The Impact of Urban Centers on American Goldfinches (Spinus Tristis) Overwintering at High Latitudes" (2017)

Home , Spinus (bird)

Eastern Michigan University DigitalCommons@EMU Master's Theses, and Doctoral Dissertations, and Master's Theses and Doctoral Dissertations Graduate Capstone Projects

2017 The impact of urban centers on American goldfinches S( pinus tristis) overwintering at high latitudes Corrie Jean Navis

Follow this and additional works at: http://commons.emich.edu/theses Part of the Biology Commons

Recommended Citation Navis, Corrie Jean, "The impact of urban centers on American goldfinches (Spinus tristis) overwintering at high latitudes" (2017). Master's Theses and Doctoral Dissertations. 908. http://commons.emich.edu/theses/908

This Open Access Thesis is brought to you for free and open access by the Master's Theses, and Doctoral Dissertations, and Graduate Capstone Projects at DigitalCommons@EMU. It has been accepted for inclusion in Master's Theses and Doctoral Dissertations by an authorized administrator of DigitalCommons@EMU. For more information, please contact [email protected].

The Impact of Urban Centers on American Goldfinches (Spinus tristis)

Overwintering at High Latitudes

Corrie J. Navis

Thesis

Submitted to the Department of Biology

Eastern Michigan University

in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

Biology

with a concentration in

Ecology, Evolution, and Organismal Biology

Thesis Committee:

Peter Bednekoff, PhD, Chair

Jamie Cornelius, PhD

Katherine Greenwald, PhD

December 7, 2016

Ypsilanti, MI

DEDICATION

This thesis is dedicated to my grandmother, Blanche VanderBent, at whose bird feeders I first encountered Spinus tristis.

ACKNOWLEDGMENTS

I would like to thank my thesis committee for their tireless support and advice during my completion of this project. Assistance with field research was provided by Giorgia

Auteri, Jeanette Bailey, Rebecca Cameron, Joseph Hill, Emily Weckesser, and Megan Wurtz, for which I am very grateful. I also appreciate the cooperation and encouragement of

Michigan Audubon, Washtenaw County Natural Areas Preservation Program, Pittsfield

Charter Township, and Mike and Susan Kielb, who granted me permission to conduct research on their properties. Thanks are due to Tony Bedogne and Caitlin Boon for assistance with GIS matters. My great thanks go to the EMU Biology Department for their support through a Meta Hellwig and Don Brown Research Fellowship, and EMU Women in

Philanthropy for their support through a research grant. Finally, I would like to thank all the members of the EMU bird lab for their feedback and support throughout this project.

iii

ABSTRACT.---American Goldfinches (Spinus tristis) are partial migrants that overwinter in northern regions where winters can be extreme. To understand the role of urban development in facilitating use of a harsh climate, I conducted winter point count surveys (n = 297) at urban and rural sites across latitudes in Michigan. The following winter

I trapped individuals in urban and rural settings to sample baseline and stress-induced plasma corticosterone (CORT) levels. Analyses revealed no interaction of latitude and urbanization in the winter distribution of American Goldfinches. Similarly, CORT concentrations did not differ between individuals trapped in urban and rural settings. It does not appear that

American Goldfinches preferentially use urban areas in winter, and urban individuals do not exhibit different CORT responses to stressors. Birds captured at sites where feeders were regularly provided did display higher baseline CORT concentrations; birds with higher

CORT may be more likely to locate supplemental food.

TABLE OF CONTENTS

Dedication ...... ii

Acknowledgements ...... iii

Abstract ...... iv

List of Figures ...... viii

Chapter 1: Do American Goldfinches (Spinus tristis) Use Cities to Facilitate Overwintering at High Latitudes?

Introduction ...... 1

Drawbacks of Cities ...... 1

Benefits of Cities...... 2

Winter Trends in Northern Cities ...... 3

Gaps in Existing Knowledge ...... 4

American Goldfinch, Spinus tristis, as Study Organism ...... 5

Objectives ...... 8

Hypothesis and Predictions ...... 8

Methods...... 9

Site Selection ...... 9

Survey Protocol ...... 11

Calculation of Urbanization ...... 13

Data Analysis ...... 14

Results ...... 15

Latitude ...... 15

Urbanization ...... 18

Interaction of Latitude and Urbanization ...... 20

Environmental Measures ...... 21

Discussion ...... 23

Latitude ...... 23

Urbanization ...... 23

Interaction of Latitude and Urbanization ...... 25

Environmental Measures ...... 26

Conclusions and Future Research ...... 27

Chapter 2: Do American Goldfinches (Spinus tristis) Overwintering in Urban and Rural Settings Experience Different Physiological Responses to Stressors?

Introduction ...... 30

Adaptive Function of Corticosterone ...... 31

Risks of Elevated Corticosterone ...... 32

Environmental Stressors ...... 34

Gaps in Existing Knowledge ...... 34

American Goldfinch, Spinus tristis, as Study Organism ...... 35

Objectives ...... 36

Hypothesis and Predictions ...... 36

Methods...... 37

Site Selection ...... 37

Calculation of Urbanization ...... 37

Timeline ...... 38

Sampling Protocol ...... 38

Laboratory Analysis ...... 41

Data Analysis ...... 42

Results ...... 43

Urban/Rural Stress Physiology ...... 43

Other Urban/Rural Differences ...... 45

Supplemental Food and Stress Physiology ...... 48

Discussion ...... 51

Urban/Rural Stress Physiology ...... 51

Other Urban/Rural Differences ...... 52

Supplemental Food and Stress Physiology ...... 53

Conclusions and Future Research ...... 54

Works Cited ...... 56

Appendix A: Bird species detected during American Goldfinch (Spinus tristis) vocalization playback surveys, 2015...... 66

Appendix B: Corticosterone (CORT) and morphological measures of American Goldfinches (Spinus tristis) captured, 2016 ...... 72

vii

LIST OF FIGURES

1.1 Established distribution of American Goldfinch, Spinus tristis...... 7

1.2 Selected survey sites for urban and rural Spinus tristis surveys ...... 10

1.3 Example of stratification by habitat type ...... 11

1.4 Seasonal Spinus tristis presence by latitudinal zone ...... 16

1.5 Seasonal Spinus tristis abundance by latitudinal zone ...... 16

1.6 Seasonal Spinus tristis presence by latitude ...... 17

1.7 Seasonal Spinus tristis abundance by latitude ...... 17

1.8 Quantified degree of urbanization of survey sites ...... 18

1.9 Abundance of Spinus tristis at urban and rural survey sites ...... 19

1.10 Spinus tristis presence by degree of urbanization ...... 19

1.11 Spinus tristis abundance by degree of urbanization ...... 20

1.12 Spinus tristis presence in urban/rural settings by latitudinal zone ...... 21

1.13 Spinus tristis abundance in urban/rural settings by latitudinal zone ...... 22

2.1 The role of corticosterone (CORT) in the hypothalamic-pituitary-adrenal (HPA) axis ...... 31

2.2 Baseline CORT concentration in Spinus tristis in rural and urban settings ...... 44

2.3 Induced (30-minute) CORT concentration in Spinus tristis in rural and urban settings ...... 44

2.4 Percent increase in CORT concentration (from baseline to 30-minute level) in Spinus tristis in urban and rural settings ...... 44

2.5 Total fat score of Spinus tristis in urban and rural settings by month ...... 46

2.6 Body molt score of Spinus tristis in urban and rural settings by month ...... 46

2.7 Total fat score of Spinus tristis by progression of body molt ...... 47

viii

2.8 Body condition of Spinus tristis by progression of body molt ...... 47

2.9 Baseline CORT concentration in Spinus tristis from settings with and without bird feeders...... 49

2.10 Induced (30-minute) CORT concentration in Spinus tristis from settings with and without bird feeders...... 49

2.11 Percent increase in CORT concentration (from baseline to 30-minute level) in Spinus tristis from settings with and without bird feeders ...... 50

2.12 Total body fat of Spinus tristis from varying capture settings ...... 50

Chapter 1: Do American Goldfinches (Spinus tristis) Use Cities to Facilitate Overwintering at High Latitudes?

As human development in North America has dramatically altered the landscape in recent centuries, resident bird species have been faced with the challenge of either living in markedly transformed surroundings or losing habitat to sprawling urban growth. While a select few species thrive around human development, many others find the expansion of cities in their range to be both a benefit and a challenge. Most bird species have an optimal level of urbanization where they can thrive; it is the minority that are “urban exploiters” that benefit most from highly developed areas (Blair 2004). In general, cities may both help and harm; much of this effect varies by species, and much remains unknown about how developed areas interact with individual species and whole bird communities.

Drawbacks of Cities

Challenges facing city-dwelling birds include predators and plant communities that differ from rural settings, as well as more fragmented habitat and reduced territory size

(McGowan 2001, Chamberlain et al. 2009). The density of birds gathering in urban and suburban areas can lead to increases in disease and parasite loads (Chamberlain et al. 2009).

Birds may sing at higher amplitudes and frequencies to overcome city background noise, with yet-unknown social and reproductive implications (Nemeth et al. 2013), while artificial light sources interfere with birds’ natural cycles (Chamberlain et al. 2009). Urban birds may have reduced clutch sizes and smaller young than those nesting in rural areas (McGowan

2001, Chamberlain et al. 2009). Overall, avian species richness, diversity, and abundance

tend to decline in urban areas, with the same small portion of species thriving in many cities

(Blair 2001).

Benefits of Cities

Urban areas are not only sources of harm for bird species, however. Chamberlain et al. (2009) found that some species exhibit earlier laying dates in cities, suggesting that birds have good access to nesting sites and food resources leading up to the breeding period. Other studies have found higher post-fledging survival among those hatched in suburban areas than among rural conspecifics (McGowan 2001). Supplemental food (importantly, that provided via bird feeders) is a valuable resource for urban birds, and some have suggested that this factor is more influential than the various drawbacks of city life (Chamberlain et al. 2009).

This food resource may be particularly impactful during winter months. Winters are a critical time for songbirds, who are faced with higher energetic demands to survive cold temperatures at the same time that food resources are most scarce (Broggi et al. 2005).

Unsurprisingly, more food is needed to maintain body temperature in frigid winter climates than to pass the season in milder areas (Carey et al. 1978, Carey and Dawson 1999, Alves et al. 2013). Even a long migration would be less energetically taxing than remaining in a colder landscape throughout the winter (Alves et al. 2013), which raises questions of why and how birds remain in the north to survive the coldest months of the year.

There is evidence to suggest that urban areas, where most North Americans now reside, have a substantial effect on songbird populations, increasing density and winter survival rates (Jokimäki and Suhonen 1998, Chace and Walsh 2006, Chamberlain et al.

2009). While refuge from harsh winter weather within urban microclimates may benefit birds, anthropogenic food sources are likely an even greater factor influencing winter avian presence in cities (Luniak 2004). Feeding of backyard birds, especially during winter months when natural food sources are scarce, has become a popular activity throughout the northern hemisphere as urban residents seek opportunities for wildlife viewing (Jones and Reynolds

2008). Some 50.2 million Americans supplied food for wild birds in 2011 alone, and the great majority of these wildlife enthusiasts reside in metropolitan areas (U.S. Fish and

Wildlife Service and U.S. Census Bureau 2014). Regular provision of seed at feeders can enable birds to meet energy needs even at the northern edge of their winter distribution, where their seasonal energetic challenges are greatest (Jokimäki et al. 2002). Perhaps realizing this, residents of northern U.S. states such as Michigan more commonly provide bird feeders, particularly in winter, than do those living farther to the south (Lepczyk et al.

2004a, 2012).

