share with the monkeys, and not to alarm calls for other, unshared predators (Rainey et al. 2004). In all cases, eavesdropping individuals are able to enhance p redator measures, usually in the form of increased vigilance. Western burrowing owls ( Athene cunicularia ) are listed as a species of special concern in several states in the Great Plains, including Montana, Oklahoma, Utah, and Wyoming (Poulin et al. 2011) and are listed as a state threatened species in Colorado (Colorado Division of Wildlife 2011). Western burrowing owls (hereafter, burrowing owls or owls) are closely associated with black tailed prairie dogs ( Cynomys ludovicianus ) in these areas, and the decline in black tailed prairie dogs (hereafter, prairie dogs) due to management and sylvatic plague has been correlated with declines in burrowing owl populations (Desmond et al. 2000). Burrowing owls have high nesting success on prairie dog colonies, a nd are found more often on than off prairie dog colonies in the Western Great Plains (Holroyd et al. 2001). Burrowing owls commonly nest on prairie dog colonies with high burrow densities (Plumpton and Lutz 1993), and show a preference for active prairie d og colonies over abandoned colonies (Restani et al. 2001, Sidle et al. 2001). Multiple hypotheses have been put forth to explain this association between owls and prairie dog colonies. The disturbances prairie dogs create provide pre made nest sites, and t he short grass provides good foraging habitat (Poulin et al. 2011). Also, having prairie dogs nearby may cause a dilution effect if prairie dogs act as an alternative prey source for predators (Desmond et al. 2000). Desmond et al. (2000) and Beradelli et a l. (2010) suggested that burrowing owls nesting on black tailed prairie dog

PAGE 40

31 colonies may eavesdrop on the prairie dogs as well: owls could attend and respond to the alarm calls that the prairie dogs emit. and offer a further explanation of increased nesting success. Burrowing owls use vocal alarm signals to alert each other to dangers, and those alerted (young or a nearby member of a pair) take appropriate actions to avoid the danger (Coulombe 1971). Burrow ing owls sharing a habitat with another species that uses an alarm calling behavior ( eg. prairie dogs or ground squirrels) may increase their vigilance when they hear an alarm from that species. Black tailed prairie dogs emit alarm calls in the presence o f predatory threats to alert relatives and distant kin of danger (Hoogland 1995). Burrowing owls, especially eggs and chicks, share predators with prairie dogs: American badgers ( Taxidea taxus ), coyotes ( Canis latrans ), red tailed hawks ( Buteo jamaicensis ) and possibly prairie rattlesnakes ( Crotalus viridis ) and bullsnakes ( Pituophis catenifer ) predate both prairie dog and burrowing owl young (Halpin 1983, Coulombe 1971). Burrowing owl adults that eavesdrop on alarming prairie dogs can potentially learn wh en predators are approaching. If they increase their vigilance so they can respond appropriately to protect their young, they may enhance their fitness. In this study I tested whether burrowing owls respond to alarm calls that prairie dogs emit. Methods W ith permission from the Central Plains Experimental Range (CPER) and the National Forest Service, I conducted playback experiments on nesting owls on 14 different prairie dog colonies on the CPER and the Pawnee National Grasslands (PNG), lo cated in Weld Co unty, Colorado (Figure III.1) Experiments took place June t hrough July, 2011. The Pawnee National Grasslands is a short grass steppe ecosystem. The

PAGE 41

32 average summer temperature is 22 o C, and annual precipitation patterns can vary greatly. The average precip itation is 32cm, most commonly occurring between April and September. Precipitation manifests primarily as soaking rains in spring and as heavy localized thunderstorms in summer (SGS LTER 2007). All experiments took place on prairie dog colonies on active ranchland, where vegetation included buffalo grass ( Buchloe dactyloides ), blue grama ( Bouteloua gracilis ), prickly pear cactus ( Opuntia Polyacantha ) s carlet globemallow ( Sphaeralcea coccinea ), saltbush ( Atriplex canescens ), and field bindweed ( Convolvul us arvensis ). I conducted audio playbacks to o wls nesting at 20 different locations on 14 active prairie dog colonies; several colonies had more than one nesting owl pair present, and when that was the case, I rotated trials between the active nests. I pla yed 5 minute recordings of a randomized sequence of sounds to each bird: a prairie dog alarm call (the experimental treatment), a small a irplane engine (a non biological control treatment), cattle mooing (a biological control treatment), and ambient noise: intermittent birdsong, insect chatter, and other typical prairie sounds. I collected digital recordings of all sounds using a MarantzPMD661 digital recorder (Mahwah, NJ) and AudioTechnica shotgun microphone (Stow, OH) With the exception of the plane s ound, which was collected near the Rocky Mountain Metropolitan Ai rport in Broomfield, Colorado, I recorded all sound stimuli on the PNG or the CPER. I recorded a herd of mooing cattle on open range on the P NG. I recorded prairie dog alarm call responses to a simulated snake predator (a toy snake pulled through the colony on a fishing line) on the CPER. I edited and combined the sounds using RavenPro 1.4 sound analysis software (Ithaca, NY). To ensure that the order the birds heard the sounds did not affect their responses, I created

