Florida State University Libraries

2016 Social Nesting Behvior of the Bahama Parrot on Abaco Island and Its Conservation Implications Caroline S. (Caroline Stahala) Walker

Follow this and additional works at the FSU Digital Library. For more information, please contact [email protected] FLORIDA STATE UNIVERSITY

COLLEGE OF ARTS AND SCIENCES

SOCIAL NESTING BEHVIOR OF THE BAHAMA PARROT ON ABACO ISLAND AND ITS

CONSERVATION IMPLICATIONS

By

CAROLINE S. WALKER

A Dissertation submitted to the Department of Biological Science in partial fulfillment of the requirements for the degree of Doctor of Philosophy

2016 Caroline Stahala Walker defended this dissertation on June 23, 2016. The members of the supervisory committee were:

Emily H. DuVal Professor Directing Dissertation

Peter Beerli University Representative

Kimberly A. Hughes Committee Member

Brian Inouye Committee Member

Frances C. James Committee Member

Thomas E. Miller Committee Member

The Graduate School has verified and approved the above-named committee members, and certifies that the dissertation has been approved in accordance with university requirements.

ii

Dedicated to Ricky Johnson

iii

ACKNOWLEDGMENTS

I have to start by acknowledging my wonderful family for their unconditional love and support. Thank you Jitka, Peter, Mike and Philipp Stahala for always being there for me and encouraging me to keep going.

I met my husband, Sandy Walker, while I was working on this project but he became the biggest supporter and encourager. Sandy was an integral part of this project, during field work he became a mechanic, driver, handyman, bag maker, field technician, logistical organizer, advocate, cheerleader and friend. Thank you Sandy for everything you have done for me.

A conservation research project of this magnitude required logistic and financial support from an enormous number of individuals and organizations. I would like to thank the following for financial assistance including The Bahamas National Trust, The Amazona Society UK, The

Amazona Society US, Antonius Roberts, FSU Dept of Biology Loftin Award and Short

Fellowship, FSU International Dissertation Research Scholarship, Friends of the Environment,

Kathryne and Richard Thorp, Lindroth Development Ltd, Mark Hagen, Michelle LeMoroux,

Islands by Design, Parrots International, Rare Species Conservation Foundation, Schooner Bay,

Shirley Cartwright, Susan Hilliard, and Suzan Payne. Kari Schmidt and George Amato, your laboratory use, analysis support and guidance were invaluable. Bill Hayes, thank you for the talks on species concervation.

Field support came from a dedicated crew of field technicians including Catrina Damrell,

Josh Kelly, Emily Davidson, Kara Cox, Amanda Hitchcock, Deanna Quinn, Fio Kerins, Lucy

Nepstad, Jenna Motz, Zach Fitzner, Alex Hughes, Nicole Napolski and Alaina Maier.

Logistical field support came from all the BNT staff including Eric Carey, Lynn Gape,

Kadie Mills, David Knowles and the one and only king of south Abaco, parrot protector Mr.

iv

Marcus Davis. Additional Abaco folks I have to thank include Anita and Stephen Knowles,

Barbara Forman, Marilla Santillo and Leanne Hopkins.

Academic support came from my major advisor Dr. Emily DuVal and committee including Peter Beerli, Tom Miller, Fran James, Kim Hughes, and Brian Inouye. Thank you all for pushing me to think beyond my main focus of conservation and making me a more well- rounded conservation biologist.

I cannot leave FSU without thanking the cohort of students I started with, particularly

Anna Strimaitis and Christina Kwapich. Thank you ladies for always being there.

v

TABLE OF CONTENTS

List of Tables ...... vii

List of Figures ...... viii

Abstract ...... x

1. INTRODUCTION ...... 1

2. GROUP NESTING BY THE BAHAMA PARROT (AMAZONA LEUCOCEPHALA BAHAMENSIS): ASSESSING EFFECTS OF CAVITY DENSITY AND MICROHABITAT ...11

3. INFLUENCE OF RELATEDNESS AND GENETIC POPULATION STRUCTURE ON NEST AGGREGATION OF THE BAHAMA PARROT (AMAZONA LEUCOCEPHALA BAHAMENSIS) ON ABACO ISLAND ...... 29

4. INTERACTIONS BETWEEN NEIGHBORS AND THE INFLUENCE OF PREDATORS ON BAHAMA PARROT NESTING ...... 48

5. RECLASSIFICATION OF THE CUBAN PARROT (AMAZONA LEUCOCEPHALA) COMPLEX ...... 66

6. CONCLUSION ...... 81

APPENDIX A: ALLELE MATCHES FOR EACH CHICK/PARENT PAIR SAMPLED INCLUDING YEAR SAMPLE WAS COLLECTED ...... 85

APPENDIX B: 154 BAHAMA PARROTS SAMPLED FOR DNA BY YEAR, NEST ID,AGE AND SEX ...... 87

APPENDIX C: ACUC ASSURANCE LETTER ...... 93

APPENDIX D: LICENCE TO REUSE FIGURES ...... 94

References ...... 95

Biographical Sketch ...... 108

vi

LIST OF TABLES

2.1 Degree of exposed rock and distance to nearest plant differed between parrot nests (n=68) and unused cavities (n=29). Wilcoxon rank-sum tests were used to compare values of characteristics between used and unused nests. Asterisk (*) indicates significance of p<0.05...... 28

3.1 Characterization of population-wide variation at microsatellite loci genotyped for the Abaco population of the Bahama Parrot (N=154), observed heterozygosity (Ho) and expected heterozygosity (He) (Russello et al. 2001, Russello, 2008)...... 47

4.1 Demographic information for the Bahama Parrot and predator control information for 2003- 2005 and 2010-2012. Demographic values include adult survival, as estimated through modified Mayfield analysis (Stahala 2005); mean breeding productivity as estimated through Mayfield analysis. Data on predator control were obtained from the Bahamas National Trust (BNT), including years in which data were collected for predator control and the number of cats removed during or prior to the breeding season...... 62

4.2 Results of predator simulation at focal nests indicate focal and nearest neighbor (NN) responded vocally during the simulation but at different rates. Neighbors did not approach focal nests during predator simulations. Asterisk denotes significant value at the 0.05 level. 63

4.3 Model parameters and results for GLMMs with the binary response of nest success N=216 nest observations. Predator control and nearest neighbor nest fate were the only measured factors had an effect on nest success of focal Bahama Parrot nests. Nests were more successful when neighbors were successful and predator control was in effect...... 63

5.1 Diagnostic characteristics of the proposed species and subspecies of Amazona leucocephala. Bahama species descriptions are from Reynolds and Hayes (2009). Cuba and Cayman species descriptions from Forshaw (2006) and Reynolds and Hayes (2009)...... 76

5.2 Criteria used to reclassify avian species in the Caribbean, including Hispaniolan Oriole (Icterus d. dominicensis), Cuban Oriole ( I. d. melanopsis), Yellow-throated Warbler (Dendroica d.dominica), and Bahama Warbler (Dendroica d. flavescens), compared to the criteria for the Cuban Amazon complex and congeners of the Puerto Rican Parrot (Amazona vittata) and Hispaniolan Parrot (Amazona ventralis). Comparisons support the validity of reclassification of the Cuban Amazon complex. CN – Indicates closest neighboring parrot population to the study population...... 77

5.3 Proposed classification of the Cuban Amazon complex in light of new information about the seven populations (based on Table 5.1)...... 78

vii

LIST OF FIGURES

1.1 Cuban Parrot (Amazona leucocephala) distribution and nesting distribution on Abaco and Inagua (points) (figure from Russello et al. 2010). Black dots indicate nests found during 2007 for Great Inagua and 2004 for Great Abaco...... 9

1.2 South Abaco distribution of parrot nests and habitat. Light blue line is a minimum convex polygon of the nesting area; black line delineates the Abaco National Park; blue and red stars denote nest sites monitored in 2003 and 2004 respectively...... 10

2.1 Ripley’s K distribution graph of Bahama Parrot nests on Abaco, The Bahamas indicate a clustered pattern of nests at all distances up to 15km. Green and blue colored lines indicate confidence interval for randomly distributed nests. Positive K(r) values above the confidence interval for random distribution indicate a clustered distribution...... 23

2.2 Number of active nest cavities in a four hundred meter radius of nest cavities (N=82) in comparison to random suitable cavities (N= 51) indicates parrots nested near other nesting individuals. Wilcoxon z = -9.98, p<0.0001. Thick horizontal line indicates median, height of box represents inter-quartile range and whiskers are 1.5 inter-quartile ranges, data points beyond this are outliers...... 24

2.3 Nests were found in areas with more available ground cavities, when the number of empty cavities in 300m2 strip transects around nests (N=82) and randomly selected cavities (N=51) was compared to the number of empty cavities found within a 400m radius around that cavity; Wilcoxon z = -2.73, p= 0.006). Thick horizontal line indicates median; length of box is inter-quartile range, and whiskers are 1.5 inter-quartile ranges...... 25

2.4 Number of active nests increased with the number of empty cavities per meter squared in the nest area (GLM, F1,132=3.27, P<0.001, Spearman’s rho=0.38 p=0.0004). Points were jittered to show individual nests...... 26

2.5 Plants grew nearer to non-nest cavities (N=29) than to either old nests found in previous years (N=30) or active nests used for the first time in a given breeding season (N=38). Wilcoxon rank-sum test z values for new vs old z=-0.19; old vs non-nest z=-2.22; new vs non-nest z=-2.51...... 27

3.1 The distance between nests is not related to pairwise relatedness among chicks.Linear (ANOVA F(1, 45)=0.35, p>0.56) and cubic (ANOVA F(3, 46)=1.85, p>0.15) fits were used with only one chick per nest (47 pairwise comparisons from 47 different nest cavities)...... 43

3.2 A comparison of relatedness between 3 distance categories Closest Nest (CN); Farthest nest (greater than 5km, FN); and Mid-distance nest (1-5km, MN) of neighboring nests (N=150 total nest pairs) shows no difference between categories...... 44

viii

3.3 Map of southern Abaco Island in, The Bahamas, showing the distribution of 76 nests where at least one individual was sampled for DNA...... 45

3.4 Map of nest locations used to compare northern (N=13) and southern (N=8) clusters within the nesting range for subpopulation structure. Fst =0 .03 (p=0.25), indicates no Substructure ...... 46

4.1 Vocal synchrony was unrelated to distance between nesting pairs (Wilcoxon Rank Sum Test; NN vs DN Z = 0.16 p = 0.16; RN vs DN Z=0.275 p=0.78; RN vs NN Z=-1.31 p=0.18). Synchrony rate was measured as the number of times nests vocalized within the same time interval, compared between a focal nest (focal) and surrounding nests at different distances; NN = Nearest Neighbor (N=30), DN = Distant Neighbor (30) over 1km, Focal and RN = Random Neighbor (N=30). Grand mean for number of synchronous vocalization intervals per observation period is indicated by horizontal line. Box plot shows median, interquartile range,15x interquartile range, and outliers beyond this range ...... 64

4.2 Focal individuals and their nearest neighbors increased their rates of vocalization during the predator treatment (GLMM, Z= -2.85, p<0.005). Neighbors responded less frequently than focal pairs. Figure shows the proportion of nests out of 39 trials that were approached by artificial predators...... 65

5.1 Distribution of Cuban Amazon complex with the current and proposed classification...... 79

5.2 Bayesian haplotype tree depicting relationships among sampled Amazona leucocephala haplotypes and two outgroups relative to their geographic and taxonomic distributions. Bayesian posterior probabilities (50%) are indicated above the branches. Reprinted with permission from Russello et al. 2010...... 80

ix

ABSTRACT

Nesting distributions of animals vary from isolated individuals to large colonies of breeding individuals and they can be influenced by numerous factors, including environmental conditions, relatedness and social interactions. Parrots are one of the most threatened orders of birds. Factors that influence their nesting success in natural habitat have been the focus of previous research. However, these past studies operated in systems limited in nest sites, which may also have influenced population dynamics. In contrast, the Bahama Parrot nests underground in abundant limestone solution cavities in the karst pine forest of Abaco Island, The

Bahamas. This situation allows exploration of environmental and social factors that may influence nest success and breeding productivity without the limitation of the availability of nesting site resources. The objective of this study is to investigate the causes and consequences of semi-colonial nesting aggregations in the Bahama Parrot, addressing the influence of local habitat, relatedness of nesting neighbors, social behavior, and success of neighboring nests. The results of this information can be used in the management of habitat and parrot populations on

Abaco Island, The Bahamas.

The research was carried out on Great Abaco Island during the parrot breeding season

(April - September) in 2010 – 2013. Nests were found by traversing logging roads. A number of ecological features were measured around nesting cavities and unused limestone cavities with the same dimensions as parrot cavities to identify favored characteristics. Nearest neighbors were identified for a subset of nests. Behavioral observations included time budgets of vocalization and movement to and from nests at focal and comparison nests (nearest neighbors and more distant nests) to determine if parrots were synchronizing behavior with close-nesting individuals.

Predator simulations were conducted to elicit defense responses from focal and nearest neighbor x nesting pairs. Using DNA collected from adults and chicks, I assessed whether spatially aggregated nests reflected kin clusters.

Bahama Parrots on Abaco were distributed in aggregated spatial patterns in the nesting area. They nested in cavities found in more open areas. One contributing factor for the nest distribution was the uneven distribution of limestone cavities. Greater numbers of nests were found in areas with higher cavity concentrations. However, areas with high cavity concentrations but no parrot nests indicated that additional factors also were involved in concentrating nests in an area.

Relatedness did not influence the aggregated nesting pattern. Close nesting neighbors were not more closely related than nesting individuals at other nest sites. No evidence of extra pair paternity was found within the small sample of nests that had full families sampled.

However, genotypes of chicks raised in the same cavity in different years did provide support for the general belief that monogamous parrot pairs often reused the same nesting cavities over multiple years.

Distances between neighbors had no effect on vocal synchrony. When a predator was introduced to a nest, the vocal response by the nesting pair and the nearest neighbor nesting pair increased, however no other behaviors provided nest defense. An effective defense against feral cats as predators was the predator control program carried out by the Bahamas National Trust.

Nest success was higher in years with the predator control program underway. I did detect a relationships between nest success at a focal nest and its neighbor.

Finally, I reviewed the taxonomic status of the Cuban parrot (Amazona leucocephala) complex. My review suggested that the two Bahama populations of the Cuban parrot (Amazona leucocephala bahamensis) should be classified as separate species (Amazona abaconensis and

xi

Amazona inaguanensis) and that the two Cuban populations should be a single subspecies

(Amazona leucocephala leucocephala).

These parrots use limestone cavities in the ground in open pine woodland as nest sites, and areas with higher densities of these cavities are used to a greater extent. As populations are expected to increase due to the success of a predator control program, sites with high cavity density are expected to be prime habitat for new nesting individuals. I therefore recommend the continuation of the prescribed burning to maintain the open understory that these parrots select.

Furthermore, given the effectiveness of a current feral cat removal program in increasing nesting success, I recommend that predator control continue.

xii

CHAPTER 1

INTRODUCTION

Social interactions among birds, such as group foraging, communal roosting, and pair bonding have been shown to influence population dynamics, sometimes to the benefit and sometimes to the detriment of species (Alfieri and Dugatkin 2009, Griesser and Nystrand 2009,

Travis et al. 1996, Monterrubio-Rico et al. 2006). For example, group behavior may be costly for individuals, as interactions expose individuals to increased risk of detection by predators, disease transmission, and competition for resources. Brown and Brown (1986) demonstrated a decrease of nestling success due to ectoparasites in larger groups of the colonial-nesting Cliff Swallows

(Petrochelidon pyrrhonota). However, other studies of this species demonstrated that individuals in large nesting aggregations benefit from increased effectiveness in detecting predators, amount of food nestlings receive, and survivorship of breeding individuals (Brown and Brown 1987,

Brown 1986). This suggests that group behavior can carry both costs and benefits to individuals.

When management decisions are being made for the protection of a social species, it would be helpful to understand how group behavior affects key aspects of individual survival and reproductive success. Nesting aggregations are particularly interesting in that these costs and benefits can be assessed for both breeding adults and their offspring, and management interventions during nesting periods can affect multiple demographic levels of the population.

The family Psittacidae (parrots) includes many highly social species, including those that are often found in flocks of various sizes for feeding, roosting and other activities (Enkerlin-

Hoeflich et al. 2006). Large flocks of parrots are typical at feeding sites (Brightsmith and

Aramburu 2004) and roost sites at night (Gnam and Burchsted 1991, Cougill and Marsden 2004,

Chapman et al. 1989). Parrots are even found to aggregate at nest sites (Monterrubio-Rico et al. 1

2006, Forshaw 1989). Monk parakeets (Myiopsitta monachus) are known for their communal

nesting behavior, building nest structures next to or alongside neighbors’ nests. Unlike other

parrots that are dependent on where cavities are, monk parakeets actually choose to build their

nest cavities right next to one another.

Parrots in the genus Amazona are 30 species of short tailed parrots native to South

America, Mexico and the Caribbean. These Amazon parrots are secondary cavity nesters, relying on preformed cavities in trees. Nesting pairs frequently reuse nest sites, providing an opportunity for stable interactions to develop not only between mates but also among neighboring nesting pairs (Snyder et al 1987). It is unclear what the cost and benefits of these behaviors are in most instances. Flock sizes of Amazon parrots (genus Amazona) are prone to reduction due to

anthropogenic effects such as habitat fragmentation, habitat loss, predation and poaching which

could in turn impact group dynamics (Wiley et al. 2004, Snyder et al. 1987). Understanding the

costs and benefits of social behavior in Amazon parrots may provide management tools to

minimize or mitigate some of these effects.

The Bahama Parrot (Amazona leucocephala bahamensis) is highly social during all times

of the year and at all life stages. In contrast to most other parrots, which use tree cavities as nest

sites, the Bahama Parrots on Abaco Island use underground nests (Snyder et al. 2000). This nesting behavior may be due to a lack of tree cavities or the beneficial features the underground limestone cavities provide (O’Brien et al. 2006). My dissertation research addresses hypotheses about the costs and benefits of social nesting behavior in this Amazon parrot. Specifically my work addresses habitat, behavioral and genetic factors associated with nesting behavior.

2

Background

Potential reasons for parrots to live in groups instead of in solitary situations include ecological benefits of resources, predator avoidance, reproduction, and energy expenditure (Horn

1968). Groups may allow foragers to take more or larger and more dangerous prey (Alexander

1974). For example, a pack of wolves may be able to take down larger game, or to find more distributed prey (Sutton et al. 2015). The resource dispersion hypothesis suggests that group size and distribution are based on the distribution and abundance of a specific resource (Horn 1968,

Johnson 2002). Such resources may include food, mates, refugia, or resources used for nesting.

Group behavior could also provide benefits attained through kin selection. Kin selection could explain social grouping if the individuals acting in groups increase the success of closely related individuals (Hamilton 1964). Prior work in parrots has investigated demography and habitat use at the population level, but variation in individual behavior and its relation to variation in demographic parameters, such as breeding productivity and survival, remain poorly understood

(Snyder et al. 1987, Thompson 2004, Enkerlin-Hoeflich 2006, Myers et al. 1996, Gilardi and

Munn 1998).

Understanding the implications of social behavior such as the effects of nesting density and interactions between nesting neighbors is important because parrot species tend to be of conservation concern. The parrot populations in the West Indies have experienced severe declines and even extinction over the last several hundred years. The 12 macaw species (genus

Ara) which were once found throughout the West Indian islands are now extinct, as are two thirds of the parakeet species (Aratinga spp.) and one third of the Amazon parrot species that once existed in the area (Wiley et al. 2004). The remaining 15 parrot species are still under natural and anthropogenic threats of extinction, as recognized by CITES and USFWS and IUCN

3

(IUCN 2009). Most of the remaining species are currently threatened due to habitat loss, predation and poaching. In addition to investigating the adaptive benefits of group nesting, this project will also inform our understanding of the threat of the continued decline in population sizes and in its contribution to the instability of Amazon parrot populations, since lower population sizes (lower nesting densities) may reduce the ability of individuals to interact (Allee

1949).

Study System The Bahama Parrot (Amazona leucocephala bahamensis) is currently considered a subspecies of the Cuban Amazon or Cuban Parrot, and is found only on Great Abaco and Great

Inagua islands, The Bahamas (Figure 1.1). Additional subspecies are found on Cuba

(A.l.leucocephala), Isle of Pines (A.l. palmarum), Grand Cayman Island (A.l. caymanensis), and

Cayman Brac (A.l. hesterna) (Figure 1.1). These subspecies are geographically isolated and

there is no evidence of recent dispersal between islands (Ottens-Wainright et al. 2004).

Fossil and documented records indicate that Bahama Parrots were found throughout the

Bahamas, possibly as recently at Columbus’s landing on San Salvador in 1492 (Olson and

Hilgartner 1982, Bond 1956a, Wiley et al. 2004). However, more recent documented history

indicates the Acklin’s Island population was extirpated by the 1950’s (Bond 1947; Bond 1956a;

Bond 1956b). This caused the two closest populations of the Bahama Parrot, Abaco and Inagua

Islands, to be separated by about 400 miles (650 km).

The parrot population on Great Abaco is found in the southern portion of the island.

Nesting in this population takes place in a discrete portion of south Abaco from the

settlements of Crossing Rocks (W 77.19o, N 26.140) and Sandy Point (W 77.40, N 26.010) to

4

Hole-in-the-Wall (W 77.20, N 25.890) (Figure 1.2). I refer to this discrete nesting range as the nesting area. Nonbreeding habitat along the eastern coast of the Southern portion of Great Abaco supports parrots as far north as the settlements of Cherokee and Little Harbour.