Winter Trends in Northern Cities

During recent decades, North American bird species have experienced a northward shift in winter distribution (La Sorte and Thompson 2007). Species that frequent bird feeders have shown a particular increase in winter abundance in northern cities (Jokimäki and

Suhonen 1998, Morneau et al. 1999, Chace and Walsh 2006). It is reasonable to suspect climate change as a driver of this northward shift, but climatic factors were indicated in only a small percentage of species studied, with anthropogenic factors such as urban development playing a larger role for most (La Sorte and Thompson 2007). Analyses have suggested that increased food availability allows birds to meet the metabolic demands of winter at higher

latitudes, even where temperatures fall below the bird’s typical limit (Middleton 1977,

Canterbury 2002).

Gaps in Existing Knowledge

Factors Influencing Winter Range Expansion.---While researchers have paid much attention in recent years to the impacts of climate change on avian range expansions, understanding the more direct impacts of human activities on expanding winter ranges is equally important (La Sorte and Thompson 2007). As winter temperatures increase at the northern edge of bird ranges, a species’ success in new territory may still be limited or facilitated by landscape features and anthropogenic influences (La Sorte and Jetz 2010).

Some research has been completed on the consequences of urban food sources on overall wild bird abundance or species richness, but more data are needed to understand how urban feeding impacts winter populations (Jones and Reynolds 2008, Robb et al. 2008).

Furthermore, little research has been conducted on the possible effects of supplemental feeding on individual bird species (Kress 2006).

North American Species.---Considerable research has been completed on differences in bird species and communities between urban and rural areas in Europe, but little has yet to focus on examining such differences in North American species (Chamberlain et al. 2009).

Given the strong influence that private landowners and urban residents may have on avian communities (Lepczyk et al. 2012), this is an area that deserves additional focus.

Limits of Existing Data Sources.---There is a long history of citizen science data collecting by amateur birders in North America, including such projects as the Christmas

Bird Count and Project Feeder Watch (Lepage and Francis 2002). While these provide large data sets for analyzing bird distributions, such methods of data collection have their limits.

Species that frequent feeders are likely over-represented in winter counts, particularly in colder northern regions, as they are highly visible to both at-home bird watchers and parties conducting field surveys (Dunn 1995). The Christmas Bird Count reports data from within circles of 24 km in diameter, unevenly distributed across the landscape, thus providing little insight into avian habitat use on a smaller scale or outside of those designated areas (Bock

1984). These geographical survey limits also exclude large swaths of rural habitat where avian communities may differ.

The eBird database is another popular citizen science database for reporting bird observations. While eBird allows researchers access to large data sets (Sullivan et al. 2009), it also has its weaknesses. User bird identification skills may vary, and there is likely a bias toward easily detectable species (Sullivan et al. 2009). Most relevantly for examining urban/rural habitat use, eBird reports tend to be clustered around areas of higher human population density (Sullivan et al. 2009). This necessarily limits a systematic comparison of trends in rural vs. urban populations of a species.

American Goldfinch, Spinus tristis, as Study Organism

One species apparently following this trend of northward expansion is the American

Goldfinch (Spinus tristis). In the late nineteenth century, these birds could be found in great numbers in the southeastern part of Michigan (Swales 1903), but a survey of wintering birds

in the eastern Upper Peninsula showed only a single record of S. tristis in the winter of 1914, and none detected in the following two decades (Beebe 1937). For most of the intervening century, the northernmost reaches of the state have been considered only summer breeding range for this small songbird (Fig. 1.1).

Most bird species have greatest population density in the center of their distribution and are less abundant at the margins (La Sorte and Thompson 2007). While continent-wide,

S. tristis follow this general trend of decreasing abundance with increasing latitude (Butler et al. 2007), population patterns at the northern limits of their winter range have begun to change during recent decades. A review of banding and Christmas Bird Count data showed a significant increase in overwintering S. tristis in southern Ontario cities over six decades

(Middleton 1977). It has been suggested that supplemental food resources provided in cities have facilitated this increasing winter abundance of S. tristis at high latitudes (Middleton

1977, Morneau et al. 1999, Robb et al. 2008).

Interestingly, a related species demonstrated a similar effect of cities on its winter distribution. Throughout the 1990s, the Lesser Goldfinch’s (Spinus psaltria) winter range expanded northward across the western United States with surprising speed, spreading across desert territory poor in food sources or shelter. As S. psaltria were found primarily overwintering in urban oases within this inhospitable landscape, it was surmised that urban feeders were the main facilitator of their range expansion (Versaw 2001). For whatever challenges and benefits birds may encounter in urban areas, supplemental feeding through backyard feeders is often the factor that first draws them in to cities from less-developed habitats (Jokimäki and Suhonen 1998). Given that S. tristis are also known to frequent

feeders (Brittingham and Temple 1989, Wilson 1994), it is reasonable to suspect that this may also encourage their overwintering in more urbanized locales.

Figure 1.1. Established distribution of American Goldfinch, Spinus tristis. The red line has been added to more clearly demarcate the breeding vs. year-round areas. Eastern portions of Michigan’s Upper Peninsula, as well as the northern Lower Peninsula, are designated as breeding range only. (Map: McGraw and Middleton 2009, from Birds of North America, maintained by the Cornell Lab of Ornithology)

Objectives

The primary goal of this study was to gain an understanding of how S. tristis use

Michigan cities and rural areas to survive throughout the winter months. While this provided a snapshot of activity at only one portion of the species’ northern distribution boundary, the effect of cities on birds at a local level is usually closely tied with continent-wide patterns in the population (Blair 2004).

Hypothesis and Predictions

Given the demands on overwintering songbirds and the strong role of urban food sources in facilitating winter survival, I hypothesized that the proportion of overwintering

S. tristis in urban versus rural areas will increase with increasing latitude in Michigan. There may still be a greater total abundance of S. tristis at more southern latitudes, but I predicted that with increasing latitude there will be an increasing relative occurrence and abundance of goldfinches in urban areas compared to rural areas.

METHODS

Site Selection

I selected survey sites from within a general study area of approximately 100 km x

530 km (Fig. 1.2). The study area was divided into five equal latitudinal zones, as shown in

Fig. 1.2; five rural and five urban survey sites were randomly selected from within each zone.

Rural sites were selected from state parks and campgrounds, county parks outside of municipal boundaries, and nature preserves operated by the Little Traverse Conservancy.

Urban sites were selected from city parks located within municipalities of ≥ 1,500 residents, and ≤ 0.5km from residential zones. This population limit allowed for inclusion of a sufficient number of communities at higher latitudes in the state, where there are fewer large urban areas.

Six of the selected sites were not large enough to accommodate three individual survey points, so they were combined with the nearest additional park(s) (within a 1 km radius) to be treated as a single survey site. In instances where a single selected site spanned greater than 1 km in either latitude or longitude, the surveys were restrained to a single 1 km x 1 km sector (randomly selected). This served to limit differences in the size of areas being surveyed. Two selected sites could not be used due to unforeseen access issues; in both cases, surveys were conducted in the nearest suitable location adjacent to the pre-selected site.

To avoid surveying in areas unsuitable for S. tristis, each selected site was stratified by habitat type. Aerial imagery was used to identify areas of edge habitat (Fig. 1.3), as

S. tristis are known to preferentially forage in forested edges (Wilson 1994, Puckett et al.

2009). Each site map was then overlaid with a grid divided into 200 m x 200 m square tracts; three tracts containing edge habitat were randomly selected to be surveyed.

Figure 1.2. Selected survey sites for urban and rural Spinus tristis surveys. The study area was divided into five equal zones running south to north, with five urban and five rural survey sites randomly selected from within each zone.

Figure 1.3. Example of site stratification by habitat type. This site shown is Milan Nature Park (Milan, MI), a city park that was selected as an urban survey site. Aerial imagery from Google Earth was used to assess habitat types within the property boundaries (a) and create a map of edge habitat suitable for conducting playback surveys (b). (Map, Fig. 1.3a: Google)

Survey Protocol

Point count surveys for S. tristis were conducted between 21 February and 15 March

2015 (“winter surveys”) and repeated at the same locations between 12 April and 4 May

2015 (“spring surveys”). Each survey consisted of broadcasting a five-minute set of S. tristis

vocalizations taken from professional recordings (Elliot 1998, Borror 2003, Tekiela 2004,

Stokes et al. 2010) from a portable speaker, at a maximum volume of 90dBA at 1m from the speaker (Kaufmann 1988, Swift et al. 1988, Mosher et al. 1990, Morrell 1991). The playback was limited to five minutes, as responses by new individuals were expected to decline after that amount of time (Mosher et al. 1990), and surveying for lengthier periods could have increased the risk of double-counting birds that had departed and returned to the survey site

(Bibby et al. 2000, Gibbons and Gregory 2006).

I surveyed three points at each site (one in each pre-selected tract), separated by a minimum of 200 m to avoid repeat counts of the same individuals (Sliwa and Sherry 1992,

Gregory et al. 2004, Gibbons and Gregory 2006). Surveys were completed throughout daylight hours (between 730 h and 1730 h), as use of recorded playbacks has been shown to eliminate the effect of time of day on detectability of many species (Sliwa and Sherry 1992,

Boscolo et al. 2006). Temperature and wind speed were measured prior to each survey using a portable weather meter (Kestrel 4500 Weather & Environmental Meter, Nielsen-Kellerman,

Boothwyn, PA), and background noise was measured using a handheld sound level meter

(BAFX3370 Decibel Meter, BAFX Products, Milwaukee, WI). Surveys were not conducted during substantial precipitation or when wind speed exceeded 3.3 mps (12 kph), as previous avian studies using playbacks have indicated that higher wind speeds may reduce detection of birds (Swift et al. 1988, Mosher et al. 1990). One planned point count from the winter surveys and two from the spring surveys were excluded due to winds exceeding that level.

When an individual was first detected, I estimated its distance from the researchers using a laser rangefinder (Bushnell Sport 850 Compact Laser Rangefinder, Bushnell Outdoor

Products, Overland Park, KS). Other bird species observed during the survey period were also noted (see Appendix A).

Calculation of Urbanization

While I designated each site as categorically urban or rural while planning the surveys, S. tristis may be impacted by details of the landscape that aren’t immediately obvious to researchers. Following completion of all surveys, the degree of urbanization was also quantified for each individual survey point. I used ArcGIS to calculate the total percentage of low density urban, high density urban, and road/parking lot land use within a

0.8 km radius of each survey point, using state land cover data with a resolution of 30 m

(Lower Peninsula Land Cover 2001 and Upper Peninsula Land Cover 2001, Michigan

Geographic Data Library). The total of these three land uses was considered the percent urbanization surrounding the survey point for statistical analyses.

The 0.8 km radius was used due to the limited information available on S. tristis movements. An extensive search of the literature revealed no firm data on the movements of individuals within a winter season. Coutlee (1967) found that individuals traveled at least 0.8 km to forage during the breeding season, when their movements are limited by proximity to their nests. While S. tristis likely cover greater distances during the winter, this distance was used as a conservative measure of the surrounding landscape likely used by an individual.

Data Analysis

Data from the three survey points within each site were combined to provide a result for the site as a whole, for each season. Results were analyzed categorically, treating sites as either urban or rural based on their initial designation and grouping sites into the five latitudinal zones. Binomial logistic regression (presence data) and three-way ANOVA

(abundance data) were used to determine if season interacted with latitudinal zone or urban/rural site designation in regards to S. tristis presence or abundance. For variables that did not interact with season, I combined winter and spring survey results for further analysis with Fisher’s exact test (presence by urban/rural site designation), binomial logistic regression (presence by latitudinal zone), and ANOVAs (abundance).

The results were also analyzed taking into consideration precise latitude and the quantified degree of urbanization calculated from surrounding land use types. These data were similarly assessed for interaction of season with latitude or degree of urbanization, using binomial logistic regression (presence) and multiple linear regression (abundance).

When season did not interact significantly with degree of urbanization, I combined winter and spring surveys for further analysis using binomial logistic and multiple linear regression.

I also examined the effect of environmental measures using binomial logistic regression (presence) and multiple linear regression (abundance).

