Faculty of Science
SYMPATRIC SONG VARIANT IN MOUNTAIN CHICKADEES (POECILE GAMBELI)
DOES NOT REDUCE AGGRESSION FROM BLACK-CAPPED CHICKADEES (POECILE
ATRICAPILLUS)
2016 | CARA LOUISE SNELL

B.Sc. Honours thesis

SYMPATRIC SONG VARIANT IN MOUNTAIN CHICKADEES (POECILE GAMBELI)
DOES NOT REDUCE AGGRESSION FROM BLACK-CAPPED CHICKADEES
(POECILE ATRICAPILLUS)

by
CARA LOUISE SNELL
A THESIS SUBMITTED IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
BACHELOR OF SCIENCE (HONS.)
in the
DEPARTMENT OF BIOLOGICAL SCIENCES
(Animal Biology)

This thesis has been accepted as conforming to the required standards by:

Matthew Reudink (Ph.D.), Thesis Supervisor, Dept. Biological Sciences
Stephen Joly (B.Sc.), Co-supervisor, Dept. Biological Sciences
Nancy Flood (Ph.D.), Co-supervisor, Dept. Biological Sciences
Tom Dickinson (Ph.D.), Examining Committee member, Dept. Biological Sciences

Dated this 29th day of June, 2016, in Kamloops, British Columbia, Canada

© Cara Louise Snell, 2016

ABSTRACT
Various species interact with one another daily and, when habitats overlap and
species compete for resources, the interactions are often negative. Character
displacements are shifts in traits that typically occur in regions of geographic overlap of
closely related species. These shifts, which act to reduce negative interactions, can be
behavioural, social or morphological. Previous research has shown that mountain
chickadees (Poecile gambeli) have an altered song structure in regions of geographic
overlap with the more dominant black-capped chickadee (Poecile atricapillus). Similar to
the situation for European and Asian tits, mountain chickadees may have changed their
song to decrease aggression from black-cappeds. To test this hypothesis, I conducted a
playback study with black-capped chickadees in Prince George, BC in which I observed
how they responded to the songs of mountain chickadees recorded in regions where the
two species were (1) sympatric and (2) allopatric. I used principal component analysis to
collapse behavioural response variables into a single “approach” variable and a single
“vocalization” variable. I then used mixed-model analysis to determine whether there was
a difference in approach or vocalization response to the two types of mountain chickadee
songs. Black-capped chickadees responded with equal intensity to both types of mountain
chickadee songs. My results demonstrate that mountain chickadee songs with the
sympatric song variant do not reduce heterospecific aggression from black-capped
chickadees. To my knowledge, this is the only instance of a character shift unassociated
with reduced aggression in the family Paridae and raises interesting questions about the
selective pressures leading to the evolution of this song divergence. It is possible that

ii

different selective pressures may result in similar evolutionary outcomes in the form of
altered songs in sympatric populations.

Thesis Supervisor: Associate Professor Matthew Reudink

iii

ACKNOWLEDGEMENTS
There are many people I need to thank for all of the support and guidance I have
received throughout this process. I would like to thank my supervisor Dr. Matthew
Reudink for his endless encouragement, patience and guidance from beginning to end. I
would also like to thank Dr. Ken Otter and Dr. Steffi LaZerte from UNBC for helping to
begin this project, assisting with field work, and for being incredibly patient and helpful
throughout the statistics. Thank you to Stephen Joly, Dr. Nancy Flood, Kristen Marini
and Jackson Kusack for providing me with advice and knowledge throughout my
research. To my friends and family, thank you for being there to listen to all of my prepresentation ramblings and to push me to succeed. Thank you to CUEF UREAP for
funding this project.

iv

TABLE OF CONTENTS

ABSTRACT .................................................................................................................................... ii
TABLE OF CONTENTS ............................................................................................................... v
INTRODUCTION ......................................................................................................................... 1
MATERIALS AND METHODS .................................................................................................. 6
Playback Construction .............................................................................................................. 6
Field Methods ............................................................................................................................ 8
Audio Analysis ......................................................................................................................... 10
Statistical Analysis ................................................................................................................... 10
RESULTS ..................................................................................................................................... 11
Principal Component Analysis ............................................................................................... 11
Approach .................................................................................................................................. 12
Song .......................................................................................................................................... 13
DISCUSSION............................................................................................................................... 14
LITERATURE CITED ............................................................................................................... 20
APPENDIX A .............................................................................................................................. 24