The distribution of parrots on Great Abaco varies depending on the time of year (Stahala

2008). During the breeding season (end of April – mid September) parrots can be found primarily within the pine forest of southern Great Abaco where limestone cavities are available

(Figure 1.2). During this time, the Caribbean Pine tree (Pinus caribaea bahamensis) produces green cones that are used by the parrots as a staple food source during the nesting season.

Pinecones are available every year; however, variability in quantity of this food source from year to year is unknown. Pine cones mature during the month of September, opening up and dispersing seeds, making this food source no longer available (Stahala 2005). During the nonbreeding period (September through April) parrots are found along the east coast of the

Southern portion of the island, which is a hardwood coppice habitat. This habitat has a greater variety of food sources available to the parrots than the pine forest during this time of year

(Stahala 2005).

The annual home range of an individual Bahama Parrot, based on minimum convex polygon telemetry analysis, is 18,000± 2000 ha (4600 ± 4800 acres) and ranges from 5568 –

30,573 ha (13758 – 75547 acres) (Stahala 2008). Individuals may cover substantial distances: one Bahama Parrot was documented traveling 28 km (17 miles) in one day and a total of 44 km

(27 miles) over a 3-day period (Stahala 2008). The range and habitat used by parrots exceeds the size of the Abaco National Park, underscoring the need for greater habitat protection or, minimally, preservation of currently protected areas. Bahama Parrots leave the park area after

5 fledging and aggregate in flocks of 5-200 individuals during the nonbreeding season. Parrot aggregations change in size as individuals move between various sites.

The Abaco National Park was established in 1994 to protect the primary breeding habitat of the Bahama Parrot on Abaco. The park covers 8,302 ha with the Caribbean pine and some

hardwood coppice stands as the dominant forest cover. However, the park does not protect the

hardwood habitat parrots migrate to after the breeding season. The park, as well as the rest of the

Island, faced logging pressure over the last century. Pine was harvested by Owens-Illinois until

the 1960’s on Abaco. Seed trees were left throughout the logging operation to allow trees to

regenerate on the Island, which they have. The impacts of this operation on the Bahama Parrot

are unknown since we don’t have records about this population prior to the logging operation.

Research into the population began in 1976 with roost counts (Snyder 1982), which continued

into the mid to late 1980’s (Gnam and Burchsted 1991). Gnam’s research also included assessing

nesting biology of the parrot. The information provided by this research culminated in the

establishment of the Abaco National Park. A more comprehensive population viability analysis

was conducted in 2005 (Stahala 2005). This research indicated that non-native predators were

negatively affecting the population. Breeding parrots and their nests are vulnerable to feral cat

predation due to being on the forest floor. The loss of breeding adults was found to be limiting

the parrot population growth. Predator control, targeting the reduction of non-native species, is

being conducted in an effort to improve survival and productivity of the species. This program is

seen as controversial by those who try to protect the feral cats, therefore understanding the

impacts of the predator control effort on the Bahama parrot is important (Peterson et al. 2012).

Threats to the parrot are not limited to feral cats. Other non-native predators include rats

and raccoons while native predators include the Bahamas boa (Epicrates exsul), Red tailed

6

Hawks (Buteo jamaicensis) and Land crabs (Cardisoma guanhumi). However, unlike feral cats that can devastate and entire nesting family including breeding adults, chicks and/or eggs, native predators are less devastating. Boas can take eggs or small chicks but don’t harm the adults.

Hawks can take adults but only one at a time and not necessarily nesting individuals.

Nonpredator threats to the Bahama Parrot include habitat loss, impacts due to hurricanes and droughts, fires and poaching. The BNT implemented a predator control program targeting non- native predators in 2005 but only started intensive control efforts in 2010. A 2012 population survey suggests the population has increased on Abaco since predator control efforts were implemented from a point estimate of 3393 parrots in 2003 to 5373 parrots in 2012. My research into nesting aggregations will provide information on how to proceed with managing nesting areas for the growing parrot population.

The Bahama Parrot on Great Abaco has been studied to a greater extent than most

Amazon parrot species, with the exception of the critically endangered Puerto Rican Parrot

(Amazona vittata), because its unusual behavior of nesting underground in limestone solution cavities makes it a particularly tractable research subject. The research conducted to date on the

Bahama Parrot has focused on population (Gnam and Burchstead 1991, Stahala 2005), habitat

(Mori 2006, Stahala 2008) and genetic level questions (Russello et al. 2010), but information about variation in individual demography or behavior, or about the mechanisms driving behavior of this species is lacking. This information is typically difficult to obtain for Amazon parrots because of the challenge of catching wild individuals for marking. Because of the Abaco population’s unusual underground nesting habits, this parrot is easier to catch during the breeding season than other Amazons, thus allowing individual marking of wild adults.

7

Significance of Work In this project I explore the role of social interactions in the breeding biology of Bahama

Parrot. Parrots are one of the most threatened orders of birds and factors that influence nesting success in natural parrot habitat have been the focus of previous research (Snyder et al. 2000,

Seixas and Mourao 2002). However, these studies focused on nest site limitation as a key factor, while Bahama Parrots have abundant available nest sites (Mori 2006). This allows us to explore factors influencing nest success and breeding productivity as related to social interactions between individuals in adjacent nest sites without the limitation of nesting resources. This work will address a lack of information on nesting group behavior in parrots and lead to improved management of this and other parrot species. Additionally, proposals are underway by the

Bahamas National Trust (BNT) to the Bahamas government to expand protected breeding areas for the parrot. Understanding important components influencing nest distribution will allow the

BNT to make its request based on the results of this research.

The main objective of this study is to determine what influences Bahama Parrots to nest in semi- colonial nesting aggregations. Chapter 2 addresses the distribution of Bahama Parrot nests and assess how habitat influence nesting distribution of Bahama Parrots on Abaco; Chapter

3 tests to what degree relatedness influences nesting aggregations; and Chapter 4 assesses the potential for information sharing among neighbors and behavioral responses to disturbances at neighbors’ nests to examine the potential for benefits of neighbors through shared predators defense; finally, in Chapter 5, I review the taxonomic status of the Abaco population of the

Bahama Parrot, currently a subspecies of Amazona leucocephala, and present the argument that it should be reclassified as a distinct species.

8

Figure 1.1 Cuban Parrot (Amazona leucocephala) distribution and nesting distribution on Abaco and Inagua (points) (figure from Russello et al. 2010). Black dots indicate nests found during 2007 for Great Inagua and 2004 for Great Abaco

9

Figure 1.2 South Abaco distribution of parrot nests and habitat. Light blue line is a minimum convex polygon of the nesting area; black line delineates the Abaco National Park; blue and red stars denote nest sites monitored in 2003 and 2004 respectively

10

CHAPTER 2

GROUP NESTING BY THE BAHAMA PARROT (AMAZONA LEUCOCEPHALA BAHAMENSIS): ASSESSING EFFECTS OF CAVITY DENSITY AND MICROHABITAT

Introduction

Nesting distributions of animals vary from isolated individuals to large colonies of breeding individuals, and can be influenced by environmental conditions or social interactions.

Environmental conditions relevant for nesting include, among other factors, the availability of suitable habitat and food abundance (Beauchamp et al. 1997, Burk and Nol 1998). Distribution based on environmental conditions can result in aggregations proportionate to the distribution of resources (Fretwell 1970). Therefore, aggregated resources can lead to an aggregated distribution of individuals (Newton 1994, Salinas-Melgoza et al. 2009). Social interactions influencing nest site selection can include age-specific behaviors (LaMontagne et al. 2013), heredity (Brown and Brown 2000), information sharing (Brown et al. 2000) or predator defense

(Hamilton 1971). Reproductive life stages are of particular importance in species conservation

(Heppell 2000, O'Brien et al. 2006), and understanding the factors influencing nesting distributions is integral for establishing appropriate and effective management strategies.

By definition, secondary cavity nesting species are dependent on existing cavities, which tend to be a limiting environmental resource (Newton 1994, Cornelius 2008). Therefore, the distribution of secondary cavity nesting species is highly influenced by the distribution of cavity resources. This dependence on available cavities makes it difficult to assess the relative importance of environmental conditions and social interactions on nesting distributions

(Eberhard 2002, Monterrubio-Rico et al. 2006). The availability and distribution of cavities has

11 been shown to affect population size (Newton 1994, Jackson 2004) and life history dynamics, including number of offspring, survival (Saab 2004), and population distribution (Brawn 1988,

Robertson 1990) of secondary cavity nesters.

In contrast to cavity-limited species, secondary cavity nesting species without resource limitations provide an opportunity to evaluate the influence of environmental and behavioral factors on nesting distribution. Parrots (Order Psittaciformes) are known to be flexible in their choice of nesting cavities. Though they have been reported to nest in cavities as varied as termitaria and holes in cliff sides, most use pre-formed cavities in trees (Brightsmith 2005).

Given removal of trees for forestry and through activities associated with chick poaching, nest sites are a limiting factor for breeding parrots (Snyder 2000, Wiley et al. 2004).

Several parrot species have been documented nesting in aggregations including the

Burrowing Parrot (Cyanoliseus patagonus) (Masello 2002), Military Macaw (Ara militaris)

(Bonilla-Ruz et al. 2007), Maroon-fronted Parrot (Rhynchopsitta terrisi) (Salinas-Melgoza et al.

2009), and Lear’s Macaw (Anodorhynchus leari) (deMenezes et al. 2006), all of which nest in

cliff faces. The aggregated nesting distributions of Thick-billed Parrots (Rhynchopsitta

pachyrhyncha) and Lilac-crowned Parrots (Amazona finschi), which are tree cavity nesters, are attributed to social behavior (Monterrubio-Rico et al. 2006, Salinas-Melgoza et al. 2009). The distributions for these species range from pairs having nests in the same tree as conspecifics to being separated by hundreds of meters. However, tree nest cavities apparently are limiting in both Thick-billed and Lilac-crowned Parrots (Monterrubio-Rico et al. 2006, Salinas-Melgoza et al. 2009). In contrast, Bahama Parrots nesting on Abaco Island are secondary cavity nesters but are not cavity-limited because they use underground limestone nest cavities (Snyder et al. 1982), and cavities with dimensions in the range of those considered suitable for parrot nesting are

12 locally abundant and apparently not limiting (Mori 2006). Therefore, this population provides an opportunity to investigate nest site selection relatively free of the resource limitations usually affecting cavity nesting birds.

The limestone cavities in which Bahama Parrots nest on Abaco Island are formed when the weathering of the limestone rock creates epikarst formations (Walker et al. 2008). Nesting area on Abaco is found only in the Caribbean pine (Pinus caribaea) ecosystem of the northern most island in the Bahamas, Abaco Island (Gnam 1991, Stahala 2005). Although the karst system forms the entire 120 mile long Island, only the southernmost 15 miles of pine forest are used by parrots for nesting. At this time, it is not understood why parrots nest exclusively in this part of the island. The pine forest is also the primary food source of the parrots during the breeding season (Gnam 1991). Based on the nesting and food resource information, The Abaco

National Park was established in 1994, setting aside 20,500 acres of south Abaco as protected area for parrot nesting (BNT 2008).

The objectives of this study were (1) to determine the nest distribution pattern of the

Bahama Parrot on Abaco Island within this protected nesting area; (2) to investigate whether nest distribution is driven by the distribution of limestone cavities throughout the landscape; (3) to assess whether parrots select nest cavities with certain ecological features. These results will be useful in management decisions affecting this and other parrot species. In particular, this information will be useful in identifying the important components of nesting areas for this species and determining whether conservation efforts targeting nest areas should focus on current nesting areas or on expanding habitat protection to new regions with suitable nest site characteristics.

13

Methods

Field

To identify the locations of active nests, nest searching was conducted May to August

2009 – 2012, throughout the parrot nesting area which includes the areas between Crossing

Rocks (W 77.19o, N 26.140), Sandy Point (W 77.40, N 26.010), and Hole-in-the-Wall (W 77.20,

N 25.890) on Great Abaco Island, The Bahamas. Searches were conducted using systematic

transects. To ensure complete coverage of area during these searches, transects were conducted

along an existing matrix of east-west logging roads, which are spaced approximately 250 m

apart, with north-south roads every 800 m. Each road (N = 30) was searched twice using this

technique during each summer. Nest searching was conducted during highest parrot activity in

the morning (0545 hours through 1030 hours) and evening (1600 hours through 2000 hours).

Nests were found by walking along roads and looking for signs of parrot nesting behavior. These

behaviors included sentinel behavior (conspicuous perching by a single individual), pair

vocalization, or nest entry (single parrot flying from tree to ground). Additional nests were found

by rechecking nest cavities used in previous years and by opportunistic observations of parrot

activity.

Estimates of the density of suitable limestone cavities were generated by running 2 meter

wide strip transects for 50 meters in three randomly-selected compass directions from 82 active

nests and 51 random sites. To select random sites, researchers followed remnant logging roads

throughout the second growth pine forest for distances chosen from a random number table (1-

800m), then entered the forest for a distance (1-400m) and compass direction (0-360 degrees)

similarly selected at random. We counted the number of suitable cavities in each transect, where

“suitable” cavities were defined as limestone holes with a cavity opening diameter of 8-51 cm

14 and depth of 47–210 cm, which were the ranges of diameter and depth for 112 active nesting cavities measured in this population in 2004 - 2005 (Mori 2006).

Nest density was estimated by determining the number of active nests within a 400m radius of each focal cavity (502,655m2 area). The nest density value was estimated for active

nests as well as sites randomly selected for estimating the density of suitable limestone cavities

mentioned above.

Ecological measurements were collected in the areas around nest cavities (N = 68) and

suitable unused cavities (N=29) that were found during strip transects. The goal of these

measurements was to determine if parrots were using cavities with certain characteristics. The

following measurements were collected: (1) radius of clearing/exposed (unvegetated) rock

surrounding nest; (2) distance from cavity to nearest shrub/plant with stem <0.5cm; (3) plant

density at .5m, 1m, and 5m from the cavity entrance in randomly selected cardinal direction,

using a Robel pole (Robel et al. 1970); (4) distance to closest pine tree; and (5) number of pine

trees within an 11.2 m radius of the nest. Radius of exposed rock and distance to the closest plant

are a measure of openness around the cavity hole. Similarly density of vegetation also confers a

measure of openness but at a greater distance from the cavity. The number of pine trees and

distance to closest pine were collected since pine trees provide a significant food source during

the breeding season. This estimate would indicate a preference to sites with high food

availability around the nest. Each of the five measures was compared between used and unused

suitable cavities.

15

Analysis

Nest distribution was established by using nearest neighbor distances in a Clark-Evans and Ripley’s K distribution model (Clark and Evans 1954). The Clark-Evans model estimates the ratio of average nearest neighbor distance to an expected nearest neighbor distance if nests were distributed uniformly for the entire nesting area. Ripley’s K tests a point distribution against a random distribution over a range of distances (Ripley 1977). Distances between nests were measured using Garmin handheld GPS 60. The distribution model also requires total nesting area which was estimated using a minimum convex polygon of all nests which determines the smallest area that encompasses every nest using acute angles (Mohr 1947,

Casiraghi et al. 2003).

Habitat characteristics (1-5 above) were compared between active nest cavities and unused suitable cavities using the nonparametric Wilcoxon rank sum test due to the non-normal distribution (for all variables Shapiro-Wilk W test W= 0.58-0.94; p<0.015) of the response variables in program JMP (JMP). Assessment of significance was based on an alpha of 0.05.

The relationship between number of active nests and the number of unused suitable cavities around a given nest was analyzed using GLM with a Poisson distribution in Program R

(R_Core_Team 2012).

Comparison of cavity density of occupied and unoccupied cavities was conducted using a

non-parametric Wilcoxon rank sum test due to a non-normal distribution of response variables in

program JMP (Shapiro-Wilks W Test; unoccupied cavities: W=0.81; p=0.015; occupied cavities:

W=0.6; p<0.001). Assessment of significance was based on an alpha of 0.05.

16

Results

Nest distribution within the protected nesting area

A total of 148 unique active nests were monitored between 2009 and 2012, with 46-68

nests monitored in each year. Total minimum convex polygon (MCP) occupied by these nests

was106km2. Nest cavities were frequently reused between years with a mean of 46.3% SE 6.1 of

nests reused in a given year (56, 48, and 35% in 2010, 2011, and 2012 were reused compared

with the previous year). Four banded adult pairs were seen reusing the same nests in two

consecutive years, suggesting that reuse of the same cavity was often by the same breeding pair.

Nests were aggregated throughout the nesting area. The Clark-Evans distribution value

for all nests monitored in 2009-2012 was 0.56 (N = 148) and ranged from 0.55-0.63 (mean 0.58+

0.02 SE) for individual years. Ripley’s K values were based the same range of nests as the

previous analysis and the K(r) values indicate a significant different from random dispersal over

distances of 0 to 15 km (Figure 2.1).

Is nest aggregation driven by the distribution of limestone cavities?

Assessment of the density of occupied and unoccupied cavities, near nests and randomly

selected points indicated that parrots tended to nest in areas occupied by other parrots. Active

nests had higher densities of other active nests within a 400m radius (nests: 4.3 + 0.20 SE, non-

nest cavities had no active nests; Wilcoxon z = -9.98, p<0.0001) (Figure 2.2). The variance was tested between active nests and random cavities for both variables, empty cavity density

(Levene’s test F=14.80 p=0.0002) and density of active nests (Levene’s test F=96.3 p=0.0001).

In both cases variance was higher around nests and lower at random sites.

Within areas occupied by nesting parrots, active nest distributions were related to cavity

17 availability. Unlike randomly selected non-nest cavities, active nests occurred in areas with greater density of suitable but unoccupied nest cavities (nests: 1.79 + 0.36 SE cavities per

300m2, non-nest cavities: 0.5 + 0.17 SE cavities per 300m2; Kruskal-Wallis Rank Sum Test, Z=-

2.73 p=0.006) (Figure 2.3). When analyses were restricted to data gathered from active nest areas, the number of empty cavities increased with increasing nest density (GLM, F1,132=3.27,

P<0.001; Figure 2.4).

Finally, fewer holes were found in the random areas, out of the 51 randomly selected empty cavities surveyed, 11 transected yielded suitable nest cavities. However, no active nests were found in the 400m radius of any of the 51 randomly selected sites. Even though number of holes at random sites was skewed towards zero, in non-zero transects empty cavity values were within the range of active nests with 1-6 empty holes mean 1.69 SE 0.51 for random sites and 1-

19 empty holes and mean of 3.96 SE 0.67 for active sites (t-test, p=0.017). This suggests that parrots prefer areas occupied by other parrots.

Ecological correlates of active nesting

To determine if elements within the environment influence cavity selection, I quantified

biotic and abiotic correlates of nesting activity. Cavities used as nests had more exposed rock

around holes than did unused cavities (Wilcoxon Rank Sums Test, W=613, p=0.006); and

correspondingly the distance to the closest stemmed plant was greater for active nests than for

unused cavities (Wilcoxon Rank Sums Test, W=559.5, p=0.001; Table 2.1). However, exposed

rock was not significantly correlated with distance to nearest stemmed plant with (Pearson R2=

0.046, p=0.09). There were no differences between nest cavities and unused cavities in

vegetation density, distance to closest pine, or density of pine trees (Table 2.1).

18

With the goal of determining if more open cavities were actually new cavities found due to increased visibility, I tested for a difference in distance from the cavity opening to the nearest plant between active nest cavities known to be used in previous years (n=30) and active nest cavities newly found that year (n= 38), assuming newly found cavities are more likely to be newly initiated nests. There was no significant difference in openness between new or old nesting cavities (Figure 2.5).

Discussion

Cavities used for nesting by Bahama Parrots were aggregated in the pine forest of south

Abaco. There can be numerous reasons for aggregated nesting, particularly by a generally

gregarious species (Goodenough et al. 2009). Here my goal was to quantify ecological factors

that may explain this pattern. Specifically, I examined whether nest aggregations were

associated with areas of higher overall limestone hole densities or with ecological variation near

cavities. My results showed nesting sites were found in areas with higher cavity densities, and

that nest cavities were associated with other active nest sites. Two environmental factors were

associated with active nest cavities: exposed rock and distance to closest stemmed plant to the

nest openning. These related measures of obstructions around the cavity opening indicated

active nests were more likely to occur in areas clear of vegetation.

The aggregated nesting of the Bahama Parrot is similar to the distribution of other parrots

(Monterrubio-Rico et al. 2006), however the homogeneous pine rockland system of Abaco

Island, which includes abundant cavities, allows for the evaluation of this distribution in a

system that is not resource limited for nesting sites (Monterrubio-Rico et al. 2006, Mori 2006,

19

BNT 2008). Even with abundant resources, this parrot species appears to distribute itself in an aggregated fashion.

Higher nest density in areas that also have a higher overall density of limestone cavities suggests that these high cavity density areas are important sites for nesting. These high density limestone cavity areas sustain more active nests than other areas. The sampling for the comparison of active nest density and density of cavities was conducted at different landscape scales. Although 300m2 were sampled around each nest site for empty cavities, this was only done to a 50m radius from the nest. However, the number of additional active nests around a given active nest was conducted out to 400m. To do this, I made the assumption that the values of these variables did not change greatly over the 400m radius. The number of active nests for a radius smaller thank 400m would have provided too many zeros making a comparison less meaningful.