PAGE 42

33 three different treatment orders: Order 1 prairie dog alarm (A), cattle mooing (biological control B), airplane (non biological control N); Order 2 B, N, A; and Order 3 N, A, B. I started every trial with two min of ambient noise, and played stimulus sounds at a consistent volume and for the sam e amount of time across each trial Each stimulus sound lasted 30 seconds, and the stimulus sounds were separated from each other by 1 min of ambient noise. I video recorded a ll trials using a JVC Everio HD camcorder (Wayne, NJ). The time frame over which I ran trials coincided with the rearing stages. I did not know if this would affect adult vigilance patterns, so I divided all trials into two categor have surfaced and are beginning to disperse to nearby burrows) I then tested for cues based on nest stage. Field Methods For each trial, I placed a SME AFS portable field speaker (Elmont, NY) within 10 m of an active nest burrow. Owls usually flushed or ducked into the nest or a nearby burrow upon my arrival. Before beginning any pla ybacks, I waited in a blind for the bird to return to the vicinity of the speaker. For most trials the owl was within 10 to 20 m of the speaker. If the owl was more than 30 m away from the speaker, I did not score the trial in data analysis. The blind was 25 to 50 m from the nest burrow and/or focal owl for all trials. I used a pop up hunting blind when it was not possible to use the truck as a blind. The truck was preferable because the owls are accustomed to seeing ranch vehicles on the prairie, so I used it as much as possible. I tested for differences in responses based on blind type when running analyses. B irds were unmarked, so it was not always possible to determine which experiments were repeated

PAGE 43

34 on the same bird on different days I rotated nest sites visited daily to ensure that I did not subject the same owl to the same tests consecutively. Data Analysis I scored video files using JWatcher1.0 software (Blumstein and Daniel 2007) using the ethogram in Table III.1 If the owl moved beyond 30 m o f the speaker during the playback, or ducked underground and did not re emerge, I included only those parts of the trial with the bird within 30 m of the speaker in the analysis. If there were prairie dogs vocalizing in the background of the video, I did n ot include the video in analyses. For analysis, I included vigilance (head turns), bobbing, locomotion, vocalizing, or stretching tall or squatting low while looking. Relaxed responses included preening, yawning, or sleeping (Table III.1 ). For the purposes of this study, I categorized foraging or looking down at and not toward external threats like pre dators. I calculated 1) latency to first alert response to presented sound cues (how quickly the owl reacted to the sound cue), and 2) the intensity of the response (how actively the owl responded) as determined by the number of alert responses within 10 s ec of the presented cue. I considered e ach alert response mutually exclusive. I used R version 2.11.1 (2010: The R foundation for statistical computing) for all statistical analyses. Analysis I Latency to Response I measured how quickly the owls resp onded to each sound stimulus and compared these measurements by sound cue To find the I subtracted the start time of a given sound cue from the start time of the first alert response after that sound began playing. Because

PAGE 44

35 results were not normally distributed, my data did not meet assumptions for parametric statistical tests. I used a Kruskal Wallis rank sum test to determine whether there were differences in median time to response in each of the sound stimulus cat egories: alarm, biological control, and non biological control. In order to determine which categories showed significant differences from one another, I ran pair wise Wilcoxon tests with Bonferroni adjustment. Analysis II Intensity of Response Here I examined the rate of response to sound cues: i.e. how many alert responses the owl showed per unit time after hearing the sound cue. I first 10 sec during the playback of any give n sound cue. I counted the number of alert responses during that 10 sec time frame and assigned the responses to one of three categories 2 alert behavioral rate: 3 4 alert behavioral responses; I administered a chi that one of the categories (Non biological control category 3) had less than 5 data points. Due to P < 0.01), I analyzed truck and pop up blind data separately Results I conducted a total of 9 3 playback experiment trials on adult owls over the two month period ( truck blind n = 51, pop up blind n = 42 ). I collect ed useful data from 58 of those trials (truck blind n = 38, pop up blind n = 20), which included 16 different nesting pairs on 14 prairie dog colonies (Appendix II) Thirty five of th e 9 3 trials were censored

PAGE 45

36 for various reasons, including excessive background noise: high winds ; interruptions such as vehicles, prairie dogs, or cattle; owls leaving the experimental trial area; or too much distance between the video camera and the bird f or accurate behavioral scoring Latency to Response Times There were no systematic differences between the way ANOVA stimulus and phenology: F 2, 151 = 0.62, P = 0.54), or by blind (2 way ANOVA st imulus and blind: F 2, 151 = 0.32, P = 0.72), so I pooled all data in subsequent analyses. Owls showed shorter median response times to the prairie dog alarm cue and the non biological control cue as compared to the biological control sound (Pairwise Wilcox on test with Bonferroni adjustment: Alarm and Non biological cue P =0.045, Alarm and Biological control cue P =0.028). There was no significant difference in the median response time to the non biological control sound cue and the prairie dog alarm sound cue ( P = 1.0) ( T able III.2 Figure III.2 ) Intensity of Response There were no significant differences in response rate category (1 = low, 2 = m oderate 3 = high) based on the order the stimuli were presented P = 0.31, Biolo gical Control order: P = 0.20, Non biological control order: P = P = 0.30). When I conducted Chi squared analysis by blind type, I found fewer category 3 responses ( n =3) than expected with the pop up blind (Fishe P < 0.01) Because of this discrepancy, I analyzed truck data and pop up blind data separately. In trials run from the truck ( n = 38), owls showed more category 3 responses and fewer category 1 responses than expected when presented with th e prairie dog alarm cue. W hen presented with the biological control sound cue t hey showed fewer category 3