RESULTS

I detected a total of 21 S. tristis at seven survey sites (nine individual surveys) during the winter round of surveys. For the most part, single birds responded to the playback surveys; however, larger groups appeared in response to two surveys. The spring surveys revealed a total of 36 S. tristis, detected singly or in pairs at 17 sites (28 individual surveys).

In nearly all cases, responding birds approached close and responded vocally to the recorded playbacks, making confident identification easy. In the one instance when a bird remained hidden, it could be clearly heard vocalizing.

Latitude

Both latitudinal zone and precise latitude interacted with season in terms of S. tristis abundance (zone: F1,96 = 5.973, P = 0.016; latitude: t = -2.311, P = 0.023), though not in regards to presence (zone: z = -1.883, P = 0.060; latitude: z = -1.842, P = 0.066). I conducted further analyses of latitudinal trends on each season of survey data separately.

Surprisingly, neither S. tristis presence (P = 0.365) nor abundance (F1,48 = 0.080,

P = 0.779) differed by latitudinal zone in the winter surveys (Fig. 1.4, Fig. 1.5). This was likely influenced by detection of two groups of S. tristis (comprising five and eight individuals, respectively) in the second most northerly zone. An analysis by precise latitude confirmed this finding: neither presence nor abundance of S. tristis varied with latitude in winter (presence: z = -1.477, P = 0.140; abundance: t = 0.108, P = 0.915) (Fig. 1.6, Fig. 1.7).

Both S. tristis presence and abundance correlated with latitudinal zone during the spring surveys, however (presence: P < 0.001; abundance: F1,48 = 12.75, P < 0.001) (Fig. 1.4,

Fig. 1.5). Similar trends emerged when examining the role of precise latitude in spring

(presence: z = -3.500, P < 0.001; abundance: t = -3.528, P < 0.001) (Fig. 1.6, Fig. 1.7).

1 Winter Spring 0.8

0.6

0.4

presence presence (proportion of sites) 0.2

tristis

. . S 0 1 2 3 4 5 Latitudinal zone (south to north)

Figure 1.4. Seasonal Spinus tristis presence by latitudinal zone. Zones run from south to north, beginning with zone 1 in the south of Michigan (see Fig. 1.3). Presence differed significantly by zone only during the spring round of surveys.

5 Winter Spring 4

abundance(√count)

tristis 1

. . S

0 1 2 3 4 5 Latitudinal zone (south to north) Figure 1.5. Seasonal Spinus tristis abundance by latitudinal zone. Abundance differed significantly by zone only in the spring. No individuals were detected in zone 5 during winter, or in zone 4 during spring.

1 Winter Spring

presence

tristis

. . S

0 41 42 43 44 45 46 47 Latitude (degrees N)

Figure 1.6. Seasonal Spinus tristis presence by latitude. A survey value of “1” indicates that one or more individuals were detected. While presence declined somewhat with increasing latitude within both seasons of survey, this relationship was significant only within the spring surveys.

3 Winter Spring

abundance(√count)

tristis . .

0 41 42 43 44 45 46 47 Latitude (degrees N)

Figure 1.7. Seasonal Spinus tristis abundance by latitude. Abundance declined significantly with increasing latitude only in the spring surveys.

Urbanization

The individual survey points (n = 150) ranged from 0 to 73.14% urban within a

0.8 km radius (Fig. 1.8). Neither a priori urban/rural site designation nor calculated degree of urbanization interacted with season to predict S. tristis presence (site designation: z = 0.153,

P = 0.878; degree of urbanization: z = 0.470, P = 0.638) or abundance (site designation:

F1,96 = 1.247, P = 0.267; degree of urbanization: t = 0.984, P = 0.327). Therefore, I combined data from both seasons of surveying for further analyses of urban/rural effects.

There was no difference in S. tristis presence between a priori designated urban and rural sites (P = 1), with individuals detected at 10 urban and 10 rural sites. Abundance likewise differed little between urban and rural sites (F1,48 = 0.467, P = 0.498; Fig. 1.9).

Neither presence (z = 0.180, P = 0.857; Fig. 1.10) nor abundance (t = 0.175, P = 0.862; Fig.

1.11) varied along a spectrum of urbanization.

Number Number of survey sites 0 0 10 20 30 50 60 70 80

Degree of urbanization (% urban land use within 0.8 km radius)

Figure 1.8. Quantified degree of urbanization of survey sites. The degree of urbanization was calculated for each survey point using land use data from within a 0.8 km radius of the point; site urbanization was taken as a mean of the values of the three survey points within the site. Fully half the sites were < 10% urban by this measure, with nearly a third (n = 16) being surrounded by < 5% urban land use.

abundance(count)

S. S. tristis 10

0 Urban Rural

Site designation

Figure 1.9. Abundance of Spinus tristis at urban and rural survey sites. There was no significant difference in numbers of S. tristis detected at urban and rural survey sites across the combined dataset. A total of 33 individuals were detected at urban sites and 24 at rural sites.

presence

tristis

. . S

0 0 20 40 60 80 Degree of urbanization (% urban land use within 0.8 km radius)

Figure 1.10. Spinus tristis presence by degree of urbanization. Considering survey results from both seasons, degree of urbanization did not significantly correlate with S. tristis presence.

abundance(√count) 1 tristis

. . S 0 0 20 40 60 80 Degree of urbanization (% urban land use within 0.8 km radius)

Figure 1.11. Spinus tristis abundance by degree of urbanization. Across both seasons of surveying, degree of urbanization did not significantly correlate with S. tristis abundance.

Interaction of Latitude and Urbanization

Spinus tristis were not associated with any interaction of latitudinal zone, urban/rural site designation, and season (presence: z = -0.388, P = 0.698; abundance: F1,92 = 0.147,

P = 0.702), nor did they follow any pattern indicating interaction of latitude, degree of urbanization, and season (presence: z = 1.132, P = 0.258; abundance: t = 0.314, P = 0.754).

Data from both seasons were therefore combined for examination of interactions between latitude and urbanization.

Neither latitudinal zone and urban/rural site designation (Fig. 1.12) nor exact latitude and calculated degree of urbanization interacted to predict S. tristis presence (zone x site designation: z = 0.291, P = 0.771; latitude x degree of urbanization: z = 0.693, P = 0.488).

Similarly, S. tristis abundance was not associated with any interaction of zone and urban/rural designation (F1,46 = 1.651, P = 0.265) (Fig. 1.13), nor of latitude and degree of urbanization (t = -0.322, P = 0.749).

Environmental Measures

Spinus tristis presence and abundance were not correlated with temperature

(presence: z = -1.059, P = 0.290; abundance: t = -0.455, P = 0.651), wind speed (presence: z = 0.879, P = 0.379; abundance: t = 0.244, P = 0.808), light level (presence: z = 1.084,

P = 0.278; abundance: t = 0.204, P = 0.839), or background noise (presence: z = 1.104,

P = 0.270; abundance: t = 0.556, P = 0.580) when adjusted for date. Neither presence

(z = -1.585, P = 0.113) nor abundance (t = -1.547, P = 0.125) varied with time of day.

6 Urban Rural

presence presence (#sites)of 2

tristis . . S 0 1 2 3 4 5

Latitudinal zone (south to north)

Figure 1.12. Spinus tristis presence in urban/rural settings by latitudinal zone. Figure shows combined results from winter and spring surveys. Site designation and latitudinal zone did not significantly interact as relates to S. tristis presence.

4 Urban Rural 3

abundance (√count) abundance

tristis

. . S

0 1 2 3 4 5 Latitudinal zone (south to north)

Figure 1.13. Spinus tristis abundance in urban/rural settings by latitudinal zone. Figure shows combined results from winter and spring surveys. Site designation and latitudinal zone did not significantly interact as relates to number of S. tristis detected.

DISCUSSION

While S. tristis were, as expected, more abundant at southern latitudes, there was no apparent interaction of latitude and urbanization. At higher latitudes, S. tristis were not disproportionately found in urban areas.

Latitude

The decline in S. tristis presence and abundance moving north through the state of

Michigan is consistent with prior findings about the patterns of songbird species near the northern limits of their distribution (Butler et al. 2007, La Sorte and Thompson 2007).

Northern Michigan overlaps with the accepted northern limits of S. tristis (Floyd 2008), though doubtless individuals may be reported outside of such boundaries. I did not detect any individuals during either round of surveys in the Upper Peninsula, and sighted groups on few instances in the northernmost portions of the Lower Peninsula.

Urbanization

In my surveys there was no effect of cities on the presence or abundance of S. tristis detected. This runs counter to Middleton’s (1977) conclusions from his study of S. tristis winter population trends in nearby southern Ontario. While Middleton had access to more

S. tristis presence records (data from several banding stations over five winter seasons,

November through May), his study did not focus on where S. tristis were absent. The banding stations in question were all located in Guelph, ON (Middleton 1977), which in 1968

was estimated to be home to some 54,000 residents (Bourne and Baker 1968). The increasing winter numbers of urban S. tristis that Middleton observed may have been paralleled by burgeoning rural populations that were not well-sampled.

It is also worth noting that Middleton’s (1977) data on winter population changes from 1915–1975 came from published results of Christmas Bird Counts (CBC) throughout

Ontario. The CBC and similar citizen science birding programs tend to be biased toward sampling of habitats that occur along roads (Sauer et al. 1995) and urban or semi-urban areas of high human population density (Drennan 1981). This is a relevant bias when attempting to answer the question of how S. tristis are impacted by urbanization, as such surveys frequently neglect less developed habitat areas (Drennan 1981, Sauer et al. 1995, Dunn et al. 2005). The data is also reported on a broad scale; CBC circles are large enough (24 km in diameter) that it is difficult to infer the specific habitat in which each bird was observed. My data, collected across both urban and rural settings, suggest that S. tristis do not show a preference for urban areas during high-latitude winters.

Temporal analysis of bird population trends from CBC data must also be undertaken with caution, as volunteer participation in the CBC tends to increase over time, resulting in more count circles and better coverage within circles (Sauer et al. 1995). Such changes in effort are important to consider when estimating populations from citizen science data

(Geissler and Noon 1981). In such programs, species lists and counts naturally tend to increase over time due to increased volunteer participation and experience, as well as more widespread access to quality optics and field guides (de Vos 1964, Sauer et al. 1995). For many species, the relationship between observer effort and bird count is not linear (Butcher and McCulloch 1990). Models have been developed to derive temporal population change

estimates from CBC or Breeding Bird Survey data (Sauer and Geissler 1990), but Middleton

(1977) appears not to have made any such adjustments when considering the increases in

CBC S. tristis counts during the time period he examined. It cannot be certain that S. tristis numbers actually increased by any significant amount in Ontario cities over the decades he considered.

It is also possible that S. tristis are influenced by a larger area than the radius of land use I calculated for each site. The 0.8 km radius (Coutlee 1967) was used as a conservative estimate of S. tristis home range due to limited data available in the literature. Middleton

(1977) noted that S. tristis moved throughout Guelph, ON, in winter, and reported that roughly half of recaptured individuals in any winter were found at a different banding station from the one at which they were initially captured. Lacking more precise estimates of winter

S. tristis home ranges, it is difficult to evaluate the spatial area by which individuals may truly be impacted. Birds may be affected by broader landscapes than captured in these analyses, or may preferentially utilize those portions of cities that best serve their resource needs. At the scale studied here, however, S. tristis do not appear to preferentially select between urban and less developed winter habitats in Michigan. This suggests they fall into

Blair’s (2004) intermediate ecological group: neither avoiding nor considerably favoring cities, but flexibly “suburban adaptable.”