High density cavity areas should be considered critical nest sites. Since this parrot is a species of conservation concern, these results suggest specific management directions. These results suggest that hole density around active nest sites can be evaluated as an index of habitat quality for Bahama Parrot nesting. A component of the Bahama Parrot management plan is to consider suitable sites for protection (BNT 2012). If a site is to be considered suitable for parrot nesting on Abaco, the site should have more cavities than would be found at random sites (i.e.

>0.5 cavities/300m2). The higher this value is, potentially the more important the site is for nesting.

Nest sites had more bare rock surrounding the cavity than did unused limestone holes.

This was expected because Bahama Parrots on Abaco remove pine straw and leaf litter from inside and around the nest cavity (O'Brien et al. 2006). One hypothesis for why parrots perform

20 this cleaning behavior is that it makes the cavity less susceptible to fires by removing these flammable materials, but it may also increase visibility of predators around the nest area, or occur as a general ritualistic nest building behavior (Barber et al. 2001, O'Brien et al. 2006). This behavior is not unlike the cleaning out of limestone cavities by Red Grouper (Epinephelus morio), like parrots removing debris from inside and outside of the cavity, the grouper remove sediment from underwater limestone cavities to manipulate the cavity and make it desirable for the occupant (Ellis 2015).

Active nests occurred in more open areas, with greater distances between the nest opening and surrounding vegetation than detected for unoccupied cavities. There was no difference between cavity types in distance to pine trees or the density of pine trees. Although pine trees provide the highest source of food during the nesting season, parrots are also found to aggregate in pine trees and other food plants away from nest areas and so immediate food resources may have little influence on cavity choice. The openness of the cavity area could be an environmental feature that existed prior to parrot use, or openness might be a result of parrots using the area. These parrots remove pine straw and leaves from inside the cavity and cavity opening (CS, personal observation), which indicates at least part of openness is due to parrot use.

On the other hand, parrots may be selecting cavities relatively far from living vegetation because these holes are expected to be more visible and hence detectable by nest-searching parrots at the time a nest is established.

Assessment of the proximity of vegetation to newly discovered vs. previously occupied nests revealed no differences, indicating that parrots either selected nests that were in more open areas or actively removed vegetation from around the site, as well as removing leaf litter as previously observed. Parrots in the study population have natural and introduced predators,

21 including land crabs, hawks and feral cats known to take adults and chicks at nests, and having a more open nest cavity could reduce predation risk by increasing visibility around the nest (Gnam

1991, Gotmark et al. 1995).

These results establish characteristics differentiating Bahama Parrot nest cavities from other suitable limestone cavities in the area. Nests also occur in clusters within Abaco National

Park. My results suggest that clusters of available holes may in part explain this distribution, but also suggest a role of conspecific interactions in nest site selection because randomly selected empty cavities, even in areas of higher cavity density, were found to have no parrot nesting in the vicinity. In other words, active nests were better predictors of other nests in the area than density of available holes. This suggests that as the population grows it will expand from the edges of currently occupied areas. In addition to habitat characteristics, other factors could be influencing the aggregations of nests, as suggested by Salinas-Melgoza et al. (2009). These include conspecific cueing, as reported in Lilac-crowned parrots (Amazona finschi) where the presence of conspecifics indicates territory quality. Additionally, individuals could be attracted to one another for enhanced predator avoidance through group vigilance, or aggregation could represent family clusters if dispersal is limited (Lessells et al. 1994, Gotmark et al. 1995).

These results underscore the importance of the current nesting area in Abaco National

Park, given the high available forested area in the region. Every effort should be made to prevent disturbance to the current nesting area, and particularly to the limestone cavities. Additionally, if the protected area of Abaco National Park is expanded to include new areas, areas with high cavity density and open understory should be considered. Since parrots show a preference for cavities that have plants farther from the opening, efforts should be made to maintain a burn

22 regime which opens up the understory of the pine forest and may make these cavities available to parrots for nesting (O’Brien et al. 2008).

Figure 2.1. Ripley’s K distribution graph of Bahama Parrot nests on Abaco, The Bahamas indicate a clustered pattern of nests at all distances up to 15km. Green and blue colored lines indicate confidence interval for randomly distributed nests. Positive K(r) values above the confidence interval for random distribution indicate a clustered distribution.

23

Figure 2.2. Number of active nest cavities in a four hundred meter radius of nest cavities (N=82) in comparison to random suitable cavities (N= 51) indicates parrots nested near other nesting individuals. Wilcoxon z = -9.98, p<0.0001. Thick horizontal line indicates median, height of box represents inter-quartile range and whiskers are 1.5 inter-quartile ranges, data points beyond this are outliers.

24

Figure 2.3. Nests were found in areas with more available ground cavities, when the number of empty cavities in 300m2 strip transects around nests (N=82) and randomly selected cavities (N=51) was compared to the number of empty cavities found within a 400m radius around that cavity; Wilcoxon z = -2.73, p= 0.006). Thick horizontal line indicates median; length of box is inter-quartile range, and whiskers are include 1.5 inter-quartile ranges.

25

Figure 2.4. Number of active nests increased with the number of empty cavities per meter squared in the nest area (GLM, F1,132=3.27, P<0.001, Spearman’s rho=0.38 p=0.0004). Points were jittered to show individual nests.

26

Figure 2.5. Plants grew nearer to non-nest cavities (N=29) than to either old nests found in previous years (N=30) or active nests used for the first time in a given breeding season (N=38). Wilcoxon rank-sum test z values for new vs old z=-0.19; old vs non-nest z=-2.22; new vs non- nest z=-2.51.

27

Table 2.1. Degree of exposed rock and distance to nearest plant differed between parrot nests (n=68) and unused cavities (n=29). Wilcoxon rank-sum tests were used to compare values of characteristics between used and unused nests. Asterisk (*) indicates significance of p<0.05.

Parameter Nest Cavity Mean Unused Cavity W p-value Mean Exposed Rock radius 25.86 ±SE 3.35 15.75 ±SE 3.28 613 0.006* (cm) Nearest Plant (cm) 26.14 ±SE 2.74 13.18 ±SE 3.27 559.5 0.001*

Robel Vegetation Density 8.43 ±SE 2.08 13.46 ±SE 3.94 1078 0.26 at 0.5m (cm) Robel Vegetation Density 18.54 ±SE 3.23 16.97 ±SE 4.39 905.5 0.75 at 1m (cm) Robel Vegetation Density 53.34 ±SE 3.43 49.58 ±SE 6.10 864 0.51 at 5m (cm) Distance to pine (cm) 245.95 ±SE 16.43 262.0 ±SE 25.08 1009.5 0.60

Number of pines in 44.02 ±SE 3.45 49.35 ±SE 5.26 1063 0.33 400m2

28

CHAPTER 3

INFLUENCE OF RELATEDNESS AND GENETIC POPULATION STRUCTURE ON NEST AGGREGATION OF THE BAHAMA PARROT (AMAZONA LEUCOCEPHALA BAHAMENSIS) ON ABACO ISLAND

Introduction

Group nesting behavior can vary considerably among species of birds and mammals, in

both spatial and social characteristics. Spatial aggregation within groups may vary from the tight-

knit nests at nesting colonies to loose aggregations referred to as semi-colonial nests. The most

extreme form of group nesting is a communal nest where multiple individuals use the same nest

such as in Smooth-billed Anis (Crotophaga ani) (Davis 1940), Gray Squirrels (Sciurus

carolinensis) (Koprowski 1996) and Monk Parakeets (Viani et al 2016). Less extreme colonial

nesting behavior is common throughout the animal kingdom, and aggregations of individual

nests are found in species as diverse as seabirds (Oro 1996), swallows (Brown 1986) and prairie

dogs (Hoogland 1981). Some nesting aggregations are more dispersed, such as the semi-colonial

nesting exhibited by the Montagu’s Harrier (Circus pygarus) (Arroyo et al. 2001). Spatial aggregation of nesting varies and yet even loose aggregations can have impacts on individuals in the group.

The direct benefits of nesting aggregations vary from predator vigilance (Clutton-Brock

2009, Elgar 1989) and thermoregulation (Edelman and Koprowski 2007, Viani et al 2016) to information transfer (Brown 1988). For example, Montagu’s Harrier benefits from the semi- colonial breeding behavior through increases in predator defense. In this species, individuals nesting near each other have higher success in detecting predators in the nesting area, and detected predators are mobbed by more individuals, increasing effectiveness of antipredator

29 response (Arroyo et al. 2001). Similarly, Black-tailed Prairie Dogs (Cynomys ludovicianus), which have dense nesting burrows, detect predators more quickly and scan less per individual than do White-tailed Prairie Dogs (Cynomys leucurus), which have relatively dispersed burrows

(Hoogland 1981). In a situation unrelated to predators, Gray Squirrels increase their communal

nesting behavior during winter months, suggesting a thermoregulatory advantage (Koprowski

1996). Monk parakeets exhibit a similar advantage of thermoregulatory advantages provided by

their communal nest sites (Viani et al. 2016). Brown (1988) demonstrated that Cliff Swallows

(Hirundo pyrrhonata) benefited from nesting in larger colonies by gaining an advantage in finding foraging sites. There are numerous benefits for animals to nest near each other, regardless of distance between them.

In addition to the direct benefits of group nesting behavior conferred to the individual, indirect benefits through kin selection can also be a driver for group nesting behavior (Hamilton

1964a&b; Alexander 1974, Hayes 2000). In some species, close kin assist in the raising of related offspring to pass along genes through common descent, as is found within a nest with helpers in cooperatively breeding taxa. Examples of kin selection favoring interactions between individuals at adjacent nests is rare. Long-tailed Tits (Aegithalos caudatus) nest in small aggregations, and if an individual breeder fails at their own nesting attempt, they can become a helper at a neighboring nest but only if they are related to the breeding pair (Russell and

Hatchwell 2001). In this case, close proximity of nests by related individuals is helpful, but only

if they are unsuccessful at their own nests. Nesting colonies can also be composed of related

individuals without any direct or indirect benefits being conferred. This type of nesting colony

would show genetic population structure without benefits. This is seen in nesting loggerhead sea

turtle colonies (Bowen and Karl 2007). Nesting females return to nest at beaches where they

30 hatched, thus increasing relatedness among the population due to lack of dispersal from nesting area.

Aggregated nesting could also be explained by cryptic kin selection, where a relatedness structure exists in a population particularly when individuals are in close spatial proximity. But overt cooperation, such as helping behavior by adult offspring, may not be detected (Hatchwell

2010). For kin selection to influence group nesting behavior, neighbors would have to be more closely related than expected at random from the population (Hatchwell 2010; Clark et al. 2012), or interactions would have to disproportionately favor neighbors who were relatives.

Genetic population structure, and more specifically genetic relationships among individuals in nesting groups, may be observed at various scales. Sea turtles exhibit increased maternal relatedness on a small regional scale. Females at particular nesting beaches exhibit increased relatedness, but interact with unrelated individuals on a global scale during other parts of their life (Bowen and Karl 2007). Some bird species, such as Acorn Woodpeckers

(Melanerpes formicivorous), are found in highly related groups with offspring cooperating with parents and potentially staying or dispersing at a later time (Koenig and Mumme 1987). In both cases individuals in groups exhibit a spatial relatedness structure, but the presence of kin affects nesting success only in the cooperative breeding groups. This indicates that a spatial relatedness structure can exist in nonadaptive as well as adaptive contexts.

Parrots (Order Psittaciformes) highly social behavior is legendary. Most species form both foraging and roosting flocks (Chapman et al. 1989, Gilardi and Munn 1988, Westcott and

Cockburn 1988, Pizo et al. 1995). One example of group nesting documented in parrots

(Monterrubio-Rico et al. 2006) explored aggregated nesting in Thick-billed Parrot

(Rhynchopsitta pachyrhyncha). In some cases nests were found in the same tree, and the

31 investigation focused on a lack of suitable nest sites as a reason for the behavior. Aggregated nesting may have a social component as well, though this hypothesis has not been investigated

(Monterrubio-Rico et al. 2006). Aggregated nesting could be a result of familiarity with a nest site, tit-for-tat between familiar neighbors, or additional benefits garnered from close neighbors.

When nesting in close proximity, the opportunity for extra pair copulations (EPCs) and therefore extra pair paternity (EPP) may occur even in a socially monogamous species (Petrie and Kempenaers 1998). Parrots are widely assumed to breed monogamously, particularly due to their long lifespan and need for bipaternal care (Masello et al. 2002). Although socially monogamous, investigation of genetic paternity has revealed mixed results in parrots. Studies on aggregated nesting Burrowing Parrots (Cyanoliseus patagonus) found no evidence of EPC

(Masello et al. 2002), however there is some evidence of EPC before mate switching in cockatiels (Spoon et al. 2007). EPP represents an important possible cost to neighbors or benefit to parents of aggregated nesting, and thus this possibility should be addressed since it may affect aggregated nesting behaviors.

The Bahama Parrot (Amazona leucocephala bahamensis) on Abaco Island nests in aggregations in the southern portion of the island’s Caribbean Pine forest habitat (Chapter I,

Snyder et al. 1982, Gnam and Rockwell 1991). Previous work demonstrated that habitat plays a role in explaining the aggregated nesting behavior of the Bahama Parrots on Abaco (Chapter I).

A relationship between suitable limestone hole density and active nests indicated that more parrot nests are found in areas that have more holes. However, this relationship was found only in transects from active nests; transects from vacant but nest-suitable cavities showed no relationship between hole density and occupancy. Many areas of the field site contained high numbers of suitable cavities but no nesting parrots. This suggests that social effects also

32 influence nesting behavior in this species. A component of presence of other parrots seems to be necessary for other individuals to nest in the area.

Molecular studies have already identified population substructure between the Abaco

Island population of the Bahama Parrot and a second population found on Great Inagua (Russello et al. 2010), but it is unknown if there is fine scale substructure within the Abaco population.

Identifying population substructure, whether it exists as spatial grouping of family members or as subpopulations within one larger population, is important for the management of this species. As in most species of conservation concern, maintaining genetic diversity is a goal for this species.

If population substructure is found, future management decisions would need to consider this pattern, for example protecting nesting areas with different genetic structures equally (Lande

1988). Such decisions would include protecting habitat for groups rather than individuals, basing conservation decisions on number of clusters versus number of individuals, or protecting entire population to maintain genetic diversity by preventing inbreeding through the focus on just one site. If this population were to become a source population to re-establish extirpated populations, understanding the substructure of the source population would also need to be considered (Miller et al. 1999). Translocation decisions currently prioritize maintenance of genetic diversity in new populations, and so knowledge of the genetic make-up of the source

(Abaco) population is needed to inform decisions on where individuals are collected to provide the new founder population with the greatest possible genetic diversity. However, genetic structure of the donor population would only be one considering for translocation; variation in age structure, existing pairs or individuals of the founder population also needs to be considered.

Social interactions between individuals and unnoticed by researchers are not uncommon

and therefore exploring reasons for aggregated nesting behavior beyond resource limitation is a

33 necessary step in understanding the causes and consequences of this behavior (Hatchwell 2010 and Williams et al. 2013). Here I (1) test a key prediction of the hypothesis that nesting aggregations are influenced by kin selection, specifically investigating whether there is a relationship between nest distance and genetic relatedness; (2) determine whether extra pair

paternity is a potential effect of aggregated nesting in this species by first assessing the

occurrence of EPP and, if detected, whether EPP is influenced by proximity to neighbors; and

(3) investigate whether there are genetically distinct subpopulations found in the Abaco nesting

population, a factor which informs future management decisions. This study addresses genetic

and relatedness components of social nesting to explore forces beyond habitat and nest site

limitation that may influence aggregated nesting in this population.

Methods

Field methods

To identify, genetically sample and identify sex of nesting parrots, nest searching was conducted May to August in 2009 – 2012, throughout the parrot nesting area including Crossing

Rocks (W 77.19o, N 26.140) and Sandy Point (W 77.40, N 26.010) to Hole-in-the-Wall (W 77.20,

N 25.890) on Great Abaco Island, The Bahamas. Searches were conducted using systematic

transects to ensure complete coverage of nesting area and rechecking nest cavities used in

previous years to increase sample size. Active nests were monitored on a weekly basis. Ledges

were found in some of the underground nest cavities which allowed chicks to hide and made

them unavailable for capture, and so 51 nests were sampled for genetic material out of the 122

active nests monitored throughout the nesting area. As many adults, chicks, and different nest

locations as possible were sampled based on feasibility of removing birds from nest.

34

Adults were captured when they entered the nesting cavity to feed chicks. Samples from both adults and nestlings were collected only after chicks in the nest were 5 weeks of age or older in order to minimize impacts to chicks. A maximum of .1mL of blood was collected from the brachial vein. All samples were collected in accordance with local, US and Bahamian regulations. Blood was stored in 2 mL cryogenic tubes with Queens lysis buffer solution (10 mM

Tris, 10 mM EDTA, 1% N-laurlsarcosine, 10 mM NaCl). Samples were stored at room temperature in field and frozen once in the lab. All sampled birds were uniquely banded with a combination of 12 gauge aluminum anodized colored band (The Ring Lord, Saskatoon, Canada) and Size 12 aluminum closed or stainless steel open numbered legs bands (L&M bird leg bands

San Bernardinao CA). Once a parrot was banded, it was identifiable by observation from a

distance to determine nest reuse by paired adults. GPS locations were recorded for all nests and

used to estimate distances between nests. All nests had documented latitude and longitude

recorded with Garmin handheld GPS 60 units.

Molecular Analysis

Genomic DNA was extracted from blood using EZNA DNA Tissue kit from Omega Bio-

Tek, Norcross GA. PCR reactions were conducted on an Applied Biosystems Veriti 7500 Fast

Thermocycler in 10L reactions containing 6-20 ng DNA, 5L Qiagen multiplex master mix

(Qiagen Inc., Valencia, CA), and 1L primer mix of 3-4 primer pairs (in 0.03-0.06 µM solution).

In addition to a sexing locus using the P2/P8 primer pair (Griffiths et al. 1998), individuals were genotyped at nine microsatellite loci previously identified as variable for this subspecies

(Russello et al. 2010). Cycling conditions for all loci were optimized to 95C for 15 min; 30

cycles of 95C for 30 sec, followed by 59C to 51C touchdown for 90 sec, and 72C for 60 sec;

35 and final step of 72C for 30min. PCR products were resolved on an Applied Biosystems 3730

Genetic Analyzer and analyzed with Genemapper 3.7 from Applied Biosystems (Foster City,

CA).

Program SPAGeDi 1.4 (Hardy and Vekemans 2002) was used to determine Queller and

Goodnight (1989) relatedness coefficients between individuals and FST (Weir and Cockerham

1984) values between groups. A relationship between nest distances and relatedness was

explored with an ANOVA fitting a linear and cubic model using program R. The linear model

was used to explore the relationship in case chicks were more related to one another close to nest

sites. A cubic relationship was chosen in case there is some natal dispersal but still within a

general vicinity of a parrots fledging site. Only one chick per nest and one nesting attempt (if

nest was sampled in more than one year) were selected at random to represent a nest cavity, if

multiple samples were available.

To determine if relatedness correlated with nesting in a general proximity range rather

than at a specific distance, three proximity categories were created for each focal nest: Closest

Nest <1km away (CN); and Mid-distance Nest 1-5km of neighboring nests (MN); and Far Nest

greater than 5km from the focal nest (FN). Relatedness values were then determined for each

pairwise comparison (Focal/CN; Focal/MN; Focal/FN) Relatedness values were then compared

between categories in an ANOVA (Figure 3.4).

Thirty-seven sampled chicks from 76 sampled nests (including same nests over multiple

years) also had an adult male and/or female sampled from the same nest, allowing limited

parentage analyses. Adults caught in a nest were considered the apparent parents of the chicks in

that nest, later confirmed as being the individuals that returned to care for chicks.

36

Parrots are difficult to capture safely and sample for DNA. In some instances I did sample entire family groups (adults and chicks) but sampling of adults and chicks may not have occurred in the same year. However, in 37 cases the chicks but not the parents at a nest were sampled, and parents attending that same cavity were captured and sampled in a subsequent year.

I assumed adult parrots reused nest cavities over multiple years; therefore, I had to use multiple lines of evidence of paternity to deal with limited information. In situations where I had two parents and chicks I tested paternity by exclusion. To test the validity of the assumption that the same parents reused the cavity between years, I observed banded parents over multiple years to verify same individuals using a nest. Finally, I compared relatedness values of chicks from same nest cavity over multiple years to the apparent parents from the nest.

Program CERVUS (Marshall et al. 1998) was used to determine paternity of a given chick at a nest location using a known mother (when available) and a candidate father. I considered the male at each nest site as the candidate father, and assessed whether he was a genotypic match with the chicks, given the known maternal genotype. A mother was assigned to

27 of the 37 chicks at sampled nest location. When a female was available, she was assigned as

the mother of a chick, the male was then tested for possible paternity or potential match for paternity (Appendix A). Mothers-chick pairs were checked for allele mismatches prior to assignment of maternity. Mismatches of mother/chick father/chick were regenotyped to verify allele identity.

To quantify genetic substructure in the Abaco parrot population, Fst and degree of significance was determined using Program SPAGeDi 1.4 (Hardy and Vekemans 2002) and used to compare between the extreme ranges of the parrot population. Based on all the nests sampled

(Figure 3.3) one individual from the 13 northern most nests and 8 southern most nests (Figure

37

3.4) was selected and compared for subpopulation structure due to the suggestion that the birds at these population extremes exhibited differences in vocalization (Gnam 1991).

