|UFDC Home||myUFDC Home | Help|
This item has the following downloads:
1 PATTERN AND PROCESS IN URBAN BIRD COMMUNITIES: WHAT MAKES THE NORTHERN MOCKINGBIRD AN URBAN ADAPTER? By CHRISTINE M. STRACEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2010
2 2010 Christine M. Stracey
3 To Tim
4 ACKNOWLEDGMENTS I thank Steve Daniels, Tim Richard, Rebecca Sandidge, Amelia Savage, Judit Ungvari Martin, Rachel Hanauer, Laura Levins, and Pepe Clavijo for assistance with data collection. I also thank Jill Jankowski, Gustavo Londoo, Judit Ungvari Martin, Carrie Vath, and Oscar Gonzalez for help with all aspects of the development of the research. Lunch Bunch and Birdlab participants also provided constructive feedback on numerous occasions. Ben Bolker and Mary Christman provided assistance with statistical analyses and Corey Tarwater assisted with Program MARK. My committee members, Doug Levey, Dave Steadm an, Michael Avery, and Katie Sieving, provided valuable feedback on the design, execution, and completion of this research. My advisor, Scott Robinson, was helpful, supportive, and patient beyond measure and I enjoyed every one of our interactions. I am also indebted to all the private landowners and businesses that permitted us to work on their property. The management at the Oaks Mall and Butler Plaza were very accommodating and this work would not have been possible without their support. Funding was provided by student research awards from the Florida Ornithological Society and the American Ornithologists Union to C.M. Stracey, a Doctoral Dissertation Improvement Grant from the National Science Foundation (grant # 0709646) to C.M. Stracey, and the K atharine Ordway Endowment of the Florida Museum of Natural History. My parents also provided limitless emotional support throughout the entire process. I also thank my husband, Tim Richard, for helping with data collection and general sanity.
5 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES ........................................................................................................ 10 LIST OF FIGURES ........................................................................................................ 10 1 GENERAL INTRODUCTION .................................................................................. 14 2 DOES NEST PREDATION SHAPE URBAN BIRD COMMUNI TIES? ..................... 21 Introduction ............................................................................................................. 21 Methods .................................................................................................................. 23 Censuses ......................................................................................................... 23 Nest Monitoring ................................................................................................ 26 Results .................................................................................................................... 29 Avian Nest Predator Abundance ...................................................................... 29 Community Composition .................................................................................. 29 Nesting Success ............................................................................................... 31 Discussion .............................................................................................................. 33 Nest Predator Abundance ................................................................................ 33 Urban Bird Community Composition ................................................................ 36 Patterns of Nest Predati on in Urban Habitats ................................................... 38 Conclusions ............................................................................................................ 41 3 CATS AND FAT DOVES: RESOLVING THE URBAN NEST PREDATOR PARADOX .............................................................................................................. 54 Introduction ............................................................................................................. 54 Methods .................................................................................................................. 58 Nest Predation Rates ....................................................................................... 58 Nest Predator Identification .............................................................................. 59 Results .................................................................................................................... 61 Nest Predation Rates ....................................................................................... 61 Nest Predator Identification .............................................................................. 61 Discussion .............................................................................................................. 62 Management Recommendations ...................................................................... 67 Conclusions ...................................................................................................... 68
6 4 FOOD LIMITATION IN NESTING URBAN AND NONURBAN MOCKINGBIRDS: RESOURCE MATCHING OR MISMATCHING? ....................... 75 Introduction ............................................................................................................. 75 Methods .................................................................................................................. 78 Study System and Territory Density ................................................................. 78 Number of Fledglings Produced Per Successful Nest and Per Hectare ........... 79 Clutch Size ....................................................................................................... 80 Hatching Suc cess and Fledging Success ......................................................... 80 Nestling Mass ................................................................................................... 81 Nestling Provisioning Rates, Average Food Size, and Food Type ................... 81 Adult Mass ........................................................................................................ 83 Results .................................................................................................................... 83 Number of Fledglings Produced Per Successful Nest and Per Hectare ........... 83 Clutch Size ....................................................................................................... 84 Hatching Success and Fledging Success ......................................................... 84 Nestling Mass ................................................................................................... 84 Nestling Provisioning Rates, Average Food Size, and Food Type ................... 84 Adult Mass ........................................................................................................ 85 Discussion .............................................................................................................. 85 5 IS AN URBAN ADAPTER, THE NORTHERN MOCKINGBIRD, MORE PRODUCTIVE IN URBAN HABITATS? .................................................................. 96 Introduction ............................................................................................................. 96 Methods .................................................................................................................. 99 Data Collection ................................................................................................. 99 Statisti cal Analyses ........................................................................................ 101 Source/Sink Threshold ................................................................................... 103 Results .................................................................................................................. 104 Nesting S eason Initiation ................................................................................ 104 Nesting Productivity ........................................................................................ 104 Estimated Survival .......................................................................................... 104 Decision Rules Governing Site Fidelity .......................................................... 105 SourceSink Analyses .................................................................................... 106 Discussion ............................................................................................................ 106 6 CONCLUSIONS ................................................................................................... 121 APPENDIX A COMMUNITY CENSUS DATA ............................................................................. 124 B MODEL RESULTS ............................................................................................... 127
7 LIST OF REFERENCES ............................................................................................. 130 BIOGRAPHICAL SKETCH .......................................................................................... 139
8 LIST OF TABLES Table page 2 1 Model selection results for the top six logistic exposure models of daily survival for the Northern Mockingbird (2005 2006), Brown Thrasher (2005 2006), and Northern Cardinal (2004 2006) ...................................................... 44 2 2 Daily survival rates (95% confidence limits [CL]) in the years indicated, number of nests monitored, and the statistical significance of landuse on daily survival rates. ............................................................................................. 46 3 1 Average proportion of different types of ground cover for four habitat types based on aerial images of each study site. ......................................................... 70 3 2 Model selection results (top six models) for the logistic exposure models of daily survival for the Northern Mockingbird. ........................................................ 70 4 1 Average (+/ standard error [ SE] ) number of fledglings produced per successful nest, proportion of eggs laid that hatched, proportion of nestlings that fledged, nestling mass at age six days, adult female and male mass, and the results of significance tests for each response variable. ............................... 89 4 2 The average (+/ SE) number of feeding trips in 2008 per nestling per hour at nestling age three, six, and ninedays in urban and non urban land use. .......... 90 4 3 The mean (+/ SE) percentage of feeding t rips consisting of fruit, larvae, and Orthoptera to nestlings age six days in urban and nonurban landuse. ............ 90 5 1 Statistical significance of generalized linear mixed models (GLMM) for average start date, total number of fledglings produced per female per year, number of successful nests per female per year, number of nesting attempts per female per year, and the proportion of birds on their territory the entire breeding season. .............................................................................................. 111 5 2 The total number of nestlings banded, the number that returned as breeding adults, and the proportion of banded nestlings that returned to breed in each habitat. .............................................................................................................. 111 5 3 Model selection results of Cormack Jolly Seber models used in program MARK. .............................................................................................................. 112 5 5 Estimation of the source/sink threshold in each habitat under three diff erent adult and juvenile survival scenarios and the number of female fledglings per female per year predicted from the GLMM. ...................................................... 113
9 A 1 Total number of individuals detected in each habitat during 100m, fixedradius point counts of each species along with their mass (Dunning 1993) and nesting guild (Ehrlich et al. 1988). ............................................................. 124 B 1 Statistical significance of models testing reproductive parameters related to food availability from Chapter 4. ....................................................................... 127
10 LIST OF FIGURES Figure page 2 1 The location of study sites for the nesting study on images modified fro m Google Earth mapping service. ....................................................................... 47 2 2 Average (+/ standard error [ SE] ) number of total avian nest predators detected per census point (five minute, 100m fixed radius) and the average n umber of each species detected. ...................................................................... 48 2 3 Average (+/ SE) number of individuals of introduced species detected per point in each habitat type based on five mi nute, 100m fixed radius poi nts ........ 49 2 4 Average (+/ SE) number of individuals detected per point in each habitat type based on five minute, 100m fixed radius points. ......................................... 50 2 5 Average (+/ SE) number per census point (five minute, 100m fixed radius) of individuals of all species combined detected in each habitat based on mass and nesting guild. ............................................................................................... 51 2 6 Daily survival rate (+/ 95% confidence limits [ CL ] ) of the Northern Mockingbird at the incubation and nestling stage in each habitat type. .............. 52 2 7 Daily survival rate (+/ 95% CL) of the B rown Thrasher in 2005 and 2006 in each habitat type. ............................................................................................... 53 3 1 Location of study sites in and around Gainesville, FL. ........................................ 72 3 2 Daily survival rate (probability of a nest escaping predation; +/ 95% C L ) of Northern Mockingbird nests in different habitat types between 2005 and 2009. .................................................................................................................. 73 3 3 The identity of nest predators in ur ban and nonurban habitats. ......................... 74 4 1 The average (+/ SE) number of mockingbird pairs per hectare in four habitat types. .................................................................................................................. 91 4 2 Ave rage number of fledglings produced per hectare. ......................................... 92 4 3 Average clutch size in four habitat types. ........................................................... 93 4 4 Average number of feedi ng trips per hour per nestling in each habitat type. ...... 94 4 5 Average food size relative to adult bill length delivered to six day old nestlings in urban and nonurban landuse in 2007 2008. ................................. 95 5 1 The average (+/ SE) ordinal date of the completion of a females first clutch in each habitat type. ......................................................................................... 114
11 5 2 The average (+/ SE) number of total fledglings produced per female per year in each habitat type. ................................................................................. 115 5 3 The average (+/ SE) number of successful nests per female per year in each habitat type. ............................................................................................. 116 5 4 A SE) in each habitat type estimated in program MARK. .............................................................................................................. 117 5 5 The proportion of pairs (+/ SE) that occupied their territory the entire breeding season by habitat type. ...................................................................... 118 5 6 The probability (+/ SE) of a female returning to a site following her success in the previous season i n each habitat type. ..................................................... 119 5 7 The probability (+/ SE) of a female returning to a site following the number of nests that fledged young in the previous season in each habitat type. ............. 120
12 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PATTERN AND PROCESS IN URBAN BIRD COMMUNITIES: WHAT MAKES THE NORTHERN MOCKINGBIRD AN URBAN ADAPTER? By Christine M. Stracey May 2010 Chair: Scott K. Robinson Major: Zoology Urbanization is one of the leading causes of species endangerment in the United States; however, certain species thrive in urban habitats. I tested the hypotheses that an increase in food resources and a decr ease in nest predation explain the success of native avian species that occur at high densities in urban habitats (i.e. urban adapters). I compared populations of Northern Mockingbirds in parking lots and residential neighborhoods to populations in pastur es and wildlife preserves during 20052009 in Florida. Data support the hypothesis that urban nest predation is lower, on average, than nonurban nest predation. Avian nest predators, however, are more abundant in urban habitats, leading to a mismatch between predation rates and predator abundance. Using video cameras on nests, I identified cats as the dominant urban nest predator and Coopers Hawks as the dominant nonurban predator. The most abundant avian nest predators accounted for none of the recorded predation events. Coopers Hawks were present at the urban sites, yet were not recorded as urban nest predators. I hypothesize that urban Coopers Hawks have undergone a dietary shift as a result of abundant alternate prey. I conclude that changes in nest predator community
13 composition and dietary shifts of predators contribute to the urban nest predator paradox. I found no significant differences between urban and nonurban habitats in the quantity or quality of fledglings produced per nesting attempt. Number of fledglings produced per hectare, however, was significantly higher in urban habitats. The higher abundances of mockingbirds and higher production of fledglings per unit area may reflect greater food abundance in urban areas, but individual s were distributed among habitats such that per capita food availability was roughly equal. Productivity of urban mockingbird pairs exceeded that of nonurban pairs and more than offset estimated adult mortality. Based on this research urban areas do not represent sink habitat for mockingbirds. By comparing populations of a native species that is successful in urban areas, as well as in natural habitats, we can begin to understand the factors that allow some species to exploit these drastically modified e cosystems and anticipate how these communities will change with time.
14 CHAPTER 1 GENERAL INTRODUCTION The goal of urban ecology is to understand how nature works in places where people live and work. As the area and density of urban environments increase, so too does the recognition amongst ecologists that urban environments represent unique opportunities to study ecosystems, communities, populations, and individual organisms in novel landscapes. While many species, termed urban avoiders, are adversely affected by urbanization, some species (urban adapters) are more abundant in urban and suburban habitats than in more natural areas and still others (urban exploiters) are essentially commensals of humans and only found in close association with people (B lair 1996, Shochat et al. 2006). The fact that certain species, including some native species, are able to use and even flourish in urban settings highlights the potential importance of urban ecosystems in the conservation of biodiversity. Urban ecosystems will never replace or surpass the value of pristine nature reserves for the conservation of biodiversity, but they do provide some unique conservation opportunities (Miller and Hobbs 2002, Rosenzweig 2003, Miller 2005). In particular, biodiversity in ur ban habitats provides vital opportunities to address the extinction of experience, in which people living in cities are disconnected from nature and thus do not value it (Miller 2005). If conservation biologists hope to conserve global biodiversity, then the general public needs to support conservation efforts. Because over 50% of the human population now resides in urban areas (Grimm et al. 2008), in order to achieve broad general support for conservation efforts, people living in cities must value nat ure. Furthermore, recent research has identified numerous links between individual human well being and a connection with nature. For example,
15 time in nature may be as effective in treating ADHD as medication (Taylor and Kuo 2009) and hospital patients w ith a view of nature recovered more quickly from surgery than patients without such a view (Ulrich 1984). Interaction with nature may even shape our values; people who had experimentally experienced immersion in nature were more likely to express values t hat were prosocial and community focused, whereas people who were not immersed in nature were more likely to express self focused values (Weinstein et al. 2009). Importantly, the psychological benefit people derive from urban greenspace appears to increase with increasing biodiversity of the greenspace (Fuller et al. 2007). Given the importance of humannature interaction, for both human well being and for conservation, it is important that ecologists understand the ecology of urban ecosystems. For urban planners to fully exploit the potential of urban ecosystems and promote urban biodiversity, ecologists must first understand the mechanisms that structure urban populations and communities. Urban bird communities, in particular, are characterized by four main patterns: a reduction in species richness, an increase in density of successful urban species, an increase in avian biomass, and an increase in abundance of introduced species (Marzluff et al. 2001, Chase and Walsh 2006). While it is critical that w e understand what changes in the urban environment exclude numerous species, it is also necessary to understand the ecology of urban adapters and exploiters. This research, therefore, focuses on urban adapters, and I attempt to determine the characteristi cs of these species, as well as the characteristics of urban habitats, that enable them to be successful in urban and suburban environments.