v

INTRODUCTION
When ranges overlap and species compete for resources, negative interactions frequently
occur. Closely related species often inhabit the same ecological niche and have similar life
history traits, which can lead to high levels of competition between them. One way of reducing
these negative interactions is through character displacement (Brown & Wilson 1956). Character
displacements involve divergences in ecological, behavioural, morphological, or physiological
traits between closely related species in regions of geographic overlap, which acts to differentiate
closely related competitors and thus reduce interspecific competition (Brown & Wilson 1956).
Character displacement has been demonstrated across a range of several taxa including several
species in the order Carnivora (Davies et al. 2007) as well as salamanders (Adams 2004), lizards
(Huey 1974; Melville 2002), fish (Schluter & McPhail 1990) and birds (Grant 1972; Doutrelant
et al. 2000; Grava et al. 2013; Hamao et al. 2015).
In carnivores, carnassial teeth are an integral part of food processing and the length of these
teeth can vary based on the prey items being consumed. Davies et al. (2007) used a phylogenetic
analysis with complete species level sampling for Carnivora, and examined all sister species
pairs, excluding marine taxa and domestic dogs and cats. In sister species for which there is
minimal geographical overlap, carnassial teeth are very similar in length. However, sister species
with the greatest geographical overlap show the most differentiation in carnassial length.
Interspecific competition for food resources between sister species in sympatry is suggested as
the cause of this morphological character displacement; divergent carnassial tooth length has
been selected for (Davies et al. 2007).

1

In a well-studied terrestrial salamander genus, Plethodon, aggressive behavior was directly
associated with morphological divergence in both P. jordani and P. teyahalee. In the Southern
Appalachian Mountains of the Eastern United States, P. teyahalee inhabits lower elevations,
whereas P. jordani inhabits higher elevations. However, there is a naturally occurring region of
range overlap where frequent aggressive interactions such as biting, occur. Head shape differs in
both allopatric and sympatric populations of P. jordani and P. teyahalee, but allopatric
populations have less morphological divergence than sympatric populations (Adams 2004).
Two subterranean skink species, Typhlosaurus gariepensis and T. lineatus, occur
sympatrically in the Kalahari Desert, with the smaller T. gariepensis living entirely within the
range of the larger T. lineatus. In this area, T. lineatus have larger snout-vent lengths, head
lengths, and head dimensions than they do in areas where their range does not overlap with that
of allopatric T. lineatus. Based on dietary and morphological evidence, female and immature T.
lineatus have undergone both a morphological and ecological character displacement that
reduces dietary overlap with sympatric T. gariepensis; adult male T. lineatus, however, have not
(Huey 1974). Rather than switching prey taxa altogether, as immature and female T. lineatus
have done, T. lineatus males switch within the same prey taxon. Ultimately, this does not result
in less diet overlap with T. gariepensis compared to the reduction in overlap that has occurred for
females and immatures; nonetheless, it is a unique behavioural change occurring only in
sympatry (Huey 1974).
Similarly, lizard species, Niveoscincus microlepidotus and N. greeni, which occur in alpine
areas, can inhabit two distinct habitat types: boulder fields and alpine heaths. Although alpine
heaths have fewer thermal opportunities (i.e., fewer basking sites) than boulder fields- both
species can use either type of habitat. In allopatry, N. microlepidotus, which is smaller and
2

subordinate to N. greeni, was found significantly more often in boulder habitats than alpine
heaths, whereas in sympatry it is restricted to the latter habitat type. Also, while in sympatry with
the dominant N. greeni, N. microlepidotus exhibits a reduced body size, indicating both a
morphological and ecological character displacement (Melville 2002).
This pattern has also been noted in two subspecies of marine three-spined sticklebacks
(Gasterosteus aculeatus), which exhibit ecological character displacement when in sympatry
with closely related subspecies. Only one subspecies of three-spined sticklebacks typically
occurs in an area, however, pairs of subspecies have been found to co-occur in the Strait of
Georgia, within four drainages around three islands. When isolated from each other, both
subspecies can inhabit the limnetic and benthic zones in a lake. In sympatry, however, one
subspecies inhabits the limnetic zone and the other remains in the benthic zone: the two occupy
separate niches. Distributions of the three-spined sticklebacks to either limnetic or benthic zones
suggest that these subspecies pairs evolved together several times (Schluter & McPhail 1992).
Character displacements have occurred in a range of avian species. Studies on closely related
species pairs, such as great tits (Parus major) and Eurasian blue tits (Cyanistes caeruleus),
mountain chickadees (Poecile gambeli) and black-capped chickadees (Poecile atricapillus), and
varied tits (Poecile varius) and Japanese tits (Parus minor) all found that when such species
lived in sympatry, the subordinate species diverged in expression of shared traits from the sister
species, undergoing a character displacement (Doutrelant et al. 2000; Grava et al. 2013; Hamao
et al. 2015). For example, Eurasian blue tits that inhabit regions where few or no dominant great
tits are present produce songs that are similar to those of great tits (Doutrelant et al 2000). In
regions of sympatry, however, Eurasian blue tits, which are the subordinate species, add a trill to
the end of their song. Great tits react less strongly to the altered Eurasian blue tit trilled songs
3