Results

Twenty one adult parrots and 133 chicks were sampled for DNA and genotyped from the nine typed loci. The 9 loci each had between 2-5 alleles (average 3.6 SE0.41 alleles/locus) (Table

3.1). Of the adults sampled, 6 pairs were sampled as parental pairs at a nest site (Appendix B).

Thirty-six nests had more than one individual sampled from the same nest cavity. This includes adults and chicks in a nest, in some instances. Thirteen nests were family groups with at least one chick and one adult sampled. This allowed me to assess paternity of sampled chicks in the nest and relatedness of chicks over multiple years to determine if the same parents used the same nest cavity over multiple years.

Are individuals at close nests more likely to be related?

An analysis of relatedness between chicks in nests at various distances was conducted to

determine if neighbors nesting more closely were more likely to be related. One randomly

selected chick from each nest cavity was used to be included in the analysis and compared to a

randomly selected nest. This analysis considered 47 individual comparisons of chicks from 47

nests, and detected no relationship between relatedness and nest distances ANOVA F(1, 45)=0.35, p>0.56 (Figure 3.1). The cubic polynomial fit used to determine if relatedness increased at an intermediate distance also showed no relationship (ANOVA F(3, 46)=1.85, p>0.15) (Figure 3.1).

Chicks from different distance categories showed no difference in relatedness (R) to one

another. Relatedness of chicks in the focal nest to the closest known nest (CN) to averaged R of

38

0.14, while R to chicks in mid distance nests (MN) averaged 0.19, and R to chicks far from focal nest (DN) averaged 0.17 ANOVA F(2,125) = 1.26, p>0.28 (Figure 3.2).

Assessing genetic subpopulations

Comparison of nests from the northern and southernmost extremes of the population resulted in an Fst of 0.03 (p=0.25; Figure 3.4). This value indicates that there is little genetic differentiation between indviduals from the extremes of this population (Wright 1978).

Evidence of EPP

No evidence of extra-pair paternity was detected in the Abaco parrot population. All 5 nests for which at least one chick and two parents were sampled in the same year had matching alleles (Appendix A).

Data over multiple years was available for several nests, therefore it was possible to establish the parentage of chicks over multiple years in a given nest. Longitudinal information on nest cavity use and reuse by breeding adults was obtained from three nests (0405, 0408 and

1101) during this study. For these nests, both parents were sampled during the study, and chicks were produced and sampled in the nest cavity over multiple years. The chicks from nest cavities

405 and 1101 matched the parents captured in the first year of monitoring over multiple years.

The parents of nest 408 were captured and sampled in 2011. The offspring in nest 408 were also captured and sampled in 2011. The results of the parentage analysis support the assumption that apparent parents were actually the parents of the offspring for that year.

Comparisons of adults sampled in one year to chicks hatched in the same cavity in previous years revealed mixed results of nest reuse by the same adults. The offspring in cavity

408 in previous years (2009 and 2010) were also captured and sampled. Comparison of

39 genotypes across years supports the male sampled in 2011 as the father of the offspring from previous years, however, the analysis does not support the 2011 female as being the mother of the 2009 and 2010 offspring (N=4 chicks). This suggests the same male bred in that cavity in all three years, but changed mates.

However, most of the evidence supports both of the individuals in a breeding pair returning to a nest cavity over multiple years. Multi-year nest site fidelity by both members of the breeding pair was provided by nest 405, 1006, 1141, 1206 and 567. In addition to DNA sampling, the adults in nest 405 were banded in 2010. These adults were monitored during the

2011 and 2012 breeding seasons and found to be reusing the same nest cavity (405). Both adults were captured and banded for nests 1006 and 1141 however I could not access the chicks in the nest cavity and therefore only the adults were captured. Nests 1206 and 567 had only one adult sampled and banded and at least one chick (Appendix II). In all four of these nests the banded adults were seen again the following year at the same nest cavity, in the case of 1141, the banded adult was seen again the following 2 years at the same nest cavity. Most of our evidence points to the same parrot pair returning to a nest site over multiple years.

Discussion

Nest aggregations do not appear to be associated with genetic relationships in the

Bahama Parrot on Abaco. Close-nesting neighbors are not any more closely related to one

another than distant neighbors. If benefits from nesting near kin or a lack of dispersal from natal

nest were responsible for observed nest aggregations, I would have expected a relationship

between nest distances and the relatedness of attending parrots. Furthermore, I found no

evidence for genetic population substructure among nesting parrots, despite the expectation that

such differentiation would exist among the tested clusters, based on previous documentation of

40 vocal differences (Gnam 1991). This suggests nest site selection is not dependent on natal nest site and the nesting Abaco parrot population should be treated as one population for conservation purposes.

The allelic diversity within this population was estimated at an average of 3.6 alleles per locus over the nine loci which were analyzed. This value is comparable to the 3.2 alleles per locus over the same nine loci in Russello et al. (2010). Considering the low effective population size of 70 individuals (Russello et al. 2010), this low diversity is expected. This low diversity could impact the paternity analysis since the probability of sharing the same alleles with other individuals in the population by chance is higher than would be expected with a higher allelic diversity. As the population increases this allelic diversity value can be monitored over time to see if the genetic population diversity increases along with the demographic increases (Schwartz et al. 2007).

Social and genetic monogamy is expected in parrot species due to their long life spans and social nature (Masello et al. 2002). Extra pair paternity in the Abaco parrot population was not detected, though the limited sample size of parentage comparisons prevents extensive generalization from this finding. Paternity, and therefore mate fidelity, was first analyzed for apparent fathers and chicks sampled in the same year. All chicks analyzed matched the apparent father who was also sampled in the same year. Because we know breeding adults tend to return to the same nest cavity year after year, chicks from the same nest cavity but different years were also analyzed and also found to match the apparent father from the nest. Although pairs may be monogamous over a single nesting season, mate fidelity over seasons was also addressed. The data from one nest (408) showed that mate fidelity is not certain over years, as males do bring new females to previous year’s nest site. However, the fate of the previous female from this nest

41 was unknown. The previous female may have died rather than been ‘replaced’ while still available as a potential mate.

The previous example from nest 408 suggests that the nest cavity may ‘belong’ to the male as this particular male returned to the same nest cavity with a new female, and patterns of cavity reuse across years is a potential direction for further behavioral research. However, observations of band combinations from parrots at the seven other nests for which cross-year data was available supports nest site fidelity by parrots. Nest site fidelity varies for different parrot species (2%-74% reuse) (Berunsky and Reboreda 2009). Although an exact value of reuse could not be determined for Bahama Parrots, all 13 banded individuals that were monitored in multiple years reused their nest cavities indicating this species has a high nest reuse rate. In other species, nest site reuse tends to be associated with limited nesting resources (Aitken et al. 2002), but nest sites do not seem to be limited for the Abaco population. Since nest cavity reuse is high it would seem plausible that parrots at a given nest site have the ability to develop long-term social associations with their neighbors, regardless of whether they are genetic relatives. This social association between nesting neighbors is explored in Chapter 4.

The lack of relatedness between close-nesting individuals and low population difference between the north and south portion of the nesting suggests the Abaco parrot population can be treated as a single population without geographic substructure for management purposes. This means that parrots from any single geographic nesting region are not any more genetically important than parrots nesting in other regions, and so suggests that nesting area protection should be afforded to all areas equally with regard to genetic structure of parrots in the area.

Prior to this study, it was thought that vocalization differences between the north and south ends of the nesting population indicated a potential split in the nesting population. However, this does

42 not appear to be the case. If nest site prioritization needs to occur, it should be based on environmental conditions such as the density of nest cavities and currently occupied habitat

(Chapter 1) rather than the local population’s genetic structure. Additionally, efforts should be made to protect used nest cavities as there is high site fidelity and it may attract additional nest as the population increases.

Figure 3.1. The distance between nests is not related to pairwise relatedness among chicks. Linear (ANOVA F(1, 45)=0.35, p>0.56) and cubic (ANOVA F(3, 46)=1.85, p>0.15) fits were used with only one chick per nest (47 pairwise comparisons from 47 different nest cavities).

43

Figure 3.2. A comparison of relatedness between 3 distance categories Closest Nest (CN); Farthest nest (greater than 5km, FN); and Mid-distance nest (1-5km, MN) of neighboring nests (N=150 total nest pairs) shows no difference between categories.

44

Figure 3.3 Map of southern Abaco Island in, The Bahamas, showing the distribution of 76 nests where at least one individual was sampled for DNA.

45

Figure 3.4. Map of nest locations used to compare northern (N=13) and southern (N=8) clusters within the nesting range for subpopulation structure. Fst =0 .03 (p=0.25), indicates no substructure.

46

Table 3.1 Characterization of population-wide variation at microsatellite loci genotyped for the Abaco population of the Bahama Parrot (N=154), observed heterozygosity (Ho) and expected heterozygosity (He) (Russello et al. 2001, Russello, 2008).

Locus #of alleles per locus Ho HE AgGT4 2 0.066 0.064 AgGT17 2 0.124 0.128 AgGT19 5 0.702 0.635 AgGT21 4 0.583 0.599 AgGT22 3 0.638 0.516 AgGT42 3 0.523 0.498 AgGT72 4 0.664 0.665 AgGT83 5 0.787 0.724 AgGT90 5 0.546 0.534

47

CHAPTER4

INTERACTIONS BETWEEN NEIGHBORS AND THE INFLUENCE OF PREDATORS ON BAHAMA PARROT NESTING

Introduction

Antipredator defense is a major driver of aggregation behaviors in many species.

Antipredator benefits may be derived from general dilution of risk, in which an individual’s risk of predation is reduced because risk is spread over more individuals in the group (Turner and

Pitcher 1986, Krause and Ruxton 2002). Further reduction of risk can be obtained based on position within a group. For example, the individuals at the center of a group are at lower risk of predation than those on the edge. (Hamilton 1971). Vigilance for predators can also be shared among group members, and therefore groups may exhibit increased overall vigilance even when each individual can decrease the amount of time they spend on being vigilant. This may allow individuals to conserve energy or spend more time on other tasks (Kenward 1978, Bertram 1980,

Roberts 1996). Finally, an individual in an aggregation can increase their likelihood of escaping predator attacks through predator confusion (Neil and Cullen 1974) and increased success of antipredator responses such as mobbing behaviors that are performed by more individuals in groups (Shields 1984).

Decreased predation risk or decreased time spent in antipredator behaviors while in social groups has been reported for species as diverse as schooling fish (Neil and Cullen 1974), prairie dogs (Hoogland 1981), and Siberian Jays (Perisoreus infaustus) (Griesser and Nystrand 2009).

Group antipredator behavior is found not only in large, mobile aggregations, but can also be seen in aggregations of more sedentary breeding individuals or nest sites. Antipredator benefits have been reported in nesting aggregations of Fieldfares (Turdus pilaris) (Andersson and Wicklund

48

1978), Least Flycatchers (Empidonax minimus) (Perry and Andersen 2003), Montagu's Harriers

(Circus pygargus) (Arroyo et al. 2001), European Starlings (Sturnus vulgaris) (Zorrato, Santucci, and Alleva 2009) and swallows (Hirundo spp.) (Moller 1987, Brown et al. 2000). For example,

Least Flycatchers nesting in clusters, particularly those at the center of clusters, experienced lower predation rates compared to those nesting in lower densities (Perry and Andersen 2003).

This lower predation rate was attributed to the higher rate of vocalization or alarm calls at higher nest densities, which draws attention to approaching predators. Fieldfares exhibited similar patterns of increased protection of colonies that included more individuals. However, in this case, mobbing attacks may be the cause for lower predation rates (Andersson and Wicklund

1978).

A strength of colonial species is the ability to use their numbers in mobbing behaviors.

Mobbing behaviors aren’t always physical attacks they can also be in the form of mobbing calls.

Montague’s Harriers have been documented exhibiting a semi-colonial breeding strategy in some instances. As more harriers were associated with a colony the rate of predator detection also increased. Additionally, more individuals were recruited into colonies to mob predators with alarm calls. The colonial nests also exhibited lower predation rates than the more isolated harrier nests (Arroyo et al. 2001). Alarm calls have been shown to include information regarding the predator (Seyfarth et al. 1980, Templeton et al. 2005). For example, Black-capped Chickadees produce various acoustic features that indicate the size of the predator (Templeton et al. 2005).

Thus hearing calls of conspecifics provides individuals with information on predators in the area.

Information Center Hypothesis (ICH) is another behavioral explanation for animal aggregations (Ward and Zahavi 1973). The ICH proposes that aggregations are a place for information to be exchanged between individuals. Interacting with other individuals or groups

49 can provide information that is not locally available (Hellman and Hamilton 2014). In the case of nesting aggregations, this is information provided by neighboring pairs at nest sites in an aggregation. In contrast, the ‘calling for help’ hypothesis for why animals breed in aggregations reverses the idea of obtaining info from neighbors but actually suggests focal nests send information to neighbors to elicit antipredator responses from these neighbors (Rohwer et al.

1976; Hurd 1996; Grim 2008). For example, Blackcaps (Sylvia atricapilla) call from nest sites when a predator is detected. Neighbors from surrounding nests will mob the predator with alarm calls and reduce predation of defended nests (Grim 2008). Thus regardless if information on a predator is sent or received, both hypotheses propose an advantage to nesting individuals in having neighbors close by.

Many of the previously mentioned hypotheses for group behavior revolve around predator defense. However predators, particularly non-native predators, can overwhelm native animal population regardless of the behavioral defenses of the native species (Buxton 2014). In these circumstances management measures need to be implemented to protect native species

(Harding et al. 2001). Eradication measures have been implemented for island species such as

Laysan Albatross (Phoebastria immutabilis), Wedge-tailed Shearwaters (Puffinus pacificus) and

Kaka (Nestor meridionalis) (Young et al. 2013, Moorhouse et al. 2003). Eradication is not

always feasible due to the inability to remove all target individuals without harming other species

or due to other logistic issues. In these cases long term predator management efforts are

implemented to manage predators (Harding et al. 2001). Such a program has been implemented

for the Bahama Parrot on Abaco Island since 2005.

The Bahama Parrot (Amazona leucocephala bahamensis) on Abaco Island nests in

aggregations in South Abaco (See Chapter 1). This population nests in underground limestone

50 solution cavities, making nests and nesting adults particularly vulnerable to predation as cats can and do enter these ground cavities to kill adults and young. Previous research (Stahala 2005) has identified the adult breeding stage as the target time period for parrot population management.

The loss of breeding adults has a twofold impact; not only are the chicks and breeding adults lost for that year but the reproductive potential is also lost. A population viability analysis was conducted by Stahala (2005) and the results showed that with the current demographic values the population was decreasing. A model revealed that mitigation of adult predation by introduced predators during the nesting period had the highest potential to reverse the current population decrease. Feral cats have been documented through cameras and fur left over inside the nest as being the main predators of nesting parrots. Predation by feral cats is a particular problem for this parrot population’s viability, which prompted the Bahamas National Trust (BNT 2012) to implement a predator control program each breeding season starting in 2008 (Stahala 2005, BNT

2008). The effectiveness of the predator control has yet to be evaluated nor have the interactions among social groups in the presence of predators.

The primary objective of this research is to investigate the role of predator defense in the nesting aggregations of Bahama Parrots. Specifically, I will test the hypothesis that neighboring nests act as information centers by testing four key predictions: that (1) there is higher synchrony in vocalization when nests are closer to each other; (2) neighbors vocalize more frequently when predators are present at a focal nest. suggesting mobbing alarm calls; (3) neighbors exhibit

mobbing behavior by approaching focal nests more frequently when predators are present; and

(4) neighbor proximity affects nest success. Finally, I will investigate the importance of ground

predators for Bahama Parrot nesting success by examining whether a government-directed

control program targeting feral cats resulted in higher nesting success.

51

Methods

Nest Searching and Monitoring

Nest searching was conducted May through August 2010 - 2013 throughout the Bahama

Parrot nesting area on Abaco Island, which extends from Crossing Rocks (W 77.19o, N 26.140),

Sandy Point (W 77.40, N 26.010) and Hole-in-the-Wall (W 77.20, N 25.890) on Great Abaco

Island, The Bahamas. Searches were conducted using systematic random transects to ensure

complete coverage of nesting area and rechecking nest cavities used in previous years. Active

nests were monitored on a weekly basis to determine status of adults, eggs, chicks, and fate of

nest. Additional nest success data from 2003-2005 were collected during a previous study using

the same monitoring and nest protocol as this study (Stahala 2005).

Nearest Neighbor Determination

The nearest neighbors of a nesting pair were determined by monitoring the focal nest site

over a four-hour period between 6-10 am or 4-8pm, when activity at nests was highest, to

identify other nesting pairs in the area. Parrots regularly vocalize during these times at nest sites

and away from nest sites. An active nearest neighbor nest was identified when an individual or

both individuals from a pair flew to the ground to enter the nest. A search of the site was then

conducted to locate the nest.

Quantifying Vocalization and Behavioral Observations

Data were collected to calculate behavioral and vocalization time budgets while broods

were in nest cavities, July through September of 2012 and 2013. Field staff collected

vocalization and behavioral data from 16:00 – 20:00 (hereafter, the observation period) at

52 approximately 50m from focal nest, nearest neighboring nest (nearest neighbor or NN), random nest (RN) and a distant nest (DN) (<1000m which is out of vocalization range) (Gnam 1991). A random nest was an active nest in the breeding population randomly selected to be monitored for vocalization during the same time period as the NN and DN as a control for vocal synchrony without the distance component. If neighbors are transmitting information I would expect

NN/Focal nests should be synchronized and others not; I expect all comparisons to be synchronized if all parrots return to nest during the same time period. Focal nests were selected randomly without replacement based on numbers from a random number table to prevent biased selection of easily accessible nests. Vocalization data included every call or ‘squawk’ from either individual in the pair at the nest site during each minute of the observation. Time budgets logged behavior (vocalizing, flying away, flying towards neighbor, stationary) of the breeding pair at focal, nearest neighbor and distant nest during each minute during the observation period.

Vocalization was also documented at one minute time intervals to later determine synchrony of vocalization between Focal/NN, Focal/RN and Focal/DN pairs.

Predator Simulation

To determine if neighbors also responded or ‘alarm called’ when a focal nest was approached by a predator, simultaneous observations were conducted at focal and NN. Parrots perceive humans as predators and respond, at least temporarily, with behaviors associated with predators, including vocalizations and mobbing behavior (Burger & Gochfeld 2003), and so human approach at the focal nest was used to stimulate antipredator response. Prior to approaching the nest, researchers sat in a covered area approx. 50 m from nests to prevent disturbance; non-focal nests were observed from similar areas and distances. Once a

53 predetermined time was reached during the observation period, the researcher stood up and began walking towards the focal nest. Synchronized timepieces were used to keep track of time for time budget data and approach time at the focal and comparison nests, as distant nests were well out of sight and earshot of the focal nest. Researcher walked up to the nest, recorded behavior in time budget (see above) of Focal and NN for that time interval and left the nesting area. The entire procedure lasted approximately 45 – 150 min depending on approach time with the approach to nest lasting approximately 1min.

Predator Removal Data

In response to low nesting adult survival rates and low productivity of the Bahama Parrot, the Bahamas National Trust (BNT) initiated a predator removal program in 2005 and 2010-2012

(Stahala 2005). The population size in 2004 was estimated between 1578 to 2600 parrots on

Great Abaco Island. Breeding productivity ranged from 0.8 – 1.2 chicks per nest, but a population viability analysis (PVA) revealed that a minimum of 1.4 chicks per nest for the annual breeding season was needed for population persistence (Stahala 2005). During predation events by cats, along with eggs and/or chicks, adult parrots were also being killed inside the nest cavity. The estimated adult survival for 2004 was 87%. Data collected during the nesting 2003-

1005 nesting period, Gnam’s research (1991), and the PVA indicate the primary cause of low nest productivity values and low adult survival during the nesting period were primarily due to the risk of predation by feral cats. During the ongoing predator control effort, a BNT warden was responsible for removing cats from the Bahama Parrot nesting area. These management efforts coincided with the current behavioral study, thus removal data was shared by the BNT with

54 researchers in this study to assess impact of this intervention. The information received included the number of cats removed from nesting area on pooled here to annual figures for analysis.

Data Analysis

To test the prediction that the occurrence of vocalizations was synchronized among neighbors, the time budget data was analyzed. Synchrony was determined by comparing vocalization at focal, NN, and DN nests at 1 minute time intervals during behavioral observations. Vocalizations were considered synchronous if parrots at different nests vocalized during the same minute within the time budget. A time interval was assigned either a synchronous or nonsynchronous value for vocalization. The number of synchronous vocalizations was calculated between a Focal nest and its 3 neighbor categories (NN, RN and

DN), for each time budget conducted. The mean number of synchronous minute time periods of calling between Focal and neighbor categories was then compared using a nonparametric

Wilcoxon rank sum test due to a non-normal distribution of synchrony values (Shapiro-Wilks W

Test p<0.0001).

I next tested the predictions that neighbors vocalize more frequently when predators are present at focal nests, and that neighbors exhibit mobbing behavior by approaching focal nests more frequently when predators are present through a predator simulation. Behavior

(vocalization, flight from nest, flight towards focal , no response) during the predator simulation was compared to the control sampling period, which was a minute interval randomly chosen from time budgets recorded within 30 min prior to the approach treatment. I examined whether neighbor behavior (vocalization or approach towards nest) was related to circumstances at the focal nest using a generalized linear mixed model (GLMM) with a binomial error distribution

55 and logit link in package lme4. The factors included treatment (simulated predator approach vs. control), and nest type (focal vs NN nest and their interaction). ‘Focal nest’ and ‘group’ was included as a random effect in the model. Some of the focal nests were monitored up to 3 times during the research period (2012 and 2013 nest seasons), and previous results suggest that the same individuals attend nest cavities for multiple broods (Chapter 3). The random effect ‘group’ consists of the focal and nearest neighbor pair in the analysis being analyzed during control and treatment periods.