16 The two traditional hypotheses that researchers have invoked to explain the success of urban adapters and exploiter s involve alterations in both topdown and bottom up factors that regulate population size. The food enhancement hypothesis posits that urban areas have more abundant and more predictable food resources than nonurban areas (Marzluff et al., 2001, Shochat et al., 2004a, Faeth et al., 2005, Shochat et al., 2006). The role of food in population regulation has long been discussed (e.g. Lack 1966) and most populations that are supplemented with food exhibit an increase in density (Boutin 1990). From a topdown perspective, researchers have also proposed that birds nesting in urban habitats experience a reduction in nest predation rates (Gering and Blair 1999, Marzluff et al. 2001, Shochat et al. 2004a, Faeth et al. 2005, Shochat et al. 2006). Nest predation is the leading cause of reproductive failure in most passerine species and can limit avian populations (Martin 1993, Newton 1998). An alternative to these hypotheses is the idea that urban areas are functioning as ecological traps (i.e. attractive sinks, G ates and Gysel 1978) for urban adapters in that individuals are attracted to cities and towns, but can not meet their requirements there and thus produce relatively few, poor quality offspring. For example, the additional anthropogenic food available in urban habitats may attract dispersing adults, but the quality of these resources may be low, especially for developing nestlings. Likewise, stable nesting substrates, such as ornamental bushes, may be attractive to breeding birds, but may be vulnerable to introduced nest predators (Schmidt and Whelan 2001, Borgmann and Rodewald 2004). It is therefore important as we attempt to promote urban biodiversity to know whether urban areas are acting as ecological traps and are drawing individuals away from more suitable rural habitats.
17 To investigate what causes certain urban species to become numerically dominant in the community and evaluate the above hypotheses, I focused on the Northern Mockingbird ( Mimus polyglottos ). The mockingbird is a highly successful urban species, particularly in the southeast, where they can reach an average of 3.0 individuals per census point in urban areas in north Florida (Chapter 2). Mockingbirds are also found in nonurban areas such as pasture, dunes, and scrub, where they reach an average of 1.0 individuals per census point (Chapter 2). Mockingbird nests are easy to locate and adults can be captured with mist nets and color marked. In addition, the public can readily identify with mockingbirds, the state bird of five states, which is important for obtaining permission to work in parking lots and peoples backyards. By working in areas where I had frequent contact with the public I was able to discuss my research and raise awareness of wildlife in cities. A comparative study of populations in the two habitats provided an opportunity to examine factors that have allowed the mockingbird to become so abundant in urban settings and further our understanding of the mechanisms that structure urban bird communities. I compared mocki ngbird populations in urban areas with populations in nonurban areas in and around Gainesville, Florida. I worked in five urban study sites (two parking lots and three residential neighborhoods) and four nonurban study sites (two pastures and two scrub habitats). To examine the role nest predation plays in structuring urban bird communities, I first used census data to establish patterns of urban bird communities in relation to factors (e.g., body size and nesting guild) that may be correlated with nest predation (Chapter 2). I also investigated nest predation rates in urban and nonurban habitats for five common species. These patterns highlighted an
18 unusual feature of urban areas: nest predation rates in urban habitats were lower than in nonurban habitats, despite a higher abundance of avian nest predators in urban habitats. This mismatch between nest predation rates and nest predator abundance is the urban nest predator paradox. I attempt to resolve the urban nest predator paradox in Chapter 3 by using video cameras to identify the main nest predators in urban and nonurban habitats. Without knowing the identity and importance of the nest predators, we can not begin to unravel this paradox. I found evidence for changes in urban predator communi ty composition, dietary shifts of urban predators and nest defense behavior of prey, which could explain the urban nest predator paradox. In addition, cats were the main urban nest predator, which has implications for management. In Chapter 4 I focused on t he food enhancement hypothesis, which posits that an increase in food resources in urban areas leads to higher reproductive success and hence higher densities of birds. Mockingbirds mostly feed their nestlings arthropods and during the winter they primari ly eat fruits (Derrickson and Breitwisch 1992). Direct measures of insect availability are difficult to obtain for mockingbirds because they forage on the ground, in flight, and in trees. I therefore did not directly quantify food availability, but rather measured reproductive parameters that are correlated with food resources including fledgling quantity and quality. I found no significant differences in the quantity or quality of fledglings produced per successful nest in urban and nonurban habitats. The ultimate determinant of the success of a population is the ability of that population to produce enough offspring to maintain its population size. Both food and
19 predation directly influence the number of fledglings produced and these factors also like ly influence the selection of breeding sites. In Chapter 5 I compared per capita production of fledglings and apparent adult survival, and estimated the source/sink threshold (the number of fledglings needed to offset adult mortality to maintain a stable population) for urban and nonurban populations. In addition, I examined the role that nest success plays in determining the decision rules used by individuals to return to a study site between years. I found that urban habitats produce fledglings in exc ess of the estimated number required to offset adult survival under a variety of adult and juvenile survival estimates. Nonurban habitats, on the other hand, were consistently below the estimated source/sink threshold, possibly as a result of increased adult dispersal. Understanding the mechanisms that underlie the patterns of community structure observed in urban environments is critical for predicting how species will respond to urbanization. Based on these results, we can predict that large species that can deter predators and small species that nest in safe locations will be more likely to persist in urban habitats, but that communities will continue to change as predator communities change. To derive a complete understanding of the consequences of ur banization on wildlife, it is important not only to study species that are negatively affected by urbanization, but also species that seemingly benefit. This research begins to tease apart the importance of food resources and predation rates in the population regulation of a successful urban species. Understanding the mechanisms that drive urban community structure will allow us to begin generalizing across cities and regions instead of being confined to case studies.
20 Furthermore, to study mockingbirds in residential areas and parking lots, I needed the permission of many property owners and businesses. I built a network of properties on which I had permission to work by distributing fliers and going door to door. This was an important aspect of my research because it brought an awareness of nature and science to peoples backyards. Raising the awareness of the general public, especially children, about the nature that surrounds them, even in parking lots and their backyards, is a first step in bringing science and an appreciation for nature to everyone.
21 CHAPTER 2 DOES NEST PREDATION SHAPE URBAN BIRD COM MUNITIES? Introduction The ways in which bird communities change in response to urbanization have been well documented (Marzluff et al. 2001, Shochat 2004). Four general patterns have emerged: (1) a reduction in species richness, (2) a decrease in evenness (urban bird communities have fewer but more abundant species), (3) an increase in abundance of large species, and (4) an increase in abundance of nonnat ive species. While these patterns have been documented in various urban areas, few studies have investigated the underlying mechanisms causing such patterns. Results to date emphasize two competing, but not mutually exclusive, mechanisms increased food av ailability and decreased predation pressure in urban habitats (Shochat et al. 2004, 2006). Here we explore the role of nest predation in structuring urban bird communities. Our working hypothesis, the urban nest predator hypothesis, is that altered communi ties of nest predators in urban habitats affect which species do and do not thrive. We refer to three groups of species: urban exploiters species that are essentially human commensals urban adapters species that are abundant in urban areas but also maintain populations in more natural habitats, and urban avoiders species that reach their highest densities in natural habitats (Blair 1996, Shochat et al. 2006). Specifically, we predict that (1) urban adapters have lower rates of nest loss to predators in urban habitats than those same species in nonurban habitats and (2) species missing from urban communities will have lifehistory traits (nest defense and placement) that make them more vulnerable to the kinds of nest predators that thrive in urban areas. Unfortunately, we are not able to test the reciprocal version of our first prediction, that
22 urban avoiders have higher rates of nest predation in urban areas because we did not locate enough nests of urban avoiders in urban habitat. Nest predation has been hypothesized to be a major force structuring bird communities (Martin 1993) and a critical determinant of the viability of populations in fragmented habitats (Robinson et al. 1995). It is now widely accepted that humancaused alterations of communities of nest predators can have cascading effects on community structure in remnant patches of natural habitat in humandominated landscapes (Chalfoun et al. 2002, Heske et al. 1999). Indeed, one of the primary changes in fragmented habitats is the loss of top predators and the resulting increase in generalist mesopredators, many of which incidentally eat bird eggs and nestlings (Soule et al. 1988, Crooks and Soule 1999). Nesting success of many birds in fragmented habitats is so low that habitat fragments have been hypothesized to be population sinks (Robinson et al. 1995) or even ecological traps that attract nesting birds but which produce too few young to compensate for adult mortality (Heske et al. 1999, Weldon and Haddad 2005). As a result, bird communi ties in humanaltered landscapes (e.g. forest fragments within an agricultural matrix) are dominated by medium sized species that can defend their nests against predators and species that nest in sites that are inaccessible to all but a small subset of nes t predators (e.g., cavities) (Whitcomb et al. 1981, Brown and Sullivan 2005). If there are comparable differences in nest predator communities in relation to urbanization, then it is possible that those differences in predation pressure could be driving the distribution and abundance of avian species.
23 To date, however, studies of urban nest predation have produced mixed results. For example, Gering and Blair (1999) found that rates of nest predation on artificial nests decreased along an urban gradient and hypothesized that urban areas may actually offer a refuge from nest predation, at least for some species. Because urban bird communities often have high populations of avian nest predators, this result has been called a predator paradox (Shochat et al. 2006). Other studies have found no relationship between urbanization and nest predation rates (e.g., Melampy et al. 1999, Haskell et al. 2001), whereas still others have found an increase in nest predation along the urban gradient (e.g., Jokimki and Huhta 2000, Thorington and Bowman 2003). Such increased nest predation rates may be especially prevalent in species that live in remnant patches of native forest within urban areas (Borgmann and Rodewald 2004). The goals of this study were to compare nesting success and composition of avian bird communities, including species that regularly act as nest predators (e.g., corvids) and introduced species, in urban and nonurban habitats. We predicted that (1) avian species that are nest predators would be more abundant in urban habitats (Marzluff et al. 2001, Sorace 2002, Shochat et al. 2006), (2) species with lifehistory traits that provide effective defenses (e.g. l arge body size, aggressive nest defense, protected nest sites) against avian predation would be more abundant in urban habitats, and (3) urban adapters would experience lower nest predation rates in urban habitats. Methods Censuses To address the hypothesis that nest predation rates account for many of the changes in urban bird communities and test our predictions, we censused urban and nonurban bird communities and measured nest predation rates in a variety of habitats
24 in Florida, USA. The goal of our censusing was to characterize the bird communities of habitats believed to be representative of norther n and central Florida. Censuses were conducted in the spring and summer of 2004. We used fiveminute, 100m fixed radius point counts, during which we recorded species and number of individuals within the 100m radius. We chose 100m rather than the traditi onal 50m (Ralph et al. 1995) because it increased sample sizes for habitats for which we had relatively few census points and because detectability was generally high in all habitats with the exception of some broadleaf forests. The comparatively dense vegetation of hammock and floodplain forests may have decreased detectability, especially of small species. However, this bias against detecting small species in forested habitat makes our results showing an increase in small bodied birds in these habitats (s ee below) a conservative test of our hypothesis. All censuses were conducted after 1 May, by which time all breeding species had returned. Censusing began at sunrise and ended no later than 1030 H. Within each census area, points were located at least 25 0 m apart. Most points were located within 100 km of Gainesville, but additional censuses were conducted in the Archibald Biological Station and the MacArthur Ranch, in coastal scrub near St. Augustine and Jacksonville, and in urban settings in Jacksonvill e, St. Augustine, and Sarasota. Points were grouped into the following habitat categories (number of points in parentheses): parking lot (42), residential neighborhood (16), urban forest fragment, as defined by intact canopy and understory (11), pasture (15), nonurban scrub/secondary growth (52), and nonurban forest (48). We calculated the average number of detections per point in the different habitats for each species for which we had nesting data. The
25 average number of avian nest predators detected per point was also calculated. We considered the following species nest predators: Fish Crow ( Corvus ossifragus ), American Crow ( Corvus brachyrhynchos ), Common Grackle ( Quiscalus quiscula), Boat tailed Grackle ( Quiscalus major ), Blue Jay ( Cyanocitta cristata) and Redshouldered Hawk ( Buteo lineatus ). Estimates of the abundance of avian nest predators are conservative because we counted flocks of the same species as one detection in the analyses. Avian nest predators were censused at the same time as the rest of the species. While this might result in an underestimate of their true abundance, due to their large home ranges, we feel this is justified because we are comparing the relative abundance of these predators across different habitats and not directly com paring their abundance with that of other species. We also calculated the average number of detections per point of introduced species. We included the following species in this category: House Sparrow ( Passer domesticus ), House Finch ( Carpodacus mexicanus ), Eurasian Collared Dove ( Streptopelia decaocto), European Starling ( Sturnus vulgaris ), Monk Parakeet ( Myiopsitta monachus ), and Rock Pigeon ( Columba livia ). To test for an effect of habitat on the average number of detections of introduced species, avian nest predators, and the different nesting species we performed Kruskal Wallis tests because our data were not normally distributed. If there was a significant effect of habitat on the average number of detections, we then conducted post hoc comparisons of the habitats using Tukeys pairwise comparisons with alpha levels adjusted for multiple comparisons. To examine how the abundance of birds of different body masses is affected by urbanization, we calculated the average number of individuals detected per point in the