than they do to either untrilled Eurasian blue tit song or the song of conspecific great tits; the
character shift in Eurasian blue tit song acts to reduce aggression from great tits (Doutrelant et al.
2000). A study by Hamao et al. (2015) of Japanese tits and varied tits also shows this pattern. In
regions where the two species live in sympatry, subordinate Japanese tits sing at a lower
frequency, acoustically diverging from the songs of varied tits, as well as from the songs of
Japanese tits living where varied tits are not present. These results suggest that the subordinate
Japanese tit alters its song characteristics to avoid harassment by the dominant varied tit (Hamao
et al. 2015); although future playback studies are needed to confirm that this results in reduced
aggression.
In North America, the black-capped chickadee is socially dominant to the mountain
chickadee (Grava et al. 2012a; Grava et al. 2012b; Grava et al. 2013). Typically, the two species
segregate due to different habitat preferences and elevation, but historically overlapping zones
exist throughout their ranges (Figure 1). With the onset of forestry practices that create habitat
mosaics that do not occur naturally as often, the two species now frequently overlap (Grava et al.
2012a).

4

Figure 1. Map of North America showing black-capped chickadee range (dark grey), mountain
chickadee range (medium grey) and the zone in which both species occur (light grey). Data
obtained from Bird Life International (Birdlife International 2015).

Mountain chickadees are subordinate to black-capped chickadees and exhibit greater rangewide variation in their song (Lohr 2008), which may be driven by dominance-mediated character
displacement. Grava et al. (2013) found differences in mountain chickadee song structure
between regions of co-occurrence and isolation, whereas black-capped chickadees have a
consistent song throughout North America. In areas of contact between mountain chickadees and
black-capped chickadees, the mountain chickadee song would shift away from the structure of
the black-capped song in some way.
While research by Grava et al. (2012a; 2013) suggests that mountain chickadees add a song
variant in areas of sympatry with black-capped chickadees, no studies have examined whether
5

these sympatric song variants result in fewer negative interactions with dominant black-capped
chickadees, as seen in European and Asian tits (Doutrelant et al. 2000; Hamao et al. 2015).
Following the study design of Doutrelant (2000), which examined the behavioural response of
great tits to Eurasian blue tit vocalizations recorded in areas of sympatry and allopatry, I tested
the hypothesis that mountain chickadees alter their vocalizations to minimize negative
interactions from black-capped chickadees. I predicted that black-capped chickadees would
respond less aggressively to mountain chickadee songs recorded in areas of sympatry (sympatric
song variant) than to allopatric mountain chickadee songs (allopatric song variant).
MATERIALS AND METHODS

Playback Construction
Conforming to playback guidelines outlined by McGregor (1992; 2000) and McGregor et
al. (1992), I created a series of 10 playback dyads (paired playbacks, one playback with
sympatric songs and one with allopatric songs) using mountain chickadee songs from
populations across British Columbia and the United States. Stimuli were obtained from the study
by Grava et al. (2013). “Allopatric” playbacks consisted of mountain chickadee songs taken from
a region in which black-capped chickadees were not present. “Sympatric” playbacks consisted of
mountain chickadee songs recorded from areas where both chickadee species co-exist. Each
playback consisted of 12 songs per minute spaced approximately 5 to 6 seconds apart, for a total
of 24 songs over a total of 2 minutes (Figure 2). The playbacks were created using Avisoft
SASLab Pro software (Specht 2012), which allowed me to standardize the volume and amplitude
of all songs. I used Audacity software (Audacity Team 2008) to select a series of 3-4 different
songs used to create one stimuli type, either allopatric or sympatric. These selected songs were
6

then randomly mixed and repeated within each playback for variation, while still representing
successive songs representing a single stimulus male. All songs used within a playback had an
amplitude ranging between -21 and -18dB, resulting in a playback of either all allopatric songs or
all sympatric songs. Low levels of Brownian noise were added to the playbacks to sound more
natural and to diminish any noises from the editing process. The final stimuli included a 30 s
blank intro followed by two minutes of stimulus broadcast. Unique dyads were created by
randomly pairing one “allopatric” stimulus with a “sympatric” stimulus. I tested each dyad to
ensure a sound level of 75dB during playback. All dyads were loaded onto an Apple iPod Touch
for field playback and broadcast through a Logitech X100 Bluetooth speaker.

Figure 2. A diagram showing the construction of each playback audio file. Every song in each
individual playback is either allopatric or sympatric, and is a combination of 3-4 mountain
chickadee songs from various populations.
7

Field Methods
Playback experiments were conducted in Prince George, B.C. (53°54'40.1"N
122°45'18.6"W), from the 21-29th of April, 2015, following the methods of Doutrelant et al.
(2000). I conducted playback experiments on a total of 22 black-capped chickadees, at 22
separated locations around the University of Northern British Columbia campus, the outskirts of
Prince George and within the Prince George city park, Forests for the World. Playbacks were
performed in the morning (07:00 to 12:00). To prevent interference, I chose focal males if they
were in a territory alone (and/or with a mate) with no competing males in the vicinity, and if the
bird was at least 250m from any previous birds tested that same day. Two to four 10m long
ropes, with 5m markers, were set up equal distances apart radiating out from the middle of each
focal bird’s location, with a Logitech X100 Bluetooth speaker hanging in the centre,
approximately 1m off the ground (Figure 3). The speaker faced the direction of the focal bird,
and was connected via Bluetooth or using a plug-in adapter directly to the iPod.