Interaction of Latitude and Urbanization

Latitude and urban settings did not interact to predict S. tristis presence or abundance in this study. Other explanations for northward winter distribution expansion must be

considered. In the middle of the last century, de Vos (1964) attributed the northward expansion of avian breeding ranges in the Great Lakes region to climate change. However,

Middleton (1977) examined winter temperature averages and found increases in S. tristis populations even during periods when winter conditions had worsened. Other species, including the Northern Cardinal (Cardinalis cardinalis) and Blue Jay (Cyanocitta cristata) have similarly expanded northward in winter, with these range expansions attributed to increasing urbanization and development (de Vos 1964, Bock and Lepthien 1976).

However, it should be noted that for many species, it may not simply be the expansion of residential zones and supplemental food provision there that allow such expansions. Bobolinks (Dolichonyx oryzivorus) and Mourning Doves (Zenaida macroura) have expanded northward in developed areas; however, their movements are associated with clearing of forests for agriculture and creation of habitat edges (de Vos 1964). As S. tristis demonstrate a similar preference for habitat edges, such as the interface of forest and fields or agricultural lands (Wilson 1994, Puckett et al. 2009), it is possible that rural and agricultural development has been a major factor in the species’ northward expansion. Residential areas are less than ideal for many avian species, with highest numbers detected there in years when the population as a whole is flourishing (Batten 1972).

Environmental Measures

Background noise might be expected to be higher in urban areas, and could limit the ability of birds to hear the broadcast playbacks, as well as the ability of researchers to detect any vocal responses. Similarly, high wind speeds or dim light could reduce detectability and

suppress S. tristis activity. No individuals were detected below -0.5 °C, at winds higher than

0.6 mps (2.16 kph) during the winter season, or at ambient light levels below 6800 lux.

However, none of the environmental variables or time of day corresponded with S. tristis presence or abundance when accounting for date, nor with urban/rural site designation. The lack of detection under such harsh weather conditions is consistent with the lower numbers detected throughout the winter season than during the spring months.

Conclusions and Future Research

The results of this study did not support the idea that the northward winter range expansion of S. tristis is due to exploitation of urban areas in order to survive under harsh conditions. While individuals were found in urban as well as rural areas, this proportion did not change with increasing latitude, indicating that cities play a similar role for S. tristis winter populations throughout Michigan.

My analysis was challenged by low rates of detection during my surveys, despite the use of recorded playbacks. Across both seasons of surveys, individuals were detected in only

37 surveys out of a total of 297. With such low rates of detection across all types of habitat, it is difficult to detect subtle S. tristis habitat preferences. Some surveys that have used recorded playbacks to prompt bird responses studied territorial species during the breeding season (e.g., Boscolo et al. 2006), when the likelihood of responses to song broadcasts is high. Spinus tristis is known to be a highly social species and will forage in flocks even during the breeding season (Coutlee 1967). The territorial displays that are uncommon during nesting are even rarer in winter (Coutlee 1967). It is possible that individuals heard the

broadcast playbacks but did not respond due to lack of a territorial defense impulse. Indeed, I found it curious that in most instances (six of nine winter surveys with S. tristis detected, and

20 of 28 in spring), only a single bird responded. In these cases, the bird typically came quite near and would at times continue watching and vocalizing even after the five-minute playback had concluded. The use of playbacks may be biased toward detection of lone or loosely-associated individuals responding socially, rather than territorially. It is worth noting that while the remainder of responses in the spring surveys were by pairs of birds, in nearly every case both individuals were male (the exception was a pair that was not clearly visible, so I was unable to determine sex). The prevalence of social males, coupled with the fact that migration and nesting would not yet have been underway during the survey timeframe

(Saunders 1947, Holcomb 1969), supports the notion that individuals were not likely responding territorially to the playbacks. This tendency toward winter sociality does complicate surveying and makes it difficult to accurately extrapolate distribution trends from point count surveys (Carrascal et al. 2006). Further research on age class, sex, and social grouping of individuals responding to playbacks would shed further light on the utility of playbacks in conducting winter S. tristis distribution surveys.

Additional research is also required to elucidate the causes of apparent expansion beyond the species’ previous northern limits. It is possible that climate change is in fact the major driver of S. tristis range expansions, with high-latitude rural and urban populations undergoing equivalent growth. It seems that both urban and rural S. tristis are able to find adequate resources for survival in the current climate despite the high energy demands of harsh winters (Carey et al. 1978, Carey and Dawson 1999, Broggi et al. 2005, Alves et al.

2013). A more intensive survey of selected urban and rural sites across latitudes, extending

over several years, would determine what changes the northernmost wintering populations are undergoing. Research examining temporal trends in S. tristis distribution may also utilize citizen science data (e.g., CBC or eBird records) but should take care in avoiding biases in such data. Recent examinations of CBC data (e.g., Link and Sauer 1999a,b; Sauer and Link

2002) have provided modeling guidelines to better account for varying effort between years and count circles.

An examination of potential body condition or survivorship differences between

S. tristis overwintering in urban and rural areas would also be of interest. If there is no difference in survival and fitness between individuals utilizing natural or developed habitats through the nonbreeding season, this would indicate that S. tristis is well situated to thrive in the face of further development.

Chapter 2: Do American Goldfinches (Spinus tristis) Overwintering in Urban and Rural Settings Experience Different Physiological Responses to Stressors?

North American birds today live in an increasingly urbanized world. In the last quarter of the twentieth century, developed lands in the contiguous United States increased substantially, while area of forest and agricultural lands declined (Sleeter et al. 2013). In the eastern temperate forests that include much of Michigan’s natural landscape, acreage of developed land increased by over 30% from 1973 to 2000 (Sleeter et al. 2013).

For a bird to take advantage of such human-modified landscapes, it must be able to handle the many disturbances and stressors that exist there. Human presence, pollutants, reduced natural foraging areas, decreased diversity of vegetation, free-roaming domestic animals, vehicles, high levels of light and noise, competition with introduced species, and limited available nesting sites can all present serious challenges for urban-dwelling birds

(Savard et al. 2000, Wingfield and Romero 2001, Ditchkoff et al. 2006, Partecke et al. 2006,

Zhang et al. 2011). Ability to cope with such unpredictable urban stressors can be essential to a bird’s future survival and fitness (Partecke et al. 2006, Fokidis et al. 2009).

Such continual challenges can eventually result in an increase of birds’ baseline levels of corticosterone (CORT), a major stress and metabolic hormone (Wingfield and Romero

2001). CORT is a glucocorticoid involved in the hypothalamic-pituitary-adrenal (HPA) axis

(Fig. 2.1) by which birds regulate metabolic functions and respond to stressors through the triggering of various physiological and behavioral changes that increase their odds of survival (Smith and Vale 2006). Baseline avian plasma CORT concentrations naturally vary according to season, life history stage, and even time of day (Blas 2015). Upon encountering an unpredictable disturbance, or stressor, however, CORT is elevated above typical levels,

triggering changes that allow the individual to meet its energy needs despite the challenges in its environment (Blas 2015). a)

Figure 2.1. The role of corticosterone (CORT) (a) in the hypothalamic-pituitary-adrenal (HPA) axis (b).

Adaptive Function of Corticosterone

Natural cycles of baseline plasma CORT allow an organism to prepare for a variety of life history stages; for instance, significantly elevated baseline CORT can be associated with fat storage prior to seasonal migrations (Piersma et al. 2000, Scanes and Braun 2013).

Moderate physiological stress responses to short-term disturbances are likewise adaptive,

allowing an individual to respond to resource scarcity and divert energy to functions related to immediate survival (Astheimer et al. 1994, Partecke et al. 2006). Such stress-induced

CORT increases may result in increased feeding behavior (Lõhmus et al. 2006) and increased nocturnal restfulness, which allows birds to save more of their energy stores for daytime activities (Buttemer et al. 1991). Experimentally elevated CORT levels in hatchling Black- legged Kittiwakes (Rissa tridactyla) resulted in intensified begging by the young and increased provision of food by parents (Kitaysky et al. 2003). Modest increases in CORT have been shown to increase food caching behaviors, boost fat accumulation, and improve performance in spatial memory and cache-retrieval tasks in Mountain Chickadees (Poecile gambeli) (Pravosudov 2003). Such behaviors may be adaptive when environmental conditions are poor or highly variable, such as during winters at high latitudes (Pravosudov

2003).

The sensitivity of response to stressors may vary in different seasons, environmental conditions, and life history stages (Wingfield et al. 1997, Romero 2006). Stress responses may be relatively suppressed in the non-breeding season (Astheimer et al. 1994, Romero et al. 2006); in birds breeding in harsh environmental conditions (Wingfield et al. 1997,

Cornelius et al. 2012); or during molt, when energy reserves are devoted to growth of new plumage (Romero et al. 1998a,b; Romero and Remage-Healy 2000).

Risks of Elevated Corticosterone

When a bird is able to escape the influence of a short-term disturbance, CORT aids in its survival and return to a normal life history stage (Blas 2015). Ongoing exposure to

stressors, however, can result in harmful physiological effects (Wingfield et al. 1997,

Partecke et al. 2006, Fokidis et al. 2009). Sustained high levels of CORT can significantly lower feather quality, stymie growth, suppress immune function, interrupt reproduction, decrease nesting success, and result in the loss of neurons and breakdown of skeletal muscle

(Silverin 1986, Wingfield et al. 1998, Wingfield and Romero 2001, DesRochers et al. 2009,

Chávez-Zichinelli et al. 2010, Kennedy et al. 2013). Experimentally elevated CORT levels in hatchling R. tridactyla doubled the time needed to learn feeding tasks and significantly decreased performance in spatial problem solving tasks, suggesting that unusually elevated

CORT may impact motor skills or memory (Kitaysky et al. 2003).

Birds may also change their behavior to avoid or mitigate the impacts of stressors

(Ditchkoff et al. 2006), including abandoning home ranges, using stored energy reserves while fleeing or taking shelter from the disturbance (Wingfield and Romero 2001), and spending increased time away from nests, leaving young dangerously exposed to predation

(Kitaysky et al. 2001b). While in the face of short-term disturbances such actions usually contribute to survival and long-term fitness, if a bird is unable to escape stressors (such as in habitats permanently modified by humans), these benefits fade (Blas 2015). Heightened triggering of stress responses due to frequent or lengthy exposure to disturbances can result in an individual using excess energy or diverting time and resources away from the functions of its expected life history stage (Romero et al. 2009).

Environmental Stressors

Not all challenging environmental factors result in a stress response. Birds can acclimatize to predictable changes, such as seasonal weather patterns, without a necessary increase in baseline CORT levels (Wingfield et al. 1997, Wingfield and Romero 2001).

Unpredictable disturbances, however, particularly those that persist over time, may result in heightened stress responses (Wingfield and Romero 2001). City-dwelling birds, faced with the frequent novel disturbances present in a developed environment, must have the ability to modify their physiological response to stressors if they are to avoid the negative effects of long-term exposure to heightened CORT levels (Partecke et al. 2006). The magnitude of

CORT response to stressors can vary a great deal between species, with domesticated and urban-adapted birds routinely exhibiting less intense CORT responses than other species

(Cockrem 2007).

Gaps in Existing Knowledge

Generalizing Trends Between Species.---There may be physiological differences between urban and rural birds of the same species, though not much research has been done to examine such impacts of urbanization (Bonier et al. 2007, Zhang et al. 2011). Among those studies that have evaluated baseline or stress-induced CORT levels in both urban and rural birds, there is no clear pattern of how urban disturbances influence CORT (Bonier

2012). Baseline CORT in House Sparrows (Passer domesticus), a species well known to thrive in developed environments, did not differ between urban and rural settings (Chávez-

Zichinelli et al. 2010), which is consistent with the notion that adaptable generalist species tend to fare well in cities (Evans et al. 2011). Other research has shown heightened baseline

CORT and intensified stress responses in urban residents of some species (Fokidis et al.

2009). Studies of multiple avian species demonstrate that the relationship between setting and baseline or induced CORT varies by species and season (Partecke et al. 2006, Fokidis et al. 2009, Bonier 2012). Given the variation in response to urban disturbances across species, a pattern cannot be generalized among species from existing research.