26 following size classes: 014.9 g, 1539.9 g, 4099.9 g, 100199.9 g, 200+ g. These size classes were chosen based on natural breaks in the data and in an effort to generate a normal distribution for body mass. Body mass estimates for each spec ies were taken from Dunning (1993). In addition, we examined how nesting guild (as defined by either open cup or nests placed in cavities and enclosed locations) differed according to habitat by calculating the average number of enclosed nesting individual s detected and the average number of opencup nesting individuals detected in each size class. To analyze the effect of habitat, mass, and nesting guild on the average number of detections we performed a quasi Poisson regression (SAS 9.1). Because there were a few treatments where we observed no individuals (e.g., we did not detect any individuals of mass 100199.9 g that were cavity nesters in nonurban forests), we added 0.001 to each data point to allow the model to converge. We checked the effect of thi s by performing the same analysis with the addition of 0.01 and comparing the model parameters for each. In both cases we obtained very similar Fstatistics and the same P values; thus the model did not seem to be sensitive to this addition. Because we are also interested in whether any differences in abundance reflect the urban versus nonurban landuse dichotomy, we tested the effect of landuse category (urban, urban forest fragment, and nonurban) in the same manner. Nest Monitoring The nesting study was conducted in 2004 2006 in areas in and around Gainesville, FL (Fig. 2 1). The following study sites were used followed by the years for which we have data from each site: the University of Florida campus (2004), downtown Gainesville (2004 2005), urban fo rest fragments within Gainesville (2004), two wildlife
27 preserves (2005 2006), three residential neighborhoods (2005 2006), two parking lots (2005 2006), and three pastures (2006). We searched each study site for nests of all common species. At least 15 nests of the following species were located and included in analyses: Northern Mockingbird ( Mimus polyglottos; N = 585) Brown Thrasher ( Toxostoma rufum ; N = 175), Northern Cardinal ( Cardinalis cardinalis ; N = 149), Mourning Dove ( Zenaida macroura; N = 55 ), House Finch ( Carpodacus mexicanus ; N = 60 ), Eurasian Collared Dove ( Streptopelia decaocto; N = 15 ), and Loggerhead Shrike ( Lanius ludovicianus ; N = 15 ). Nests were monitored approximately every third day following discovery. Daily survival rates were calc ulated using Shaffers logistic regression model (Shaffer 2004, SAS 9.1). Because we are interested only in predation rates, we only considered nests that failed due to predation as unsuccessful. Therefore, the daily survival rates we report are not a me asure of overall nest survival, but rather the probability of a nest escaping predation. For some species we had enough data to model survival based on habitat type (e.g., parking lot, residential, etc.), but for other species for which we found fewer nest s we combined habitat types and compared urban landuse, including parking lot, residential, town, and campus, to nonurban landuse, including pasture and wildlife preserve. Data for the Mourning Dove and Loggerhead Shrike, for which we have relatively small sample sizes, were pooled across all three years to compare survival rates between urban and nonurban landuse. To test for an effect of landuse on daily survival we modeled survival based on urban versus nonurban landuse using the logistic regres sion model. Because House Finch and Eurasian Collared Dove nests were not located at any of the nonurban study sites, we calculated the daily survival
28 rate in urban landuse (parking lots, residential neighborhoods, downtown) and b ecause of low sample siz e we used data pooled from 2004 2006. For the Northern Mockingbird and Brown Thrasher we had at least 15 nests from 2005 2006 combined in each habitat, allowing us to model daily survival rates as a function of urban versus nonurban landuse and also as a function of habitat type (e.g., parking lot, residential, etc). In 2004 we only had data from urban habitat types, predominantly from campus, and so we report 2004 daily survival rates separately. For both mockingbirds and thrashers we used informationt heoretic methods (Burnham and Anderson 2002) to evaluate candidate models. We compared models that included the following variables (with and without interactions): year, nest stage (incubation, nestling), and habitat type or land use, as well as a model t hat assumed constant survival (Table 21 ) For the Northern Mockingbird we proceeded to compare the confidence intervals for estimates of daily survival between the five habitat types for each nest stage and then for urban and non urban land use at each nest stage. For the Brown Thrasher we compared confidence intervals of daily survival rates between habitat types and landuse, but not between nest stages. We examined confidence intervals for 2005 and 2006 separately. We did not include parking lot or town in our comparisons for thrashers because sample sizes for those two categories were limited to three and five nests, respectively. The Northern Cardinal was the only species for which we had data from urban forest fragments (2004 only), which allowed us t o compare daily survival in urban forest fragments with urban and nonurban landuse. We compared candidate models that
29 included landuse, year, and nest stage (with and without interactions) and a model that assumed constant survival (Table 21) Results Avian Nest Predator Abundance Avian nest predators were detected far more often in parking lots and residential neighborhoods than in urban forest fragments, pastures, nonurban scrub, and nonurban forests (Fig. 2 2; 2 = 73.68, df = 5, P <0.0001). Becaus e we recorded flocks as one detection the relative abundance of avian predators is even greater, given that American Crows typically traveled in groups of 36 individuals, Fish Crows typically occurred in groups of 10 30 individuals, and both species of gr ackles traveled in flocks of up to 30 individuals. However, all six species of predators do not mirror the general pattern (Fig. 22). There was no difference in the average number of detections of American Crows ( 2 = 9.17, df = 5, P = 0.10) or Redshould ered Hawks ( 2 = 4.14, df = 5, P = 0.53) in different habitats. There was a significant effect of habitat on the average number of detections of Blue Jays ( 2 = 12.16, df = 5, P = 0.03), however none of the Tukeys pair wise comparisons were significant at the corrected alphalevel, so this trend is weak at best. The remaining species, Fish Crow ( 2 = 62.31, df = 5, P < 0.0001), Common Grackle ( 2 = 38.23, df = 5, P < 0.0001), and Boat tailed Grackle ( 2 = 61.03, df = 5, P < 0.0001), all show the same pattern: they were significantly mo re abundant in parking lots and residential neighborhoods based on post hoc comparisons. Community Composition Urban and nonurban communities differed greatly in detections of most species (Appendix A). The effect of habitat on introduced birds is especially pronounced (Fig. 2-
30 3; 2 = 111.12, df = 5, P < 0.0001), with these species restricted to parking lots and residential areas, with the exception of the Eurasian Collared Dove, which also occurred around barns in pastures. Even the House Sparrow ( Passer domesticus ) and Eurasian Starling ( Sturnus vulgaris ) were not recorded in agricultural areas (Fig. 23). Other urban adapters included the Northern Mockingbird (Fig. 24a; 2 = 93.66, df = 5, P < 0.0001) and Mourning Dove (Fi g. 24b; 2 = 49.59, df = 5, P < 0.0001). Brown Thrashers and Loggerhead Shrikes were also detected in urban habitats, but their detection probabilities were so low as a result of their extremely infrequent singing that we could not test for any patterns. Northern Cardinals (Fig. 24c; 2 = 40.59, df = 5, P < 0.0001) and Carolina Wrens ( Thryothorus ludovicianus ) (Fig. 24d; 2 = 27.25, df = 5, P < 0.0001) tended to be slightly more common in nonurban habitats, however both species were abundant in all habi tats with any woody vegetation. The lifehistory traits of urban and nonurban birds also differed considerably (Fig. 2 5). There was a significant threeway interaction between habitat, mass, and nesting guild (F20,1820 = 3.39, P < 0.001) and between landuse, mass, and nesting guild (F8,1820 = 6.34, P < 0.0001). Small bodied cupnesting individuals were far more abundant in nonurban habitats (Fig 25). Conversely, largebodied individuals many of which are also nest predators, were far more abundant in urban habitats (Fig. 25). In the smallest size class (0 14.9 g), open cup nesters made up roughly half of the urban detections whereas in nonurban habitats, nearly 100% were open cup nesters (Fig. 25). Interestingly, the urban forest fragments appear more similar to the two nonurban habitats than to the urban habitats. The same pattern holds for the next size class (1539.9 g) there were more cavity nesting individuals detected in urban areas than in
31 nonurban areas, with urban forest fragments falling roughly in between. Indeed, the only abundant urban birds of less than 40 g were the House Sparrow and Carolina Wren, both of which place most of their nests in human structures in urban areas (Robinson and Stracey, unpubl. data). Nesting Success A total of 1,044 nests were monitored during the three years of the study. For both the Mourning Dove ( 2 = 3.43, df = 1, P = 0.06) and Loggerhead Shrike ( 2 = 1.32, df = 1, P = 0.25) the daily survival rate was higher in urban landuse (parking lots, residential, campus, town), yet neither of these differences was significant (Table 22 ). Among the introduced species restricted to urban landuse, the daily survival rate of House Finches (99.68%) was very high (Table 22 ). The survival rate of the Eurasian Collared Dove was similar to the survival rate of the Mourning Dove in urban landuse (Table 22 ). We have no data for the small bodied open cup nesters in either urban or nonurban landuse. During 2005 2006, w e monitored 461 Northern Mockingbird nests and 144 Br own Thrasher nests for an effective sample of 5863 and 1447, respectively. For the mockingbird at the landc < 2 all included landuse, year, and nest stage and differed in the interactions included between these three variables (Table 21 ). In urban landuse there were no differences in survival rates between the incubation and nestling stage in either 2005 or 2006 (Table 22 ). However, in 2006 in nonurban landuse the nestling stage had lower rates of nest survival t han the incubation stage. Survival rates were consistently lower in nonurban landuse relative to urban landuse across years and nest stage, although in some cases the
32 95% confidence intervals of the daily survival rates overlapped slightly. At the habit at level, the model that included habitat and nest stage and their interaction was the only c < 2 (Table 21 ). Nests in pastures and wildlife preserves tended to have lower survival rates in the nestling stage, however the confidence inter vals for these survival rates are much wider and overlap with those for the incubation period (Fig. 26). In parking lot, town, and residential areas the estimates of daily survival rates for incubation and nestling stages were much more similar to each ot her (Fig. 26). There was also a trend for parking lot and town to have higher rates of survival than pastures and wildlife preserves in both the incubation and nestling stages (Fig. 26). Survival rates in residential tended to be intermediate between those in parking lot and town and those in pastures and wildlife preserves (Fig. 26). For the Brown Thrasher the top model included landuse, year, and their interaction (Table 21 c = 0.235) included only landuse. In both 2005 and 2006 survival rates were lower in nonurban landuse than in urban landuse, however in 2006 the difference was greatly reduced because survival rates in nonurban landuse were higher than those in 2005 (Table 2 2 ). At the habitat level, there were three model c values less than two, all of which included habitat as a parameter (Table 21 ). Year was included in two of the three models. In 2005 residential nests had higher survival rates than wildlife preserves (Fig. 27). In 2006 the 95% confidence intervals for all three habitat types (residential, pasture, and wildlife preserve) overlapped, although there was a trend for higher survival rates in residential habitat than in pastures and wildlife preserves (Fig. 27).
33 For the Northern Cardinal (effecti ve sample size = 1534) the model with landuse c < 2 (Table 21 ). Daily survival rates in urban landuse were higher than those in nonurban landuse (Table 22 ). Daily survival rates in urban forest fragments were similar to those in nonurban landuse and tended to be lower than survival rates in urban landuse (Table 22 ). Discussion Our results are generally consistent with the hypothesis that differences in abundances of nest predators drive many of the changes in bird community composition ofte n documented in urban habitats. We found much higher abundances of avian nest predators in urban habitats (Fig. 22). Urban bird communities consisted mostly of largebodied individuals ( Fig. 2 5), many of which aggressively defend their nests against predators (C. Stracey, pers. obs.), and smaller bodied individuals t hat nest in inaccessible sites. The near absence of small bodied opencup nesters in urban areas may reflect the vulnerability of these species to avian predators. As predicted, in general urban adapters experienced lower nest predation rates in urban habitats relative to nonurban habitats (Tables 2 1 and 22 ). Nest Predator Abundance There is little dispute that urban areas host altered predator communities. There are two important pieces of information we need to assess how this change in predator community may affect urban bird communities 1) how the abundance of different nest predators changes with urbanization and 2) the relative importance of different types of nest predators. Avian nest predators were more frequently detected i n the urban communities we censused, similar to other areas where comparable census data exist (Gregory and Marchant 1996, Jokimki and Huhta 2000, Marzluff et al. 2001, Sorace
34 2002). This greater abundance may r esult from increased anthropogenic food resources and decreased home range size (Marzluff and Neatherlin 2006). In addition, these increased food resources are typically concentrated at point sources that are highly predictable, which larger birds may be m ore effective at exploiting (Daily and Ehrlich 1994, Shelley et al. 2004, French and Smith 2005, MacNally and Timewell 2005). Such large species often incidentally take eggs and nestlings (Marzluff et al. 2001). We do not, however, know the relative import ance of these species as nest predators and if that importance varies by habitat. We have evidence that the predator communities likely differ : in urban areas there was no difference in predation rates at different nest stages, but there was a difference in nonurban areas. These data suggest that different predators were responsible for the majority of predation events in different habitats We have no data from mammalian predators or snakes and the majority of previous studies concerning the abundance of mammalian nest predators relate to their abundance in habitat fragments within either an agricultural or urban matrix, and rarely to their abundance in the urban areas themselves (Chalfoun et al. 2002). Within urban forest fragments rats, foxes, cats, dog s, and opossums have been shown to occur at higher densities than in nonfragmented areas (Crooks 2002, Sorace 2002). Raccoons ( Procyon lotor ) are positively affected by urbanization and reach higher densities in both urban forest fragments and the urban environment itself (Prange et al. 2003, Randa and Yunger 2006). The effect of urbanization on snakes is likewise not well known. Snakes were more abundant in fragmented habitat surrounded by agriculture (Weatherhead and Charland 1985, Durner and Gates 1993) However, Patten and Bolger (2003) found that within urban fragments of coastal scrub snake abundance was lower. In urban
35 areas themselves, it is likely that snake abundance is further reduced owing to persecution by people. If there are fewer snakes in urban areas, then the high nest predation rates in rural areas may be largely a result of increased abundance of snakes, which are potentially very important nest predators (Weatherhead and BlouinDemers 2004). Schmidt et al. (2001) found that weak predator s ( species that only occasionally depredate nests ) did not fully compensate for strong predators ( species that dr i ve patterns of nest predation) when the strong predator was removed. Therefore, increases in avian predators in urban areas may be more than offset by a decrease in snakes. However, snake abundance and diversity varies from one geographic region to the next and we do not know how this relates to the effects of urbanization on snakes. Future studies should explicitly consider the effects of urbanization on snakes and the importance of snake predation on nests. The way in which these changes in the predator community translate into changes in nest predation rates is not likely to be straightforward because different species are likely to be differentially affected by predation, depending on their life histories. For example, nesting guild is an important determinant of the susceptibility of nests to different predators (Patten and Bolger 2003, Schmidt et al. 2006) and thus differences in predator community composition in urban areas might favor the persistence of certain species over others. To fully understand the role of nest predation in structuring urban bird communities we need to determine how the abundances of all nest predators change, not si mply avian predators. Moreover, we need to explore how these changes in abundance translate into changes in nest predation rates, paying particular attention to their relative effect on species with varying life histories.