To test whether the government control program to remove introduced predators had an influence on nesting success, and whether the effects of this program were mediated by parrots’ interactions with neighbors, I tested for differences in nest success across active nests considering effects of proximity to other nests, fate of nearest neighbor, predator control in previous year, and management technique in R v. 3.2.2, using a generalized linear mixed model

(GLMM) with a binomial error distribution and logit link in package lme4. The response variable in the model was a binary success variable based on whether a nest fledged or failed. I chose to use a binary variable of successful/unsuccessful since those where the primary outcomes I was interested in evaluating, however for a more nuanced model, a logistic regression could have been used to determine a probability of success on a continuum (Shaffer 2004). The model included fixed effects of nearest neighbor distance, whether a predator control program was in effect in the nest year, and fate of closest neighboring nest. Data on the number of prior predator control years and number of cats removed was also collected, however these values were highly correlated with predator control program currently in effect (Tabachnick & Fidell 1996). Year

(N=6) and nest ID (164) were included as random effects in the model.

56

I used a variation of the Mayfield method to estimate a breeding period survival for tending adults instead of nest survival (Mayfield 1961). In a standard Mayfield analysis, a fate is assigned to each nest in the found along with the number of days observed, or exposure days, until it reached this fate. By analyzing nest fate using exposure days, results remove the bias of not including nests which have already been lost, therefore unavailable to be included in the sample. In my analysis I was interested in adult survival estimates while they were incubating and feeding chicks in the nest. To estimate adult survival, I used the same Mayfield approach but instead of calculating exposure days of a nest, I calculated exposure days of adults in the nest

(Heisey and Fuller 1985). A successful nesting adult (instead of a successful nest) was defined as not having an adult killed by a predator inside the nest. The Mayfield approach was used to remain consistent with previously analyzed data using this method (Stahala 2005). Also, since the primary results were survival estimates, this approach is appropriate (Lloyd and Tewksbury

2007). Information on which year predator control was conducted in the parrot nesting area was obtained from the Bahamas National Trust, the agency responsible for management of the

Bahama Parrot (Table 4.1).

Results

Data on vocal synchrony and predator response were collected for a total of 30 focal nests with associated group time budgets (i.e. simultaneously monitored NN/DN/RN nests).

Additional nests had time budget and predator simulations conducted without information on distant or random nests. I found no support for an influence of proximity between nests on the vocal synchrony at those nests, as neighbors nearest to focal nests did not have levels of synchronous vocalization significantly different from those of distant or random neighbors

57 compared to the focal nest (Figure 4.1; Wilcoxon Rank Sum Test; NN vs DN Z = 0.16 p = 0.16;

RN vs DN Z=0.275 p=78; RN vs NN Z=-1.31 p=0.18

Predator simulation trials revealed that parrots at both focal and neighboring nests increased their vocalization rates when a predator was present (GLMM, Z= -2.85 p<0.005, N=

153 individual observations representing 75 controls and 78 predator simulation treatments), but neighbors responded at a significantly lower rate than the focal nest being approached (GLMM,

Z=-2.66, p<0.01, N=153 representing 76 Focal nests and 77 NN) (Figure 4.2 & Table 4.2).

I tested whether neighbors participated in nest defense by examining whether they approached a focal nest more frequently than at random when a predator was near , but found no difference in approach rates (GLMM, Z=-0.075, p=0.94, N=78 observations) (Table 4.2).

However, neighbors only rarely approached focal nests, and the only time such an approach was documented was when a predator treatment was being conducted (n=3 flights towards focal nest). Parrots from focal nests were documented flying closer to nest cavity when simulated predator approached the nest site while calling loudly. This response indicated that focal parrots seemed to consider at researcher approaching the nest as a potential predator.

Across all samples (N=216 observations of 164 nests over 6 years) the likelihood of nest success was higher during predator control years (GLMM, Z= 3.05, p<0.002, N=216) and when nearest neighbors were also successful (GLMM, Z= 3.36, p>0.0007, N=216). There was no support for proximity of neighbors influencing nesting success (GLMM, Z=0.76, p =0.45,

N=216). The number of cats removed and years prior predator control were removed from the model due to high correlation between these variables and the predator control variable

(correlation of fixed effects ‘years prior predator control year’ and ‘predator control’ = -0.821;

‘number of cats removed’ and ‘predator control year’ = -0.713 ) (Table 4.3).

58

Discussion

I found no support for the hypothesis that nesting aggregations of the Bahama Parrot act as information centers or that nest aggregations provide predator defense benefits to close neighbors. Although the presence of simulated predators elicited vocalizations by parrots at nests and those nesting nearby, physical approach to the nest area by parrots other than the local pair was rare. Increased vocal behavior could warn others of the potential predator, but seems unlikely to deter predation. In contrast, increases in nesting success were observed in years with a predator control program against feral cats. In this case I was interested in the measure of successful vs unsuccessful nests and thus used a simple but appropriate binary approach

(Hellevik 2009). However if temporal variation in survival is suspected such as differences in survival during the different nest stages or seasonal differences, other models using logistic exposure are more useful in elucidating these patterns in probability of survival (Dinsmore et al.

2002).

The lack of spatial effects on vocal synchrony of nesting parrot pairs suggests that parrot pairs do synchronize but there is no difference at geographic scales. It appears that parrots vocalize at nest sites during evening at about the same time throughout the nesting area, making up the dawn and evening chorus. Without synchronous vocalization between close-nesting individuals, this temporal vocalization could still be used to transmit information during these times (Hutchinson 2002). The evening chorus could be synchronized to allow all parrots to display territorial calling at the same time (Baker 2004). However, the information transmission analysis in this study focuses solely at nest sites. Parrots are a gregarious species and are known to interact during in other situations such as at feeding sites. Information transmission can still occur in other ways to find these food sites or at the food sites themselves.

59

Parrots at focal nest sites and their nearest neighbors increased their vocalization when a simulated predator was present, suggesting that information may be transmitted that could alert a neighbor to the risk of a predator in the area. However I found only limited evidence of any mobbing behavior at neighbors’ nests. I do have evidence that neighbors occasionally approach a threatened nest. In 78 observations of simulated predator trials, I observed only 3 approaches by neighbors towards the focal nest. Approaches to neighboring nests were not documented outside of the experimental periods. However the behavior was so infrequent that I was not able to determine whether it was regular antipredator behavior.

Nesting individuals increased vocalization at the nest site may target the predator itself rather than the neighbor, as vocalizations could draw the attention of an unwanted predator rather than transmitting information to neighboring congeners that a predator is present. In species nesting in close proximity, mobbing responses at nearby nests are common and the benefit derived from group defense is potentially an important driver of aggregated nesting behavior

(Curio 1978). In the Bahama Parrot population, defensive interactions by neighbors appear to be limited to a vocal mobbing behavior which is characterized by increasing vocalization in response to the calls of adults at the focal nest (Curio 1978,Grimm 2008, Naguib et al. 1999).

The mobbing behaviors and any other benefits conferred from alarm calls or eavesdropping on neighbors were apparently insufficient to protect nests and nesting individuals from non-native predators. Given previously documented declines in nesting success when introduced predators were present, it appears the Bahama Parrot is not able to defend itself against these predators. Having neighbors at close distances was not an advantage to nest success. A similar result was found in communal nesting Monk parakeets where no relationships were found between communal nest sizes and productivity (Navarro et al. 1992). However, this

60 study suggests that ongoing predator control efforts do have a positive impact on parrot nest survival. Nest survival increased to a viable level of greater than 1.4 chicks per initiated nest after the introduction of predator control efforts in 2 out of 4 years and adult survival during the nesting period was found to be in viable levels in 2 out of 3 years post predator control initiation.

Feral cats have home ranges which vary in size (0.2km2 – 7.6km2) depending on the sex of the cat and resource availability (Barratt 1997, Fitzgerald and Karl 1986, Knoecny 1987). The overlap of cat territories tends to be minimized in food limited situations, which describes the

Bahamas’ limestone island habitat (Konecny 1987). Konecny (1987) estimated 2.2 -2.5 cats per km2 in on limestone islands with similar vegetation and no native mammals. At this density, a feral cat territory holder would become familiar with the area and know the locations of food resources (including parrot nests). On Great Abaco Island, the parrot nest density in high density nesting areas is approximately the same as the cat density reported by Konecny (1987).

However, the fate of neighboring parrot nests is related, which suggests that cats may have more than one nest in a territory and a resident cat may affect both nests in an area.

Lethal control of feral cat populations is controversial, as it is difficult to remove all individuals. Some animal groups advocate leaving territory-holding cats in areas to prevent more cats from moving in (Alley Cat Allies 2011) or filling a void. This view is at odds with my finding that the removal of cats (range 9-55 cats/year), in a given year was effective at increasing nest success regardless of the number removed. I think the territorial nature of cats may largely explain these results: territorial cats may be familiar with nest locations and revisit them until they are able to catch adults or young. Removing these territory-wise individuals delays discovery of active parrot nests by any recent immigrant cats. This possibility is supported by the fact that a trail cam focused on a parrot nest (as part of another study) documented a single

61 cat visiting a Bahama Parrot nest multiple times without predating it. This observation could be tested with further camera monitoring to see whether feral cats that visit nest sites actually predated those nests, or by using artificial nests or surrogate prey rather than actual parrot nests.

Apparently group nesting by this parrot population is not driven by antipredator behavior.

Instead, it suggests that mitigating the effects of introduced predators requires the ongoing predator control efforts. Conservation recommendations for optimal nest productivity therefore include protection of preferred nesting habitat, based on previous studies (Chapter 2), and continued predator control to prevent ‘wise’ cats in a territory during parrot breeding season.

Since eradication of feral cats does not seem feasible in the Abaco National Park (BNT 2005,

Nogales et al. 2004). Further research on the use of the nesting area by predators could suggest

ways to make the control effort more targeted and efficient.

Table 4.1. Demographic information for the Bahama Parrot and predator control information for 2003-2005 and 2010-2012. Demographic values included adult survival, as estimated through modified Mayfield analysis (Stahala 2005); mean breeding productivity as estimated through Mayfield analysis. Data on predator control were obtained from the Bahamas National Trust (BNT), for years in which data were collected for predator control and data for the number of cats removed during or prior to the breeding season.

2003 2004 2005 2010 2011 2012 N of nests monitored 38 77 69 56 55 65 % Nest Predation by Cats 28% 27% 22% 15% 2% 2% Adult Survival NA 0.87 NA 0.85 1 1 Breeding Productivity 1.23 0.8 1.26 1.4 1.68 1.22 +S.E. +0.22 +0.13 +0.20 +0.16 +0.21 +0.13 Predator Control No No Yes Yes Yes Yes N Cats removed 0 0 55 9 11 48

62

Table 4.2 Results of predator simulation at focal nests indicate focal and nearest neighbor (NN) respond vocally during the simulation but at different rates. Neighbors did approach focal nests during predator simulations. Asterisk denotes significant value at the 0.05 level.

Predator Simulation: vocal response Estimate Std. Error z value Pr(>|z|) (Intercept) 1.78 0.457 3.89 0.0001 * Treatment vs Control -1.159 0.406 -2.85 0.004 * Neighbor vs Focal -1.085 0.407 -2.66 0.007 *

Predator Simulation: approach response by neighbor Estimate Std. Error z value Pr(>|z|) (Intercept) -13.195 5.223 -2.526 0.0115 * Treatment vs Control -30.93 411.734 -0.075 0.9401

Table 4.3 Model parameters and results for GLMMs with the binary response of nest success N=216 nest observations. Predator control and nearest neighbor nest fate were the only measured factors had an effect on nest success of focal Bahama Parrot nests. Nests were more successful when neighbors were successful and predator control was in effect.

Std. Z- Estimate Error value Pr(>|z|) (Intercept) -1.2 0.43 -2.76 0.005 * Distance 0.48 0.63 0.76 0.445 Predator Control 2.07 0.67 3.05 0.002 * 0.0007 NN Fate 2.43 0.72 3.36 * Predator Control*Distance -1.11 0.8 -1.38 0.167 Predator Control*NN Fate -1.52 0.84 -1.8 0.07

63

Figure 4.1. Vocal synchrony was unrelated to distance between nesting pairs (Wilcoxon Rank Sum Test; NN vs DN Z = 0.16 p = 0.16; RN vs DN Z=0.275 p=78; RN vs NN Z=-1.31 p=0.18). Synchrony rate was measured as the number of times nests vocalized within the same time interval, compared between a focal nest (focal) and surrounding nests at different distances; NN = Nearest Neighbor (N=30), DN = Distant Neighbor (30) over 1km, Focal and RN = Random Neighbor (N=30). Grand mean for number of synchronous vocalization intervals per observation period is indicated by horizontal line. Box plot shows median, interquartile range,15x interquartile range, and outliers beyond this range.

64

Figure 4.2 Focal individuals and their nearest neighbors increased their rates of vocalization during the predator treatment (GLMM, Z= -2.85, p<0.005). Neighbors responded less frequently than focal pairs. Figure shows the proportion of nests out of 39 trials that were approached by artificial predators.

65

CHAPTER 5

RECLASSIFICATION OF THE CUBAN PARROT (AMAZONA LEUCOCEPHALA) COMPLEX

Description of the Problem

The Cuban parrot complex (currently Amazona leucocephala) is composed of six extant populations (Cuba, Isla del la Juventud, Cayman Brac, Grand Cayman, Great Abaco, and Great

Inagua) of five subspecies (Figure 5.1) with similar morphological features, including red throat and foreneck, white forehead and forecrown, red abdominal patch and pale bill (Forshaw 2006).

The classification of the complex as species or subspecies has historically depended on qualitative plumage characteristics and geographic boundaries. However, significant morphological, behavioral, vocal and genetic differences have recently been identified among three of the six populations. Based on this new information, I recommend that the Cuban parrot complex should be reclassified. In this paper, I outline the history of the classification of the

Cuban parrot complex and the new information on genetic, morphological, vocal and behavioral differences among the populations, which indicates the existing classification is in need of revision.

History

The Cuban parrot complex has had a long history of name changes (Peters 1928). As early as 1731, Catesby documented the parrots in Cuba as Psittacus paradisi. In 1886, Cory documented Chrysotis caymenensis on a trip to Grand Cayman as a parrot distinct from the parrot population in Cuba. The parrot populations in the Bahamas were classified with the Cuban population as Psittacus collarus but recognized as a distinct variety, bahamensis (Bryant 1867).

66

Around the turn of the century, several records referred to the Bahama population as a distinct species either by the name Chrysotis bahamensis or Amazona bahamensis (Allen 1905, Bonhote

1903).

Peters (1928) formally examined the various populations of the Cuban parrot complex and

suggested a formal classification for the parrots in Cuba, the Cayman Islands and the Bahamas

based on specimens in the Museum of Comparative Zoology (number of specimens unknown).

He recognized four subspecies within a single species. This classification included Amazona

leucocephala leucocephala composed of all Cuban parrot populations, A.l. hesterna on Cayman

Brac and Little Cayman, A.l. caymenensis on Grand Cayman and A.l. bahamensis in the

Bahamas. Unlike for the other Amazona leucocephala parrot subspecies, the description of

A.l.bahamensis did not indicate from which island the specimens used for the Bahamas were collected, and Peters excluded the Abaco population from the final description (Peters 1928).

Except for the extirpation of individual parrot populations on Little Cayman and Acklins

Island in the Bahamas in the mid 1900’s, the distribution of the Cuban parrot complex has remained relatively constant (Bond 1956 & Bond 1964). The range continues to encompass the area from the Cayman Islands (Grand Cayman and Cayman Brac) to Cuba (Mainland Cuba and

Isla de la Juventude), and The Bahamas (Great Abaco and Great Inagua Islands). The American

Ornithologists’ Union currently recognizes six extant populations as Amazona leucocephala

(AOU 1998). Five subspecies are currently described for A. leucocephala, which include A. l.

leucocephala (Cuba), A. l. palmarum (Isla de la Juventud), A. l. bahamensis (Great Abaco and

Great Inagua), A. l. hesterna (Cayman Brac), and A. l. caymanensis (Grand Cayman).

67

New information about these parrot populations has been obtained with more sophisticated methods than were available to Peters; thus, I suggest a new classification of the Amazona

leucocephala parrot complex.

New Information

Morphology

Peters based his 1928 classification of the Cuban parrot complex on morphological and

plumage color traits. Reynolds and Hayes (2009) conducted a quantitative reassessment of

morphological and color differences among the six extant populations of the Cuban parrot

complex and the extinct Acklins population using museum specimens. There were significant

differences in all traits measured between islands including culmen, wing chord, tail, and foot.

Additionally, they examined the amount and extent of white on head, amount of red on throat

and belly, and color on inner and outer eye (Table 5.1). No single character could distinguish

one population from another, however, when Reynolds and Hayes (2009) used two or more

characters in a discriminant function analysis, individuals were assigned to correct populations

an average of 81% of the time, with a 95% correct assignment of Cayman populations and an

89% correct assignment of Abaco populations when compared to all other populations. The

authors concluded that the three Bahama populations were as distinguishable, if not more so,

than other populations in the current Cuban parrot complex. Comparison of neighboring

populations showed that Abaco/Inagua populations can be differentiated 100% of the time. The

population with the lowest degree of differentiation was the Cayman Brac population with 67%

differentiation from the Cuban population.

68

Behavior

Vocalizations

In parrots, vocalizations can be horizontally transmitted within a group or vertically transmitted from parents to offspring (Berg et al. 2011); specific calls and vocalizations are learned from other individuals in the population. The vocalizations can become population specific when learned in isolation from other populations. While vocalizations cannot be used as a single indicator of species differences, they may be indicators of the cultural structure of a population (Wright and Dorin 2001). Vocalizations in addition to variation in other characteristics (morphological, plumage, genetics) can contribute to tan understanding of social organization and differentiation.

A quantitative assessment of the flight calls of the six extant populations of the Cuban parrot complex supports divergence of the individual island populations with reduced differentiation between the two populations on Cuba (Reynolds et al. 2010). Abaco parrots are distinguishable from all other populations in the complex by their flight calls, which have a low fundamental frequency and few notes. Inagua parrot flight calls are also unique due to their high fundamental frequency and a frequency jump bifurcation. Cayman Brac population calls have long syllable duration and syllable interval length. The Cayman population calls have short syllables. The two Cuban populations’ calls share subharmonic features, but, sample sizes were small. Reynold et al. (2010) did find distinct dialects among the Cuban populations.

Breeding

Breeding segregation is not a diagnostic feature of different Amazona species. In fact, the introduction and subsequent interbreeding between different species of Amazona parrots is

69 considered a major conservation threat to endemic/resident parrot species (Nichols 1980).

Nevertheless, private alleles found in the Abaco and Inagua populations tell us there is probably no interbreeding between the two populations. Additionally, a temporal segregation in breeding has developed in the Abaco population compared to the Inagua population. On Abaco, breeding beginsat the end of April while the Inagua population begins nesting at least 5 weeks earlier

(Snyder et al. 1982).

Habitat use

The habitats of each Cuban parrot population vary greatly. There may be underlying behavioral or genetic adaptations to these differences in the current populations. Local adaptation might be expected if traits for specific habitats are selected. Low to non-existent dispersal and therefore low to no gene flow among populations should increase the chances of adaptation to the local environments.

Abaco

Parrots on the island of Abaco are the only Amazona parrots to successfully nest in

underground limestone solution cavities in Caribbean pine forests (Snyder 1982). This

population is also the latest nesting Amazon parrot population. The nesting season begins

towards the end of May and ends mid-September (Gnam 1991). During the non-breeding season,

the birds population move into the hardwood forests on the island (Stahala 2008).

70

Inagua

The Great Inagua population uses the dry and wet hardwood hammocks on the island.

The Inagua parrot is only known to successfully nest in tree cavities although ground cavities are available and are explored by Inagua parrots. Some of the tree nest cavities are low to the ground with nest entrances 24 cm from the ground. Nesting occurs March through July (Stahala 2007).

Cuba

The Cuban populations of A. leucocephala inhabit areas with mature trees and snags including remote woodlands in mountains and lowlands. Palm groves are also used for nesting by parrots in savannas and wetland areas. Nesting occurs March through July, as in the Inagua parrot population. Foraging occurs in the same habitats within tree canopies (Wiley et al. 2004).

Cayman Islands

A. leucocephala parrots nest in beach ridge scrubland, dry hardwood forests, black mangrove habitats and urban areas. These trees and nests cavity entrances vary in height. Nesting occurs from March through June (Wiley et al. 2004).

Genetics

An in-depth genetic assessment of the Cuban Amazon complex shows divergence of the extant and one extinct A. leucocephala populations, with uncertainty remaining in the two Cuban and Cayman Brac populations (Russello et al. 2010). Russello et al. (2010) sampled the six extant populations and the recently extirpated Acklins Island parrot population using DNA collected from field and museum specimens. The results of the Bayesian haplotype tree indicates

71 the three Bahama populations (Abaco, Inagua, Acklins) and the Grand Cayman population have high posterior probabilities (93-99%) that they form distinct monophyletic groups (share a common recent ancestor), whereas the two Cuban populations formed one phylogenetic group

(Figure 5.2). Surprisingly, the two parrot populations from the Cayman Islands are not sister taxa and form a paraphyletic group with the Cuban parrots. There is a low posterior probability (52%) for the differentiation of the Cayman Brac and the Cuban populations; thus I am treating the taxonomy of this group (Cayman Brac and Cuban populations) as unresolved at this point.