8

Figure 3. Diagram of the typical playback set-up. All playback trials were set-up based on these
parameters, using two to four 10m ropes for distance estimates.

Each focal chickadee was given a randomly selected playback stimulus (either allopatric
or sympatric), followed by a second playback stimulus of the alternative type 1-2 hours later at
the same location, resulting in a total of two behavioural recordings per bird. The next focal bird
was given a set of playback stimuli in the opposite order to randomize trials and avoid playback
bias resulting from order effects. In each playback trial, focal birds were first called in with a
“primer” call (a short playback of mobbing chick-a-dee calls), and were then given the selected
playback recording. Behavioural responses were observed and recorded vocally using an
Audiotechnica AT8015 microphone and a Marantz PMD661 MKII Professional Portable Flash
Field Recorder. I verbally recorded the focal male’s distance from the speaker, all short and long
9

flights taken, whether there was a mate present, vocalizations made (gargles, calls, or songs), and
potential interruptions (e.g., intruding male, human disturbances such as construction, dogs, etc.).

Audio Analysis
Recordings were individually annotated using Avisoft SASLab Pro bioacoustics
software, taking note of focal bird location and type of responses. These annotations were then
exported into text files. The observations noted in the text files were processed and summarized
using R statistical software (R Development 2014). Specifically, I extracted information on the
focal male’s distance from the speaker playback, time spent at each distance, how many
vocalizations were made and level of aggressiveness with relation to the stimuli they were
exposed to (whether the focal male song was non-overlapping, overlapping, or overlapped in
relation to the playback stimuli), and time spent within 10 m of the speaker after the playback
ended. Only birds with complete recording files of the trial were used in analysis; incomplete
recordings were excluded from analysis (nincomplete = 2).

Statistical Analysis
I used R statistical software to conduct iterative principal component analysis to collapse
the variables into two holistic measures, one describing approach behavior and a second
describing vocal response. Next, I constructed generalized mixed models in JMP 12 (SAS
Institute 2015) to determine whether approach response or vocal responses varied with respect to
the playback presented (allopatric/sympatric). In addition, I included playback order and starting
distance (as well as an interaction term) as covariates. Male ID was included as a random factor
to account for the paired design.

10

RESULTS

Principal Component Analysis
I collected 65 variables for each trial, which were associated with both individual
approach and vocal responses to playback (Table A1). To collapse these variables, we conducted
an iterative principal component analysis, using a broken stick approach to find principal
components that best explained the data. Because many of the distance variables were related,
we collapsed several of these variables (e.g., distance from 0-5m and distance from 5-10m were
collapsed to distance from 0-10m) to derive a total of six variables, for which the first two
principal components explained 70% of the variance (Table 1). PC1 (hereafter called “approach
response”) was most strongly associated with variables related to focal male distance from
speaker during the playback period, and the association was negative (i.e., lower PC1 values
indicate that birds approached more closely [minimum distance], spent more time close to the
speaker [time 0-10m], spend less time far [time >20m], with a weak effect of singing fewer
overlapping songs). PC2 (hereafter called “song response”) was most strongly associated with
variables related to focal male vocalization responses to the stimuli during the playback period
(i.e., spending less time far [time >20m], singing and overlapping songs more; Table 1).

11

Table 1. Principal component analysis results of measures of distance from the speaker and song
response variables. PC1 was most strongly associated with the approach variables, while PC2
was most associated with vocal responses. Bolded values reflect variables with contributions of
greater than 0.33 or less than -0.33, which are considered to make a substantial contribution to
the axis (Ho 2006).

Approach
I analyzed paired responses from 20 individuals (40 observations total). Black-capped
chickadees did not differ in approach response when presented with allopatric vs. sympatric
mountain chickadee song (t19 = -0.58, p = 0.57). Because the approach response may have been
influenced by the initial starting location of the black-capped chickadee, we included starting
distance as a covariate and an interaction between starting distance and playback in the model.
As anticipated, starting distance had a strong effect, but there remained no effect of playback
type, nor was there an interaction between starting distance and playback type (starting distance:
t17 = 7.02, p < 0.0001; playback type: t17 = 1.00, p = 0.33; starting distance*playback type: t17 = 1.31, p = 0.21). Because the interaction term was not significant, it was removed from the model.