Inclusion of Multiple Localities.---Many studies have examined CORT levels in birds from only a single urban and a single rural location; such data must be interpreted carefully, as there can be a great deal of variation between the conditions of different urban sites

(Zhang et al. 2011). No known studies have yet examined CORT in an avian species sampled over multiple urban and corresponding rural areas (Bonier 2012). In this study, I seek to sample individuals from multiple urban and rural locations in southeastern Michigan.

American Goldfinch, Spinus tristis, as Study Organism

Overwintering American Goldfinch (Spinus tristis) populations have apparently increased in the northern Great Lakes area during the past century (Middleton 1977), and are known to utilize both forest edges (Wilson 1994, Puckett et al. 2009) and residential bird feeders (Brittingham and Temple 1989, Wilson 1994). This makes them an ideal species for studying the physiological differences between individuals relying primarily on wildland resources and those making use of urban landscapes. Furthermore, there have not yet been any studies comparing S. tristis CORT levels in urban and rural environments (Bonier 2012).

Objectives

The primary aim of this study was to assess how S. tristis living at high latitudes during winter respond, physiologically, to unpredictable stressors in their environment. Of particular interest was understanding how urban areas impact S. tristis during this season.

While this study provided data on physiological responses occurring during only the winter, non-breeding season, this filled an important gap; little was yet known about S. tristis hormone activity during this life stage.

Hypothesis and Predictions

As most species previously studied show some variation in baseline and/or induced

CORT levels between urban and rural settings, I hypothesized that there would be a difference in these measures for S. tristis sampled in urban versus rural environments. I predicted that individuals in urban settings would display higher baseline CORT levels but experience a smaller increase of CORT when exposed to an acute stressor, due to more regular exposure to disturbances. Such findings would suggest that the regular disturbances of urban dwelling result in higher baseline CORT levels, but also reduced sensitivity to acute disturbances.

METHODS

Site Selection

Sampling sites were selected to maximize the number of S. tristis likely to be captured. I selected urban sites from residential locations known to have high use of bird feeders by S. tristis, and municipal parks with relatively high frequency of recent S. tristis observations as reported in the eBird citizen science database. Rural sites comprised county nature preserves, a bird sanctuary, and a rural residence. Landowner permissions were obtained prior to trapping at any site.

Calculation of Urbanization

I categorized each trapping site as urban or rural based on calculation of surrounding land use. ArcGIS was used to calculate the total percentage of low density urban, high density urban, and road/parking lot land use within a 0.8 km radius of each trapping location, using state land cover data with a resolution of 30 m (Lower Peninsula Land Cover 2001 and

Upper Peninsula Land Cover 2001, Michigan Geographic Data Library). The total of these three land uses was considered the percent urbanization of the trapping locale.

The 0.8 km radius was based on the best available information on S. tristis movements. An extensive search of the literature revealed no firm data on the movements of individuals within a winter season. Coutlee (1967) found that individuals traveled at least 0.8 km to forage during the breeding season, when their movements are limited by proximity to their nests. While S. tristis likely cover greater distances during the winter, this distance was used as a conservative measure of the surrounding landscape likely used by an individual.

I was able to successfully collect blood samples from individuals at three rural locations (< 5% urban) and two urban sites (> 45% urban) (see Table 1 for site descriptions).

Timeline

Sampling occurred from 18 January to 11 April 2016. This allowed for assessment of

CORT levels during the nonbreeding season (Holcomb 1969), well before the start of spring migration (Saunders 1947). All birds were captured between 0750 h and 1110 h. Trapping concluded by noon each day to minimize the impact of normal changes in plasma CORT concentrations throughout the day.

Sampling Protocol

At each site, a live trap was placed around a bird feeder baited with Guizotia abyssinica seed (often marketed as niger or Nyjer® seed). In cases where a homeowner had already installed a bird feeder, that feeder was used inside the trap; in other instances, a mesh

“sock” feeder was used. I also erected a mist net at sites where birds were not readily attracted to the feeder trap and where site conditions allowed installation of nets. The feeder trap allowed rapid removal of the bird; however, at rural sites where there was not typically a feeder already installed, individuals appeared more wary of the trap and feeder. A single individual at a rural location (Brauer Preserve) was captured via mist net; all other samples were taken from birds captured in a feeder trap. Birds were lured to the nets and trap using ongoing playback of S. tristis calls and songs taken from several professional recordings

(Elliot 1998, Borror 2003, Tekiela 2004, Stokes et al. 2010).

Blood

minutesample.

sampledto refers the number

ine and ine a 30

number

The The

f urbanization, f were sites designated categorically either as

birds were,however, includedin analyses of demographic and morphometric differences.

capture sites. Asthere a was clear differentiation in o degree

Spinustristis

Bird feeders had Bird feeders been installed prior tothis study by at residents two urbanand one rural location.

Descriptionof

Table Table urbanor rural. individuals of trapped at thatlocation which from blood samples were collected; not all individualscontributed both basel a could samples not obtained be several from additional birds;these

I removed trapped birds from the net or trap and collected a blood sample within three minutes after capture (that is, the time when the bird first entered the trap or net). The alar vein was pierced with a 26-gauge needle and up to 70 μl of blood drawn into a heparinized capillary tube, following the protocol established by Wingfield et al. (1983). This sample allowed ascertainment of the bird’s baseline CORT level (Wingfield et al. 1982).

Each bird was then placed into a cloth bag as a way of standardizing the stressors experienced after capture (Wingfield et al. 1992). While capture stress is not precisely equivalent to any stressor the bird would face in nature, it is a good approximation of a moderately stressful event (Wingfield et al. 1995). I collected a second blood sample at 30 minutes after capture to assess stress-induced CORT levels.

After both blood samples were collected, each bird was weighed to the nearest 0.1 g using a spring scale. I determined the age class and sex of each individual based on plumage and body size. Age class determinations were limited to second year (SY), those who had hatched the previous summer and had not yet gone through the first pre-alternate molt, and after second year (ASY), which included all others. I recorded each bird’s molt status and morphometric measures, including wing chord, tarsus length, keel length, and length of the cloacal protuberance in males or brood patch stage in females (using the scale developed by

Bailey 1952; though, due to capture during the non-breeding season, none of the females displayed an active brood patch). I used the 0–5 scale described by Helms and Drury (1960) to assess furcular and abdominal fat deposits; total fat (the sum of the two scores) was used for statistical analyses. Progression of body molt was scored on a 0–3 scale, indicating no, low, moderate, or heavy body molt (S. tristis do not replace flight feathers during their spring

“pre-alternate” molt). Each bird was fitted with a U.S. Geological Survey numbered

aluminum band prior to release. Previously-banded birds were not sampled if they had last been blood sampled within the prior two weeks.

I centrifuged blood samples following each daily trapping session, at 10,000 rpm for eight minutes. The plasma layer was extracted by syringe (Wingfield et al. 1983) and stored in microcentrifuge tubes at -80 °C until analysis.

The above sampling protocol was covered under USGS Federal Bird Banding Permit number 23949 and Michigan DNR Scientific Collector’s Permit number SC 1550. Eastern

Michigan University’s Institutional Animal Care and Use Committee (IACUC) also reviewed and approved the protocol (approval number 2015-072) prior to the commencement of research.

Laboratory Analysis

To analyze total CORT from the collected plasma samples, I conducted enzyme immunoassay analysis using a Corticosterone EIA Kit from Enzo Life Sciences (cat. no.

ADI-900-097; Plymouth Meeting, PA). All samples were analyzed in triplicate, except for those samples where low volume of collected plasma precluded it. The assay was validated using a parallelism test and optimized for S. tristis samples at a 1:15 dilution and 1% SDB.

Minimum CORT detectability was calculated as 2 SDs from the mean Bo (maximum binding) on the standard curve. Any samples reading below the detectability limit were set to the minimum detectable CORT concentration for the assay plate. Sensitivities ranged from

0.150 to 0.375 ng/ml. Interplate variability was 7.08%.

Data Analysis

A principal component analysis was conducted to account for measures of body size.

PCA1 included wing chord, tarsus, and keel, and accounted for 87.977% of variance. PCA1 was plotted against body mass, and the residuals treated as a size-adjusted body condition score variable.

I assessed the distributions of CORT and morphometric data using the Shapiro-Wilk test of normality and transformed data as needed to meet statistical test assumptions. Baseline

CORT and percent increase of CORT were log-transformed to obtain normal distributions. A log-modulus transformation (John and Draper 1980) was conducted to achieve normal distribution of the body condition scores, which contain negative and positive values.

Differences in baseline, 30-minute, and % increase in CORT were compared between urban and rural sites using Welch two sample t-tests. I also used this test to examine other potential differences between urban and rural birds (such as physical characteristics), as well as between individuals captured at sites with and without regularly provided bird feeders.

Correlations between other physical and environmental characteristics were examined via linear regression and ANOVAs. Differences in categorical factors such as age class, sex, or body molt status were assessed using Fisher’s exact test. Standard error of the mean is reported as the measure of variability for all descriptive statistics.

RESULTS

A total of 33 individuals were captured during the trapping period. Of these, baseline blood samples were collected from 30 birds, and 30-minute samples were collected from 28 birds. Baseline sample collection was completed a mean of 144.6 ± 7.28 seconds after capture. Two baseline samples were completed after the three-minute time goal (205 and 270 seconds after capture, respectively); however, the data showed no correlation between length of time until baseline sample completion and baseline CORT concentration (t = 0.851,

P = 0.404). Stress-induced samples were completed within a mean of 101.25 ± 24.42 seconds of 30 minutes after capture. A single sample was not completed until 600 seconds after the

30-minute mark; however, there was no correlation between deviation from the 30-minute point and stress-induced CORT concentration (t = -0.292, P = 0.773).

Urban/Rural Stress Physiology

Birds from urban and rural capture locales did not differ in baseline plasma CORT concentrations (t = -0.664, df = 22.988, P = 0.514) (Fig. 2.2), induced CORT at 30 minutes

(t = 0.461, df = 22.972, P = 0.649) (Fig. 2.3), or percent increase in CORT from baseline to

30 minutes (t = 0.655, df = 19.44, P = 0.520) (Fig. 2.4).

1.0 12

Plasma[CORT] (ng/mL) 2 Plasma[CORT] (ng/mL)

0.1 0 Rural Urban Rural Urban Capture setting Capture setting Figure 2.2. Baseline CORT concentration in Spinus tristis Figure 2.3. Induced (30-minute) CORT concentration in in rural and urban settings. Birds from rural (n = 15) and Spinus tristis in rural and urban settings. Individuals urban (n = 12) capture settings did not significantly from rural (n = 14) and urban (n = 11) capture settings differ in baseline plasma CORT concentration. CORT did not have significantly different CORT concentrations concentrations are shown on a log scale. Lines when measured 30 minutes after capture. Lines represent ± 1SEM. represent ± 1SEM.

3000

% % increase plasmain[CORT] 1 Rural Urban Capture setting

Figure 2.4. Percent increase in CORT concentration (from baseline to 30-minute level) in Spinus tristis in urban and rural settings. Individuals from rural (n = 12) and urban (n = 10) capture settings did not display significantly different percent increases in CORT concentration, taken as percentage of baseline level. Percentages are shown on a log scale. Lines represent ± 1SEM.

Other Urban/Rural Differences

While S. tristis trapped at urban and rural sites displayed similar CORT responses, there were some differences between the two sets of birds. All individuals captured at urban sites were after second year (ASY), whereas rural captures included five second year (SY) birds; however, this did not represent a significant difference in age class distribution

(P = 0.137). Second year birds exhibited a higher percent increase in CORT concentrations between capture and 30 minutes (t = 21.277, df = 42.285, P < 0.001).