However, the Grand Cayman population is monophyletic.

In addition to distinct phylogenetic groups, the three Bahamas populations exhibit mitochondrial haplotype diversity, suggesting divergence. Russello et al. (2010) identified Abaco and Acklins as each having three unique mitochondrial haplotypes and the Inagua population having six unique haplotypes as well as private alleles for each population.

The estimated sequence divergence between various populations ranges from 1.2% to

4.8%. These differences correspond well to those expected of full species (Johns and Avise

1998). The Grand Cayman and mainland Cuba populations of A. leucocephala show the largest divergence of 4.8%, approximating divergence between the Puerto Rican Parrot (A. vittata) and

Hispaniolan Parrot (A. ventralis) (4.5%), which are recognized as distinct species. Interestingly, the genetic distance between the Abaco and Inagua populations which are currently considered one subspecies (A.l.bahamensis), is higher (2.8% sequence divergence of mtDNA CR1 haplotypes) than divergence between two currently recognized Cuban subspecies (A. l. leucocephala and A. l. palmarum, 1.2%). However, the genetic differences between the populations in the Cuban parrot complex are all higher than differences between sister taxa of recently recognized bird species in the Bahamas (0.7% - 1%; Table 5.3). The genetic regions

72 assessed for sequence divergence in the Cuban Parrot complex are appropriate and robust for the species-level inference made in this study. The same mitochondrial region (control region, CR1) that was used in the Cuban parrot study was also used in the analysis of divergence in Yellow- throated Warblers (Russello et al. 2010, McKay et al. 2010) that elevated the Bahama’s population to species status (American Ornithologists’ Union 2010). Additional sequence coding genes (tRNA-Thr, pND6, pGLU) used in the Cuban parrot complex study add greater confidence to the resolution (Russello et al. 2010).

Recommendation

I propose, based on comparison of the Amazona leucocephala populations to other

known species from the area that have been recently upgraded to distinct species (Table 5.2), the

Great Abaco, Great Inagua, and Grand Cayman populations of A. leucocephala bahamensis be

elevated to three distinct full species. The genetic differences summarized here correspond

remarkably well with the differences in morphology and vocalizations found by Reynolds and

Hayes (2009) and Reynolds et al. (2010). Although on their own the vocal, behavioral and

morphological characteristics may not be reason enough for species classification in parrots, the

congruent genetic evidence shows a strong underlying divergence among the populations that are

proposed here to be distinct species (Alstrom et al. 2008, Johnson et al. 1999). Additionally, the

private alleles found in the Abaco and Inagua populations suggest divergence.

The data presented also suggest combining the two Cuban populations (Cuba and Isla del

la Junventud) as a single subspecies (Amazona leucocephala leucocephala). Finally the Cayman

Brac population appears to be lacking data and is unresolved at this time, thus I suggest no

73 change to the classification of this population (Table 5.3; Figure 5.1). The proposed revisions follow the suggested divisions of Russello et al. (2010).

Personal Views on Species Concepts

There is no one-size-fits-all definition of a species, and because of this there are well over a dozen different suggestions for defining a species (or species concepts). Even when trying to define a species concept within the class Aves, it is difficult. Some birds are more mobile than others, for example migrating over long distances across continents while others are flightless and isolated on an island. Other differences such as habitat fragmentation or habitat specificity further complicate establishing relationships between individuals. Ernst Mayr (1942) stated that

“species are groups of actually or potentially interbreeding natural populations, which are reproductively isolated from other such groups”. From Mayr’s definition it is easy to conclude that the migratory individuals breed only with other individuals because it is easy to document that they interact with other individuals across the globe but only mate and reproduce with individuals that look like them on the breeding grounds. For the non-migratory species, it is easy to say they reproduce with other similar sympatric individuals, but unclear that they would be unable reproduce with other individuals from other parts of the world. The Mayr (1942) definition or biological species definition seems to be the most useful definition of species, as it leaves room for interpretation of terms such as the definition of ‘isolation’ and ‘natural populations’. This room for interpretation gives it the needed flexibility to allow us to define groups of animals as species even in situations with incomplete information. For example, this may be applied even in the absence of such as genetic or other cladistics data required by the phylogenetic species concept, or recognition information on whether an individual recognizes

74 another individual as a potential mate (Cracraft 1989, Paterson 1985). Parrots actually make a good case studies of species concepts. Parrots at the genus level are pretty clearly different categories. Macaws (Ara, Anodorhynchus, Cyanopsitta, Primolius, Orthopsittaca, and

Diopsittaca genera) and Amazons (Amazona) do live sympatrically and do not interbreed. Once we start looking within the genus level it gets more complicated. The Black-billed (Amazona agilis) and Yellow-billed (Amazona collaria) Parrots are two Amazona species that live sympatrically on Jamaica. These two groups of birds are clearly two different species. However, both of these species have a common threat: introduced parrots. This isn’t because they are at risk of being killed by these other parrots but they are at risk of interbreeding with these other

Amazona parrots. Then how can these two groups be defined as species if they are threatened by the possibility of breeding with another ‘species’? In this case, the ‘natural population’ is a situation where other Amazon parrots are not present and the ‘reproductive isolation’ is assisted by a geographic barrier. Much in the same way, the Bahama Parrot population on Abaco, in a natural situation, is reproductively isolated from the Inagua population in the same way Jamaican

Amazona species are isolated from other introduced Amazona species.

Real world species are not as easy as a text book definition. Part of this is because

species are continually changing, being formed or being lost. Species are much more on a

continuum than discreet definitions. However, my opinion is that Mayr’s definition is still the

most useful for defining a species. Based on this the Abaco and Inagua populations of the

Bahama Parrot should be considered separate species.

75

Table 5.1: Diagnostic characteristics of the proposed species and and subspecies of Amazona leucocephala. Bahama species descriptions are from Reynolds and Hayes (2009). Cuba and Cayman species descriptions from Forshaw (2006) and Reynolds and Hayes (2009).

76

Table 5.2: Criteria used to reclassify avian species in the Caribbean, including Hispaniolan Oriole (Icterus d. dominicensis), Cuban Oriole ( I. d. melanopsis), Yellow-throated Warbler (Dendroica d.dominica), and Bahama Warbler (Dendroica d. flavescens), compared to the criteria for the Cuban Amazon complex and congeners of the Puerto Rican Parrot (Amazona vittata) and Hispaniolan Parrot (Amazona ventralis). Comparisons support the validity of reclassification of the Cuban Amazon complex. CN – Indicates closest neighboring parrot population to the study population

77

Table 5.3. Proposed classification of the Cuban Amazon complex in light of new information about the seven populations (based on Table 5.1).

78

Figure 5.1: Distribution of Cuban Amazon complex with the current and proposed classification.

79

Figure 5.2. Bayesian haplotype tree depicting relationships among sampled Amazona leucocephala haplotypes and two outgroups relative to their geographic and taxonomic distributions. Bayesian posterior probabilities (50%) are indicated above the branches. Reprinted with permission from Russello et al. 2010.

80

CHAPTER 6

CONCLUSION

The results of this study provide insight into the Bahama Parrots nesting habitat, genetic and

behavioral reasons for group nesting and the population’s survival. The findings from the habitat

portion of this study conclude that Bahama Parrots select nest sites in areas where other

conspecifics are nesting, particularly if the site has a high density of suitable cavities. These

parrots also have nest cavities with openings free of encroaching vegetation. The results of the

genetic portion of the study conclude that chicks in close nests are not any more likely to be

closely related to each other than other nests farther away. Genotyping adults and chicks over

multiple years supports the idea that nests are reused by the same nesting pair in successive

years, in most cases. The predator response study showed that vocalization of nesting parrots and

their neighbors increases when nesting areas are intruded upon. However, a more effective

method to prevent predation is the predator control effort implemented by the Bahamas National

Trust which is having a positive impact on the Bahama Parrot on Abaco. Finally, based on a

review of the entire Cuban Amazon complex, the Abaco and Inagua populations of the Bahama

Parrot should be considered separate species.

A goal of this study was to determine what influences Bahama Parrots to nest in a semi- colonial nesting aggregations. This study established that these parrots do nest in semi-colonial aggregations in south Abaco. Assessment of resource distribution suggests the reason for this is primarily habitat-based, as parrots nested in higher concentrations when an area had a higher

density of limestone cavities. However, the aggregation behavior is not just a function of nesting

habitat density, as it appears to be also influenced by the presence of other parrot nests. Genetic

testing showed that birds in close nests were not more closely related to each other than birds in

81 more distant nests, so I could not detect kin selection or indirect benefits between close nests through this study. Neighbors do not seem to be vocalizing with each other more frequently than with others in the nesting area. However, parrots did vocalize in responses to perceived predators near their nest sites. Although due to various influences of sampling and inference some of these results have greater support than others. For example, the measures of openness around nest cavities have high support however whether parrots look for open cavities versus manipulating a nest site to meet this openness criterial is only supported by an analysis that compares already established nests. A comparison of nest openness earlier in the season when nests are first found could provide more insight into this part of the question. Also, no support for genetic structure, either around nest sites or over the whole nesting area, was detected. The limit of only nine microsatellites and low diversity was certainly not helpful in elucidating any structure. Although these results don’t show any support for genetic structure this does not mean that further higher resolution analysis, such as SNP, couldn’t find genetic structure within this population. Still, this

new information on habitat use, relatedness within and between nests, and response to predators

can be used to influence management of this species to increase its likelihood of persistence.

Habitat influences where the Bahama Parrot nests on Abaco. This includes density of

cavities, how open the cavities are, and if other parrots are in the area. Based on these results,

my recommendations for habitat management include the protection of sites in and around the

current parrot nesting area that have a high density of suitable limestone cavities particularly if

other parrots are nesting in the area. Even if a site does not currently have nesting parrots, sites with limestone cavities could be potential expansion sites if the parrot population increases.

Parrots seem to be selecting nest cavities based on the openness of the area around a cavity opening. Although we measured this on the micro-habitat level of individual cavities, it

82 can be extended to a larger landscape level. Based on O’Brien et al. (2008) we know that parrots

can survive moderate intensity forest fires during the nesting season. These forest fires can

potentially also open the understory to make nesting cavities available and potentially suitable

for parrots to use. I recommend continuing plans for a prescribed burning regime for south

Abaco. Currently the there is no official burning program on the Island. Fires are primarily set

by locals during the dry period to clear out the pine understory for hunting. This is usually done

during dry periods and ends up burning in the hardwood coppice habitat. Hardwood coppice

habitat does not tolerate fire and when fire does encroach it causes the loss of this important

parrot wintering habitat.

A long term goal of the BNT is to reintroduce parrots to islands within their former

range. If this plan is implemented, information from this study can be used to inform future

management decisions. Nesting neighbors do not seem to be related, in fact, I detected no

significant genetic structure in this population, based on nine microsatellite loci. This suggests

that, if pairs or family groups of parrots from Abaco were to be used for translocation purposes,

these groups could come from anywhere on the island.

Close neighbors do exhibit defensive strategies when a predator appears at a nest site,

including increases in vocalization and some minimal approaches towards focal nest. These behaviors can be perceived as assistance from close neighbors. However the predator control effort implemented by the Bahamas National Trust has been effective. I recommend this program continue as a management strategy for effective conservation of the Bahama Parrot.

Information from this study can be applied to other parrot species in the region, particularly on islands. This study showed the importance of microhabitat with nest cavity openness and macrohabitat with hole density across the landscape. With information about

83 habitat at various scales, provides the insight into making sure the proper areas can be protected and managed. Additionally, if cavities are available over multiple years, they are reused by parrots. Protecting existing habitat, including nest sites, is key to parrots being able to reuse nesting areas successfully. This is important because parrots tend to reuse cavities that have proven to be successful and it does not reduce the number of suitable cavities available. Islands are particularly susceptible to invasive species. The Bahama Parrot population was negatively impacted by introduced mammals, particularly feral cats and potentially raccoons, until a predator control program was initiated. This study shows that the current control program works when implemented continuously. Managing predators is certainly feasible on other islands with predator problems if implemented regularly. My hope is that with this study we can now validate and target our management for the Bahama Parrot and other parrot species more effectively.

84

APPENDIX A

ALLELE MATCHES FOR EACH CHICK/PARENT PAIR SAMPLED INCLUDING YEAR SAMPLE WAS COLLECTED

Chick ID Nest Year Chick Year Year Female Match Male Match Sampled Mother Father Sampled Sampled 9/9 Allele 9/9 Allele 0405A2010C 405 2010 2010 2010 match match 9/9 Allele 9/9 Allele 0405B2010C 405 2010 2010 2010 match match 9/9 Allele 9/9 Allele 0405A2011C 405 2011 2010 2010 match match 9/9 Allele 9/9 Allele 0405B2011C 405 2011 2010 2010 match match 9/9 Allele 9/9 Allele 0405A2012C 405 2012 2010 2010 match match 9/9 Allele 9/9 Allele 0405C2012C 405 2012 2010 2010 match match 8/9 Allele* 9/9 Allele 0408A2009C 408 2009 2011 2011 match match 8/9 Allele* 9/9 Allele 0408B2009C 408 2009 2011 2011 match match 8/9 Allele* 9/9 Allele 0408A2010C 408 2010 2011 2011 match match 8/9 Allele* 9/9 Allele 0408B2010C 408 2010 2011 2011 match match 9/9 Allele 9/9 Allele 0408A2011C 408 2011 2011 2011 match match 9/9 Allele 9/9 Allele 0408B2011C 408 2011 2011 2011 match match 9/9 Allele 9/9 Allele 0408C2011C 408 2011 2011 2011 match match 9/9 Allele 9/9 Allele 0554C2012C 554 2012 2012 2012 match match 9/9 Allele 0567A2009C 567 2009 - 2012 - match 9/9 Allele 0567A2010C 567 2010 - 2012 - match 9/9 Allele 0567A2011C 567 2011 - 2012 - match 9/9 Allele 0567B2012C 567 2012 - 2012 - match 9/9 Allele 0567A2013C 567 2013 - 2012 - match

85

Chick ID Nest Year Chick Year Year Female Match Male Match Sampled Mother Father Sampled Sampled 9/9 Allele 0567B2013C 567 2013 - 2012 - match 9/9 Allele 9/9 Allele 1101A2011C 1101 2011 2011 2011 match match 9/9 Allele 9/9 Allele 1101A2012C 1101 2012 2011 2011 match match 9/9 Allele 9/9 Allele 1101B2012C 1101 2012 2011 2011 match match 9/9 Allele 1113A2011C 1113 2011 - 2011 - match 9/9 Allele 1114B2011C 1114 2011 - 2011 - match 9/9 Allele 1114C2011C 1114 2011 - 2011 - match 9/9 Allele 1114D2011C 1114 2011 - 2011 - match 9/9 Allele 1114E2011C 1114 2011 - 2011 - match 9/9 Allele 1114A2012C 1114 2012 - 2011 - match 9/9 Allele 1114B2012C 1114 2012 - 2011 - match 9/9 Allele 1114C2012C 1114 2012 - 2011 - match 9/9 Allele 1138A2011C 1138 2011 - 2011 - match 9/9 Allele 9/9 Allele 1139B2012C 1139 2012 2012 2012 match match 9/9 Allele 9/9 Allele 1139C2012C 1139 2012 2012 2012 match match 9/9 Allele 9/9 Allele 1139D2012C 1139 2012 2012 2012 match match 9/9 Allele 1206B2012C 1206 2012 - 2011 - match

Asterisk (*) indicates mismatch in alleles between chicks and female. Female and chicks sampled in different years.

86

APPENDIX B

154 BAHAMA PARROTS SAMPLED FOR DNA BY YEAR, NEST ID, AGE AND SEX

Age at Nest Individual ID Year Sampled Sampling Group Sex 0301A2012C 2012 Chick 301 _ 0301B2012C 2012 Chick 301 F 0301C2012C 2012 Chick 301 F 0304A2009C 2009 Chick 304 M 0304B2009C 2009 Chick 304 M 0305A2010C 2010 Chick 305 - 0305B2010C 2010 Chick 305 M 0305A2011C 2011 Chick 305 F 0305B2011C 2011 Chick 305 M 0305C2011C 2011 Chick 305 M 0320A2012C 2012 Chick 320 M 0320B2012C 2012 Chick 320 M 0320C2012C 2012 Chick 320 M 0320A2013C 2013 Chick 320 M 0320B2013C 2013 Chick 320 M 0322A2010C 2010 Chick 322 - 0322A2011C 2011 Chick 322 M 0405A2010A 2010 Adult 405 M 0405B2010A 2010 Adult 405 F 0405A2010C 2010 Chick 405 M 0405B2010C 2010 Chick 405 M 0405A2011C 2011 Chick 405 F 0405B2011C 2011 Chick 405 M 0405A2012C 2012 Chick 405 M 0405C2012C 2012 Chick 405 F

87

Age at Nest Individual ID Year Sampled Sampling Group Sex 0408A2011A 2011 Adult 408 M 0408B2011A 2011 Adult 408 F 0408A2009C 2009 Chick 408 F 0408B2009C 2009 Chick 408 F 0408A2010C 2010 Chick 408 F 0408B2010C 2010 Chick 408 F 0408A2011C 2011 Chick 408 - 0408B2011C 2011 Chick 408 M 0408C2011C 2011 Chick 408 M 0415A2011C 2011 Chick 415 M 0421A2009C 2009 Chick 421 F 0421A2013C 2013 Chick 421 F 0421B2013C 2013 Chick 421 M 0421C2013C 2013 Chick 421 M 0422A2009C 2009 Chick 422 M 0422B2009C 2009 Chick 422 F 0422A2010C 2010 Chick 422 M 0422B2010C 2010 Chick 422 F 0422A2011C 2011 Chick 422 F 0422B2011C 2011 Chick 422 M 0422A2013C 2013 Chick 422 F 0422B2013C 2013 Chick 422 M 0422C2013C 2013 Chick 422 M 0443A2012C 2012 Chick 443 M 0443B2012C 2012 Chick 443 M 0443C2012C 2012 Chick 443 F 0449A2009C 2009 Chick 449 F 0449A2011C 2011 Chick 449 F 0469A2012C 2012 Chick 469 F

88

Age at Nest Individual ID Year Sampled Sampling Group Sex 0469B2012C 2012 Chick 469 F 0469C2012C 2012 Chick 469 M 0507A2011C 2011 Chick 507 F 0523A2010C 2010 Chick 523 F 0523B2010C 2010 Chick 523 F 0537A2010C 2010 Chick 537 F 0537A2011C 2011 Chick 537 F 0537A2012C 2012 Chick 537 M 0554A2012A 2012 Adult 554 F 0554B2012A 2012 Adult 554 M 0554C2012C 2012 Chick 554 M 0567A2012A 2012 Adult 567 M 0567A2009C 2009 Chick 567 F 0567A2010C 2010 Chick 567 F 0567A2011C 2011 Chick 567 M 0567B2012C 2012 Chick 567 F 0567A2013C 2013 Chick 567 M 0567C2013C 2013 Chick 567 M 0801A2009C 2009 Chick 801 M 0801B2009C 2009 Chick 801 M 0904A2010C 2010 Chick 904 M 0904B2010C 2010 Chick 904 M 0904C2010C 2010 Chick 904 F 0915A2009C 2009 Chick 915 M 0922A2009C 2009 Chick 922 F 0922B2009C 2009 Chick 922 M 0922C2009C 2009 Chick 922 F 0922A2010C 2010 Chick 922 M 0922B2010C 2010 Chick 922 M

89

Age at Nest Individual ID Year Sampled Sampling Group Sex 1002A2012A 2012 Adult 1002 F 1002A2010C 2010 Chick 1002 F 1002B2012C 2012 Chick 1002 M 1002C2012C 2012 Chick 1002 M 1003A2010A 2010 Adult 1003 F 1003A2010C 2010 Chick 1003 M 1003B2010C 2010 Chick 1003 F 1005A2010C 2010 Chick 1005 F 1005B2010C 2010 Chick 1005 M 1006A2012A 2012 Adult 1006 M 1008A2011C 2011 Chick 1008 M 1010A2010C 2010 Chick 1010 F 1010B2010C 2010 Chick 1010 F 1014A2010C 2010 Chick 1014 F 1017A2010C 2010 Chick 1017 F 1020A2010C 2010 Chick 1020 F 1101A2011A 2011 Adult 1101 M 1101B2011A 2011 Adult 1101 F 1101A2011C 2011 Chick 1101 F 1101A2012C 2012 Chick 1101 M 1101B2012C 2012 Chick 1101 M 1102A2011C 2011 Chick 1102 F 1103A2011C 2011 Chick 1103 F 1103B2011C 2011 Chick 1103 F 1103C2011C 2011 Chick 1103 M 1112A2012C 2012 Chick 1112 F 1113B2011A 2011 Adult 1113 M 1113A2011C 2011 Chick 1113 M 1114A2011A 2011 Adult 1114 M

90

Age at Nest Individual ID Year Sampled Sampling Group Sex 1114B2011C 2011 Chick 1114 M 1114C2011C 2011 Chick 1114 M 1114D2011C 2011 Chick 1114 F 1114E2011C 2011 Chick 1114 F 1114A2012C 2012 Chick 1114 M 1114B2012C 2012 Chick 1114 F 1114C2012C 2012 Chick 1114 F 1118A2011C 2011 Chick 1118 M 1118B2011C 2011 Chick 1118 M 1118C2011C 2011 Chick 1118 F 1119A2011C 2011 Chick 1119 M 1119B2011C 2011 Chick 1119 - 1119C2011C 2011 Chick 1119 M 1126A2011C 2011 Chick 1126 F 1128A2011C 2011 Chick 1128 F 1136A2011C 2011 Chick 1136 F 1136B2011C 2011 Chick 1136 F 1136C2011C 2011 Chick 1136 M 1138A2011A 2011 Adult 1138 M 1138A2011C 2011 Chick 1138 M 1139E2012A 2012 Adult 1139 M 1139F2012A 2012 Adult 1139 F 1139A2012C 2012 Chick 1139 F 1139B2012C 2012 Chick 1139 F 1139C2012C 2012 Chick 1139 F 1139D2012C 2012 Chick 1139 F 1141A2011A 2011 Adult 1141 M 1206A2012A 2012 Adult 1206 F 1206B2012C 2012 Chick 1206 F

91

Age at Nest Individual ID Year Sampled Sampling Group Sex 1206C2012C 2012 Chick 1206 M 1211A2012C 2012 Chick 1211 F 1216A2012C 2012 Chick 1216 F 1221A2012C 2012 Chick 1221 F 1221B2012C 2012 Chick 1221 F 1229A2012A 2012 Adult 1229 F 1229B2012A 2012 Adult 1229 M 1229C2012C 2012 Chick 1229 F 1229D2012C 2012 Chick 1229 M 1229E2012C 2012 Chick 1229 M 1316A2013C 2013 Chick 1316 F 1319A2013C 2013 Chick 1319 M 1319B2013C 2013 Chick 1319 M

92

APPENDIX C

ACUC LETTER OF APPROVAL

93

APPENDIX D

LICENCE TO REUSE FIGURES

94

REFERENCES

Aitken, K.E.H., K.L. Wiebe, and K. Martin. 2002. Nest-site reuse patterns for a cavity-nesting bird community in interior British Columbia. Auk 119(2):391-402.