36 Urban Bird Community Composition Several patterns in bird community composition in Florida were consistent with the hypothesis that nest predation is shaping urban communities. The reduced abundance of small bodied birds in urban environments (Fig. 25) may at least partly reflect their v ulnerability to avian nest predators. Some of the most successful urban species defend their nests very aggressively against nest predators. Mockingbirds, in particular, mob nest predators when they are still far from their nests (C. Stracey, pers. obs.). In addition, because of higher mockingbird density in urban habitats, urban mockingbirds may be more successful at deterring avian predators before they reach the nest due to greater overall mobbing behavior (Robinson 1985, Wiklund and Anderson 1994, Picman et al. 2002). Indeed Breitwisch (1988) found that mockingbirds that mobbed more aggressively were also less likely to lose their nests to predators. Brown Thrashers and Loggerhead Shrikes were also extremely aggressive when predators or nest checkers approached the nest. If successful nest defense is necessary to offset increases in avian nest predators, then it is possible that nest predation could explain the pattern of increased average body size in urban areas. Among small bodied individuals that do thrive in urban environments, the vast majority nest in cavities or other enclosed places (Fig. 25). Even opencup nesters, such the House Finch, usually place their nests in enclosed places in human structures in urban environments. Thus, small bodied b irds can thrive in urban environments, but only if they place their nests in sites that are inaccessible to largebodied avian nest predators. This same pattern also holds for forest fragments, which tend to retain their smallbodied cavity nesters but los e many or even most opencup nesters (Whitcomb et al. 1981, Brown and Sullivan 2005).
37 An alternative explanation for the increased abundance of largebodied birds in urban areas is what we term the interspecific dominance hypothesis. Larger species are typ ically competitively dominant to smaller species and thus are better able to exploit point sources of food (Daily and Ehrlich 1994, French and Smith 2005, Shelley et al. 2004, MacNally and Timewell 2005) that are frequently found in urbanized settings (e.g ., bird feeders, garbage bins, fruiting ornamental shrubs and trees), in addition to being better competitors for limited nest sites (Ingold 1994). In edges of Australian urban forest fragments abundance of the hyper aggressive Noisy Miner ( Manorina melan ocephala) is negatively correlated with species richness of smaller bodied birds (Piper and Catterall 2003). Thus, the interspecific dominance hypothesis is consistent with Brown and Sullivans (2005) metaanalysis of the differential response of birds of varying body sizes to fragmentation. They found that large birds (excluding the largest species) were more abundant in small woodlots and that small species were more common in large tracts of forest. They attributed this pattern to the competitive dominance of large species. The interspecific dominance hypothesis expands Shochats (2004) hypothesis that the predictability of food resources in time leads to changes in population structure. The interspecific dominance hypothesis and the urban nest predator hypothesis are not mutually exclusive and may even act synergistically to favor largebodied species in urban areas. Some of the differences we observed in abundance could be due to the structural suitability of the habitat or habitat selection. We do not evaluate these possibilities here, although we attempted to control for them on a gross level by restricting our comparisons to urban settings that most closely resemble nonurban counterparts (e.g.,
38 parking lots compared with nonurban open habitat and treed residential areas compared with forests). In addition, we restricted the species that we looked at to those that occurred in both urban and nonurban settings (with the exception of a few introduced species found only in urban settings). Future studies should examine the role of habitat structure more carefully. Patterns of Nest Predation in Urban Habitats For most species for which we were able to compare nesting success in urban and nonurban habitats, rates of nest predation were lower in urban habitats, although this difference was not always significant, likely due to small sample sizes (Tables 2 1 and 22 ). U rban areas appear to offer a refuge from predation for many of these species and their increased abundance and status as urban exploiters and adapters may result from their high nesting success. An exception to this generalization is the cardinal, which was slightly more abundant in nonurban areas in spite of having higher nesting success in urban habitats (Fig. 24d; Table 2 2 ). Differences in nest predation rates for the cardinal tended to be less extreme than for the other species for which we have data. Leston and Rodewald (2006) found slightly higher rates of nest predation in urban forest fragments relative to nonurban habitats, but the di fferences were insufficient to cause a change in seasonlong productivity. Perhaps cardinals, which are abundant in both urban and nonurban habitats, may nest successfully in both settings. We do not have data on nesting success of Carolina Wrens, which s how a similar pattern to cardinals in terms of their relative abundance in urban and nonurban areas (Fig. 24c). Unfortunately, this species hides its nests so well in nonurban settings that even our most skilled nest searchers were unable to find more than a few nests. Nesting data for
39 this species would be particularly useful in further testing the link between nest predation, abundance of the species, and habitat. Nest Predation and the Success of Introduced Species Nest predation pressure may also help explain why introduced species in Florida tend to be restricted to urban areas (Fig. 23) and why some introduced species succeed while others fail to become established in the urban areas where they are usually released. To our knowledge, none of the opencup nesting species that have been released in Florida have succeeded and spread. The Bluegray Tanager ( Thraupis episcopus ), for example, became extinct after a brief period of apparent success and the Redwhiskered Bulbul ( Pycnonotus jocosus ) has not spread beyond a restricted area in south Florida (Robertson and Woolfenden 1992, Stevenson and Anderson 1994). The birds that have succeeded are overwhelmingly enclosed nesters such as parrots, House Sparrows, and starlings. For at least one of these species, the House Finch, human structures such as parking garages provide nest sites that are essentially invulnerable to predation (Table 2 2 ). The House Finch is usually considered an open cup nesting species; however, the majority of nests (55 of 60) that we found and monitored were located in spaces that were functionally cavities. For example, many nests were located in small (approximately 15 cm x 6 cm) openings in the concrete ceiling of parking garages and in narrow spaces above lights in breezeways. Of all the species we studied, the House Finch had the highest daily nest survival rate (99.68%). The availability of such safe nest sites has likely contributed as much to their successful invasion of eastern North America as their ability to exploit bird feeders. The lack of
40 such enclosed nest sites in nonurban areas may explain why they continue to be mostly restricted to human settlements in their introduced range. There are, however, two opencup nesters that have been successful invasives, the Eurasi an Collared Dove and the Rock Pigeon ( Columba livia ). Both species are relatively large and defend their nests to some extent against predators. The Rock Pigeon nests on buildings in loose colonies where access by predators is limited, but the Eurasian Col lared Dove nests in similar places to the native Mourning Dove, including in trees in parking lots and residential neighborhoods. Although the Eurasian Collared Dove does lose nests to predators, their nesting success is comparable to urban Mourning Doves and is quite high compared to Mourning Doves nesting in rural areas. Thus, the relatively high nesting success of Eurasian Collared Doves may be fuelling their phenomenal population increase and range expansion. Alternative Hypotheses An alternative explanation for the reduction in nest predation observed in urban areas is the predator satiation hypothesis or the french fry hypothesis (fide J. Marzluff). With the predator satiation hypothesis the increase in abundance of avian predators does not increase nest predation rates because the avian predators are satiated by alternative food sources provided by humans in urban areas. An increase in food, however, often results in an increase in foragers, which typically leads to increased competition over that resource (Sol et al. 1998, Shochat 2004). While most avian nest predators are generalists and no doubt take advantage of alternative food sources, it is possible that incidental predation (sensu Vickery et al. 1992) would actually increase,
41 especially over t ime as generalists become more abundant; this hypothesis remains to be tested. Another possible explanation for the reduction in nest predation rates is the suitability of the nest site. Nest sites may be inherently safer in urban settings as a result of differences in the structure of the nest site. Many of the nests in town were located in thick dense shrubs that were maintained with constant trimming. Such sites might be inherently safer than those available in nonurban settings. Conclusions One of the goals of our study was to determine the mechanisms underlying the urban nest predator paradox, which is the phenomenon of reduced rates of nest predation in urban areas despite conspicuously higher abundances of many nest predators, especially avian pred ators (Shochat 2004). Indeed, for species that are more abundant in urban habitats, we did find lower rates of nest predation in urban than nonurban areas. For at least this subset of species, urban areas may offer a refuge from nest predation as hypothes ized by Gering and Blair (1999). As we have argued above, however, this result may only hold for species that have effective defenses against the kinds of predators that thrive in urban areas. For those species that cannot defend themselves effectively against avian predators, such as most small bodied opencup nesters, their absence from urban areas may reflect an inability to nest successfully in the face of abundant flocks of crows and grackles. These opencup nesters were too rare to study in urban areas, which prevented us from making any comparisons. Thus, there may not be an urban nest predator paradox ; urban areas may only be a refuge for species that are less vulnerable to the kinds of predators that are also urban adapters. Conversely, the scarci ty of many urban adapters in rural areas may reflect
42 their vulnerability to predators such as snakes, which are likely more abundant in nonurban areas. An additional test of the effects of nest predators on urban bird communities could come from studies of species nesting in small urban fragments (e.g., Leston and Rodewald 2006). These fragments may have the worst of both worlds high populations of both avian nest predators and of other predators such as snakes, which may persist in small urban woodlots. Urban forest fragments had high levels of nest predation in this study, but we lack data from small bodied species that sometimes persist in these small fragments to make the comparison. The generality of our results will depend in large extent to how different predator communities in different geographic regions respond to urbanization. There is some evidence that avian nest predators thrive in urban habitats worldwide (Gregory and Marchant 1996, Jokimki and Huhta 2000, Marzluff et al. 2001, Sorace 2002), but the responses of mammals to urbanization may vary much more among regions. Clearly, much work remains to be done in our efforts to understand the role of predation in structuring urban bird communities. It will be important to design experiments that distinguish between the above explanations of the predator paradox. In addition, it is necessary to experimentally assess the relationship between body size, nest defense, nesting guild, and nesting success in urban areas. Also, incorporating the interaction between food and predation, both of which differ between urban and nonurban areas, will be critical for advancing our understanding of urban ecology. One such interaction may be between food and nest defense. Increased food resources have been show n to lead to increased levels of nest defense (Komdeur and Kats 1999, Duncan Rastogi et
43 al. 2006). Therefore birds in urban areas, which are thought to have increased food resources, may show increased nest defense resulting in reduced nest predation rates Until we are able to understand the mechanistic role of predation in urban areas, we will be left with numerous, and often conflicting, case studies on predation rates.
44 Table 21. Model selection results for the top six logistic exposure models of dail y survival for the Northern Mockingbird (2005 2006), Brown Thrasher (2005 2006), and Northern Cardinal (2004 2006) Model Log(L)a Kb AICc c c d wi e Northern Mockingbird Habitat Type Habitat stagef -730 10 1480.04 0 0.584 Habitat stagef + year -729.05 12 1482.16 2.13 0.202 Habitat year stagef -725.29 16 1482.68 2.64 0.156 Habitat + stagef -737.15 6 1486.32 6.29 0.025 Habitat + year + stagef -737.02 7 1488.05 8.01 0.011 Habitat -739.19 5 1488.39 8.36 0.009 Land-use Category Category year stagef -735.43 8 1486.88 0 0.342 Category year + stagef -738.1 6 1488.22 1.34 0.175 Category + year + stagef -738.1 6 1488.22 1.34 0.175 Category + year stagef -738.46 6 1488.94 2.06 0.122 Category + stagef -741.82 3 1489.64 2.76 0.086 Category year -741.15 4 1490.31 3.43 0.062 Brown Thrasher Habitat Type Habitat -234.43 5 478.9 0 0.432 Habitat year -231.93 8 479.95 1.05 0.255 Habitat + year -234.15 6 480.36 1.46 0.208 Habitat + stagef -234.33 7 482.73 3.83 0.064 Habitat + year + stagef -234.04 8 484.18 5.28 0.031 Habitat year + stagef -231.2 12 486.62 7.72 0.009 Land-use year -232.85 4 473.73 0 0.479 Land-use -234.98 2 473.97 0.24 0.426 Land-use + stagef -234.88 4 477.78 4.05 0.063 Land-use year + stagef -232.34 8 480.78 7.05 0.014 Land-use + year + stagef -232.34 8 480.78 7.05 0.014 Land-use + year stagef -233.33 9 484.79 11.05 0.002
45 Table 21. Continued Model Log(L)a Kb AICc c c d wi e Northern Cardinal Land-use Category Land-use -208.97 3 423.96 0 0.54 Land-use + year -208.43 5 426.91 2.95 0.124 Land-use + stagef -208.61 5 427.26 3.3 0.104 Land-use year -207.61 6 427.28 3.12 0.103 Constant -213.25 1 428.51 4.55 0.056 Land-use + year + stagef -208.04 7 430.16 6.20 0.024 Notes: a: value of the maximized loglikelihood function. b: number of parameters in the model c: Akaike s information criterion for small samples d: scaled value of AICc e: Akaike weight f: stage of the nesting cycle (incubation or nestling) interaction terms with that variable included in the model + no interactions with that variable included in the model
46 Table 22 Daily survival rates (95% confidence limits [ CL ] ) in the years indicated, number of nests monitored, and the statistical significance of landuse on daily survival rates. Landuse Daily Survival Rate N Chisquare df P Mourning Dove (20042006) Urban 96.27 (94.2397.61) 44 Non Urban 92.00 (84.7695.93) 11 3.43 1 0.064 Loggerhead Shrike (20042006) Urban 97.80 (94.8099.08) 11 Non Urban 94.95 (85.4598.36) 4 1.32 1 0.25 House Finch (20042006) Urban 99.68 (99.169 9.98) 60 Non Urban n/a 0 n/a Eurasian Collared Dove (20042006) Urban 96.89 (90.7497.60) 15 Non Urban n/a 0 n/a Northern Cardinal (20042006) Urban 97.00 (95.5498.04) 65 UFF 93.50 (86.9696.87) 11 Non Urban 94.13 (92.2295.59) 73 n/a Northern Mockingbird (2004) Urban 95.50 (94.3096.46) 124 Non Urban n/a 0 n/a Northern Mockingbird (20052006) Urban 314 Incubation 96.94 (96.1997.55) Nestling 96.81 (95.9497.49) Non Urban 147 Incubation 94.53 (93.0795.70) Nestling 90.75 (87.7593.07)
47 Table 22 Continued Landuse Daily Survival Rate N Chisquare df P Brown Thrasher (2004) Urban 95.12 (92.2996.95) 31 Non Urban n/a 0 n/a Brown Thrasher (20052006) Urban 50 2005 97.41 (94.6798.76) 2006 96.31 (93.6297.90) Non Urban 94 2005 89.70 (86.9591.93) 2006 92.43 (89.0394.84) Note : UFF: Urban Forest Fragment. Figure 21. The location of study sites for the nesting study on images modified from Google Earth mapping service. A C: pastures; D E: wildlife preserves; F H: residential neighborhoods; I: University of Florida campus; J: downtown Gainesville; KL: parking lots.