Urban and rural birds also differed in total fat score (sum score of furcular and abdominal fat) adjusted for month of capture, with urban birds declining in total fat over the course of the season (F3,29 = 4.602, P = 0.029) (Fig. 2.5). There was, however, no interaction of capture setting and month in regards to body condition score (F3,27 = 0.033, P = 0.858).

When considering all S. tristis sampled, body condition did not correlate with any CORT measure (baseline: t = -0.70, P = 0.492; induced: t = -1.218, P = 0.237; percent increase: t = -0.365, P = 0.719). Capture setting and month interacted in regards to body molt status

(F3,27 = 15.1, P = 0.044) (Fig. 2.6). Body molt status predicted total fat (F1,29 = 8.462,

P = 0.007) (Fig. 2.7), though not size-adjusted body condition score (F1,28 = 2.378,

P = 0.134) (Fig. 2.8).

5 Urban

Rural 4

2 Totalscore fat

0 January February March April Month

Figure 2.5. Total fat score of Spinus tristis in urban and rural settings by month. Capture setting and month significantly interacted in regards to total fat score, with urban birds showing a decline in total fat score over the course of the winter. No urban individuals were captured in January, and only a single rural individual in April. Lines represent ± 1SEM.

3 Urban

Rural

1 Body Body moltscore

0 March April

Month

Figure 2.6. Body molt score of Spinus tristis in urban and rural settings by month. Urban captures were observed to have greater active body molt than rural individuals. Body molt was not observed in any birds until March (one rural individual was captured in April, but was not undergoing active body molt). Lines represent ± 1SEM.

Totalscore fat 1

0 None Low Moderate Heavy

Body molt

Figure 2.7. Total fat score of Spinus tristis by progression of body molt. Total fat (furcular + abdominal fat scores) declined significantly with increasing levels of active body molt. Birds in all categories of active molt carried a notably lower mean total fat than did birds not undergoing body molt.

0.8

0.6

0.4

0.2

0 modulustransformed)

- Body Body conditionscore

(log -0.2

-0.4 None Low Moderate Heavy

Body molt

Figure 2.8. Body condition of Spinus tristis by progression of body molt. Size-adjusted body condition score did not correspond with active body molt status.

Supplemental Food and Stress Physiology

One important habitat characteristic was not represented consistently across urban and rural trapping sites. All urban sites where S. tristis were captured included a bird feeder that had been provided by residents prior to trapping (see Table 2.1). One of the sites categorized as rural, and where some 80% of rural captures took place, also had a feeder present. As food availability and foraging demands are important factors in predicting baseline avian CORT (Kitaysky et al. 2001a, Pravosudov et al. 2001), I also analyzed the data for differences between individuals captured at sites with and without established feeders.

Baseline CORT concentrations differed between sites with feeders and those without

(t = -2.742, df = 5.357, P = 0.038), with higher baseline concentrations found in birds captured at sites with established feeders (Fig. 2.9). There was no significant difference in stress-induced CORT at 30 minutes (t = 0.632, df = 3.862, P = 0.563) (Fig. 2.10) or in the percent increase in CORT (t = -0.212, df = 2.144, P = 0.851) (Fig. 2.11).

Additionally, capture setting—considering both the urban/rural nature and feeder provision—predicted total fat score (F1,24 = 9.310, P = 0.005). This result held true when adjusting for month and time of capture, both of which may impact the size of fat deposits that a bird carries. The three types of setting (urban with feeder, rural with feeder, and rural with no feeder) were compared using pairwise Welch two sample t-tests with a Bonferroni correction. Birds at urban sites with a feeder carried significantly smaller fat stores than did birds at rural sites with no feeder (t = 2.829, df = 12.033, P = 0.046). Fat score did not differ between rural sites with and without feeders (t = 1.747, df = 9.577, P = 0.338) nor between urban and rural sites that had feeders (t = 1.675, df = 21.128, P = 0.326).

1.0

Plasma [CORT] (ng/mL)

0.1 Feeder No feeder Feeder status at capture setting Figure 2.9. Baseline CORT concentration in Spinus tristis from settings with and without bird feeders. Individuals captured in settings with a regularly provided feeder (n = 23) showed higher baseline CORT levels than those captured in settings with no established feeder (n = 4). CORT concentrations are shown on a log scale. Lines represent ± 1SEM.

2 Plasma[CORT] (ng/mL)

0 Feeder No feeder Feeder status at capture setting

Figure 2.10. Induced (30-minute) CORT concentration in Spinus tristis from settings with and without bird feeders. Induced CORT concentrations in birds trapped at settings with a provided feeder (n = 21) did not differ significantly from those captured at settings with no established feeder (n = 4). Lines represent ± 1SEM.

10000

% % increase plasmain[CORT] 1 Feeder No feeder Feeder status at capture setting

Figure 2.11. Percent increase in CORT concentration (from baseline to 30-minute level) in Spinus tristis from settings with and without bird feeders. Individuals captured at settings with a provided feeder (n = 19) and those without an established feeder (n = 3) did not display different increases in CORT concentration, taken as percentage of baseline level. Percentages are shown on a log scale. Lines represent ± 1SEM.

Totalscore fat

0 Rural (no feeder) Rural (feeder) Urban (feeder) Capture setting Figure 2.12. Total body fat of Spinus tristis from varying capture settings. Individuals from rural sites with no regularly provided bird feeder (n = 7) carried significantly greater total fat stores than did birds trapped at urban sites with a provided feeder (n = 12).

DISCUSSION

Spinus tristis sampled in this study showed no difference in baseline or stress-induced

CORT levels between urban and rural capture settings. However, a number of interesting findings suggest areas for further research.

Urban/Rural Stress Physiology

Urban and rural birds sampled in this study did not differ in any CORT measure

(baseline, stress induced, or percent increase). This may be due to generally low sensitivity to stressors during the non-breeding season. Numerous passerine species are known to modulate

CORT responses, with less sensitivity to disturbances during the non-breeding season and molt than during breeding (Romero et al. 1998a,b, 2000, 2006; Astheimer et al. 1994;

Breuner and Wingfield 2000; Romero and Remage-Healy 2000; Pryke et al. 2012). It is possible that this seasonal modulation of the HPA axis supersedes any habitat-specific differences that may exist during other life history stages. Romero et al. (2006) conducted a similar comparative study of House Sparrow (Passer domesticus) CORT responses in benign and harsh habitats. While the birds displayed seasonal CORT cycles, those living in an urban area surrounded by harsh environmental conditions did not differ from those in a more natural setting (Romero et al. 2006). My data suggest that S. tristis may share a similar commensal relationship with humans, wherein provision of supplemental food is sufficient to mitigate any negative effects of residing in developed areas.

Other Urban/Rural Differences

There were some differences between the urban and rural birds sampled, however.

Second-year (SY) birds were identified only at rural locales, though in small numbers

(n = 4). That these individuals experienced a greater mean percentage increase in CORT following the stress-inducing protocol indicates that there may be important age class differences in sensitivity to stressors.

Also of interest are the findings regarding body condition score. Knutie and Pereyra

(2012) found that both baseline and maximum CORT responses in S. tristis and Pine Siskins

(Spinus pinus) were negatively correlated with body condition. If body condition had differed significantly between urban and rural S. tristis in this study, a corresponding variation in induced CORT response would be expected. Neither body condition nor induced CORT responses differed between capture settings. I also did not detect any correlation between body condition and CORT measures, however, which differs from Knutie and Pereyra’s

(2012) finding.

Total fat did differ between capture settings, with urban birds declining in total fat throughout the winter while rural birds maintained similar fat scores month to month. Some researchers have found increasing fat stores in CORT-supplemented birds, though no change in overall mass; this may be due to a combination of CORT’s behavioral effects, prompting additional feeding, and CORT-triggered gluconeogenesis causing the simultaneous breakdown of muscle protein (Wingfield and Silverin 1986, Wingfield et al. 1998). However, other studies of free-ranging birds have found no correlation between fat and baseline or induced CORT (Romero et al. 1998a,b). In nature, feeding quickly reduces CORT levels, providing negative feedback for a CORT-induced drive to take in additional food (Romero et

al. 1998b). A more likely explanation for the declining fat scores in these urban birds is the correlation between fat stores and molt. Birds tend to carry less fat when undergoing molt, which is an energetically demanding process (Romero et al. 1998a). Urban birds sampled in this study showed more progressed body molt in March and April than did rural birds captured in the same period, and this negatively correlated with total fat. What remains unanswered is the cause of molt differences between the rural and urban populations. It is possible that the dependable food sources provided by bird feeders allow molt to commence earlier or proceed more quickly in urban populations. Consistent access to food at urban feeders may also allow urban birds to carry smaller fat stores, as energy stores can be reliably replenished each day.

Supplemental Food and Stress Physiology

The provision of bird feeders did predict baseline CORT concentrations. It is noteworthy that individuals captured at settings where a feeder had been provided prior to trapping sessions had higher baseline CORT levels than did those at non-feeder locales, regardless of urban or rural designation. Typically, birds with unpredictable or intermittent access to food display higher baseline CORT concentrations than those with unrestricted food resources (Pravosudov et al. 2001). Low-quality diets are also associated with high

CORT levels (Pryke et al. 2012); it is possible that seed provided in feeders represents lower- quality food than that naturally available to S. tristis. However, this is unlikely, considering that this CORT difference extended to birds trapped at rural settings with a provided feeder, who presumably had access to natural food resources nearby if those were preferable. It is more likely that the heightened baseline CORT measured at feeder sites is a result of which

birds locate and use those feeders. It is possible that birds locate provided feeders at times of natural resource shortages, when their CORT levels and foraging activity would be high

(Wingfield and Silverin 1986, Wingfield et al. 1998). CORT plays an important role in recovery from such stressors, prompting intensified feeding once food supplies are available again (Wingfield et al. 1998). It is likely, not that feeders cause birds to have heightened baseline CORT, but that birds with higher CORT are more likely to happen upon provided food sources.

It is also of interest that both urban/rural setting and access to bird feeders appear to be factors in predicting fat storage. Birds using feeders in urban areas likely face higher density of predators such as domestic cats (Lepczyk et al. 2004b). Both direct selection by predators and fat reductions by individual birds may contribute to smaller songbird fat stores when predation risk—and therefore, the need to maintain maneuverability for escape—is high (Gosler et al. 1995). While my attempts to capture S. tristis at urban settings without a provided feeder were not successful, a focused effort to obtain such samples would be useful in further elucidating the roles of setting and feeder provision in predicting the size of fat stores that birds carry.

Conclusions and Future Research

I found no difference in winter CORT responses between Spinus tristis captured in urban and rural environments. This study did reveal some differences between individuals overwintering in different habitat types, however, that warrant further exploration.

One limitation to this study is the complexity of the HPA axis and the multiple stages at which modulation can occur (Astheimer et al. 1994; see Fig. 2.1b). Not only do birds modulate hormonal responses to stress on a variety of levels during winter, but they also modulate the resulting behavioral responses (Breuner and Wingfield 2000). Thus, solely sampling total plasma CORT levels cannot predict the way in which a bird will behaviorally respond to a stressor (Breuner and Wingfield 2000). The fact that urban and rural birds did not differ in plasma CORT concentrations does not directly reveal what impact those CORT levels may have on individuals.

Further consideration of characteristics of urban-dwelling S. tristis could shed additional light on differences between those birds that utilize feeders and those using natural food resources within an urban setting. I was unable to capture any urban S. tristis at non- feeder sites; however, it would be interesting to know whether their baseline CORT is lower than birds using provided feeder sites, similar to the trend amongst rural S. tristis. There is also much room for research on molt phenology between urban and rural birds.

These initial findings suggest that S. tristis is able to cope with the stressors of urban environments in winter, prompting a cautiously optimistic outlook for the species’ continued survival alongside humans as developed areas continue to expand. The suggested areas for further study would aid in a more complete understanding of the species’ relationship to human development and ability to respond to a changing landscape.