Allen, G. M. 1905. Summer birds in the Bahamas. Auk 22: 113-133.

Alley Cat Allies. 2011. The vacuum effect: why catch and kill doesn’t work. Bethesda, MD: Alley Cat Allies. Retrieved May 5, 2016 from http://www.alleycat.org/document.doc?id=534

Alexander, R. D. 1974. The evolution of social behavior. Annual Review of Ecology and Systematics. 325-383.

Alferi, M.S. and L.A.Dugatkin. 2009. Increased survival as a function of interactions with experienced partners. Ethology Ecology and Evolution 21(2):101-109.

Allee, W.C. 1949. Principles of Animal Ecology. Philadelphia: Saunders. Alstrom, P., P.C. Rasmussen, U. Olsson, and P. Sundberg. 2008. Species delimitation based on multiple criteria: the Spotted Bush Warbler Bradypterus thoracicus complex (Aves: Megaluridae) The Linnaean Society of London, Zoological Journal of the Linnaean Society 154:291–307.

American Ornithologists’ Union. 1998. Check-list of North American Birds, 7th ed. American Ornithologists’ Union, Washington, D.C.

American Ornithologists’ Union. 2010. Fifty-first supplement to the American Ornithologists’ Union Check-list of North American birds. Auk 127:726–744.

Andersson, M. and C.G. Wiklund. 1978. Clumping versus spacing out: experiments on nest predation in fieldfares (Turdus pilaris). Animal Behaviour 26:1207-1212.

Arroyo, B., F. Mougeot, and V. Bretagnolle. 2001. Colonial breeding and nest defense in Montagu's harrier (Circus pygargus). Behavioral Ecology and Sociobiology 50(2):109- 115.

Barratt, D.G. 1997. Home range size, habitat utilisation and movement patterns of suburban and farm cats (Felis catus). Ecography 20(3):271-280.

Bahamas National Trust (BNT). 2008. General Management Plan: Abaco National Park. Bahamas National Trust.

Bahamas National Trust, (BNT). 2012. Bahama Parrot (Amazona leucocephala bahamensis): Species Management Plan. Bahamas National Trust.

95 Baker, M.C. 2004. The chorus song of cooperatively breeding laughing kookaburras (Coraciiformes, Halcyonidae: Dacelo novaeguineae): characterization and comparison among groups. Ethology 110(1):21-35.

Barber, I., D. Nairn, and F.A. Huntingford. 2001. Nests as ornaments: revealing construction by male sticklebacks. Behavioral Ecology 12(4):390-396.

Beauchamp, G., M. Belisle, and L. A. Giraldeau. 1997. Influence of conspecific attraction on the spatial distribution of learning foragers in a patchy habitat. Journal of Animal Ecology 66:671-682.

Berg, K.S., S. Delgado, K.A. Cortopassi, S.R. Beissinger. and J.W. Bradbury. 2011. Vertical transmission of learned signatures in a wild parrot. Proceedings of the Royal Society B, doi: 10.1098/rspb.2011.0932

Berkunsky, I., & J.C. Reboreda. 2009. Nest site fidelity and cavity reoccupation by Blue fronted Parrots Amazona aestiva in the dry Chaco of Argentina. Ibis 151(1): 145-150. ‐ ‐ Bertram, B.C.R. 1980. Vigilance and group sizes in ostriches. Animal Behaviour 28:278–286.

Bond, J. 1947. Field guide to birds of the West Indies; a guide to all the species of birds known from the Greater Antilles, Lesser Antilles and Bahama Islands. New York, NY: Macmillan

Bond, J. 1956(a). Check-list of birds of the West Indies. Philadelphia, PA: Academy of Natural Sciences of Philadelphia.

Bond, J. 1956(b). Supplement to the Check-list of birds. Philadelphia, PA: Academy of Natural Sciences of Philadelphia.

Bond, J. 1964. Ninth supplement to the Check-list of the West Indies. Philadelphia, PA: Academy of Natural Sciences of Philadelphia.

Bonhote, J. L. 1903. The Bahama Amazon Parrot. Avicultural Magazine 2:239-241.

Bonilla-Ruz, C., G. Reyes-Macedo, and R. García. 2007. Observations of the Military Macaw (Ara militaris) in northern Oaxaca, Mexico. Wilson Journal of Ornithology 119:729-732.

Bowen, B. W., and S.A. Karl. 2007. Population genetics and phylogeography of sea turtles. Molecular Ecology 16(23):4886-4907.

Brawn, J. D. and R.P. Balda. 1988. Population biology of cavity nesters in northern Arizona: Do nest sites limit breeding densities? Condor 90:61-71.

96 Brightsmith, D. J. 2005. Competition, predation and nest niche shifts among tropical cavity nesters: phylogeny and natural history evolution of parrots (Psittaciformes) and trogons (Trogoniformes). Journal of Avian Biology 36:64-73.

Brightsmith, D.J., and R. Aramburú. 2004. Avian geophagy and soil characteristics in southeastern Peru. Biotropica 36:534-543.

Brown, C. R. 1984. Laying eggs in a neighbor's nest: benefit and cost of colonial nesting in swallows. Science 224(4648):518-519.

Brown C.R. and M.B. Brown. 1986. Ectoparasitism as a cost of coloniality in Cliff Swallows (Hirundo Pyrrhonota). Ecology 67(5):1206-1218.

Brown, C. R. 1988. Enhanced foraging efficiency through information centers: a benefit of coloniality in cliff swallows. Ecology 602-613.

Brown, C. R. and M. B. Brown. 2000. Breeding habitat selection in cliff swallows: The effect of conspecific reproductive success on colony choice. Journal of Animal Ecology 69:133- 142.

Brown, C.R., M. B. Brown, and E. Danchin. 2000. Breeding habitat selection in cliff swallows: the effect of conspecific reproductive success on colony choice. Journal of Animal Ecology 69(1):133-142.

Bryant, H. 1967. Additions to a list of birds seen at the Bahamas. Proceedings of the Boston Society of Natural History 11:63-70.

Burger, J. and M. Gochfeld. 2003. Parrot behavior at a Rio Manu (Peru) clay lick: temporal patterns, associations, and antipredator responses. Acta Ethologica 6(1):23-34.

Burk, D.M. and E. Nol. 1998. Influences of food abundance, nest-site habitat, and forest fragmentation on breeding ovenbirds. Auk 115:96-104.

Buxton R.T., C. Jones , H. Moller, and D.R. Towns. 2014. Drivers of seabird population recovery on New Zealand islands after predator eradication. Conservation biology 28(2):333-44.

Casiraghi, M., C. Polidori, P. Ferreri, D. G. Preatoni, F. Andrietti, and A. Martinoli. 2003. Does the distance between nest clusters affect reporoductive success in Ammophila sabulosa. Ethology, Ecology and Evolution 15:329-341.

Catesby, M. 1731. Habitat in America. Natural History of the Carolinas, Florida and the Bahamas. London, UK: Marin, Wilcox & Stichall.

Chapman, C. A., L. J. Chapman, and L. Lefebvre. 1989. Variability in parrot flock size: possible functions of communal roosts. Condor 842-847.

97

Clark, P. J. and F. C. Evans. 1954. Distance to nearest neighbor as a measure of spatial relationships in populations. Ecology 35:445-453.

Clark, R. W., W.S. Brown, R. Stechert, and H.W. Greene. 2012. Cryptic sociality in rattlesnakes (Crotalus horridus) detected by kinship analysis. Biology Letters, rsbl20111217.

Clutton-Brock, T. 2009. The cooperation between non-kin in animal society. Nature 462:51-57.

Cornelius, C., K, Cockle, N. Politi, I. Berkunsky, L. Sandoval, V. Ojeda, L. Rivera, M. Hunter Jr, and K. Martin. 2008. Cavity-nesting birds in neotropical forests: Cavities as a potentially limiting resource. Ornitologia Neotropical 19:253-268.

Cory, C. B. 1886. Descriptions of thirteen new species of birds from the island of Grand Cayman, West Indies. Auk 3:497-501.

Cougill, S. and S. J. Marsden. 2004. Variability in roost size in an Amazona parrot: implications for roost monitoring. Journal of Field Ornithology 75(1):67–73

Cracraft, J. 1989. Speciation and its ontology: the empirical consequences of alternative species concepts for understanding patterns and processes of differentiation. Speciation and its Consequences 28-59.

Curio, E. 1978. The adaptive significance of avian mobbing I. Teleonomic hypotheses and predictions. Zeitschrift für Tierpsychologie 48:175-183.

Davis, D. E. 1940. Social nesting habits of the Smooth-billed Ani. Auk, 179-218.

deMenezes, A. C., H. F. P. d. Araujo, J. L. X. doNascimento, A. C. G. Rego, A. A. Paiva, R. N. Serafim, S. D. Bella, and P. C. Lima. 2006. Monitoramento da população de Anodorhynchus leari (Bonaparte, 1856) (Psittacidae) en la naturalesa. Ornithologia 1:109-113.

Dinsmore, S.J., G.C. White, and F.L. Knopf. 2002. Advanced techniques for modeling avian nest survival. Ecology, 83(12):3476-3488.

Edelman, A. J., and J.L. Koprowski. 2007. Communal nesting in asocial Abert's squirrels: the role of social thermoregulation and breeding strategy. Ethology 113(2):147-154.

Eberhard, J. R. 2002. Cavity adoption and the evolution of coloniality in cavity-nesting birds. The Condor 104:240-247.

Elgar, M. A. 1989. Predator vigilance and group size in mammals and birds: a critical review of the empirical evidence. Biological Reviews 64(1):13-33.

98 Ellis, R.D. 2015. Ecological Effects of Red Grouper (Epinephelus Morio) in Florida Bay. Dissertation. Florida State University, Tallahassee, FL.

Enkerlin-Hoeflich, E.C., N.F.R. Snyder, and J.W.Wiley. 2006. Behavior of Wild Amazona and Rhynchopsitta Parrots, with comparative insights from other psittacines. In Manual of Parrot Behavior. ed, Andrew Luescher. Ames, IA: Wiley Blackwell.

Fitzgerald, B.M. and B.J. Karl. 1986. Home range of feral house cats (Felis catus) in forest of the Orongorongo Valley, Wellington, New Zealand. New Zealand Journal of Ecology 9:71-82.

Forshaw, J.M. 2006. Parrots of the World. Princeton, NJ: Princeton University Press.

Fretwell, S.D. and. H.L. Lucas. 1970. On territorial behaviour and other factors influencing habitat distribution in birds. Acta Biotheoretica 19:16-36.

Gilardi, J.D., and C.A. Munn. 1998. Patterns of activity, flocking, and habitat use in parrots of the Peruvian Amazon. Condor 641-653.

Gill, F.B. 1990. Ornithology, 2nd ed. New York, NY: W.H. Freeman.

Giraldeau, L.A., and T. Caraco. 2000. Social foraging theory. Princeton, NJ: Princeton University Press.

Gnam, R.S. 1991. Breeding biology of the Bahama Parrot Amazona leucocephala bahamensis. Dissertation. City University New York, New York.

Gnam, R.S., and A.E. Burchsted. 1991. Population estimates for the Bahama Parrot on Abaco Island, Bahamas. Journal of Field Ornithology 62:139-146.

Gnam, R., and R.F. Rockwell. 1991. Reproductive potential and output of the Bahama Parrot Amazona leucocephala bahamensis. Ibis 133(4): 400-405.

Gotmark, F., D. Blomqvist, O.C. Johansson, and J. Bergkvist.1995. Nest site selection: A trade- off between concealment and view of the surroundings. Journal of Avian Biology 26:305-312.

Granovetter, M. 1978. Threshold models of collective behavior. American Journal of Sociology 1420-1443.

Griesser, M. and M. Nystrand. 2009. Vigilance and predation of a forest-living bird species depend on large-scale habitat structure. Behavioral Ecology 20(4):709-715.

Griffiths, R., M.C. Double, K. Orr, and R.J. Dawson. 1998. A DNA test to sex most birds. Molecular Ecology 7(8):1071-1075.

99

Grim, T. 2008. Are blackcaps (Sylvia atricapilla) defending their nests also calling for help from their neighbours? Journal of Ornithology 149(2): 169-180.

Hamilton, W. D. 1964a. The genetical evolution of social behaviour. I. Journal of Theoretical Biology 7(1):1-16.

Hamilton, W. D. 1964b. The genetical evolution of social behaviour. II. Journal of Theoretical Biology 7(1):17-52.

Hamilton, W.D. 1971. Geometry for the selfish herd. Journal of Theoretical Biology 31(2):295-311.

Hardy, O.J. and X. Vekemans. 2002. SPAGeDi: a versatile computer program to analyse spatial genetic structure at the individual or population levels. Molecular Ecology Notes 2:618-620.

Harding, E.K., D.F. Doak, and J.D. Albertson. 2001. Evaluating the Effectiveness of Predator Control: the Non Native Red Fox as a Case Study. Conservation Biology 15(4):1114-22.

Hayes, L.D. 2000. To nest‐ communally or not to nest communally: a review of rodent communal nesting and nursing. Animal Behaviour 59:677-688.

Hatchwell, B.J. 2010. Cryptic Kin Selection: Kin Structure in vertebrate populations and opportunities for kin directed cooperation. Ethology 116(3):203-216.

Heisey, D.M., and T.K. Fuller. 1985.‐ Evaluation of survival and cause-specific mortality rates using telemetry data. The Journal of wildlife management 1:668-74.

Hellevik O. 2009. Linear versus logistic regression when the dependent variable is a dichotomy. Quality & Quantity 43(1):59-74.

Hoogland, J.L. 1981. The evolution of coloniality in white-tailed and black-tailed prairie dogs (Sciuridae: Cynomys leucurus and C. ludovicianus). Ecology 252-272.

Horn, H. 1968. The adaptive significance of colonial nesting in the Brewer’s blackbird (Euphagus cyanocephalus). Ecology 49(4):682-694.

Hurd, C.R. 1996. Interspecific attraction to the mobbing calls of blackcapped chickadees (Parus atricapillus). Behavioral Ecology and Sociobiology 38:287–292.

Hutchinson, J.M. 2002. Two explanations of the dawn chorus compared: how monotonically changing light levels favour a short break from singing. Animal Behaviour 64(4):527- 539.

100 Jackson, J.A. and B.J.S. Jackson. 2004. Ecological relationships between fungi and woodpecker cavity sites. Condor 106:37-49.

JMP, Version Pro 11. SAS Institute Inc., Cary, NC, 1989-2007.

Johns, G.C. and J.C. Avise. 1998. A comparative summary of genetic distances in the vertebrates from the mitochondrial Cytochrome b gene. Journal of Molecular Biology and Evolution 15(11):1481–1490.

Johnson, D.D.P., Kays, R., Blackwell, P.G. & Macdonald, D.W. 2002. Does the resource dispersion hypothesis explain group living. Trends in Ecology and Evolution 17, 563– 570.

Johnson, N.K., J.V. Remsen, and C. Cicero. 1999. Resolution of the debate over species concepts in ornithology: a new comprehensive biological species concept. in Proceedings 22nd International Ornithological Congress, Durban. Johannesburg, South Africa: BirdLife.

Kenward, R.E. 1978. Hawks and doves: factors affecting success and selection in goshawk attacks on wild pigeons. Journal of Animal Ecology 47:449–460.

Kerth, G., N. Perony, and F. Schweitzer. 2011. Bats are able to maintain long-term social relationships despite the high fission-fusion dynamics of their groups. Proceedings of the Royal Society B: Biological Sciences, rspb20102718.

Konecny, M.J. 1987. Home range and activity patterns of feral house cats in the Galapagos Islands. Oikos 50(1):17-23.

Koenig, W.D. and R.L. Mumme. 1987. Population ecology of the cooperatively breeding Acorn Woodpecker. Princeton, NJ.: Princeton University Press, Princeton.

Krause, J. and G.D. Ruxton. 2002 Living in groups. Oxford, UK: Oxford series in ecology and evolution.

Lande, R. 1988. Genetics and demography in biological conservation. Science 241(4872):1455-1460.

Koprowski, J. L. 1996. Natal philopatry, communal nesting, and kinship in fox squirrels and gray squirrels. Journal of Mammalogy 77(4):1006-1016.

LaMontagne, J.M., C.T. Williams, J.L. Donald, M.M. Humphries, A.G. McAdam, and S. Boutin. 2013. Linking intraspecific variation in territory size, cone supply, and survival of North American red squirrels. Journal of Mammalogy 94:1048-1058.

Lessells, C.M., M.I. Avery, and J.R. Krebs. 1994. Nonrandom dispersal of kin: why do European Bee-eater (Merops apister) brothers nest close together? Behavioral Ecology 5:105-113.

101 Lloyd J.D. and J.J. Tewksbury. 2007. Analyzing avian nest survival in forests and grasslands: a comparison of the Mayfield and logistic-exposure methods. Studies in avian biology 34:96.

Marshall, T. C., J. Slate, L. Kruuk. And J.M. Pemberton. 1998. Statistical confidence for likelihood-based paternity inference in natural populations. Molecular Ecology 7: 639-655.

Marzluff, J.M., B. Heinrich, and C.S. Marzluff. 1996. Raven roosts are mobile information centres. Animal Behaviour 51:89–103.

Mayfield, H. 1961. Nesting success calculated from exposure. Wilson Bulletin 73:255-261.

Mayr, E. 1942. Systematics and the Origin of Species. Columbia University Press, New York, NY.

Masello, J.F. and. P. Quillfeldt. 2002. Chick growth and breeding success of the Burrowing Parrot. The Condor 104:574-586.

Masello, J.F., A. Sramkova, P. Quillfeldt, J.T. Epplen, and T. Lubjuhn. 2002. Genetic monogamy in burrowing parrots Cyanoliseus patagonus? Journal of Avian Biology 33 (1):99- 103.

Mayer, P.M. and M.R. Ryan. 1991. Survival rates of artificial Piping Plover nests in American Avocet colonies. Condor 93:753-755.

McKay, B.D., M.B.J. Reynolds,W.K. Hayes, and D.S. Lee. 2010. Evidence for the species status of the Bahama Yellow-throated Warbler (Dendroica “dominica” flavescens). Auk 127(4):932-939.

Miller, B., K. Ralls, R.P. Reading, J.M. Scott and J. Estes. 1999. Biological and technical considerations of carnivore translocation: a review. Animal Conservation 2(1):59-68.

Mohr, C.O. 1947. Table of equivalent populations of north american small mammals. American Midland Naturalist 37:223-249.

Møller, A.P. 1987. Advantages and disadvantages of coloniality in the swallow, Hirundo rustica. Animal Behaviour 35(3):819-832.

Monterrubio-Rico, T.C., J. Cruz-Nieto, E. Enkerlin-Hoeflich, D. Venegas-Holguin, L. Tellez-Garcia, and C. Marin-Togo. 2006. Gregarious nesting behavior of Thick-billed Parrots (Rhynchopsitta pachyrhyncha) in aspen stands. The Wilson Journal of Ornithology 118(2): 237-243.

102 Moorhouse R., T. Greene, P. Dilks, R. Powlesland , L. Moran, G. Taylor, A. Jones, J. Knegtmans, D. Wills, M. Pryde, I. Fraser. 2003. Control of introduced mammalian predators improves kaka Nestor meridionalis breeding success: reversing the decline of a threatened New Zealand parrot. Biological Conservation 110(1):33-44.