12

Starting distance remained significant, but there was no significant effect of playback type
(starting distance: t18 = 7.88, p < 0.0001; playback type: t18 = -0.17, p = 0.87; Figure 4).
Next, we generated a model with an additional effect of playback order (i.e., whether
allopatric was presented first or second) and an interaction term to determine whether there was
an order effect (i.e., whether the song played first [allopatric or sympatric] influenced the blackcapped chickadee responses). Neither the interaction term nor the order effect were significant
(first trial: t18 = 0.62, p = 0.54; playback type: t18 = -0.06, p = 0.96; first trial*playback type: t18 =
-0.49, p = 0.63). Upon removal of the interaction term, there was still no effect of either playback
order or playback type (first trial: t18 = 0.44, p = 0.67; playback type: t19 = -0.58, p = 0.57). When
we included starting distance in the model, there was a significant effect of starting distance,
however playback trial and order were not significant (first trial: t18 = 0.51, p = 0.62; playback
type: t18 = -0.17, p = 0.87; starting distance: t18 = 7.77, p = < 0.0001).

Song
We found no difference in song response to the type of playback presented to blackcapped chickadees (allopatric vs. sympatric; t19 = -0.93, p = 0.36). As with approach response,
we included starting distance as a covariate and an interaction between starting distance and
playback type in the model. There was no effect of starting distance, playback type, or the
interaction term (starting distance: t17 = -0.06, p = 0.95; playback type: t = 0.55, p = 0.59;
playback type*starting distance: t17 = -1.33, p = 0.20). Consequently, the interaction term was
removed from the model. There remained no effect of either starting distance or playback type
(starting distance: t18 = -1.14, p = 0.26; playback type: t18 = -1.01, p = 0.33; Figure 4).

13

Next, we produced models that included playback order and an interaction term.
Playback order, playback type, and the interaction term were all not significant (first trial: t18 = 0.69, p = 0.50; playback type: t18 = -0.92, p = 0.37; first trial*starting distance: t18 = 0.39, p =
0.70), and remained non-significant after removal of the interaction term (first trial: t18 = -0.60, p
= 0.56; playback type: t19 = -0.93, p = 0.37). An additional model was created to examine
playback order as well as starting distance, however, there were still no significant effects (first
trial: t18 = -0.55, p = 0.60; playback type: t18 = -1.00, p = 0.33; starting distance: t18 = -1.10, p =
0.29).

Figure 4. Analysis of response to allopatric vs. sympatric playback. Black-capped chickadee
PC1 approach (A) or PC2 song (B) did not differ between allopatric and sympatric trials.

DISCUSSION
In contrast to studies on European and Asian tits, I did not observe that the song variant
of subordinate mountain chickadees in regions of sympatry reduced aggressive responses from
black-capped chickadees. Black-capped chickadees did not differentiate between sympatric and
14

allopatric mountain chickadee songs in any of their behavioural responses. During both
sympatric and allopatric playback trials, black-capped chickadees reacted by either approaching
the speaker or vocalizing during the trial, suggesting that black-capped chickadees recognize
both song types of mountain chickadees, and respond to them as a potential threat. The sympatric
song variant, or character shift, exhibited by mountain chickadees in populations where the two
species are sympatric, did not appear to minimize negative interactions with black-capped
chickadees. This result is unexpected due to the findings of studies on pairs of tit species (which
are closely related to chickadees) in which the subordinate species do display character shifts that
reduce negative interactions with the more dominant form (Doutrelant et al. 2000; Hamao et al.
2015).
Black-capped chickadees respond less aggressively to heterospecific calls than to
conspecific calls (Grava et al. 2012b). However, when I presented black-capped chickadees with
only heterospecific mountain chickadee songs, there was an equal reaction to both the allopatric
and sympatric versions of this song. One possibility is that the sympatric populations of
mountain chickadees are in the early stages of character displacement; the structure of their
songs may still be very similar to that of conspecifics in allopatric populations. It may continue
to change until it becomes less recognizable to black-capped chickadees, and at that point, it may
be perceived as less of a threat. In other words, the two species may not have co-occurred long
enough for the dominant black-capped chickadee to recognize the sympatric mountain chickadee
songs as indicative of a lower threat, and for the mountain chickadee to differentiate its song
enough to reduce aggression from the dominant species. These small alterations may represent
only the beginning stages of adjustment, on the way to increasing the complexity of mountain
chickadee songs in regions of overlap with the black-capped chickadees.
15

Due to the natural segregation of both species based on habitat type, there tend to be
small areas of overlap rather than large zones. The existence of these isolated areas may result in
independent character shifts within each of the distinct sympatric populations (Grava et al.
2013), that is, there may be a variety of different character shifts. Because chickadees rely on
learning for song development, cultural evolution and the rapid emergence of local dialects
within sympatric populations of mountain chickadees may limit the ability of black-capped
chickadees to recognize and respond to non-local mountain chickadee song variants.
Another possible explanation for why mountain chickadees exhibit character shifts in
sympatry without reduced aggression from black-capped chickadees might relate to female matechoice. Both mountain chickadees and black-capped chickadees are socially monogamous, and
songs are thought to have evolved primarily through sexual selection. Female birds tend to prefer
long, complex and variable songs, leading to selection pressure on males to lengthen or alter
their songs (Catchpole 1980). In black-capped chickadees, females can learn the relative rank of
the male based on non-pitch-based cues in vocalizations, and respond more to dominant male
songs (Hoeschele et al. 2010). However, Ratcliffe and Otter (1996) suggest that females exhibit
reduced response to songs with incorrect internote intervals and flattened within-note frequency
ratios in the fee note. Additionally, Christie et al. (2003) shows that songs are highly stabilized
and that the ability to maintain consistency is a dominance-related characteristic. In regions of
co-occurrence, female mountain chickadees occasionally mate with the dominant black-capped
chickadee males (Grava et al. 2012a), which could act as selective pressure for the male
mountain chickadees to alter their song structures to secure mates. In other words, mountain
chickadee males in the presence of dominant black-capped males may alter songs, such as by