48 Figure 22. Average (+/ s tandard error [ SE] ) number of total avian nest predators detected per census point (five minute, 100m fixed radius) and the average number of each species detected ( 2 = 73.68, df = 5, P < 0.0001). Letters refer to significant post hoc comparisons of total number of predators. The following species were considered avian nest predators: American Crow (AMCR), Fish Crow (FICR), Common Grackle (COGR), Boattailed Grackle (BTGR), Blue Jay (BLJA), and Redshouldered Hawk (RSHA). PL: parking lot, RES: residential neighborhood, UFF: urban forest fragment, PAST: pasture, NO: nonurban scrub, NF: nonurban forest.
49 Figure 23. Average (+/ SE) number of individuals of introduced species detected per point in each habitat type based on five minute, 100m fixed radius points ( 2 = 111.12, df = 5, P < 0.0001). Letters refer to significant post hoc comparisons. HOSP: House Sparrow, ECDO: Eurasian Collared Dove, EUST: European Starling, HOFI: House Finch, MOPA: Monk Parakeet, ROPI: Rock Pigeon. PL: parking lot, RES: residential neighborhood, UFF: urban forest fragment, PAST: pasture, NO: nonurban scrub, NF: nonurban forest.
50 Figure 24. Average (+/ SE) number of individuals detected per point in each habitat type based on five minute, 100m fixed radius points. Letters refer to significant post hoc comparisons. a) Northern Mockingbird ( 2 = 93.66, df = 5, P < 0.0001). b) Mourning Dove ( 2 = 46.59, df = 5, P < 0.0001). c) Northern Cardinal ( 2 = 40.59, df = 5, P < 0.0001). d) Carolina Wren ( 2 = 27.25, df = 5, P < 0.0 001). PL: parking lot, RES: residential neighborhood, UFF: urban forest fragment, PAST: pasture, NO: non urban scrub, NF: nonurban forest.
51 Figure 25. Average (+/ SE) number per census point (five minute, 100m fixed radius) of individuals of all species combined detected in each habitat based on mass and nesting guild. a) parking lot. b) residential neighborhood. c) pasture. d) urban forest fragment. e) nonurban scrub. f) nonurban forest.
52 Figure 26. Daily survival rate (+/ 95% CL) of the North ern Mockingbird at the incubation and nestling stage in each habitat type with number of nests indicated below each habitat in parentheses. PL: parking lot, TOWN: downtown Gainesville, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve.
53 Figure 27. Daily survival rate (+/ 95% CL) of the Brown Thrasher in 2005 and 2006 in each habitat type with number of nests indicated below each habitat in parentheses. PL: parking lot, TOWN: downtown Gainesville, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve.
54 CHAPTER 3 CATS AND FAT DOVES: RESOLVING THE URBAN NEST PREDATOR PARADO X Introduction The amount of land area that is converted to urban land use is increasing even more rapidly than the urban human population (Marzluff et al. 2001) and by 2050, the United Nations estimates that the global urban population will equal ~6.5 billion (United Nations 2006). Urbanization is one of the leading causes of species endangerment in the United States (Czech and Krausman 1997) and although many native species are extirpated from urban areas, some adapt well (Blair 1996, Shochat et al. 2006). Native species that are successful in both urban and nonurban habitats are termed urban adapters (Blair 1996, Shochat et al. 2006). Populations of urban adapters that reside in towns and cities can form a significant proportion of the global population of these species (Mason 2000, Bland et al. 2004, Chamberlain et al. 2009, Fuller et al. 2009). Understanding the factors that promote the success of urban adapters is necessary if we are to gain a complete understanding of the processes that shape urban wildlife communities and provide meaningful recommendations to urban planners and concerned citizens that will allow us to retain or even enhance urban bird communities. The predator refuge hypothesis (Gering and Blair 1999, Marzluff et al. 2001, Shochat et al. 2004a, Faeth et al. 2005, Shochat et al. 2006, Anderies et al. 2007, Chamberlain et al. 2009) proposes that the success of avian urban adapters i s the result of a reduction in nest predation rates in urban habitats. Evidence for this hypothesis is mixed. Some studies have documented reduced rates of urban nest documented increased urban nest predation rates (e.g., Jokimki and Huhta 2000,
55 Thorington and Bowman 2003) or no differences (e.g. Melampy et al. 1999, Haskell et al. 2001). Thus, the general applicability of the predator refuge hypothesis as an explanation for the success of urban adapters is far from clear. Our ability to interpret these conflicting results is hampered by our lack of knowledge about 1) how different potential predator populations respond to urbanization and 2) the relative importance of these di fferent potential predators in determining nest predation rates. One possible explanation for the urban predator refuge is that there are simply fewer nest predators in urban areas (the reduced urban predator abundance hypothesis). Although we lack data on how most groups of nest predators respond to urbanization (Chapter 2), the abundance of at least some nest predators is actually higher in urban habitats. For example, avian nest predators (Gregory and Marchant 1996, Jokimki and Huhta 2000, Marzluff et al. 2001, Sorace 2002), raccoons (Prange et al. 2003, Randa and Yunger 2006), and cats (Sorace 2002) are consistently more abundant in urban than in nonurban habitats. This mismatch between predation rates, which are often lower in urban areas, and predator abundance, which is often higher in urban areas, has been termed the urban nest predator paradox (Shochat et al. 2006, Chapter 2). While we can generate a laundry list of species that have been documented as nest predators, in most cases the species included are based upon anecdotal eyewitness accounts of predation events and not on systematic studies of nest predator identity. Not all predators are created equal: certain species likely only infrequently depredate nests, whereas other species are consistently important nest predators (Schmidt et al. 2001). Furthermore, t he predator paradox has been primarily based
56 upon the abundance of avian nest predators, which are only a subset of the entire predator community. If important nest predators are replaced with species that only occasionally depredate bird nests (weak predators), then the absolute abundance of potential nest predators may not be a good predictor of actual predation rates (Schmidt et al. 2001). Instead, predation rates may reflect t he abundance of a few strong predators (species that are responsible for the majority of predation events and thus drive rates of nest predation; sensu Schmidt et al. 2001). I define the strong predator abundance hypothesis as the loss of strong predators in urban habitats, which leads to a reduction in nest predation rates. Until we know more about the actual identity of the predators that attack nests and on how this varies between urban and nonurban habitats, we cannot evaluate this hypothesis. Alterna tively, whether a species acts as a strong predator of nests could be context dependent (Navarrete and Menge 1996). If urban predators are satiated on anthropogenic food, then predation rates on bird nests might be reduced the French fry hypothesis (J Marzluff, pers. comm.) and otherwise strong predators may become weak predators in urban environments. Yet this additional food may attract additional nest predators thereby increasing incidental predation (sensu Vickery et al. 1992), which could in tur n offset the reduction in the importance of eggs and nestlings in the diets of individual nest predators (the incidental predation hypothesis). An additional factor that must be considered in studies of nest predation in urban communities is nest defense behavior. Some species may be able to persist or even thrive in the face of high predator populations because they are able to defend their nests effectively (Breitwisch 1988) against the dominant predators in a habitat (the
57 predator defense hypothesis: C hapter 2). Mobbing may be especially effective in urban areas because many of the numerically dominant predators are birds such as crows and blackbirds that can potentially be chased away by aggressively mobbing parents (Marzluff et al. 2001, Chace and Walsh 2006). Before we can test these hypotheses, we need data on which of the many potential predators are responsible for the majority of nest predation events and on how the identity of the predators changes in urban relative to rural communities. In this study I attempt to resolve the urban nest predator paradox for one urban adapter, the Northern Mockingbird ( Mimus polyglottos ), by using video cameras to compare the identity of nest predators at nests in urban and nonurban habitats. Data on mockingbi rds from 2005 2006 support the hypothesis that urban nest predation rates are lower than nonurban nest predation rates (Chapter 2); yet many potential avian nest predators (Fish Crow, Common Grackle, Boat tailed Grackle) are more abundant in urban habitat s in North Florida (Chapter 2), which suggests that the reduced urban predator abundance hypothesis does not apply. The urban nest predator paradox (e.g. lower rates of nest predation in urban habitats that have more abundant nest predators) assumes that there is a tight correlation between nest predation rates and predator abundance. This assumption, however, may not reflect the disproportionate affect of a few strong, but less abundant, predators. If the predator paradox can be explained by the strong predator abundance hypothesis, then I predicted that the most abundant avian nest predators in urban habitats would account for relatively few predation events. If, on the other hand, the predator paradox is being driven by the French Fry hypothesis,
58 then I would expect to see higher relative rates of predation in nonurban habitats by predators that are common to both habitats Methods Nest Predation Rates I collected data on nest predation rates at seven study sites (two parking lots, three residential neighborhoods, two pastures, and one wildlife preserve; Table 31) in areas in and around Gainesville, FL (Fig. 31) between February and August of 2007 2008. In 2009 I did not collect data from one of the residential neighborhoods, one of the parking lots and one of the pastures. T he data in this chapter extend the number of years of data collection in Chapter 2 and correspond with the years in which the cameras were de ployed. I located nests and checked their contents every one to four days. For each year I calculated habitat specific nest survival rates and 95% confidence limits using the logistic exposure method (Shaffer 2004, SAS 9.1). Because I was interested in predation rates, I only considered nests that failed due to predation as unsuccessful Therefore, the daily survival rates I report are not a measure of overall nest survival, but rather the probability of a nest escaping predation. I evaluated the importance of habitat (e.g., parking lot, residential neighborhood), nest stage (laying, e gg, and nestling), and ordinal date (number of days since January 1) on the probability of nest predation according to Akaikes information criteria (AICc, Burnham and Anderson 1998). Each year was analyzed separately. I evaluated twelve candidate models (Table 42) including a constant survival model and the global model, which contained ordinal date, nest stage, and habitat, plus all twoway interactions. I c < 2 to be well supported and used nonoverlapping 95%
59 confidence intervals to assess differences between treatments levels for variables in those models. Nest Predator Identification I placed video cameras on nests after clutches were complete in two residential neighborhoods, one pasture, and one wildlife preserve to com pare nest predator identity between urban (residential) and nonurban (pasture and wildlife preserve) landuse. I did not place cameras in parking lots for fear of theft and I excluded one of the pastures for logistical reasons. In 2007 I used two Recony x cameras (Model #RM30) on nests in two residential neighborhoods and one wildlife preserve. In 2008 and 2009 I switched to a different camera set up (modified from Pierce and Pobprasert 2007) that consisted of a small security camera (OPCM Weatherproof S ecurity Camera BS08; 6cm x 4cm x 4cm) with six IR emitting diodes for night vision. Each camera was connected (with a 25m cable) to a DVR (Archos 504) that continuously recorded nest activity on an internal hard drive. Both the DVR and security camera were powered with a 12V car battery that I placed in a plastic container and changed every other day. All videos were uploaded and saved onto external hard drives. I spray painted the camera housing and plastic container with the DVR and battery for camouf lage. The camera was affixed to a small stick that was tied to a branch near the nest. For nests that did not have any nearby branches (e.g., nests in vine tangles), I drove a 2m piece of rebar into the ground near the nest and attached the camera to the rebar. The camera was placed within 1m of the nest to ensure the nest was illuminated at night. I performed minor trimming of leaves on the inside of the nest bush that were blocking the cameras view of the nest in a manner that decreased nest concealm ent as little as possible. Most cameras were placed slightly above the nest affording a view of the nest contents, but in
60 some cases I had to place the camera eyelevel with the nest to obtain a clear view. I used a total of 18 cameras in both 2008 and 2009. I identified nests that were candidates for a camera based on the height of the nest (nests >4 m were excluded) and the availability of a place to hide the DVR and battery (in urban areas). The average nest height at our study sites was approximately 2m. Twenty one out of 298 nests were excluded because they were too high. In the residential neighborhoods, I had to hide the plastic container to prevent theft of the DVR. For example, some nests were located in trees that were right on the edge of a sidewalk or street that did not afford a place to hide the plastic container; 18 out of 124 nests were excluded for this reason. When the contents of a nest disappeared I watched the videos to determine the fate of the nests, identified nest predators to s pecies, and noted whether the predator removed eggs or nestlings. I considered multiple predation events on the same nest by different species of predators to be independent, whereas I considered multiple visits by the same species of predator to the same nest to be the same individual and thus only counted them once in my analyses. Because in some cases I was unable to identify snakes to species from night videos, I grouped all snakes into one category for my analyses. I also grouped all mammals (other than house cats) into another category because each individual species accounted for relatively few predation events; all other species were analyzed as their own category. I compared the frequency of predation by each species/group among study sites with a Fishers Exact Test (SAS 9.2). Because there was a significant association between predator and study site (P < 0.0001, Fishers exact test), I conducted post hoc contrasts within urban (two residential sites)
61 and nonurban (one pasture and one wildlif e preserve sites) areas, as well as between urban and nonurban areas with Fishers exact tests. I also tested for an effect of camera on nest predation rates in 2008 and 2009 using the logistic exposure method (Shaffer 2004). Results Nest Predation Rates c < 2) for two models that included nest stage and ordinal date (Table 32). In 2007, nest survival probability increased in the laying and nestling stage as the breeding season progressed but there was little change in nest survival across the breeding season for the egg stage. In 2008 there were five models with strong support (Table 32). All five models included nest stage as a variable, with eggs surviving at a higher rate than nestlings (based on nonoverlappi ng 95% confidence intervals). Ordinal date was included in three of the five top models (Table 32) with survival at the nestling and laying stages decreasing as the season progressed. Habitat was included in two of the top five models (Table 32); howev er, 95% confidence intervals overlapped for all habitat types (Fig. 32). In 2009 the two c < 2), included the variable habitat (Table 32). In 2009 nests in parking lots survived better than nests in the other three habitats (Fig. 32). Nest Predator Identification Cameras were placed on a total of eight nests in 2007, 52 nests in 2008, and 84 nests in 2009. Eight nests were abandoned shortly after I set up the camera and 22 predation events were missed due to problems with the camera s et up, including drained batteries and gnawed wires. A total of 58 predation events were recorded (Fig. 3 3). I excluded one predation event by a Blue Jay ( Cyanocitta cristata) from our