Mori, G.P. 2006. Physical characteristics of nests and habitat selection of breeding Bahama Parrots (Amazona leucocephala bahamensis). Master's Thesis. University of Maryland, Eastern Shores, Maryland.

Naguib, M., R. Mundry, R. Ostreiher, H. Hultsch, L. Schrader, and D. Todt. 1999. Cooperatively breeding Arabian Babblers call differently when mobbing in different predator-induced situations. Behavioral Ecology 10(6):636-640.

Neil, S.R., and J.M.Cullen. 1974. Experiments on whether schooling by their prey affects hunting behaviour of cephalopods and fish predators. Journal of Zoology 172:549–569.

Newton, I. 1994. The role of nest sites in limiting the numbers of hole-nesting birds. Biological Conservation 70:265-276.

Nichols, T. D. 1980. An unusual amazon hybrid. In Conservation of New World parrots, ed. R.F. Pasquier. ICBP Tech. Pubis, No. 1. Smithsonian Institution Press, Washington, DC.

Nogales, M., A. Martín, B.R. Tershy, C. Donlan, D. Veitch, N. Puerta, B. Wood, and J. Alonso. 2004. A review of feral cat eradication on islands. Conservation Biology 18(2):310-319.

O'Brien, J. J., C. Stahala, G.P. Mori, M. A. Callaham, and C.M. Bergh. 2006. Effects of prescribed fire on conditions inside a Cuban Parrot (Amazona leucocephala) surrogate nesting cavity on Great Abaco, Bahamas. The Wilson Journal of Ornithology 118:508- 512.

O'Brien, J.J., J.K. Hiers, M.A. Callaham Jr, R.J. Mitchell, and S.B. Jack. 2008. Interactions among overstory structure, seedling life-history traits, and fire in frequently burned neotropical pine forests. AMBIO: A Journal of the Human Environment. 37(7):542-547.

Olson, S.L. and W.B. Hilgartner. 1982. Fossil and subfossil birds from the Bahamas. Pp 22-60. in S. Olson (ed) Fossil vertebrates from the Bahamas. Washington D.C.: Smithsonian Institute Press.

Oro, D. 1996. Colonial seabird nesting in dense and small sub-colonies: an advantage against aerial predation. Condor 848-850.

Ottens-Wainright P, K.M. Halanych, J.R. Eberhard, R.I. Burke, J.W. Wiley, R.S. Gnam, X.G. Aquilera. 2004. Independent geographic origin of the genus Amazona in the West Indies. Journal of Caribbean Ornithology 17:23–49.

103 Paterson, H. E. H. (1985) in Species and Speciation, ed. Vrba, E. S. Transvaal Museum, Pretoria, South Africa 21–29.

Peters, J. 1928. The races of Amazona leucocephala (LINN). Auk 45:342-344.

Peterson, M.N. B. Hartis, S. Rodriguez, M. Green, and C.A.Lepczyk. 2012. Opinions from the front lines of cat colony management conflict. PLoS One 7(9):44616.

Petrie, M., and Kempenaers, B. 1998. Extra-pair paternity in birds: explaining variation between species and populations. Trends in Ecology & Evolution 13(2):52-58.

Pizo, M.A., I. Simáo, and M. Galetti. 1995. Diet and flock size of sympatric parrots in the Atlantic forest of Brazil. Ornitologia Neotropical 6:87-95.

Price, M.R., and W.K. Hayes. 2009. Conservation taxonomy of the Greater Antillean Oriole (Icterus dominicensis): Diagnosable plumage variation among allopatric populations supports species status. Journal of Caribbean Ornithology 22:19–26.

R_Core_Team. 2012. R: A language and environment for statistical computing.in R. F. f. S. Computing, editor., Vienna, Austria.

Reyer, H.U. 1984. Investment and relatedness: a cost/benefit analysis of breeding and helping in the pied kingfisher (Ceryle rudis). Animal Behaviour. 32(4):1163- 1178.

Reynolds, M.B.J., and W.K. Hayes. 2009. Conservation taxonomy of the Cuban Parrot (Amazona leucocephala): Variation in morphology and plumage. Journal of Caribbean Ornithology 22:1–18.

Reynolds, M.B.J., W.K. Hayes, and J.W. Wiley. 2010. Geographic variation in the flight call of the Cuban Parrot (Amazona leucocephala) and its taxonomic significance. Journal of Caribbean Ornithology 23:4-18.

Ripley, B.D. 1977. Modelling spatial patterns. Journal of the Royal Statistical Society. Series B 172-212.

Rivera-Milán, F.F., J.A. Collazo, C. Stahala, W.J. Moore, A. Davis, G. Herring, M. Steinkamp, R. Pagliaro, J.L. Thompson, and W. Bracey. 2005. Estimation of density and population size and recommendations for monitoring trends of Bahama Parrots on Great Abaco and Great Inagua. Wildlife Society Bulletin 33(3):823-834.

Robel, R. J., J.N. Briggs, A.D. Dayton, and L.C. Hulbert.1970. Relationships between visual obstruction measurements and weight of grassland vegetation. Journal of Range Management 295-297.

104 Roberts, G. 1996. Why individual vigilance declines as group size increases. Animal Behavior 51:1077-1086.

Robertson, R.J., and W.B. Rendell.1990. A comparison of the breeding ecology of a secondary cavity nesting bird, the Tree swallow (Tachycineta bicolor), in nest boxes and natural cavities. Canadian Journal of Zoology 68:1046-1052.

Rohwer, S., S.D. Frettwell, R.C. Tuckfield. 1976. Distress screams as a measure of kinship in birds. American Midland Naturalist 96:418–430.

Russello M, D. Calcagnotto , R. DeSalle, G. Amato. 2001. Characterization of microsatellite loci in the endangered St. Vincent Parrot, Amazona guildingii. Mol Ecol Notes 1:162–164.

Russello M.A., K. Lin, G. Amato, A. Caccone. 2005. Additional microsatellite loci for the endangered St. Vincent Parrot, Amazona guildingii. Conservation Genetics 6:643–645.

Russello, M. A., C. Stahala, D. Lalonde, K.L Schmidt, and G. Amato. 2010. Cryptic diversity and conservation units in the Bahama Parrot. Conservation Genetics 11(5):1809-1821.

Saab, V.A., J. Dudley, and W.L. Thompson. 2004. Factors influencing occupancy of nest cavities in recently burned forests. Condor 106:20-36.

Salinas-Melgoza, A., V. Salinas-Melgoza, and K. Renton. 2009. Factors influencing nest spacing of a secondary cavity-nesting parrot: habitat heterogeneity and proximity of conspecifics. Condor 111:305-313.

Schwartz, M.K., G. Luikart, and R.S. Waples. 2007. Genetic monitoring as a promising tool for conservation and management. Trends in Ecology & Evolution 22(1):25-33.

Seixas, G.H and G.M. Mourao. 2002. Nesting success and hatching survival of the Blue-fronted Amazon (Amazona aestiva) in the Pantanal of Mato Grosso do Sul, Brazil. Journal of Field Ornithology 73(4):399-409

Seyfarth, R.M., D.L. Cheney, P. Marler. 1980. Monkey responses to three different alarm calls: Evidence of predator classification and semantic communication. Science 210(4471):801-803.

Shaffer, T.L. 2004. A unified approach to analyzing nest success. The Auk 121(2):.526-540.

Shields, W.M. 1984. Barn swallow mobbing: self-defence, collateral kin defence, group defence, or parental care? Animal Behaviour 32(1):132-148.

Snyder, N. 2000. Parrots: Status survey and conservation action plan 2000-2004. IUCN.

105 Snyder, N.F.R., J.W. Wiley, and C.B. Kepler. 1987. The parrots of Luquillo. Los Angeles, CA: West. Found. Vert. Zool.

Snyder, N. F. R., W.B. King, and C.B. Kepler. 1982. Biology and conservation of the Bahama Parrot. Living Bird 19:91-114.

Spoon, T. R., J.R. Millam, and D.H. Owings. 2007. Behavioural compatibility, extrapair copulation and mate switching in a socially monogamous parrot. Animal Behaviour 73(5):815-824.

Stahala, C. 2005. Demography and conservation of the Bahama Parrot on Great Abaco Island. Master's Thesis. North Carolina State University, Raleigh, NC.

Stahala, C. 2007. Breeding Biology of the Bahama Parrot on Great Inagua Island. Report to the Bahamas National Trust.

Stahala, C. 2008. Seasonal movements of the Bahama Parrot (Amazona leucocephala bahamensis) between pine and hardwood forests: implications for habitat conservation. Ornitologia Neotropical (Special Supl) 19: 165-171.

Sturge, R. J., F. Jacobsen, B. B. Rosensteel, R. J. Neale, and K. E. Omland. 2009. Colonization of South America from Caribbean islands confirmed by molecular phylogeny with increased taxon sampling. Condor 111:575–579.

Sutton, G.J., A.J. Hoskins, and J.P. Arnould. 2015. Benefits of group foraging depend on prey type in a small marine predator, the little penguin. PloS one 10(12), p.e0144297.

Tabachnick, B.G., and L.S. Fidell. 1996. Using multivariate statistics. Harper Collins, New York, New York, USA

Templeton, C.N., E. Greene, and K. Davis. 2005. Allometry of Alarm Calls: Black-capped Chickadees encode information about predator size. Science 308:1934-1937.

Thompson, J J. 2004. An age-structured population model of the Puerto Rican Parrot (Amazona vittata). Ornitologia Neotropical 15:289-297.

Travis, S.E., C.N. Slobodchikoff, and P. Keim. 1996. Social assemblages and mating relationships in prairie dogs: a DNA fingerprint analysis. Behavioral Ecology 7(1):95- 100.

Turner, G.F., and T.J. Pitcher. 1986. Attack abatement: a model for group protection by combined avoidance and dilution. American Naturalist 128: 228–240.

Walker, L.N., J.E. Mylroie, and A.D. Walker. 2008. The caves of Abaco Island, Bahamas: keys to geologic timelines. Journal of Cave and Karst Studies 70:108-119.

106 Weatherhead, P. J. 1983. Two principal strategies in avian communal roosts. American Naturalist 121:237–243.

Westcott, D.A., and A. Cockburn. 1988. Flock size and vigilance in parrots. Australian Journal of Zoology 36(3):335-349.

Wiley, J.W., R.S. Gnam, S.E. Koenig, A. Dornelly, X. Galvez, P.E. Bradley, T. White, M. Zamore, P.R. Reillo, and D. Anthony. 2004. Status and conservation of the family Psittacidae in the West Indies. Journal of Caribbean Ornithology 17:94-154.

Wright, S. 1978. Evolution and the genetics of populations. Vol. 4. Variability Within and Among Natural Populations. Chicago IL: University of Chicago Press.

Wright, T. F., and M. Dorin. 1996. Regional dialects in the contact call of a parrot. Proceedings of the Royal Society of London Series B-Biological Sciences 263:867-872.

Young, L.C., E.A. VanderWerf, M.T. Lohr, C.J. Miller, A.J. Titmus, D. Peters, L. Wilson. 2013 Multi-species predator eradication within a predator-proof fence at Ka ‘ena Point, Hawai ‘i. Biological Invasions 15(12):2627-38.

Zoratto, F., D. Santucci and E. Alleva. 2009. Theories commonly adopted to explain the antipredatory benefits of group life: the case of starling (Sturnus vulgaris). Rendiconti Lincei, 20(2): 163-176.

107 BIOGRAPHICAL SKETCH

EDUCATION______August 2008 PhD candidate – Dept of Biological Science; Ecology and Evolution, - present Florida State University, Tallahassee, FL. Social Nesting Behavior of Amazon Parrots. Expected Graduation summer 201\6.

2005 Masters of Science - Major Zoology, Minor Statistics North Carolina State University, Raleigh, NC Demography and Conservation of the Bahama Parrot (Amazona leucocephala bahamensis) on Great Abaco Island. . http://repository.lib.ncsu.edu/ir/bitstream/1840.16/2789/1/etd.pdf

1999 Bachelors of Science – Major Biology University of North Carolina, Chapel Hill, NC

PROFFESIONAL EXPERIENCE______

2014- Panhandle Program Manager, Audubon Florida, Panama City, FL PRESENT •Implement Audubon's NFWF GEBF grant for the Florida Panhandle Shorebird Monitoring and Conservation Program. •Supervise staff of five full time and three part time employees. •Oversee implementation of new rooftop nesting shorebird program including monitoring of sites, bird protection initiatives, and developing management policies with FWC. •Function as liaison between University PIs, State Park PIs and Audubon FL on implementation of contracts and research efforts. •Develop conservation measures at nest sites to protect imperiled species and increase reproductive efforts. •Leading restoration efforts at St George Island Causeway Critical Wildlife Area, which includes selecting and coordinating engineers and implementing restoration efforts. •Coordinate survey and monitoring protocols for shorebird species throughout the panhandle during breeding and nonbreeding periods.

2008- Bahama Parrot Research and Conservation, Florida State University 2014 Tallahassee, FL Project Management •Designed ecological, behavioral and genetic research program for imperiled species. •Conducted field work including vegetation assessments, nest searching, bird capture and handling, bird banding, population surveys and blood collection. •Supervised a team of up to 5 research technicians per year.

108 •Acquired annual funding through NGO, private, in-kind & grant sources. •Responsible for all project and budget reports/updates. •Managed budget for 6 year program. •Compiled and analyzed microsatellite data, ecological data and behavioral data using various statistical programs including ArcMap, Program R and other models. •Responsible for acquiring all necessary permits from US and international agencies for field work and importation of samples. •Reported data, results and conclusions via scientific papers, reports to agencies, popular articles for the general public and oral presentations to general public and other scientists. •Provided management recommendations for species of concern based on habitat, behavioral and genetic results.

Teaching •Conducted class lectures and hands on laboratories for undergraduate biology students. •Lectures included units on conservation biology including the use of policy and laws such as the ESA, NEPA, RAMSAR, Clean Water Act and voluntary actions such as Safe Harbor and conservation easements.lectures and hands on laboratories for undergraduate biology students. •Lectured on conservation biology including the use of policy and laws such as the ESA, NEPA, RAMSAR Clean Water Act and voluntary actions such as Safe Harbor, conservation easements and applied research.

Jan -Sept 2012 Biological Consultant, Bahamas National Trust, Abaco, The Bahamas •Consulted on management of the Bahama Parrot •Drafted Species Management Plan for the Bahamas National Trust •Assessed and evaluated impacts of management and development activities on habitat and the persistence of the species. •Drafted and submitted reclassification proposal to AOU classification and nomenclature committee to redesignate the Cuban Amazon complex as separate species and subspecies based on new evidence from research findings. •Oversaw implementation of management activity by working with local park wardens. •Analyzed data for effectiveness of management strategies such as predator control •Advised and provided technical assistance to Bahamas government, local government and NGO’s on management of avian species and improvement to park plan. •Advised BNT on impacts to species and habitat from proposed development and land use actions. •Conducted outreach events to showcase management efforts and educate about the species and system ecology.

109 •Assisted in banding program for Caribbean Greater Flamingo conservation program through WCS. •Co-Advisor to Master’s student.

2005- US Fish and Wildlife Service, Panama City Field Office, Panama 2008 City, FL. •Regulatory Activities - Implemented Section 7 Endangered Species Act, Section 10 Habitat Conservation Plan, and Migratory Bird Act and Clean water act sec 404. Projects included Communication Towers, Pipelines and Power Corridors. •Reviewed project proposals for NEPA compliance. •Reviewed EIA/EA for projects to determine impact to endangered species and environmentally sensitive systems. •Wrote Biological Opinions for projects involving endangered species with federal nexus through the Corp of Engineers and the US Forest Service. •Carried out recovery actions for Red-cockaded woodpeckers (RCW). •Facilitated voluntary conservation efforts for beach nesting birds and other at risk species with local partners including the State of Florida Wildlife Conservation Commission; •Coordinated mitigation efforts for dissplaced rooftop nesting Least terns. •Coordinated with Corporate, Government and NGO Partners to facilitate management of endangered species including Wood Storks, Bald Eagles, RCWs, shorebirds, and migratory birds. •Managed several habitat improvement grants for uplands restoration project through the 'Partners for Fish and Wildlife' program; •Oversaw research grants for studies on species of conservation interest.

2004- US Fish and Wildlife Service, Student Career Experience Program, 2005 Raleigh Field Office. Raleigh, NC. Participated in all aspects of the USFWS area of responsibility including Ecological Services and Wildlife Refuges through a shadowing and hands on program.

2002- Bahama Parrot Population Viability and Habitat Use Project, North 2005 Carolina State University, Raleigh NC •Designed population viability and habitat use research program for a threatened species. •Drafted species management recommendations based on research results. •Supervised a team of up to 5 research technicians. •Analyzed data using population viability (PVA), ArcGIS and other statistical models. •Advised and provided technical assistance to international government and NGO’s on management of avian species.

2001- Field Technician NCSU/USGS Cooperative Unit, North Carolina State

110 2002 University, Raleigh NC •Assisted in pre-release behavioral training and releases of captive reared Puerto Rican parrots for reintroduction. •Radio tracked released parrots and monitored nests. •Collected data on aquatic and terrestrial food sources of wading birds at Merritt Island National Wildlife Refuge's natural and artificial impoundments. •Conducted wading birds censuses. •Captured and banded shorebirds (Piping Plovers, Snowy Plovers, Wilson’s Plovers, Semipalmated Plovers). •Resighted bands on shorebirds.

2000- Duke University Primate Center, Animal Care Technician; Durham, NC. 2001 •Responsible for care and health of primates in a captive setting. •Assisted in behavioral research project •Monitored animals during pre-release care for reintroducing Lemurs to Madagascar.

1999- Mason Farms Biological Reserve, Research Assistant; Chapel Hill, NC. 2002 •Captured, banded, analyzed crest color and home range of Northern Cardinal (Cardinalis cardinalis).

1999 Bowdoin National Wildlife Refuge, Research Assistant; Malta, MT •Nesting demography of Baird’s sparrow (Ammodramus bairdii). •Collected vegetation data for nesting habitat assessment.

GRANTS_AND AWARDS______2013 Canadian Parrot Conference Grant $3,500 2013 FSU Dissertation Research Improvement Grant $500 2012 FSU International Dissertation Research Fellowship $7,000 2012 Schooner Bay Ltd Grant $5,000 2011 Florida State University Dept of Biological Science Short scholarship $800 2010 Florida State University Dept of Biological Science Loftin travel grant $1000 2009 Florida State University Dept of Biological Science Loftin travel grant $1000 2008 World Parrot Trust Post. Hurricane Ike Assessment and Mitigation $5,000 2007 Disney Wildlife Grant. Great Inagua Parrot Outreach Festival $5,000 2006 Disney Foundation Conservation Grant. Using Breeding Ecology to establish a long term conservation strategy for the Bahama Parrot $10,000 2005 Sigma Xi Aid in Research. Post hurricane assessment of Abaco parrot population $1,000 2005 Amigos de las Aves. Post hurricane assessment of Abaco parrot population $2,000 2004 Parrots International. Post hurricane assessment of Abaco parrot population $4,000 2004 The Nature Conservancy. Genetics of Cuban Parrot complex study $30,500.

111

RELEVANT PUBLICATIONS______PEER-REVIEWED Russello, M. A., C. Stahala, D. LaLonde, K.L. Schmidt, and G. Amato. 2010. Cryptic diversity and conservation units in the Bahama parrot. Conservation Genetics 11(5):1809-1821.

Hayes, W.K., E.D. Bracey, M.R. Price, V. Robinette, E. Gren, C. Stahala. 2010. Population Status of Chuck-will's-widow (Caprimulgus carolinensis) in the Bahamas. Wilson Journal of Ornithology 122(2) 381-384.

Stahala, C. 2008. Seasonal movements of the Bahama Parrot (Amazona leucocephala bahamensis) between pine and hardwood forests: Implications for habitat conservation. Ornitologia Neotropical (Special Supl) 19: 165-171

O’Brien J.J., C. Stahala, G.P. Mori, M.A. Callaham, Jr and C. M. Bergh. 2006. Effects of prescribed fire on conditions inside a Cuban Parrot (Amazona leucocephala) surrogate nesting cavity on Great Abaco, Bahamas. The Wilson Journal of Ornithology 118(4):508–512.

Rivera-Milan ,Frank F; J.A. Collazo; C. Stahala; W.J. Moore; A. Davis; G. Herring; R. Pagliaro; J.L. Thompson; W. Bracey. 2005. Estimation of density and population size and recommendations for monitoring trends of Bahama parrots on Great Abaco and Great Inagua. Wildlife Society Bulletin. 33(3):823-834.

OTHER PUBLICATIONS AND REPORTS Reclassification of the Cuban Parrot (Amazona leucocephala) Complex. Submitted to the American Ornithologists Union. February 2012 Stahala, C. 2012. Bahamas National Trust. Bahama Parrot: Amazona leucocephala bahamensis Species Management Plan. Stahala, C. 2011. Amazona Society UK Stahala, C. 2008. Nature Conservancy Rapid Ecological Assessment, Bird Assessment of the Ragged Island. Stahala, C. 2007. Habitat use and population genetic analysis of the Bahama Parrot on Great Abaco and Great Inagua Islands, Bahamas. Report to The Nature Conservancy, Bahamas Country Program. FWS. 2007. Biological Opinion on impacts to the Bald Eagle (Haliaeetus leucocephalus) FWS Panama City Ecological Services Field Office. Stahala, C. 2004. Bahama Parrot. Psittascene 16(4): 2-5.

112