16

adding a song variant in regions of sympatry, to effectively compete for mates with black-capped
males, rather than to reduce aggression from black-capped chickadees.
In black-capped chickadees, social hierarchies govern access to resources, as well as
social and extra-pair mate choice (Ramsay et al. 2000). Based on the social hierarchies that
govern social behaviour, interspecific hierarchies could potentially drive hybridization, if
females are choosing dominant males as extra-pair partners regardless of species (Grava et al.
2012a). Both black-capped and mountain chickadees are monogamous and form season-long
pair-bonds; however, Grava et al. (2012a) found that in regions of overlap, black-capped males
sire the majority of mountain chickadee nestlings through extra-pair copulations, rather than
through social pairings (Grava et al. 2012a). This pattern can also be seen in hybrid zones
between Carolina chickadees (Poecile carolinensis) and black-capped chickadees, where female
black-capped chickadees prefer dominant male Carolina chickadees (Bronson et al. 2003;
Reudink et al. 2006). As the subordinate species, male mountain chickadees may have reduced
expression of a favored dominance trait, leading to mating of the female mountain chickadees
with the dominant male black-capped chickadees. This could reduce the potential reproductive
output of male mountain chickadee (Grava et al. 2012a; Grava et al. 2013). It would thus be
advantageous for the subordinate mountain chickadee to undergo a character shift that could
potentially increase attractiveness.
My results do not follow the pattern of character shifts acting to reduce aggression from
dominant species, as observed in Asian and European tits, close relatives of North American
chickadees. Repeating this study on other black-capped chickadee populations across the shared
range may prove informative, as it is possible that this allopatric population of black-capped
chickadees in Prince George, BC provided an atypical response. The form of character
17

displacements in terms of song divergence varies across sympatric populations. Because altered
songs vary across populations, this may suggest black-capped chickadees require a period of
song learning to recognize local song variants. As such, I would predict that only black-capped
chickadees in sympatric populations may exhibit a differentiated response, and that black-capped
chickadees may only reduce aggression towards the local mountain chickadee variant. In
addition, it may be beneficial to control for black-capped chickadee response to black-capped
songs (i.e., conspecific song). Presenting a male black-capped chickadee with both a blackcapped song (control) and either an allopatric or sympatric mountain chickadee songs would
allow black-capped chickadees to compare between conspecific and heterospecific songs,
potentially resulting in greater differentiation between sympatric and allopatric songs.
Furthermore, incorporation of Carolina chickadee vocalizations, to which black-capped
chickadees in Western Canada have not been exposed, into a playback presented to black-capped
chickadees may allow us to interpret whether black-capped chickadees are merely responding to
a ‘chickadee-like’ vocalization, or if they do in fact differentiate between species. Finally, mate
choice trials in which female mountain chickadees are presented songs from sympatric and
allopatric populations could indicate whether song shifts exhibited by mountain chickadees are
driven by sexual selection pressures.
My results demonstrate that mountain chickadee songs with the sympatric song variant
do not reduce heterospecific aggression from black-capped chickadees. This finding is
particularly interesting as, to my knowledge, within the family Paridae it is the only instance of a
character shift being unassociated with reduced aggression. It thus raises interesting questions
about the selective pressures leading to the evolution of this song divergence. It is possible that

18

different selective pressures (e.g. social, sexual) may result in similar evolutionary outcomes in
the form of altered songs in sympatric populations.

19

LITERATURE CITED
Adams, DC. 2004. Character displacement via aggressive interference in Appalachian
salamanders. Ecology, Evolution and Organismal Biology 85(10):2664-2670.
Audacity Team. 2008. Audacity (Version 1.3.4-beta) [Computer program]. Retrieved May 5,
2008, from http://audacityteam.org/
BirdLife International and NatureServe [Internet]. 2015. Bird species distribution maps of the
world. Cambridge, UK and NatureServe, Arlington, USA: BirdLife International; [cited
2016 Feb 16]. Available from: http://www.birdlife.org/
Bronson CL, Grubb Jr TC, Sattler GD, Braun MJ. 2003. Mate preference: a possible causal
mechanism for a moving hybrid zone. Animal Behaviour 65:489–500.
Brown JR, Wilson EO. 1956. Character displacement. Systematic Zoology 5:49-64.
Catchpole CK. 1980. Sexual selection and the evolution of complex songs among European
warblers of the genus Acrocephalus. Behaviour 74:149–165.
Christie PJ, Mennill DJ, Ratcliffe LM. 2003. Pitch shifts and song structure indicate male quality
in the dawn chorus of black-capped chickadees. Behav Ecol Sociobiol 55:341–348.
Dhondt AA. 1989. Ecological and evolutionary effects of interspecific competition in tits.
Wilson Bulletin. 101:198–216.
Doutrelant C, Leitao A, Otter K, Lambrechts AA. 2000. Effect of Eurasian blue tit song syntax
on great tit territorial responsiveness- an experimental test of the character shift
hypothesis. Behavioural Ecology and Sociobiology. 48:119-124.