WORKS CITED

Alves, J. A., T. G. Gunnarsson, D. B. Hayhow, G. F. Appleton, P. M. Potts, W. J. Sutherland, and J. A. Gill. 2013. Costs, benefits, and fitness consequences of different migratory strategies. Ecology 94:11–17.

Astheimer, L. B., W. A. Buttemer, and J. C. Wingfield. 1994. Gender and seasonal differences in the adrenocortical response to ACTH challenge in an Arctic passerine, Zonotrichia leucophrys gambelii. General and Comparative Endocrinology 94:33–43.

Bailey, R. E. 1952. The incubation patch of passerine birds. The Condor 54:121–136.

Batten, L. A. 1972. Breeding bird species diversity in relation to increasing urbanisation. Bird Study 19:157–166.

Beebe, R. 1937. Some cold-weather birds of the Upper Peninsula of Michigan. The Wilson Bulletin 49:34–36.

Bibby, C. J., N. D. Burgess, D. A. Hill, and S. Mustoe. 2000. Bird census techniques. Second edition. Academic Press, London, United Kingdom.

Blair, R. B. 2001. Creating a homogeneous avifauna. Pages 459–486 in Avian ecology and conservation in an urbanizing world (J. M. Marzluff, R. Bowman, and R. Donnelly, Editors). Springer Science+Business Media, New York, USA.

Blair, R. 2004. The effects of urban sprawl on birds at multiple levels of biological organization. Ecology and Society 9:2.

Blas, J. 2015. Stress in birds. Pages 769–807 in Sturkie’s avian physiology (C. G. Scanes, Editor). Academic Press, Cambridge, Massachusetts, USA.

Bock, C. E. 1984. Geographical correlates of abundance vs. rarity in some North American winter landbirds. The Auk 101:266–273.

Bock, C. E., and L. W. Lepthien. 1976. Changing winter distribution and abundance of the Blue Jay, 1962-1971. American Midland Naturalist 96:232–236.

Bonier, F. 2012. Hormones in the city: endocrine ecology of urban birds. Hormones and Behavior 61:763–772.

Bonier, F., P. R. Martin, and J. C. Wingfield. 2007. Urban birds have broader environmental tolerance. Biology Letters 3:670–673.

Borror, D. J. 2003. Common bird songs. Dover Publications, Mineola, New York, USA.

Boscolo, D., J. P. Metzger, and J. M. E. Vielliard. 2006. Efficiency of playback for assessing the occurrence of five bird species in Brazilian Atlantic Forest fragments. Anais da Academia Brasileira de Ciências 78:629–644.

Bourne, L. S., and A. M. Baker. 1968. Urban development, Ontario and Quebec: outline and overview. Department of Geography, University of Toronto, Ontario, Canada.

Breuner, C. W., and J. C. Wingfield. 2000. Rapid behavioral response to corticosterone varies with photoperiod and dose. Hormones and Behavior 37:23–30.

Brittingham, M. C., and S. A. Temple. 1989. Patterns of feeder use by Wisconsin birds: a survey of WSO members. Passenger Pigeon 51:321–324.

Broggi, J., E. Hohtola, M. Orell, and J.-Å. Nilsson. 2005. Local adaptation to winter conditions in a passerine spreading north: a common-garden approach. Evolution 59:1600–1603.

Butcher, G. S., and C. E. McCulloch. 1990. Influence of observer effort on the number of individual birds recorded on Christmas Bird Counts. Pages 120–129 in Survey designs and statistical methods for the estimation of avian population trends (J. R. Sauer and S. Droege, Editors). USDOI, Fish and Wildlife Service, Biological Report 90. Washington, D.C., USA.

Butler, J. R., D. P. MacMynowski, C. Laurent, and T. L. Root. 2007. Temperature-associated dynamics of songbird winter distributions and abundances. AMBIO 36:657–660.

Buttemer, W., L. Astheimer, and J. Wingfield. 1991. The effect of corticosterone on standard metabolic rates of small passerine birds. Journal of Comparative Physiology B 161:427– 431.

Canterbury, G. 2002. Metabolic adaptation and climatic constraints on winter bird distribution. Ecology 83:946–957.

Carey, C., and W. R. Dawson. 1999. A search for environmental cues used by birds in survival of cold winters. Pages 1–31 in Current ornithology. Volume 15 (V. Nolan, Jr., E. D. Ketterson, and C. F. Thompson, Editors). Springer Science+Business Media, New York, USA.

Carey, C., W. R. Dawson, L. C. Maxwell, and J. A. Faulkner. 1978. Seasonal acclimatization to temperature in cardueline finches. Journal of Comparative Physiology 125:101–113.

Carrascal, L. M., J. Seoane, D. Palomino, C. L. Alonso, and J. M. Lobo. 2006. Species- specific features affect the ability of census-derived models to map winter avian distribution. Ecological Research 21:681–691.

Chace, J. F., and J. J. Walsh. 2006. Urban effects on native avifauna: a review. Landscape and Urban Planning 74:46–69.

Chamberlain, D. E., A. R. Cannon, M. P. Toms, D. I. Leech, B. J. Hatchwell, and K. J. Gaston. 2009. Avian productivity in urban landscapes: a review and meta-analysis. Ibis 151:1–18.

Chávez-Zichinelli, C. A., I. MacGregor-Fors, P. T. Rohana, R. Valdéz, M. C. Romano, and J. E. Schondube. 2010. Stress responses of the House Sparrow (Passer domesticus) to different urban land uses. Landscape and Urban Planning 98:183–189.

Cockrem, J. F. 2007. Stress, corticosterone responses and avian personalities. Journal of Ornithology 148:169–178.

Cornelius, J. M., C. W. Breuner, and T. P. Hahn. 2012. Coping with the extremes: stress physiology varies between winter and summer in breeding opportunists. Biology Letters 8:312–315.

Coutlee, E. L. 1967. Agonistic behavior in the American Goldfinch. The Wilson Bulletin 79:89–109.

DesRochers, D. W., J. M. Reed, J. Awerman, J. A. Kluge, J. Wilkinson, L. I. van Griethuijsen, J. Aman, and L. M. Romero. 2009. Exogenous and endogenous corticosterone alter feather quality. Comparative Biochemistry and Physiology A 152:46–52.

Ditchkoff, S. S., S. T. Saalfeld, and C. J. Gibson. 2006. Animal behavior in urban ecosystems: modifications due to human-induced stress. Urban Ecosystems 9:5–12.

Drennan, S. R. 1981. The Christmas Bird Count: an overlooked and underused sample. Pages 24-29 in Studies in avian biology. Number 6. Estimating numbers of terrestrial birds (C. J. Ralph and J. M. Scott, Editors). Allen Press, Inc., Lawrence, Kansas, USA.

Dunn, E. H. 1995. Bias in Christmas Bird Counts for species that visit feeders. The Wilson Bulletin 107:122–130.

Dunn, E. H., C. M. Francis, P. J. Blancher, S. R. Drennan, M. A. Howe, D. Lepage, C. S. Robbins, K. V. Rosenberg, J. R. Sauer, K. G. Smith, and S. G. Sealy. 2005. Enhancing the scientific value of the Christmas Bird Count. The Auk 122:338–346.

Elliott, L. 1998. Common birds and their songs. Houghton Mifflin Harcourt, Boston, Massachusetts, USA.

Evans, K. L., D. E. Chamberlain, B. J. Hatchwell, R. D. Gregory, and K. J. Gaston. 2011. What makes an urban bird? Global Change Biology 17:32–44.

Floyd, T. 2008. Smithsonian field guide to the birds of North America. HarperCollins, New York, USA.

Fokidis, H. B., M. Orchinik, and P. Deviche. 2009. Corticosterone and corticosteroid binding globulin in birds: relation to urbanization in a desert city. General and Comparative Endocrinology 160:259–270.

Geissler, P. H., and B. R. Noon. 1981. Estimates of avian population trends from the North American Breeding Bird Survey. Pages 42-51 in Studies in avian biology. Number 6. Estimating numbers of terrestrial birds (C. J. Ralph and J. M. Scott, Editors). Allen Press, Inc., Lawrence, Kansas, USA.

Gibbons, D. W., and R. D. Gregory. 2006. Birds. Pages 308–350 in Ecological census techniques (W. J. Sutherland, Editor). Cambridge University Press, Cambridge, United Kingdom.

Gregory, R. D., D. W. Gibbons, and P. F. Donald. 2004. Bird census and survey techniques. Pp. 17–56 in Bird ecology and conservation: a handbook of techniques (W. J. Sutherland, I. Newton, and R. Green, Editors). Oxford University Press, Oxford, United Kingdom.

Gosler, A. G., J. J. D. Greenwood, and C. Perrins. 1995. Predation risk and the cost of being fat. Nature 377:621–623.

Helms, C. W., and W. H. Drury. 1960. Winter and migratory weight and fat field studies on some North American buntings. Bird-Banding 31:1–40.

Holcomb, L. C. 1969. Breeding biology of the American Goldfinch in Ohio. Bird-Banding 40:26–44.

John, J. A., and N. R. Draper. 1980. An alternative family of transformations. Journal of the Royal Statistical Society C-Applied Statistics. 29:190–197.

Jokimäki, J., P. Clergeau, and M.-L. Kaisanlahti-Jokimäki. 2002. Winter bird communities in urban habitats: A comparative study between central and northern Europe. Journal of Biogeography 29:69–79.

Jokimäki, J., and J. Suhonen. 1998. Distribution and habitat selection of wintering birds in urban environments. Landscape and Urban Planning 39:253–263.

Jones, D. N., and S. J. Reynolds. 2008. Feeding birds in our towns and cities: a global research opportunity. Journal of Avian Biology 39:265–271.

Kaufmann, G. W. 1988. The usefulness of taped Spotless Crake calls as a census technique. The Wilson Bulletin 100:682–686.

Kennedy, E. A., C. R. Lattin, L. M. Romero, and D. C. Dearborn. 2013. Feather coloration in museum specimens is related to feather corticosterone. Behavioral Ecology and Sociobiology 67:341–348.

Kitaysky, A., E. Kitaiskaia, J. Piatt, and J. Wingfield. 2003. Benefits and costs of increased levels of corticosterone in seabird chicks. Hormones and Behavior 43:140–149.

Kitaysky, A., E. Kitaiskaia, J. Wingfield, and J. Piatt. 2001a. Dietary restriction causes chronic elevation of corticosterone and enhances stress response in red-legged kittiwake chicks. Journal of Comparative Physiology B 171:701–709.

Kitaysky, A. S., J. C. Wingfield, and J. F. Piatt. 2001b. Corticosterone facilitates begging and affects resource allocation in the black-legged kittiwake. Behavioral Ecology 12:619– 625.

Knutie, S. A., and M. E. Pereyra. 2012. A comparison of winter stress responses in cardueline finches. The Auk 129:479–490.

Kress, S. W. 2006. The Audubon Society guide to attracting birds: creating natural habitats for properties large and small. Second edition. Comstock Publishing Associates, Ithaca, New York, USA.

La Sorte, F. A., and W. Jetz. 2010. Avian distributions under climate change: towards improved projections. Journal of Experimental Biology 213:862–869.

La Sorte, F. A., and F. R. I. Thompson. 2007. Poleward shifts in winter ranges of North American birds. Ecology 88:1803–1812.

Lepage, D., and C. M. Francis. 2002. Do feeder counts reliably indicate bird population changes? 21 years of winter bird counts in Ontario, Canada. The Condor 104:255–270.

Lepczyk, C. A., A. G. Mertig, and J. Liu. 2004a. Assessing landowner activities related to birds across rural-to-urban landscapes. Environmental Management 33:110–125.

Lepczyk, C. A., A. G. Mertig, and J. Liu. 2004b. Landowners and cat predation across rural- to-urban landscapes. Biological Conservation 115:191–201.