PAGE 46

37 responses and more cat egory 1 responses than expected. O wls showed expected responses to the non biological sound cue P = 0.024 ; Fig ure III 3 ). In trials run from the pop up ( n = 20), there were no significant differences in observed P = 0.15). However, I recorded only three category 3 responses in trials run from the po p up blind, and all of those were in response to the prairie dog alarm cue This low number of responses makes statistical analysis difficult Discussion Animals can indirectly receive information about their environment by observing information exchanges between other animals. The eavesdropping animal can obtain this information from visual, auditory, or chemical cues given by the donor. Eavesdroppers get information about traits such as size or strength of the donor, or they can gain information about res ources and predation threats (Sih et al. 2009). For these reasons, eavesdropping can be useful in mate finding and competition, and is especially valuable in predator evasion. Animals that use an anti predator alarm calling system are an important indirec t information source for eavesdropping individuals in their environments. When animals emit alarm calls, the calls may be urgency based (Furrer and Manser 2009), they may have predator specific information encoded (Seyfarth and Cheney 2003), and/or they ma y induce group members to mob a potential threat (Morton 1977). Animals that share a habitat, have similar predators, and are similar in body size to the alarm caller are likely to pay attention to those alarms because doing so can increase their fitness ( Templeton and Greene 2007). Some eavesdroppers are able to accurately interpret the information

PAGE 47

38 the alarms encode (Seyfarth and Cheney 1984, 2003 ; Rainey 2004; Hetrick and Sieving 2012), and some eavesdroppers show more anti predator behaviors in response t o a heterospecific alarm call than they do to direct evidence of a predator (Schmidt et al. 2008). Eavesdropping behavior is common between species of birds and between species of mammals, and cross taxa eavesdropping is effective in many systems as well. My study shows an example of a cross taxa eavesdropping network that likely enhances the fitness of the eavesdropper by alerting them to potential threats in their environments. I found that burrowing owls living on prairie dog colonies increased their vig ilance behaviors in response to prairie dog alarm calls: the owls reacted more quickly to the prairie dog alarm sound cue than to a non alarm biological sound cue (cattle mooing), and continued to show increased vigilance behaviors after their initial resp onse. Upon hearing the prairi e dog alarm call, the owls had fewer category 1 ( low rate ) responses than expected based on chance alone; they showed significantly more category 2 ( moderate rate ) or category 3 (high rate ) response rates than expected. Howeve r, the owls did not flee in response to the alarms and very rarely ( n = 2) vocalized in the 10 seconds following the prairie dog alarm call. This potentially indicates that the owls were increasing their vigilance, but needed first hand information before they would act on those cues. Although the owls reacted equally quickly to the non biological sound cue, they did not continue to show increased vigilance. The non biological sound cue was a sudden loud airplane engine noise that did not particularly resem ble any sounds that the owls were accustomed to hearing at the study site, with the possible exception of occasional passing trucks. It appears that this sudden noise startled the owls into a quick alert

PAGE 48

39 response, but since the noise came from a single dir ection and was not accompanied by any physical apparition, it was likely not seen as an immediate threat. The owls did not flee nor increase their vigilance over the following 10 second interval. The owls nesting on this study site are very accustomed to cattle, encountering the biological control sound cue. Cattle moo to communicate with each other, and although some cow vocalizations may be in response to threats (W atts and Stookey 2000), cattle do not share predators with burrowing owls, and the presence of cattle on burrowing owl and prairie dog habitat is relatively recent from an evolutionary perspective. Owls and prairie dogs have shared habitat for thousands of years, whereas organized cattle ranching on the Great Plains did not begin until the late 1860s (ERS 2000), so even if eavesdropping made sense in this case, it might not have had time to evolve. Also, it may not be necessary for owls to become or to rema in highly vigilant in the presence of cattle: it is possible that with the cattle nearby there is a reduced risk to the owls of predation as well as reduced foraging capabilities, both of which could result in lower vigilance over all. Owls in this area ha ve a very close association with prairie dogs; the owls live in abandoned burrows on active colonies and occasionally interact with foraging or cover seeking prairie dogs that have gotten close to their nests by flying at them or vocalizing toward them (pe rsonal observation). Foraging black tailed prairie dogs are generally quiet and intermittently vigilant. Black tailed prairie dogs do not rely on a single sentinel keeping watch for the entire group; instead all individuals regularly survey their surroundi ngs, and when a threat is detected, the individual detecting the threat will alert

PAGE 49

40 others in the group with a bark alarm (Hoogland 1995). Like prairie dogs in the alarming owls in the area also increase their vigilance behaviors in respo nse to the alarm in an attempt to ascertain the threat. My data suggest that owls nesting on prairie dog colonies likely gain some benefit from eavesdropping on prairie dog alarm calls. Owls increasing their vigilance in response to a prairie dog alarm ma y be more aware of threats in their environment as a result. However, whether the owls are reacting to the urgency of the alarm or semantic information contained in the alarm is still unknown and deserves further exploration. In many eavesdropping species, eavesdroppers show some ability to differentiate alarm calls from non alarm vocalizations ( Vitousek et al. 2007, Goodale et al. 2010 Ito and Mori 2010 ), as well as predator specific alarm calls from each other (Rainey 2004). Eavesdroppers able to differe ntiate would have the advantage of saving energy required for increased vigilance by not responding to calls that do not indicate a threat to the eavesdropper. Rainey (2004) demonstrated that yellow casqued hornbills are able to distinguish between predato r specific calls given by Diana monkeys, and only show increased anti predator behaviors when they hear the alarm for a shared predator. Like Diana monkeys, prairie dogs also emit predator specific calls which differ in response to predator type and mode of approach (Fredrickson and Slobodchikoff 2007). Burrowing owls may possess the ability to differentiate between types of alarm calls. In this study we only looked at reactions to the snake predator alarm F urther studies of owl reactions to specific alar m and non alarm calls could show whether owls are able to decipher the information coded in prairie dog alarms.