62 analysis because it occurred after the nest had been abandoned. The presence of a 2 = 0.03, df = 1, P = 0.86), 2 2 = 0.21, df = 1, P = 0.64). There was a significant association between study site and the identity of predators (P < 0.0001, Fishers exact test). The two urban sites, however, did not differ in predators (P > 0.99, Fishers exact test), nor did the two nonurban sites (P = 0.07, Fishers exact test). The contrast between the urban sites and the two nonurban sites was significant (P = 1.42 x 1010, Fishers exact test). In urban habitats cats were responsible for over 70% (17 out of 24) of predation events, whereas Coopers Hawks ( Accipiter cooperii ) were the most frequent nonurban nest predator (15 of 33 predation events, Fig. 3 3). Discussion Is there an urban refuge from nest predation? Urban areas are not always a refuge from nest predation relative to nonurban areas for the Northern Mockingbird. Nest predation rates did not differ by habitat in 2007 and 2008 (Table 32, Fig. 32). These data contrast with our previous results from 2005 and 2006 when we documented reduced nest predation in urban habitats at many of the same study sites (Chapter 2). Across the three years of this study and the tw o years from the previous study, predation rates in urban habitats, and parking lots in particular, were consistent from year to year (Fig. 32). Yearly variation in nonurban nest predation rates therefore accounted for the different levels of statistical significance between urban and nonurban areas among years. Shochat et al. (2006 and references therein) found that variation in resources, including abiotic factors that influence resource availability, are dampened in urban habitats. Predator populat ions in nonurban habitats thus may be more likely to experience population fluctuations as a result
63 of natural variations in their resources, which could lead to variability in nest predation rates. Such annual variation may in part explain why no clear pattern has emerged from previous studies on the effects of urbanization on nest predation rates. Clearly, urban areas can be a refuge from nest predation for some species in some years, but, overall, our data add to the growing consensus that the urban refuge hypothesis alone is not a general explanation for understanding patterns of urban community composition. Hypothesized mechanisms underlying changing nest predation pressures in urban bird communities In the introduction, I developed six hypotheses to explain changes in nest predation in urban areas (Table 33). Two of these relate to the abundance of predators, either all nest predators considered together (the reduced predator abundance hypothesis) or of strong predators (the strong predator abundanc e hypothesis). The next three hypotheses relate to the foraging ecology of the predators themselves, none of which specialize on bird eggs and nestlings. Two predict that nest predation rates would be reduced by the availability of alternative foods, eit her for omnivores such as grackles or crows (the French Fry hypothesis), or for top predators (the Fat Dove hypothesis). These hypotheses further predict that the strength of nest predation pressure exerted by some predators may be context dependent. The other hypothesis related to predator foraging ecology predicts increased nest predation rates by nest predators that have increased populations in the area as a result of the availability of alternative foods (the incidental nest predation hypothesis). T he last of the six hypotheses predicts that nest predation rates by the different predators in any
64 habitat will be affected by nest defense behavior. I have already attributed the lack of small, open cup nesters in urban habitats to be a result of their i nability to defend their nests against abundant avian nest predators (Chapter 2). This hypothesis further predicts that mockingbirds, an aggressive mobber and an urban adapter, would only be vulnerable to a subset of the nest predators in urban habitats, specifically those that cannot be driven away by aggressive nest defense. Perhaps the most striking pattern in these data was the absence of nest predation by the most abundant avian nest predators in northern Florida. Predation by omnivorous Fish Crows, jays, and grackles was essentially absent. Superficially, these data support the strong predator hypothesis because the absolute number of nest predators is not necessarily as important as the identity of those nest predators. There is no evidence, however, that urban areas provide a refuge from strong nest predators. The strongest predator, cats, were essentially only found in urban areas, and Coopers Hawks, the dominant predator in rural habitats, were found at roughly the same abundance in the two habitats (pers. obs.). Snakes, which have been documented as strong predators in other studies (reviewed in Weatherhead and BlouinDemers 2004), were recorded as nest predators in both urban and rural areas in this study, although they were observed somewhat more frequently in non urban habitats. The predators that were only recorded in rural areas (e. g., flying squirrels, raccoons, opossums) were not strong predators in that habitat. The results of this study suggest that simple considerations of predator abundance, either all predators or just strong ones, do not explain habitat specific patterns of nest predation. Therefore, I next consider how the foraging ecology of the
65 predators themselves might account for changes in nest predation rates by the dif ferent predators. Again, on a superficial level, the results of this study support the French Fry hypothesis because few of the omnivorous birds were recorded at mockingbird nests. Perhaps Fish Crows, Boat tailed Grackles, and Common Grackles are satiated on anthropogenic food and thus do not have to attack nests. I have no data to evaluate this hypothesis. American Crows, which were found in both urban and nonurban habitats, were equally recorded as nest predators in both habitats, including rural areas where there is no supplemental food available. The sample sizes, however, were very small and the American Crow does not appear to be driving patterns of nest predation in either habitat, at least for the Northern Mockingbird. The evidence for the Fat Dove Hypothesis, an extension of the French Fry hypothesis to the predator trophic level, is somewhat stronger, although still indirect. Coopers Hawks were only observed as nest predators in rural areas, reflecting a habitat specific dietary shift, pos sibly as a result of abundant alternative prey in the form of urban exploiters and doves in particular (the fat dove hypothesis, Table 33). The diet of urban Coopers Hawks in Indiana reflected the relative abundance of urban Rock Pigeons and Mourning Doves, which, along with European Starlings, made up 91% of their diet (Roth and Lima 2003). Doves comprised 54% of prey delivered to nestling Coopers Hawks in urban areas of Arizona, but only 4% of prey in rural areas (Estes and Mannan 2003). While it ap pears that doves and other urban exploiters may be an important alternative food source for Coopers Hawks, further data on their diet in relation to urbanization are needed. Still, these data provide strong support for the hypothesis that the strength of nest predation by some predators is habitat dependent.
66 I found evidence both for and against the incidental nest predation hypothesis, depending upon the predator species considered. The cat predation data are consistent with the predictions of the incidental predation hypothesis: many of the cats that were present in the urban study sites were being fed and even housed, yet they still supplemented their diet with the eggs and nestlings of birds in their neighborhood. Data on nest predation by cats in ar eas with and without supplemental cat feeding would provide an additional test of this hypothesis. On the other hand, there is no evidence that incidental predation by avian predators attracted to food resources in urban areas led to an increase in nest predation rates by these species given the low levels of nest predation by omnivorous Corvids and grackles. My results also provide indirect evidence for the nest defense hypothesis. Aggressive nest defense by mockingbirds could account for the lack of nes t predation by crows, blue jays, and grackles, all of which were aggressively attacked by mockingbirds (pers. obs.). The only predation event I recorded by a Blue Jay occurred on a nest that had been abandoned and thus lacked nest defense. Aggressive nes t defense in urban habitats therefore might be a critical determinant of which species become urban adapters and which of the potential predators in a community may actually drive patterns of nest predation. Among opencup nesting species, relatively larg e, aggressive species such as the Northern Mockingbird may be more likely to become urban adapters than smaller species, which are largely absent from urban communities where avian nest predators abound (Chapter 2). Therefore, just because I did not recor d predation events by most avian nest predators does not mean that they are not important determinants of avian community structure in urban areas. Future
67 studies are needed on species that have little nest defense, as well as urban avoiders, to determine the role of these avian predators i n structuring urban bird communities. Further support for the predator defense hypothesis comes from the identity of the two dominant predators in both habitats. Cats and Coopers Hawks are species that are very difficu lt to deter by mobbing. Cats are large enough that they are unlikely to be injured by mockingbirds and they also pose a potential threat to the adults themselves. In addition many cats attack at night when mobbing may be even more dangerous for a diurnal bird. During nocturnal predation events mockingbirds flushed from the nest when disturbed and only scolded once or twice, if at all. Coopers Hawks, which feed mostly on adult birds for much of the year (Roth and Lima 2003), are potentially even more dangerous to the adults. During predation events by Coopers Hawks the mockingbirds could be heard scolding in the distance, but never approached the hawks. This is in stark contrast to American Crow predation events in which I could see the mockingbirds hitting the crows in the back of the head. For these reasons, predation rates by these two predators are not likely to be affected by mobbing and habitat differences in nest predation are likely to reflect other factors such as habitat specific differences in abundance (cats) or differences in the availability of alternative foods (Coopers Hawks). For the other predators, predation rates may be more dictated by the defensive behavior of the prey than by their abundance or the availability of alternative f oods. Management Recommendations House cats, the dominant urban nest predator (Fig. 33), consumed both nestlings (12 out of 17 predation events) and eggs (5 out of 17 predation events). Researchers have documented extensive predation of adult birds by cats (Woods et al. 2003,
68 Lepczyk et al. 2004, Baker et al. 2008, van Heezik et al. 2010) and our data indicate that cats in residential areas are also a significant threat to bird eggs and nestlings. At least some of the cats on the videos had collars and i n two instances I knew to whom the cats belonged and that they were being fed. Because all but one cat predation event occurred at night, I recommend that cat owners keep their cats indoors at night during the breeding season to reduce nest predation rates in suburban habitats. While mockingbird nesting success in residential habitats is apparently high enough to sustain populations (Chapter 4 and 5), there are other species (e.g. Eastern Towhee) that nest in similar locations to mockingbirds whose populations may be particularly vulnerable to cat nest predation. Conclusions I conclude that changes in nest predator community composition and dietary shifts of predators are responsible for the apparent urban nest predator paradox in at least one species, the Northern Mockingbird. I found at least some evidence that was consistent with most of the hypothesized mechanisms for the urban nest predator paradox the French Fry hypothesis and the related fat dove hypothesis, the incidental predation hypothesis, and the nest defense hypothesis (Table 33). Urban areas clearly do not provide a generalized refuge from nest predators Rather the success of species in the urban environment may be determined by the trophic structure of the community and on the nest defense behavior of the prey, both of which may ultimately determine which species do and do not act as nest predators. Before we can fully understand the role of nest predation in shaping urban bird communities, we need studies of a suite of species that do and do not thrive in urban environments and we
69 need to study how predator diets change on a rural urban gradient and manipulate food availability to predators.
70 Table 3 1. Average proportion of different types of ground cover for four habitat types based on aerial images of each study site. Habitat (ha) Proportion cover Pavement Building Grass/Bare ground Trees/ shrubs Pine Water Other Parking lot (128.4) 0.47 0.24 0.09 0.18 0.00 0.01 0.01 Residential (96.3) 0.12 0.15 0.15 0.57 0.00 0.00 0.01 Pastu re (321) 0.001 0.003 0.71 0.27 0.00 0.00 0.01 Wildlife preserve (224.7) 0.00 0.00 0.21 0.43 0.13 0.02 0.21 Table 32. Model selection results (top six models) for the logistic exposure models of daily survival for the Northern Mockingbird 2007, 2008, and 2009. Log (L): value of the maximized loglikelihood function; K : number of parameters in the model; AICc: Akaikes information criterion for small s c: scaled value of AICc; wi: the Akaike weight; stage: the stage of the nesting cycle (incubation or nestling). indicates interaction terms included in the model and + indicates no interactions included in the model. Model Log(L) K AIC c AIC c w i 2007 Stage + Date 351.52 4 711.05 0 0.445 Stage*Date 350.08 6 712.16 1.14 0.252 Date -354.93 2 713.87 2.82 0.109 Stage 354.51 3 715.03 3.98 0.016 Stage + Date + Habitat 350.76 7 715.56 4.51 0.047 Constant Survival 357.00 1 716.00 4 .95 0.037 2008 Stage + Date 424.59 4 857.18 0 0.245 Stage 425.79 3 857.58 0.40 0.201 Stage*Date 422.85 6 857.72 0.53 0.188 Stage + Habitat -423.46 6 858.94 1.75 0.102 Stage + Date + Habitat 422.54 7 859.12 1.93 0.093 Date 427.89 2 859.79 2 .60 0.067 2009 Habitat -271.19 4 550.40 0 0.423 Date + Habitat 270.95 5 551.93 1.54 0.196 Stage*Habitat 264.14 12 552.45 2.05 0.152 Stage + Habitat 270.56 6 553.16 2.77 0.106 Stage + Date + Habitat -270.48 7 555.01 4.61 0.042 Date*Habitat 2 69.85 8 555.77 5.37 0.029
71 Table 33. Hypothesized causes of changes in nest predation pressure in urban habitats and evidence for or against it from this study. Hypothesized factors affecting nest predation Overview Predicted Nest Predation Rates Results from this study Reduced urban predator abundance Fewer overall nest predators in urban habitats Reduced None: Most predators as abundant or more abundant in urban areas (the predator paradox) Reduced strong predator abundance Strong nest predators repl aced by weak predators in urban habitats Reduced None: The main nest predators were both abundant in town French Fry Hypothesis Urban nest predators satiated on anthropogenic food (e.g., French fries) Reduced No definitive test: No predation by Fish Crows Boattailed and Common Grackles, but potentially confounded with nest defense Fat Dove Hypothesis Higher trophic level predators are satiated by urban exploiters (e.g., introduced doves) Reduced Indirect support: Urban Cooper's Hawks not observed as nes t predators in urban areas; need data on urban diet Incidental Predation Predators attracted to additional urban food incidentally consume eggs and nestlings Increased Supported: House cats documented as predators Nest Defense Urban adapters can defend themselves against abundant urban predators Reduced Indirect support: Species that can not be effectively mob were the most common recorded nest predators
72 Figure 31. Location of study sites in and around Gainesville, FL. PL: parking lot, RES: res idential neighborhood, PAST: pasture, WP: wildlife preserve.