Lepczyk, C. A., P. S. Warren, L. Machabée, A. P. Kinzig, and A. G. Mertig. 2012. Who feeds the birds? Pages 267–282 in Studies in avian biology. Number 45. Urban bird ecology and conservation (C. A. Lepczyk and P. S. Warren, Editors). University of California Press, Berkeley, California, USA.

Link, W. A., and J. R. Sauer. 1999a. Controlling for varying effort in count surveys: An analysis of Christmas Bird Count data. Journal of Agricultural, Biological, and Environmental Statistics 4:116–125.

Link, W. A., and J. R. Sauer. 1999b. On the importance of controlling for effort in analysis of count survey data: modeling population change from Christmas Bird Count data. Pages 119–124 in Bird numbers 1998: where monitoring and ecological research meet: proceedings of the 14th International Conference of the European Bird Census Council (EBCC) in Cottbus (Brandenburg), Germany, 23-31 March 1998 (A. J. Helbig and M. Flade, Editors). AULU-Verlag, Wiebelsheim, Germany.

Lõhmus, M., L. F. Sundström, and F. R. Moore. 2006. Non-invasive corticosterone treatment changes foraging intensity in red-eyed vireos Vireo olivaceus. Journal of Avian Biology 37:523–526.

Luniak, M. 2004. Synurbization - adaptation of animal wildlife to urban development. Pages 50-55 in Proceedings of the 4th International Symposium on Urban Wildlife Conservation (W. W. Shaw, L. K. Harris, and L. Vandruff, Editors). University of Arizona, Tuscon, USA.

McGowan, K. J. 2001. Demographic and behavioral comparisons of suburban and rural American Crows. Pages 365–381 in Avian ecology and conservation in an urbanizing world (J. M. Marzluff, R. Bowman, and R. Donnelly, Editors). Springer Science+Business Media, New York, USA.

McGraw, K. J., and A. L. Middleton. 2009. American Goldfinch (Spinus tristis). Birds of North America (P.G. Rodewald, Editor). Cornell Lab of Ornithology, Ithaca, NY, USA https://birdsna.org/Species-Account/bna/species/amegfi DOI: 10.2173/bna.80 Accessed 29 Nov 2016.

Middleton, A. L. A. 1977. Increase in overwintering American Goldfinch, Carduelis tristis, in Ontario. Canadian Field Naturalist 91:165–172.

Morneau, F., R. Décarie, R. Pelletier, D. Lambert, J.-L. DesGranges, and J.-P. Savard. 1999. Changes in breeding bird richness and abundance in Montreal parks over a period of 15 years. Landscape and Urban Planning 44:111–121.

Morrell, T. E. 1991. Factors affecting detection of Great Horned Owls by using broadcast vocalizations. Wildlife Society Bulletin 19:481–488.

Mosher, J. A., M. R. Fuller, and M. Kopeny. 1990. Surveying woodland raptors by broadcast of conspecific vocalizations. Journal of Field Ornithology 61:453–461.

Nemeth, E., N. Pieretti, S. A. Zollinger, N. Geberzahn, J. Partecke, A. C. Miranda, and H. Brumm. 2013. Bird song and anthropogenic noise: vocal constraints may explain why birds sing higher-frequency songs in cities. Proceedings of the Royal Society B 280:20122798.

Partecke, J., I. Schwabl, and E. Gwinner. 2006. Stress and the city: urbanization and its effects on the stress physiology in European Blackbirds. Ecology 87:1945–1952.

Piersma, T., J. Reneerkens, and M. Ramenofsky. 2000. Baseline corticosterone peaks in shorebirds with maximal energy stores for migration: a general preparatory mechanism for rapid behavioral and metabolic transitions? General and Comparative Endocrinology 120:118–126.

Pravosudov, V. V. 2003. Long-term moderate elevation of corticosterone facilitates avian food-caching behaviour and enhances spatial memory. Proceedings of the Royal Society B 270:2599–2604.

Pravosudov, V. V., A. S. Kitaysky, J. C. Wingfield, and N. S. Clayton. 2001. Long-term unpredictable foraging conditions and physiological stress response in mountain chickadees (Poecile gambeli). General and Comparative Endocrinology 123:324–331.

Pryke, S. R., L. B. Astheimer, S. C. Griffith, and W. A. Buttemer. 2012. Covariation in life- history traits: differential effects of diet on condition, hormones, behavior, and reproduction in genetic finch morphs. American Naturalist 179:375–390.

Puckett, H. L., J. R. Brandle, R. J. Johnson, and E. E. Blankenship. 2009. Avian foraging patterns in crop field edges adjacent to woody habitat. Agriculture, Ecosystems and Environment 131:9–15.

Robb, G. N., R. A. McDonald, D. E. Chamberlain, and S. Bearhop. 2008. Food for thought: supplementary feeding as a driver of ecological change in avian populations. Frontiers in Ecology and the Environment 6:476–484.

Romero, L. M., N. E. Cyr, and R. C. Romero. 2006. Corticosterone responses change seasonally in free-living house sparrows (Passer domesticus). General and Comparative Endocrinology 149:58–65.

Romero, L. M., M. J. Dickens, and N. E. Cyr. 2009. The reactive scope model — a new model integrating homeostasis, allostasis, and stress. Hormones and Behavior 55:375– 389.

Romero, L. M., J. M. Reed, and J. C. Wingfield. 2000. Effects of Weather on Corticosterone Responses in Wild Free-Living Passerine Birds. General and Comparative Endocrinology 118:113–122.

Romero, L. M., and L. Remage-Healey. 2000. Daily and seasonal variation in response to stress in captive starlings (Sturnus vulgaris): corticosterone. General and Comparative Endocrinology 119:52–59.

Romero, L. M., K. K. Soma, and J. C. Wingfield. 1998a. Changes in pituitary and adrenal sensitivities allow the snow bunting (Plectrophenax nivalis), an Arctic-breeding song bird, to modulate corticosterone release seasonally. Journal of Comparative Physiology B 168:353–358.

Romero, L. M., K. K. Soma, and J. C. Wingfield. 1998b. Hypothalamic-pituitary-adrenal axis changes allow seasonal modulation of corticosterone in a bird. American Journal of Physiology-Regulatory Integrative and Comparative Physiology 274:R1338–R1344.

Sauer, J. R., and P. H. Geissler. 1990. Estimation of annual indices from roadside surveys. Pages 58–62 in Survey designs and statistical methods for the estimation of avian population trends (J. R. Sauer and S. Droege, Editors). U.S. Fish and Wildlife Service Biological Report 90.

Sauer, J. R., and W. A. Link. 2002. Using Christmas Bird Count data in analysis of population change. Pages 10–14 in American Birds: the 102nd Christmas Bird Count (G. S. LeBaron and C. Capobianco, Editors). National Audubon Society, Ivyland, Pennsylvania, USA.

Sauer, J. R., G. W. Pendleton, and S. Orsillo. 1995. Mapping of bird distributions from point count surveys. U.S. Forest Service General Technical Report PSW-GTR-149.

Saunders, A. A. 1947. The seasons of bird song the beginning of song in spring. The Auk 64:97–107.

Savard, J.-P. L., P. Clergeau, and G. Mennechez. 2000. Biodiversity concepts and urban ecosystems. Landscape and Urban Planning 48:131–142.

Scanes, C. G., and E. Braun. 2013. Avian metabolism: its control and evolution. Frontiers in Biology 8:134–159.

Silverin, B. 1986. Corticosterone-binding proteins and behavioral effects of high plasma levels of corticosterone during the breeding period in the pied flycatcher. General and Comparative Endocrinology 64:67–74.

Sleeter, B. M., T. L. Sohl, T. R. Loveland, R. F. Auch, W. Acevedo, M. A. Drummond, K. L. Sayler, and S. V. Stehman. 2013. Land-cover change in the conterminous United States from 1973 to 2000. Global Environmental Change 23:733–748.

Sliwa, A., and T. W. Sherry. 1992. Surveying wintering warbler populations in Jamaica: Point counts with and without broadcast vocalizations. The Condor 94:924–936.

Smith, S. M., and W. W. Vale. 2006. The role of the hypothalamic-pituitary-adrenal axis in neuroendocrine responses to stress. Dialogues in Clinical Neuroscience 8:383–395.

Stokes, D., L. Stokes, and L. Elliot. 2010. Stokes field guide to bird songs: eastern region. Hachette Audio, New York, USA.

Sullivan, B. L., C. L. Wood, M. J. Iliff, R. E. Bonney, D. Fink, and S. Kelling. 2009. eBird: a citizen-based bird observation network in the biological sciences. Biological Conservation 142:2282–2292.

Swales, B. H. 1903. Notes on the winter birds of Wayne County, Michigan. The Wilson Bulletin 15:20–24.

Swift, B. L., S. R. Orman, and J. W. Ozard. 1988. Response of Least Bitterns to tape- recorded calls. The Wilson Bulletin 100:496–499.

Tekiela, S. 2004. Birds of Michigan audio CDs. Adventure Publications, Cambridge, Minnesota, USA.

U.S. Department of the Interior, U.S. Fish and Wildlife Service, and U.S. Department of Commerce, U.S. Census Bureau. 2014. 2011 national survey of fishing, hunting, and wildlife-associated recreation.

Versaw, A. E. 2001. Winter range expansion and identification of the Lesser Goldfinch (Carduelis psaltria). Great Basin Birds 4:6–11. de Vos, A. 1964. Range changes of birds in the Great Lakes region. American Midland Naturalist 71:489–502.

Wilson, W. H. J. 1994. The distribution of wintering birds in central Maine: the interactive effects of landscape and bird feeders. Journal of Field Ornithology 65:512–519.

Wingfield, J. C., C. Breuner, and J. Jacobs. 1997. Corticosterone and behavioral responses to unpredictable events. Pages 267–278 in Perspectives in avian endocrinology (S. Harvey and R. J. Etches, Editors). Journal of Endocrinology, Bristol, United Kingdom.

Wingfield, J. C., D. L. Maney, C. W. Breuner, J. D. Jacobs, S. Lynn, M. Ramenofsky, and R. D. Richardson. 1998. Ecological bases of hormone—behavior interactions: the “emergency life history stage.” American Zoologist 38:191–206.

Wingfield, J. C., M. C. Moore, and D. S. Farner. 1983. Endocrine responses to inclement weather in naturally breeding populations of White-crowned Sparrows (Zonotrichia leucophrys pugetensis). The Auk 100:56–62.

Wingfield, J. C., K. M. O’Reilly, and L. B. Astheimer. 1995. Modulation of the adrenocortical responses to acute stress in arctic birds: a possible ecological basis. American Zoologist 35:285–294.

Wingfield, J. C., and L. M. Romero. 2001. Adrenocortical responses to stress and their modulation in free-living vertebrates. Pages 211–234 in Comprehensive physiology (R. Terjung, Editor). John Wiley & Sons, Inc., Hoboken, New Jersey, USA.

Wingfield, J. C., and B. Silverin. 1986. Effects of corticosterone on territorial behavior of free-living male song sparrows Melospiza melodia. Hormones and Behavior 20:405– 417.

Wingfield, J. C., J. P. Smith, and D. S. Farner. 1982. Endocrine responses of White-crowned Sparrows to environmental stress. The Condor 84:399-409.

Wingfield, J. C., C. M. Vleck, and M. C. Moore. 1992. Seasonal changes of the adrenocortical response to stress in birds of the Sonoran desert. Journal of Experimental Zoology 264:419–428.

Zhang, S., F. Lei, S. Liu, D. Li, C. Chen, and P. Wang. 2011. Variation in baseline corticosterone levels of Tree Sparrow (Passer montanus) populations along an urban gradient in Beijing, China. Journal of Ornithology 152:801–806.

APPENDIX A: Bird species detected during American Goldfinch (Spinus tristis) vocalization playback surveys, 2015.

APPENDIX B: Corticosterone (CORT) and morphological measures of American Goldfinches (Spinus tristis) captured, 2016.