PAGE 50

41 Figures and Tables Figure III 1 Map of Burrowing Owl Study Area Pawnee National Grasslands CO Trials took place on the Pawnee National Grasslands and Central Plains Experimental Range in Northeastern Colorado. I conducted a total of 9 3 playback experimental trials on adult owls over the two month period of June and July 2011 I was able to c ollect useful data from 58 of those trials, which incl uded 16 different nesting pairs. Approximate locations of those nesting pairs are indicated by stars.

PAGE 51

42 Figure III 2 Latency to Response results Lat ency to response times showed a skewed distribution. Non parametric analysis of data showed owls had shorter median response times to the prairie dog alarm cue and the non biological control cue as compared to the biological control sound ( Pairwise Wilcoxo n test with Bonferroni adjustment: Alarm vs Non biological cue P =0.045, Alarm vs Biological control cue P =0.028). There was no significant difference in the median response time to the non biological control sound cue and the prairie dog alarm sound cue R esponses to the biological sound cue showed more variance in the length of time from sound cue to first alert response.

PAGE 52

43 Figure III 3 Intensity of Response Results I presented owls with three different sound stimuli: the experimental treatment : a prairie dog barking in response to a snake predator, a biological control : cattle mooing, and a non biological control : a low flying airplane engine. All responses are categorized as 1 low (0 2 vigi lant behaviors in the first 10 seconds of playback), 2 moderate (3 4 vigilant behaviors in the first 10 seconds of playback), or 3 high rate (5 or more vigilant behaviors in the first 10 seconds of playback). Owls showed more moderate and high rate respo nses and fewer low rate responses to the prairie dog alarm call than to the biological or non biological controls. They showed more low rate response s to the P= 0.024

PAGE 53

44 Table III 1 Ethogram of Scored Burrowing Owl Behaviors Ethogram of behaviors scored in J Watcher 1.0. Each of these behaviors was scored Table III 2 Me dian Times to Alert Response I measured the latency from stimulus presentation to alert behavioral response in three categories, Alarm (experimental treatment), Biological Control (cattle mooing), and Non biological control (airplane engine). The median t ime to reaction was faster for the Alarm and the Non biological control than it was for the Biological control. Different superscript letters indicate significant differences in medians. MEDIAN TIME TO ALERT BEHAVIOR S OUND MEDIAN TIME ( MS ) ( RANGE ) P rairie dog Alarm a 1299.5 (216 32912) Biological Control b 2620.0 (250 125824) Non biological Control a 748.0 (279 35417) Kruskal Wallace 2 8.697 P value 0.012 A LERT BEHAVIORS DESCRIPTION R ELAXED BEHAVIORS DES CRIPTION vigilance standing or perching and looking in any direction. Each time bird re focuses (turns head) is a new vigilance bout. sleeping eyes closed or heavily lidded stretching up tall preening actively preening feathers or stretching wings squatting low yawning locomotion walking, running, or flying foraging looking at the ground or following insects vocalizing cooing or alarming bobbing full body or head bob

PAGE 54

45 DATA APPENDIX APPENDIX I: PRAIRIE DOG ALARM CALL ANALYSIS Results from Balloon Trials On 13 of the 17 bird balloon presentations prairie dogs ran to nearby burrows. They proceeded underground into burrows on three of those o ccasions On the other 10 occasions they remained at the lip of a burrow. At least one prairie dog responded vocally to the bird stimulus in 15 of the 17 trials The number of barks in the subsequent alarm call s ranged from 1 250 during a 3 minute period B ird balloon alarm calls were significantly different in duration and change in frequency than those emitted in the presence of the snake stimulus, ( duration : F 3, 95 = 6.5 P =0.0 005 ; delta frequency : F 3, 95 = 3.4 P= 0.0 2 ). Post hoc comparisons using Tukey were on average significantly shorter in duration ( P = 0.01 ) and higher in delta frequency ( P =0.04) than snake alarms. Values did not differ significantly from any of the other calls or in any of the other measurement catego ries. We cannot be sure whether the responses elicited by the bird stimulus were in response to the shape of the balloon, or in response to the loud noise made by the Mylar material in the wind

PAGE 55

46 Table A I 1 Structural attr ibutes of alarm s pectro grams Mean standard error for the first alarm bark in response to a predator stimulus. Stimuli that share the same superscript letters are not significantly different ( P > 0.05). Different superscript letters indicate significant differences in the means ( P < 0.05). dog ( n = 15) 0.074 0.0035 ab 6314.1 221.58 ab snake ( n = 33) 0.076 0.0029 a 7126.5 188.81 b human ( n = 34) 0.063 0.0022 b 6267.1 302.83 a b bird ( n = 17) 0.063 0.0027 b 6008.1 233.04 a mean IQR bandwidth (Hz) mean average power (db) dog ( n = 15) 1986.8 282.24 ab 64.93 2.47 ab snake ( n = 33) 1247.6 178.99 b 59.03 1.14 b human ( n = 34) 1889.9 162.71 a 66.66 1.67 a bird ( n = 17) 1256.5 200.05 ab 60.60 2.20 ab mean high frequency mean central frequency dog ( n = 15) 7967.3 230.57 a 4054.0 156.8 a snake ( n = 33) 8572.2 195.82 a 3659.3 89.02 a human ( n = 34) 7912.0 298.06 a 3668.2 145.4 a bird ( n = 17) 7559.13 246.20 a 3992.5 143.4 a mean low frequency mean 1 st quartile frequency d og ( n = 15) 1653.3 104.7 a 3353.4 177.0 a snake ( n = 33) 1481.2 51.35 a 3053.8 131.0 a human ( n = 34) 1644.9 79.87 a 2898.1 150.7 a bird ( n = 17) 1551.1 35.55 a 3414.9 203.2 a mean 3 rd quartile freqency dog ( n = 15) 4880.9 267.2 a snake ( n = 33) 4301.4 144.7 a human ( n = 34) 4788.0 171.2 a bird ( n = 17) 4761.4 218.6 a