20

Davies JT, Meiri S, Barraclough TG, Gittleman JL. 2007. Species co-existence and character
divergence across carnivores. Ecology Letters 10:146–152.
Dawson MRW, Bloomfield LL, Charrier I, Sturdy CB. 2006. Statistical classification of blackcapped (Poecile atricapillus) and mountain chickadee (Poecile gambeli) call notes.
Journal of Comparative Psychology 120(2):147-153.
Grant PR. 1972. Convergent and divergent character displacement. Biological Journal of the
Linnean Society 4:39–68.
Grava A, Grava T, Didier R, Lait LA, Dosso J, Koran E, Burg TM and Otter KA. 2012a.
Interspecific dominance relationships and hybridization between black-capped and
mountain chickadees. Behavioural Ecology 23:566-572.
Grava A, Grava T, Otter K. 2012b. Differential response to interspecific and intraspecific signals
amongst chickadees. Ethology 118(8):711-720.
Grava A, Otter KA, Grava T, LaZerte SE, Poesel A, Rush AC. 2013. Character displacement in
dawn chorusing behaviour of sympatric mountain and black-capped chickadees. Animal
Behaviour 86:177-187.
Hamao S, Sugita N, Nishiumi I. 2015. Geographic variation in bird songs: examination of the
effects of sympatric related species on the acoustic structure of songs. acta ethol 19:81–
90.
Ho, R. 2006. Handbook of univariate and multivariate data analysis and interpretation with
SPSS. Chapman and Hall/CRC.

21

Hoeschele M, Moscicki MK, Otter KA, van Oort H, Fort KT, Farrell TM, Lee H, Robson SWJ,
Sturdy CB. 2010. Dominance signalled in an acoustic ornament. Animal Behaviour
79:657–664.
Huey RB, Pianka ER, Egan ME, Coons LW. 1974. Ecological shifts in sympatry: Kalahari
fossorial lizards (Typhlosaurus). Ecology 55:304–316.
Lohr B. 2008. Pitch-related cues in the songs of sympatric mountain and black-capped
chickadees. Behavioural Processes 77:156–165.
McGregor PK, Catchpole CK, Dabelsteen T, Falls JB, Fusani L, Gerhardt HC, Gilbert F, Horn
AG, Klump GM, Kroodsma DE, Lambrechts MM, McComb KE, Nelson DA,
Pepperberg IM, Ratcliffe L, Searcy WA, Weary DM. 1992. Design of playback
experiments: the Thornbridge hall NATO ARW consensus. In: McGregor PK, editor.
Playback studies in animal communication. New York: Plenum Press. p. 1-10.
McGregor PK. 1992. Quantifying responses to playback: one, many, or composite multivariate
measures? In: McGregor PK, editor. Playback studies in animal communication. New
York: Plenum Press. p. 1-10.
McGregor PK. 2000. Playback experiments: design and analysis. Acta Ethologica 3:3-8.
Melville J. 2002. Competition and character displacement in two species of scincid lizards.
Ecology Letters 5:386–393.
Ratcliffe L, Otter KA. 1996. Ecology and evolution of acoustic communication in birds.
Kroodsma DE, Miller EH, editors. Ithaca: Comstock Publishing Associates. p 339-355.

22

R Development Core Team. 2014. R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria. Available from: http://www.Rproject.org/

Ramsay SM, Otter KA, Mennill DJ, Ratcliffe LM, Boag PT. 2000. Divorce and Extrapair Mating
in Female Black-Capped Chickadees (Parus atricapillus): Separate Strategies with a
Common Target. Behavioral Ecology and Sociobiology 49:18–23.
Reudink MW, Mech SG, Curry RL. 2006. Extrapair paternity and mate choice in a chickadee
hybrid zone. Behavioral Ecology 17:56–62.
SAS Institute. 2015. JMP® Version 12. Cary (NC): SAS Institute Inc.
Schluter D, McPhail JD. 1992. Ecological character displacement and speciation in sticklebacks.
The American Naturalist 140:85–108.
Specht, R. 2012. Avisoft-SASLab Pro. Avisoft Bioacoustics. http://www.avisoft.com/