73 Figure 32. Daily survival rate (probability of a nest escaping predation; +/ 95% CI) of Northern Mockingbird nests in different habitat types between 2005 and 2009 with the number of nests at the base of each bar. Data from 2005 2006 are adapted from Chapter 2. PL = Parking Lot, RES = Residential, PAST = Pasture, WP = Wildlife Preserve.
74 Figure 33. The identity of nest predators in urban and nonurban habitats (total number of predation events in each habitat indicated in parentheses) as determined by video cameras on nests.
75 CHAPTER 4 FOOD LIMITATION IN N ESTING URBAN AND NONURBAN MOCKINGBIRDS: RESOURCE MATCHING OR MISMATCHING? Introduction Alterations of ecosystems caused by urbanizat ion have profound consequences for the structure of animal communities. Urban bird communities, for example, experience a decrease in species richness, but an increase in the average abundance of species that are commensals of humans (urban exploiters) and of native species that benefit from humangenerated changes (urban adapters: Blair 1996, Marzluff et al. 2001, Faeth et al. 2005). Two hypotheses have been advanced to explain why some species become urban adapters: the predator refuge hypothesis and the urban food enhancement hypothesis (Gering and Blair 1999, Marzluff et al. 2001, Shochat et al. 2004a, Faeth et al. 2005, Shochat et al. 2006, Anderies et al. 2007, Chamberlain et al. 2009). Support for the predator refuge hypothesis has been mixed some s pecies do indeed appear to find a refuge from nest predation in urban environments, but this appears to depend on the life history traits of the species (Chapter 2). The role of food resources in contributing to the success of urban adapters has received c onsiderably less attention than the role of nest predation. Urban areas may have increased availability of certain types of food (e.g. bird seed, fruiting shrubs, refuse), which is consistent with the finding that urban exploiters tend to be granivores and omnivores (Marzluff et al. 2001, Chace and Walsh 2006, Kark et al. 2007). The urban food enhancement hypothesis posits that the higher density of urban birds is a result of enhanced food resources leading to higher fledgling quality, quantity, and surviva l relative to nonurban areas. In a recent metaanalysis of the effects of urbanization on bird populations, however, Chamberlain et al. (2009) found that urban
76 areas tended to have lower clutch sizes, lower fledgling production, and lower quality (as meas ured by mass) nestlings than nonurban areas. An alternative possibility, therefore, is that the increased abundance of urban adapters may cause a mismatch between bird densities and food resources (Shochat 2004, Faeth et al. 2005, Shochat et al. 2006). Su ch a mismatch between food resources and population densities has been hypothesized to arise when food resources are highly predictable in space and time, which can lead to artificially high adult densities near these sources of food. The over matching hypothesis predicts that adults attracted to these resources experience reduced nestling quality and productivity as a result of negative density dependence. Furthermore, there should be greater differences between individuals in these measures as i ndividuals who are competitively superior gain extra resources and can live off their credit', whereas inferior competitors obtain just enough food to scrape by and survive, but not enough food to invest in reproduction (i.e. the credit card hypothesis, Shochat 200 4). Such a mismatch would be especially problematic for reproduction if the kinds of food the adults are responding to are different from the food that nestlings need for growth (Schoech et al. 2004). In this context, urban areas may actually become ecolog ical traps where nesting success is below that in natural areas (sensu Gates and Gysel 1978, but see Leston and Rodewald 2006). While individuals may be overmatching food resources, in choosing where to settle birds may be responding to other factors in t he urban environment such as reduced nest predation rates, higher overwinter survival, or a higher density of nesting sites. These other factors may keep populations above the foodbased carrying capacity
77 during the breeding season. The resulting decrease in fledgling number (or quality) per nesting attempt may be balanced by an increase in fledgling production per season or per lifetime. An alternative to the overmatching hypothesis is that birds are resource matching (an ideal free distribution: Fretwell and Lucas 1969). In areas with greater food resources, there will be greater numbers of birds. Because birds are matching their resources, however, per capita food availability will be equivalent across habitats and individuals will produce equal numbers of equal quality offspring (Leston and Rodewald 2006). Therefore, the ideal free distribution hypothesis predicts that there will be no negative consequences for individuals nesting at higher population densities in urban areas. In this paper we compare densities of the Northern Mockingbird ( Mimus polyglottos ) in urban and nonurban areas and examine evidence for differential availability of food during the breeding season. M ockingbirds are one of the most abundant urban birds in the southern US, but are al so found in nonurban areas (Chapter 2). The mockingbird is primarily a territorial insectivore during the breeding season and thus direct measurements of food availability are difficult to obtain. We therefore examine a suite of reproductive parameters th at should reflect food availability (Newton 1998). We compared clutch size, hatching success, prey delivery rates, food size and composition, nestling mass, and nestling survival. We control for nest predation by only examining nests that were not lost to predators. We do not examine fledgling production per female per year here (but see Chapter 5) because it is highly dependent on predation rates. In addition, we examine adult mass as a function of
78 habitat. If enhanced food resources cause the increased mockingbird abundance in urban areas, then we predicted that urban mockingbirds would produce more fledglings per nesting attempt than nonurban mockingbirds and that those nestlings would be of higher quality. If, on the other hand, there is a mismatch betw een food resources and population densities leading to resource overmatching, then we predicted that urban mockingbirds would produce fewer fledglings of lower quality per nesting attempt than nonurban mockingbirds. Furthermore, if this overmatching resul ts in a few winners in a populations of many losers (the credit card hypothesis), then we predicted that there would be higher variance in the quality and quantity of fledglings per nesting attempt as well as higher variance in adult mass in urban habitat s. The ideal free distribution hypothesis, on the other hand, predicts that individuals are distributed evenly with respect to food resources and that per capita productivity will not vary among habitats. Methods Study System and Territory Density Mockingbird nests were located in four habitat types (two parking lots, three residential neighborhoods, two pastures, and two wildlife preserves) in and around Gainesville, FL during 2005 2008 (Fig. 31). Nest searching began at all study sites in late February t o early March when the first birds started nest building, and ended in late July to early August when most birds had stopped nesting. The pasture sites were not added until 2006 and no nests were located in one of the wildlife preserves (Paynes Prairie St ate Park) after 2005. These habitats differed greatly in the average proportion of ground cover consisting of pavement and buildings (Table 31): we considered the parking lots and residential neighborhoods as representative of urban landuses and the past ures and wildlife preserves as representative of nonurban landuses. When sample
79 sizes were sufficient, we tested for effects of habitat type. When we did not have sufficient sample sizes, however, we made comparisons between urban and nonurban landuses When habitat type was significant we conducted post hoc comparisons using least squares mean differences with a Tukey adjustment. We used generalized linear mixed models ( proc mixed and proc glimmix SAS 9.2) using pair and study site as random effects and landuse or habitat as fixed effects for 2006 2008 data. When individuals were unbanded we used the location of the nest and the timing of egglaying to assign nests to different females. Data from 2005 were analyzed separately because a different subs et of study sites was used in 2005 and we could not use pair as a random effect because we were unable to reliably assign nests to different females because few birds were banded. We divided the number of females in each year at each study site by the area of the study site to estimate territory density in 2006 2008. In 2005 we estimated minimum territory densities at each study site by dividing the maximum number of simultaneous nests at each study site by the area of the study site. Because mockingbird nests are easy to find and we believe we found most nests at each site, this is a consistent, albeit conservative, estimate of population densities. For the 2005 data we tested for an effect of habitat type (parking lot, residential, etc.) on the density of mockingbirds. For 2006 2008 we tested for an effect of habitat type, year, and an interaction between the two on territory density. Number of Fledglings Produced Per Successful Nest and Per Hectare We calculated the average number of fledglings produced i n two ways. We calculated the average number of fledglings produced for nests that fledged at least one young at each study site to control for losses due to nest predation and complete clutch
80 abandonment, which may occur for a variety of reasons, includin g adult mortality. We also calculated the average number of fledglings produced per hectare as a measure of food availability at the habitat level. We used aerial images to calculate the number of hectares of each study site to examine the cumulative production of young from all pairs in each site during the entire nesting season. We compared each measure of fledging success across habitat types and used ordinal date (number of days since January 1) as a covariate for the number of successful nests per female. In 2005 and 2006 nest predation rates were significantly higher in nonurban landuses (Chapter 2), but in 2007 and 2008 there were no differences in nest predation rates (Chapter 3). Thus we can examine differences in fledgling production per unit area with and without the effects of nest predation. Clutch Size We collected data on clutch size from 2005 2008. We compared clutch size as a function of habitat (2005 2008) and year (2006 2008) with pair and site as random effects and date as a covariate. H atching Success and Fledging Success We compared the proportion of eggs laid that hatched and the proportion of nestlings that fledged across habitats. Maximum likelihood estimation was based on a binomial distribution (events/trial) with pair and site as random effects and date as a covariate. Data from 2005 were analyzed separately for both variables. Nestlings may fail to fledge for multiple reasons including brood reduction and nest predation. In 2007 and 2008 we used our data on nestling weight from m arked nestlings and assumed that when the smallest nestling disappeared it was a result of brood reduction and when any of the other nestlings disappeared it was a result of partial predation. We did not have
81 nestling mass data from 2005 2006 and so we used notes on the appearance of nestlings and on the pattern of nestling loss to estimate if missing nestlings were the result of brood reduction or partial predation. We did not use nestlings that were lost as a result of partial predation in this analysis. Nestling Mass We weighed nestlings to the nearest 0.05g when they were six days old in 2007 and 2008 and calculated the average nestling mass per nest. We compared average nestling mass across habitat types with date and year as covariates and site and pai r as random effects. We also tested for a correlation between nestling mass and number of nestlings. We did not calculate body condition indices because such measures, when unverified, do not improve precision, and in some cases decrease it, relative to body mass alone (Schamber et al. 2009 and citations therein). We also tested for homogeneity of variance in mass across habitats using Levenes test ( proc glm ) to assess the credit card hypothesis. Nestling Provisioning Rates, Average Food Size, and Food Type We observed nests in 2008 from a distance that did not affect the parents behavior for one hour when nestlings were six days old (approximately half way through the nestling phase) and placed video cameras at a subset of nests. We recorded the following data: number of nestlings in the nest, number of feeding trips, and, when possible, food size in relation to the birds bill and food type (fruit, Orthoptera, larvae, other). Habitat based differences in food size and food type c ould indicate differences in food quality. We conducted nest observations from video recordings for nests with video cameras. We also conducted nest observations from video recordings when nestlings
82 were three days old and again at nine days old, approximately three days prior to fledging. For each nest age (3, 6, and 9 days post hatching) we calculated the average number of nestlings per nest, feeding trips per hour per nestling, and food size. Because there was a significant interaction between nestling age and landuse on number of feeding trips (F4,40 = 24.46, P < 0.0001), we analyzed each age separately. We only analyzed food size at day six because of small sample sizes at ages three and nine. We also tested for a correlation between number of feeding trips and food size at a ge six to ensure that birds did not compensate for smaller food items by increasing prey delivery rates. To compare the proportion of different types of food delivered between landuse categories we calculated the number of trips for each of the following food types: fruit, Orthoptera, larvae (presumably mostly Lepidoptera), and other. We only used nest observations with 10 or more feeding trips because the proportion of food type can be highly skewed at smaller values. Because there was no correlation bet ween food size and number of feeding trips (see Results), this should not bias our estimates of food type. Furthermore, we restricted our analysis to observations in which we were able to identify at least 50% of the food items with no less than nine items identified. We compared the proportion of trips for each food type between landuses. Maximum likelihood estimation was based on a negative binomial distribution. W e restricted our analysis to the first observation for the pairs for which we observed mult iple nests (N=2) and did not use pair as a random effect in our mixed models
83 Adult Mass We captured adult mockingbirds at the nest with mist nets during incubation and feeding during 2005 2008 and sexed the birds based on the presence or absence of a brood patch as only females incubate. We weighed each bird to the nearest 0.5 g with a spring scale in 2005 2007 and to the nearest 0.05 g with a digital scale in 2008. We analyzed males and females separately and compared body mass across habitat types (for females) and land use (for males) with year and date as covariates and site as a random effect. We also tested for homogeneity of variance in female mass across habitats using Levenes test ( proc glm ) to test the credit card hypothesis. We restricted our analysis to 2006, when sample sizes were the highest, to avoid introducing variance between years. For males we could only compare variance in mass between urban and nonurban landuses because of small sample sizes. Results We located 970 mockingbird nests during the course of the study. In 2005, parking lots and residential neighborhoods had significantly higher mockingbird densities than the wildlife preserve (Fig. 41a). In 2006 2008, residential neighborhoods had higher territory density than any of the other habitats (Fig. 41b). Territory densities in parking lots were also higher than territory densities in pastures and wildlife preserves (Fig. 41b). Number of Fledglings Produced Per Successful Nest and Per Hectare There were no significant differenc es in the number of fledglings per successful nest based on habitat type for both 2005 and 2006 2008 (Table 41). Habitat had a significant effect on the number of fledglings produced per unit area in both 2005 (Fig. 4 2a) and 2006 2008 (Fig. 42b). Residential neighborhoods produced significantly