PAGE 56

47 APPENDIX II: BURROWING OWL RESPONSE ANALYSIS Playback Trial Locations I ran playback trials to 16 different nesting pairs of owls and their chicks on the Pawnee Nation al Grasslands and Central Plains Experimental Range. GPS coordinates and c olony names o f nests are listed below in alphabetical order: Central Plains Experimental Range: N40 o 48.001 W104 o 42.293 N40 o 48.115 W104 o 41.426 N40 o 48.059 W104 o 41.278 N40 o 48.074 W104 o 46.255 N40 o 49.032 W104 o 45.564 Pawnee National Grasslands: Carroll : N40 o 42.850 W104 o 36.884 East Stoneham : N40 o 38.732 W103 o 34.882 N40 o 38.934 W103 o 34.914 Halter: N40 o 47.911 W104 o 35.031 N40 o 48.138 W104 o 35.062 Keota: N40 o 38.770 W104 o 08. 036 Reno: N40 o 44.506 W104 o 26.993 N40 o 44.461 W104 o 26.895 Tappy: N40 o 49.546 W103 o 38.664 N40 o 49.475 W103 o 38.611 West Stoneham: N40 o 38.447 W103 o 44.122

PAGE 57

48 Table A II 1 Order of s ound cue presentation effect on time to alert behavior The order stimuli were presented did not affect response times Owls showed similar response times to each sound stimulus regardless of where it occurred in the order. M edian time to alert behavior (ms) (range) Prairie dog Alarm Biol ogical Control Non biological Control Sound order A B N 1129.0 (216 32912) 3216 (344 42489) 735 (461 8948) B N A 1597.5 (296 12874) 2369 (250 125824) 876 (279 35417) N A B 1190.5 (221 19432) 1865 (293 30328) 748 (310 1171 3) K W chi squared 1.2569 2.2518 0.0271 p value 0.5334 0.3244 0.9866 Table A II 2 Pair wise Wilcoxon test median time to alert behavior Bonnferroni adjusted pairwise comparisons of median times to alert behav ior show that the response to the prairie dog alarm was significantly shorter than the response to the biological control, but not the non biological control. Sound Prairie dog Alarm Biological Control Biological Control P = 0.028 Non biological Control P = 1 P = 0.045

PAGE 58

49 Table A II 3 Order of s ound cue presentation effect on response intensity The order in which stimuli were presented did not affect the intensity of response Observed values of intensity did no t differ from expected intensity values regardless of where the sound cue occurred in the order observed/expected (components) stimulus A B N N A B B N A Alarm: Fisher's Exact test p value = 0.3066 category 1 10/7.30 (1.00) 3/5.08 (0.85) 7/7.62 (0.05) category 2 6/9.85 (1.51) 9/6.85 (0.67) 12/10.28 (0.29) category 3 7/5.84 (0.23) 4/4.06 (0.00) 5/6.10 (0.20) Biological control: Fisher's Exact test p value = 0.19562 category 1 14/13.61 (0.01) 10/89.72 (0.01) 11/11.67 (0.04) catego ry 2 6/5.83 (0.00) 2/4.17 (1.13) 7/5.00 (0.80) category 3 1/1.56 (0.20) 3/1.11 (3.21) 0/1.33 (1.33) Non biological Control: Fisher's Exact test p value = 0.3287 category 1 9/9.3 (0.01) 11/7.85 (1.26) 6/8.83 (0.91) category 2 7/6.81 (0.01) 3/5 .74 (1.30) 9/6.45 (1.01) category 3 3/2.87 (0.01) 2/2.42 (0.07) 3/2.71 (0.03) Table A II 4 Chi squared comparison of response strength by stimulus Owls showed fewer mild responses and more moderate and strong responses th an expected in response to the prairie dog alarm call. The showed more mild responses than expected in response to the biological control sound (mooing cattle). Responses were as expected for the non biological control sound (airplane engine). We ran a Fis test because there were fewer than 5 responses in the non biological control category 3. Category observed/expected (components) Alarm Biological Control Non biological Control 1 minimal 13/20 (2.82) 24/17 (2.70) 18/17.19 (0.04) 2 moder ate 16/12 (1.33) 7/10 (1.8) 9/10 (0.10) 3 strong 13/9 (1.40) 4/7.8 (1.86) 8/7.8 (0.00) Fisher's Exact test p value = 0.02449