23

APPENDIX A
Table A1. List of all raw data variables obtained from R statistical software, used for principal
component analysis. Bolded values represent variables condensed using PCA in my data
analysis.
ID
Trial
Allo/sym
First trial
Dist_start
Playback_c_n
Playback_g_n
Playback_c_length
Playback_c_dees
Playback_g_length
Playback_dist_min
Playback_l_min
playback_0 - 1m
Playback_>1-5m
Playback_>5-10m
Playback_>10-20m
Playback_>20m
Playback_s_non-overlapping_n
Playback_s_non-overlapping_length
Playback_s_non-overlapping_amt
Playback_s_non-overlapping_perc
Playback_s_non-overlapping_freq
Playback_s_overlapped_n
Playback_s_overlapped_length
Playback_s_overlapped_amt
Playback_s_overlapped_perc
Playback_s_overlapped_freq
Playback_s_overlapping_n
Playback_s_overlapping_length
Playback_s_overlapping_amt
Playback_s_overlapping_perc
Playback_s_overlapping_freq

The focal black-capped chickadee
Trial number (1 or 2)
Stating whether the trial was an Allopatric or Sympatric playback
First playback trial (allopatric or sympatric)
Distance of focal bird at beginning of playback
Number of calls during the playback
Number of gargles during the playback
Average length of calls during the playback
Total number of dees in a call during the playback
Average length of gargles during the playback
Minimum distance from the speaker during the playback
Latency to the minimum distance during the playback
Time spent between 0 and 1 metres of the speaker during playback
Time spent between 1 and 5 metres of the speaker during playback
Time spent between 5 and 10 metres of the speaker during playback
Time spent between 10 and 20 metres of the speaker during playback
Time spent greater than 20 metres from the speaker during playback
Total number of songs that don’t overlap the stimuli (non-overlapping)
during the playback.
Average length of non-overlapping songs during the playback
Amount of non-overlapping of the stimuli on the focal male songs during
the playback
Percent of non-overlapping of the stimuli on the focal male responses
(songs) during the playback
Frequency of non-overlapping songs during the playback
Total number of songs overlapped by stimuli during the playback
Length of songs overlapped by stimuli during the playback
Amount of overlap of the stimuli on the focal male songs during the
playback
Percent of overlap of the stimuli on the focal male responses (songs)
during the playback
Frequency of the focal male’s overlapped song during the playback
Number of songs that overlap stimuli during the playback
Length of songs that overlap stimuli during the playback
Amount of overlap of the focal male’s song on the stimuli during the
playback
Percent of overlap of the focal male’s song on the stimuli during the
playback
Frequency of the focal male’s song when overlapping the stimuli during
the playback
24

Playback_s_total_n
Playback_s_total_length
Playback_s_total_freq
Post_c_n
Post_g_n
Post_c_length
Post_c_dees
Post_g_length
Post_dist_min
Post_l_min
Post_0-1m
Post_>1-5m
Post_>5-10m
Post_>10-20m
Post_>20m
Post_s_non-overlapping_n

Post_s_overlapping_freq
Post_s_total_n

Total number of focal male song responses during the playback
Total length of focal male song responses during the playback
Total frequency of focal male song responses during the playback
Number of calls during the post period
Number of gargles during the post period
Length of calls during the post period
Number of dees during the post period
Length of gargles during the post period
Minimum distance to the speaker during the post period
Latency to the minimum distance during the post period
Time spent between 0 and 1 metres of the speaker during post period
Time spent between 1 and 5 metres of the speaker during post period
Time spent between 5 and 10 metres of the speaker during post period
Time spent between 10 and 20 metres of the speaker during post period
Time spent greater than 20 metres from the speaker during post period
Number of songs that don’t overlap the stimuli (non-overlapping) during
the post period
Average length of non-overlapping songs during the post period
Amount of non-overlapping of the focal male’s song on the stimuli
during the post period
Percent of non-overlapping of the focal male’s song on the stimuli during
the post period
Frequency of non-overlapping songs during the post period
Number of songs overlapped by stimuli during the post period
Average length of overlapped songs during the post period
Amount of overlapped songs of the focal male on the stimuli during the
post period
Percent of overlapped of the focal male’s song on the stimuli during the
post period
Frequency of overlapped songs during the post period
Number of songs that overlap stimuli during the post period
Average length of overlapping songs during the post period
Amount of overlap of the focal male’s song on the stimuli during the post
period
Percent of overlap of the focal male’s song on the stimuli during the post
period
Frequency of overlapping songs during the post period
Total number of focal male song responses during the post period

Post_s_total_length

Total length of focal male song responses during the post period

Post_s_total_freq

Total frequency of focal male song responses during the post period

Post_s_non-overlapping_length
Post_s_non-overlapping_amt
Post_s_non-overlapping_perc
Post_s_non-overlapping_freq
Post_s_overlapped_n
Post_s_overlapped_length
Post_s_overlapped_amt
Post_s_overlapped_perc
Post_s_overlapped_freq
Post_s_overlapping_n
Post_s_overlapping_length
Post_s_overlapping_amt
Post_s_overlapping_perc

25