PAGE 59

50 REFERENCES A CKERS S. H. AND C. N. S LOBODCHIKOFF 1999. Communication of stimulus size and shape in alarm calls of Gunnison Cynomys gunnisoni Etholog y, 105:149 162. A GNEW W D.W. U RESK AND R.M. H ANSEN 1986. Flora and fauna associated with prairie dog colonies and adjacent ungrazed mixed grass prairie in western South Dakota. Journal of Range Management 39 (2):135 139. A NTOLIN M F P. G OBER D. L UCE D.E. B IGGINS W.E. V AN P ELT D.B. S EERY M. L OCKHART AND M. B ALL 2002. The influence of sylvatic plague on North American wildlife at the landscape level, with special emphasis on black footed ferret and p rairie dog conservation. Transactions of the 67 th North American Wildlife and Natural Re sources Conference 67:104 127. B ARKO V A 1997. History of policies concerning the black tailed prairie dog: a review. Proceedings of the Okahoma Academy of Science s 77:27 33. B ARKO V A J.H. S HAW D.M. L ESLIE J R 1999. Birds associated with black tailed prairie dog colonies in southern shortgrass prairie. The Southwestern Naturalist 44(4):484 489. B ELL M. B. V., A. N. R ADFORD R. R OSE H. M. W ADE AND A. R. R I DLEY 2009. The value of constant surveillance in a risky environment. P roceedings of the Roy al Soc iety B iological Sciences, 276: 2997 3005. B ERADELLI D., M.J. D ESMOND AND L. M URRAY 2010. Reproductive success of burrowing owls in urban and grassland ha bitats in southern New Mexico. The Wilson Journal of Ornithology 122(1):51 59. B EUCHAMP G. AND G. D. R UXTON 2007. False alarms and the evolution of antipredator vigilance. Anim al Behav ior 74:1199 1206. B ISHOP N G AND J.L. C ULBERTSON 1976. Decline o f prairie dog towns in southwestern North Dakota. Journal of Range Management 29:217 220. B LUMSTEIN D.T C. E VAN S AND J.C. D ANIEL 2006 J watcher 1.0. See http://www.jwatcher.ucla.edu/ C ATLIN D H AND D.K. R OSENBERG 2008 Breeding dispersal and ne sting behavior of burrowing owls following experimental nest predation. American Midland Naturalist 159:1 7.

PAGE 60

51 C HU M. 2001 Heterospecific responses to scream calls and vocal mimicry by Phainopeplas ( Phainopepla nitens ) in distress. Behaviour 138(6): 775 7 87. C OLORADO D IVISION OF W ILDLIFE 2011. [Internet]. Wildlife species profiles: burrowing owl. Colorado Department of Natural Resources, Denver, CO U.S.A. http://wildlife.state.co.us/WildlifeSpecies/Profiles/Birds/Pages/BurrowingOwl.asp x (last accessed 2 4 January 2012). C OULOMBE H.N. 1971. Behavior and population ecology of the burrowing owl, Speotyto cunicularia in the Imperial Valley of California. The Condor 73(2):162 176. C ULLY J F S.K. C OLLINGE R.E. V AN N IMWEGEN C. R AY W.C. J OHNSON B. T HIAG ARAJAN C ONLIN AND B.E. H OLMES 2010. Spatial variation in keystone effects: small mammal diversity associated with black tailed prairie dog colonies. Ecography 33(4): 667 677. D AVIES J M AND M. R ESTANI 2006. Survival and movements of juvenile burrowi ng owls during the postfledging period. The Condor 108(2):282 291. D ESMOND M.J., J.A. S AVIDGE AND K.M. E SKRIDGE 2000. Correlations between burrowing owl and black tailed prairie dog declines: a 7 year analysis. The Journal of Wildlife Management 64(4): 1067 1075. E CONOMIC R ESEARCH S ERVICE (ERS) 2000. [Internet] A history of American agriculture, 1607 2000 (ERS POST 12.) Washington, DC. U.S.A. http://www.agclassroom.org/gan/timeline/index.htm (last accessed 1 March 2012). F ALLOW P.M J.L. G ARDNER AND R.D. M AGRATH 2011. Sound familiar? Acoustic similarity provokes responses to unfamiliar Heterospecific alarm calls. Behavioral Ecology 22(2):401 410. F ORREST S. 2005. Getting the story right: a response to Vermier and collegues. Bioscience 55(6): 52 6 530. F REDRICKSON J K AND C N S LOBODCHIKOFF 2007. Referential specificity in the alarm calls of the black tailed prairie dog. Ethology Ecology & Evolution 19: 87 99. F URRER R D AND M B M ANSER 2009. The evolution of urgency based and functionally referential alarm calls in ground dwelling species. American Naturalist 173(3): 400 410. G OODALE E G. B EAUCHAMP R.D. M AGRATH J.C. N IEH AND G.D. R UXTON 2010. Interspecific information transfer influences animal community structure. Trends in Ecology and Evolution 25(6): 354 361.

PAGE 61

52 H ALPIN Z T. 1983 Naturally occurring encounters between black tailed prairie dogs (Cynomys ludovicianus) and snakes. American Midland Naturalist 109(1): 50 54. H ETRICK S.A. AND K.E. S IEVING 2012. Antipredator calls of t ufted titmice and interspecific transfer of encoded threat information. Behavioral Ecology 23(1): 83 92. H OLLEN L I AND A.N. R ADFORD 2009. The development of alarm call behavior in mammals and birds. Animal Behavior 78:791 800. H OLROYD G.L., R. R ODRIGU EZ E STRELLA AND S.R. S HEFFIELD 2001. Conservation of the burrowing owl in western N orth America: issues, challenges, and recommendations. Journal of Raptor Research 35(4): 399 407. H OOGLAND J L. 1 995. The black tailed prairie dog: social life of a burro wing mammal. University of Chicago Press, Chicago. I TO R. AND A. M ORI 2010. Vigilance against predators induced by eavesdropping on heterospecific alarm calls in a non vocal lizard Oplurus cuvieri cuvieri (Reptilia:Iguana). Proceedings of the Royal Soci ety Biological Sciences 277(1685):1275 1280. J ACHOWSKI D S J.J. M ILLSPAUGH D.E. B IGGINS T.M. L IVIERI AND M.R. M ATCHETT 2008. Implications of black tailed prairie dog spatial dynamics to black footed ferrets Natural Areas Journal 28(1): 14 25 J OHN SON F.R., E.J. M C N AUGHTON C.D. S HELLEY AND D.T. B LUMSTEIN 2003. Mechanisms of Heterospecific recognition in avian mobbing calls. Australian Journal of Zoology 51(6) 577 585. J ONES C G J.H. L AWTON AND M. S HACHAK 1994. Organisms as ecosystem engin eers. Oikos 69:373 386. K IRIAZIS J. AND C. N. S LOBODCHIKOFF 2006. Perceptual specificity in the alarm calls of Behav ioral Process es 73:29 35. K NOPF F L. 2008. [Internet]. Mountain plover studies Pawnee National Grassland 1 98 5 2007 Littleton(CO): The Audubon Society of Greater Denver; 2009. 74p. Available from: http://www.denveraudubon.com/documents/Knopf%202008%20PNG%20Plover %20Studies.pdf (Last accessed 21 April 2011). K OTLIAR N B B.W. B AKER A.D. W HICKER AND G. P LUMB 1999. A critical review of assumptions about the prairie dog as a keystone species. Environmental Management 24:177 192.

PAGE 62

53 K OTLIAR N B. 2000. Application of the new keystone species concept to prairie dogs: how well does it work? Conservation Biology 14:1715 1721. L ANTZ S J C.J. C ONWAY AND S.H. A NDERSON 2007. Multiscale habitat selection by burrowing owls in black tailed prairie dog colonies. Journal of Wildlife Management 71(8):2664 2672. L EA A.J., J.P. B ARRERA L.M. T OM AND D.T. B LUMSTEIN 2 008. Heterospecific eavesdropping in a nonsocial species. Behavioral Ecology 19(5):1041 1046. L OMOLINO M V AND G. S MITH 2001. Dynamic biogeography of prairie dog ( Cynomys ludovicianus ) towns near the edge of their range. Journal of Mammalogy 82(4): 937 945. L ORD K., M. F EINSTEIN AND R. C OPPINGER 2 009. Barking and Mobbing. Behav ioral Process es 81:358 368. L OUGHRY W. J 1987. The dynamics of snake harassment by black tailed prairie dogs. Behaviour 103:27 48. M ACEDONIA J.M AND C.S. E VANS 1993. Va riation among mammalian alarm call systems and the problem of meaning in animal signals. Ethology 93:177 197. M AGRATH R.D., B.J. P ITCHER AND J.L. G ARDNER 2007. A mutual understanding? Behavioral Ecology 18(5):944 951. -----. 2009. An avian eavesdropping network: alarm signal reliability and heterospecific response. Behavioral Ecology 20(4):745 752. M ARTIN D J. 1973. Selected aspects of burrowing owl ecology and behavior. The Condor 75(4):446 456. M C D ONALD D N.M. K ORFANTA AND S.J. L ANTZ 2004. [Internet]. The Burrowing Owl ( Athene cunicularia ): a technical conservation assessment. USDA Forest Service, Rocky Mountain Region. Available from: http://www.fs.fed.us/r2/projects/scp/as sessments/burrowingowl. pdf ( last accessed 10 Oct 2010 ) M ILLER B G.C. C EBALLOS AND R. R EADING 1994. The prairie dog and biotic diversity. Conservation Biology 8(3):677 681. M ORTON E S 1977. On the occurrence and significance of motivation structur al rules in some bird and mammal sounds. The American Naturalist 111:855 869.

PAGE 63

54 M UELLER C.A., AND M.B. M ANSER 2008 The information banded mongooses extract from Heterospecific alarms. Animal Behaviour 75(3):897 904. N OCERA J.J. AND L.M. R ATCLIFFE 2010 Migrant and resident birds adjust antipredator behavior in response to social information accuracy. Behavioral Ecology 21:121 128. O RTH P B AND P L K ENNEDY 2001. Do land use patterns influence nest site selection by burrowing owls ( Athene cuniculari a hypugaea ) in northeastern Colorado? Canadian Journal of Zoology 79 : 1038 1045. O WINGS D.H AND W.J. L OUGHRY 1985. Variation in snake elicited jump yipping by black tailed prairie dogs: ontogeny and snake specificity. Z eitschrift fuer Tierp sychologie 70:177 200. O WINGS D. H. AND S. C. O WINGS 1979. Snake directed behavior by black tailed prairie dogs ( Cynomys ludovicianus ). Z eitschrift fuer Tierp sychologie 49:35 54. O WINGS D H M.P. R OWE AND A.S. R UNDUS 2002. The rattling sound of rattlesnake s ( Crotalus viridis ) as a communicative resource for ground squirrels ( Spermophilus beecheyi ) and burrowing owls ( Athene cunicularia ). Journal of Comparative Psychology 116(2):197 205. P AYS O., P. C. R ENAUD P. L OISEL M. P ETIT J. G ERARD AND P. J. J ARM AN 2007. Prey synchronize their vigilant behavior with other group members. P roceedings of the Roy al Soc iety B iological Sciences, 274:1287 1291. P LUMPTON D.L. AND S.R. L UTZ 1993. Influence of vehicular traffic on time budgets of nesting burrowing owls. The Journal of Wildlife Management 57(3): 612 616. P OULIN R. L. D. T ODD E.A. H AUG B.A M ILLSAP AND M.S. M ARTELL 2011. [Internet]. Burrowing Owl ( Athene cunicularia ), The Birds of North America Online (A. Poole, Ed.). Cornell Lab of Ornithology, Itha ca, NY, U.S.A. http://bna.birds.cornell.edu/bna/species/061 (last accessed 24 January 2012). P OWER M E D. T ILMAN J.A. E STES B.A. M ENGE W.J. B OND L.S. M ILLS G. D AILY J. C. C ASTILLA J. L UBCHENCO AND R.T. P AINE 1996. Challenges in the quest for keystones. BioScience 46:609 620 R AINEY H., J. K. Z UBERBUEHLER AND P.J.B. S LATER 2004. Hornbills can distinguish between primate alarm calls. Proceedings of the Royal Society Biological Sciences 271(1540): 755 759. R ANDLER C. 2005. Red squirrels ( Sc iurus vulgaris ) respond to alarm calls of Eurasian jays ( Garrulus glandarius ). Ethology 112:411 416.

PAGE 64

55 R ASA O.A.E .1983. Dwarf mongoose and hor n bill mutualism in the Taru Desert, Kenya. Behavioral Ecology and Sociobiology 12:181 190. R ESTANI M., L.R. R AU AND D.L. F LATH 2001. Nesting ecology of burrowing owls occupying black tailed prairie dog towns in southeastern Montana Journal of Raptor Research 35(4):296 303. S CHMIDT K.A., E. L EE R.S. O STFELD AND K. S IEVING 2008. Eastern chipmunks increase thei r perception of predation risk in response to titmouse alarm calls. Behavioral Ecology 19(4): 759 763. S CHNEIDER S C M.B. W UNDER AND F.L. K NOPF 2006. Relationship between shrubs and foods in mountain plover habitat in Park County, Colorado. The South western Naturalist 51(2):197 202. S ENKIW R W. 2007. The jump yip display, vigilance, and foraging behavior of the black tailed prairie dog ( Cynomys ludovicianus ). M.Sc. thesis, University of Manitoba, Winnipeg. S EYFARTH R M AND D.L. C HENEY 1984. The acoustic features of vervet monkey grunts. Journal of the Acoustic Society of America 75(5):1623 1628. -----. 2003. Signalers and receivers in animal communication. Annual Review of Psychology 54:145 173. SGS LTER. 2007 [Internet]. About the shortgrass steppe, Colorado State University, Ft. Collins CO, U.S.A. http://sgs.cnr.colostate.edu/about_location.aspx (last accessed 24 January 2012). S HIPLEY B K R.P. R EADING AND B.J. M ILLER 2008. Capture rates of reptiles and amphibians on black tailed prair ie dog ( Cynomys ludovicianus ) colonies and on uncolonized prairie in Colorado. Western North American Naturalist 68(2):245 248. S IDLE J.G., M. B ALL T. B YER J.J. C HYNOWETH G. F OLI R. H ODORFF G. M ORAVEK R. P ETERSON AND D.N. S VINGEN 2001. Occurrenc e of Burrowing Owls in Black Tailed Prairie Dog colonies on Great Plains National Grasslands. Journal of Raptor Research 35:316 321. S IH A S.F. H ANSER AND K.A. M C H UGH 2009. Social network theory: new insights and issues for behavioral ecologists. Beh avioral Ecology and Sociobiology 63: 975 988. S LOBODCHIKOFF C. N., J. K IRIAZIS C. F ISCHER AND E. C REEF 1991. Semantic information distinguishing individual predators in the alarm calls of Gunnison's prairie dogs. Animal Behav ior, 42 :713 719.

PAGE 65

56 S MITH G AND M.V. L OMOLINO 2004. Black tailed prairie dogs and the structure of avian communities on the shortgrass plains. Oecologia 138(4):592 602. S MITH J. W., S. L. S MITH J. G. D EVILLA AND E. C. O PPENHEIMER 1976. The Jump yip display of the black taile d prairie dog Cynomys ludovicianus. Anim al Behav ior 24:609 621. S MITH J. W., S. L. S MITH E. C. O PPENHEIMER AND J. G. D EVILLA 1977. Vocalizations of the black tailed prairie dog ( Cynomys ludovicianus ). Anim al Behav ior 25:152 164. S TAPP P. 1998 A reevalutaion of the role of prairie dogs in Great Plains grasslands. Conservation Biology 12:1253 1259. T AYLOR A. M. AND D. R EBY 2010. The contribution of source filter theory to mammal vocal communication research. J ournal of Zool ogy 280:221 236. T E MPLETON C.N. AND E. G REENE 2007. Nuthatches eavesdrop on variations in heterospecific chickadee mobbing alarm calls. Proceedings of the National Academy of Sciences 104(13): 5479 5482. T REFRY S .A. AND D.S. H IK 2009. Eavesdropping on the neighbourhood: collared pika ( Ochotona collaris ) responses to playback calls of conspecifics and heterospecifics. Ethology 115: 928 938. U NITED S TATES F ISH AND W ILDLIFE S ERVICE : black tailed prairie dog. [Internet]. Mountain prairie region Lakewood, CO, USA updated 2011 May 19. Available from: http://www.fws.gov/mountain prairie/species/mammals/btprairiedog/ (last accessed 24 Feb 2012). V ITOUSEK M. N., J.S. A DELMAN N.C. G REGORY AND J.J.H. S T C LAIR 2007. Heterospecific alarm call recognition in a non voc al reptile. Biology Letters 3:632 634. W ATTS J.M. AND J.M. S TOOKEY commentary on its biological processes and welfare. Applied Animal Behavio u r Science 67:15 33. W HICKER A D AND J.K. D ETLING 1988. Ecolo gical consequences of prairie dog disturbances. Bioscience 38:778 785. W INCHELL C S. 1994. Natural History of burrowing owls Proceedings of the Sixteenth Vertebrate Pest Conference, Lincoln, NE 83 86.

PAGE 66

57 W UERTHNER G. 1997. The black tailed prairie dog: headed for extinction? Journal of Range Management 50(5): 459 466.