84 more fledglings per unit area than wildlife preserves in 2005 (Fig. 42a) and both parking lots and residential habitats produced significantly more fledglings per unit area than pastures and wildlife preserves in 2006 2008 (Fig. 42b). Clutch Size In 2005, there were no differences in clutch size among habitat types (F2,98 = 0.73, P = 0.49; Fig. 43a). There was a significant difference among habitats in clutch size in 2006 2008 (F3,281 = 5.81, P = 0.0007; Fig. 43b). This difference in clutch size resulted from more eggs being laid in pastures than in any other habitat. The greatest difference in clutch size was between pastures and wildlife preserves, with pastures having 0.34 more eggs per clutch. Hatching Success and Fledging Success Hatching success did not differ among habitat types in 2005 or in 2006 2008 (Table 41). There were also no significant differences among habitats in the proportion of nestlings that fledged (Table 41). Nestling Mass There was no correlation between the number of nestlings in a nest and average nestling mass per nest (r = 0.05, P = 0.64). There was no interaction between year and habitat (F3,10 = 0.64, P = 0.60) on nestling mass at age six, nor was there an effect of year (F1,10 = 0.20, P = 0.66). Likewise, there was no relationship between nestling mass and habitat type (Table 41). There was no significant difference in the variance of mass between habitats (F3,77 = 0.74, P = 0.53). Nestling Provisioning Rates, Average Food Siz e, and Food Type At nestling ages three, six, and nine days there were no differences in the number of feeding trips per nestling per hour between urban and nonurban landuse (Table 4-
85 2). There was a significant effect of habitat on number of feeding trips per nestling per hour at age six days (Fig. 44) with more trips per nestling in residential than in pasture. Average food size was greater in nonurban landuse at age six (Fig. 45). There was no correlation between food size and number of trips per nestling per hour at age six days (r = 0.04, P = 0.93). The proportion of feeding trips with fruit, Orthoptera, and larvae were not significantly different at age six days in urban and non urban land use (Table 43). There was, however, a trend for more Orthoptera brought to nests in nonurban landuse than in urban land use (Table 43). Adult Mass There were no differences in female mass among habitats (Table 41). Nor was there a difference in variance in female mass in 2006 among habitats (F3,57 = 1.00, P = 0.40). Likewise, there was no difference between the mass of urban males (47.47+/ 1.09) and nonurban males (47.67+/ 1.00; F1,34 = 0.02, P = 0.89) or the variance of male mass in 2006 between urban and nonurban landuse (F1,26 = 3.57, P = 0.07). Discussion Nesting success was strikingly similar in urban and non urban habitats even though non urban birds brought significantly larger food to their young at age 6 days and pastures had larger clutch sizes. Thus, our results provide no support for the urban f ood enhancement hypothesis, at least at the individual level, and no evidence of a mismatch between food resources and mockingbird abundance in urban areas. We also find no evidence for the credit card hypothesis (a few winners and lots of losers in a population) as variances in both nestling and adult body masses were equivalent across habitats. These results are not consistent with the majority of urbanization studies,
86 which encompass urban avoiders, adapters, and exploiters, that found fewer and smaller fledglings were produced per nesting attempt in urban areas (reviewed in Chamberlain et al. 2009, but see Rodewald and Shustack 2008a). In this study fledgling production per unit area was significantly higher in urban areas, even in years when there were no differences in losses to nest predators (2007 2008, Chapter 3), which suggests that food resource availability was likely higher in urban areas at the population level. These extra resources, however, did not translate into higher nesting success per nesting attempt of individual pairs. Previous studies of the effects of urbanization on the diets and nesting success of species along an urban gradient have shown that, in general, urban nestlings are in worse condition than their nonurban counterparts, particularly when adults incorporate anthropogenic food into their nestlings diet (Shawkey et al. 2004, Mennechez and Clergeau 2006, Newhouse et al. 2008, Chamberlain et al. 2009, but see Rodewald and Shustack 2008a). While we never observed mockingbirds f eeding anthropogenic food to their nestlings, prey delivered when nestlings were six days old in urban areas was significantly smaller than prey in nonurban habitats, perhaps as a result of frequent mowing and/or pesticides, which might kill larger Orthopterans, in particular. Nevertheless, the number of nestlings being produced in nests that escaped predation and the proportion of young hatched that fledged was high in all habitats, which suggests that food quality was not markedly lower in urban landuses. Alternatively, mockingbirds might compensate for a reduction in food quality by increasing the number of feeding trips per nestling per hour (as seen in the residential habitats at nestling age six) and by feeding their nestlings later into the night i n the
87 parking lots as a result of artificial light (unpubl. data). Anecdotally, we observed nestlings being fed at 10:30pm in one of the parking lots. Other urban management practices such as regular watering of lawns during periods of drought may partiall y compensate for reductions in Orthopterans by increasing populations of other less drought tolerant arthropods; very little, however, is known about arthropod responses to urbanization (McIntyre 2000, Ishitani et al. 2003, Shochat et al. 2004b, Hartley et al. 2007). Our lack of evidence for a mismatch between food resources and mockingbird populations may reflect their strong territory defense, which gives them access to a reliable supply of insects during the breeding season and which could be adjusted to reflect variation in food supply resulting in different sized territories (e.g., Smith and Shugart 1987). Birds that are unable to defend territories may exist at inflated population densities, and hence exhibit a resource mismatch close to sites with ref use and bird feeders where the costs of territory defense outweigh the benefits (Brown 1964). Many mockingbirds, however, also aggressively defend territories around fruiting shrubs in winter (Moore 1978, Hedrick and Woody 1983), which might lead to a resource mismatch if population densities reflect winter resource availability rather than breeding season availability. Under these circumstances, territories with abundant winter fruit might become ecological traps during the breeding season if individuals f ail to emigrate in response to lower breeding season resources ; our results, however, show no evidence for such a mismatch. The Northern Cardinal ( Cardinalis cardinalis ), which uses bird feeders extensively in winter but is territorial during the breeding season, also shows little evidence of a resource mismatch during the breeding season
88 when they feed their young mainly arthropods (Rodewald and Shustack 2008a). Species that feed on defendable resources may therefore be less vulnerable to resource mismatches, provided they do not incorporate anthropogenic food into their nestlings diet, than species that do not defend territories. The vulnerability of species to resource mismatches may thus depend upon their territoriality Understanding why certain speci es are distributing themselves according to an ideal free distribution or are mismatching their resources will likely require data on more than just food. Researchers should consider the role of food during the breeding season in conjunction with additiona l factors such as nest predation, nest site availability, winter food abundance and distribution, survival of fledglings and adults, dispersal, and predation on adults. Examining each factor independently may leave us with incomplete explanations for the abundance of urban adapters and exploiters as individuals may be forced to make tradeoffs between multiple factors affecting lifetime reproductive success.
89 Table 41. Average (+/ SE) number of fledglings produced per successful nest, proportion of eggs l aid that hatched, proportion of nestlings that fledged, nestling mass at age six days, adult female and male mass, and the results of significance tests for each response variable. PL = parking lot, RES = residential, PAST = pasture, WP = wildlife preserve PL RES PAST WP Significance Test Number of fledglings produced per successful nest (2005) 2.69 (0.20) 2.70 (0.18) n/a 3.01 (0.37) F 2,43 = 0.33, P = 0.72 Number of fledglings produced per successful nest (2006 2008) 2.73 (0.09) 2.65 (0.10) 2.9 3 (0.12) 2.64 (0.20) F 3,78 = 1.25, P = 0.30 Proportion of eggs laid that hatched (2005) 0.75 (0.06) 0.81 (0.05) n/a 0.87 (0.06) F 2,67 = 0.81, P = 0.45 Proportion of eggs laid that hatched (2006 2008) 0.87 (.02) 0.87 (0.02) 0.88 (0.02) 0.92 (0.03) F 3, 140 = 0.75, P = 0.52 Proportion of nestlings that fledged (2005) 0.93 (0.03) 0.97 (0.02) n/a 0.86 (0.07) F 2,43 = 1.64, P = 0.21 Proportion of nestlings that fledged (2006 2008) 0.94 (0.02) 0.92 (0.02) 0.89 (0.03) 0.83 (0.07) F 3,79 = 1.41, P = 0.25 Nestling mass 25.95 (0.71) 25.79 (0.73) 26.87 (0.44) 24.51 (0.48) F 3,10 = 0.49, P = 0.70 Adult female mass 47.39 (1.14) 46.44 (1.01) 45.68 (0.62) 46.23 (1.51) F 3,100 = 0.28, P = 0.84
90 Table 4 2. The average (+/ SE) number of feeding trips in 2008 per nestling per hour at nestling age three, six, and ninedays in urban and non urban land use and the results of ANOVAs for each age. Urban Non urban Significance Test Age 3 2.35 (0.22) 1.73 (0.17) F 1,3 = 4.81, P = 0.12 Age 6 3.93 (0.30) 3.40 (0.38) F 1 ,13 = 1.21, P = 0.29 Age 9 4.79 (0.74) 5.52 (0.66) F 1,17 = 0.53, P = 0.48 Table 4 3. The mean (+/ SE) percentage of feeding trips consisting of fruit, larvae, and Orthoptera to nestlings age six days in urban and nonurban landuse and the results of significance tests. Urban Non urban Significance Test Fruit 5.47 (2.38) 1.27 (0.86) F 1,18 = 0.48, P = 0.50 Larvae 9.04 (2.52) 10.91 (3.78) F 1,18 = 0.12, P = 0.74 Orthoptera 5.39 (2.93) 26.09 (7.08) F 1,18 = 3.38, P = 0.08
91 Figure 41. The average (+ / SE) number of mockingbird pairs per hectare in four habitat types in a) 2005 (F2,4 = 27.64, P = 0.005) and b) 20062008 (F3,10 = 45.32, P < 0.0001). Letters indicate significance at Tukey PL = parking lot, RES = residential, PAST = pas ture, WP = wildlife preserve.
92 Figure 42. Average number of fledglings produced per hectare in a) 2005 (F2,4 = 8.65, P = 0.035) and b) 20062008 by habitat type (F3,78 = 15.42, P = 0.0004). Letters indicate significance at Tukey parking lot, RES = residential, PAST = pasture, WP = wildlife preserve.
93 Figure 43 Average clutch size in four habitat types in a) 2005 (F2,98 = 0.73, P = 0.49) and b) 20062008 (F3,98 = 5.81, P = 0.0007). Letters indicate significance at Tukey adjus WP = wildlife preserve.
94 Figure 44. Average number of feeding trips per hour per nestling in each habitat type (F3,13 = 4.02, P = 0.032). Letters indicate significance at Tukey adjusted 0.05. PL = parking lot, RES = residential, PAST = pasture, WP = wildlife preserve.
95 Figure 45. Average food size relative to adult bill length delivered to six day old nestlings in urban and nonurban landuse in 2007 2008 (F1,12 = 5.74, P = 0.033).
96 CHAPTER 5 IS AN URBAN ADAPTER, THE NORTHERN MOCKING BIRD, MORE PRODUCTIVE IN URBAN HABITATS? Introduction Urbanization profoundly alters ecosystem structure and function with numerous consequences for wildlife communities. Such habitat alteration frequently leads to a loss of species and many native species are replaced by nonnative urban exploiters (Blair 1996, Marzluff et al. 2001, Shochat et al. 2006). On the other hand, some native species, termed urban adapters, are able to survive in both urban areas and more natural habitats (Blair 1996, Shochat et al. 2006). Understanding the factors that enable these species to flourish in habitats where many species are lost is important if we are to understand the factors that promote biodiversity in urbani zed habitats. There are at least four processes that can lead to higher densities of birds in cities: (1) urban birds may produce more offspring (either because of higher food [Chapter 4] or reduced nest predation [Chapter 23]), (2) urban birds may have higher survival, (3) urban birds may be more site faithful, and (4) urban areas may be more attractive to dispersing birds. In these latter two cases, urban areas may attract more birds even in the absence of high productivity or survival in which case t hey may act as an ecological trap that attracts birds but fails to provide the conditions necessary to maintain viable populations ( Gates and Gysel 1978). The ecological trap scenario is especially plausible in urban areas because conditions are extremely different from those prevailing in natural environments. Urban areas have additional point sources of food such as bird feeders and ornamental fruiting shrubs, extra water, and vegetation that is held in a permanently disturbed state as a result of lawn mowing and pruning. Such artificial environments may be extremely attractive to birds that use food or vegetation
97 as cues for selecting habitats. These habitats, however, also have novel communities of nest predators (Chapters 2 and 3) and sources of mortality such as cars, picture windows, and toxins. Before we can tease apart the factors that lead to higher densities of urban birds, we need to determine season long productivity, estimate survival, and examine the cues used by adults to return to a sit e (site fidelity). Few studies have addressed productivity, survival, and site fidelity in urban landscapes (but see Rodewald and Shustack 2008b). Urban birds tend to have lower productivity per nesting attempt than their nonurban counterparts (reviewed in Chamberlain et al. 2009). On the other hand, urban birds begin breeding earlier than nonurban birds (reviewed in Chamberlain et al. 2009) and may be able to compensate for reduced productivity per nesting attempt. Making generalizations from previous studies, however, is challenging because studies differ in the point along the urbanization gradient that they compare (e.g., city center versus rural and city center versus low density suburban), as well as in the species studied (e.g., urban adapters and urban avoiders). Species that are apparently successful urban dwellers should be more likely to have higher urban productivity than species that are urban avoiders unless urban areas are acting as an ecological trap. Post fledging survival and recruitment may also be reduced in urban areas because of increased mortality at roads (Burger and Gochfield 1992, Mumme et al. 2000) and buildings (Klem 1990, Hager et al. 2008) and predation by introduced predators such as cats (Crooks and Soule 1999, Lepzcyk et al. 2003). Studies that have used apparent survival estimates of juveniles have found reduced juvenile survival in urban habitats (Rollinson and Jones 2002, Beck and Heinsohn 2006). The only study
98 to quantify post fledging survival with telemetry, howeve r, found species specific patterns of survival in relation to urbanization (Whittaker and Marzluff 2009). Two of the four species studied had lower survival in more urbanized landscapes whereas the other two species showed no relationship between survival and urbanization. Similarly, adult mortality may also be higher in urban habitats as a result of collisions with cars and other human structures (Klem 1990, Burger and Gochfield 1992, Mumme et al. 2000, Hager et al. 2008), and losses to urban predators (S oule et al 1988, Crooks and Soule 1999, Sorace 2002). Increased food resources in winter, on the other hand, may increase survival. Adult survival, though, is difficult to quantify, which has led researchers to quantify either perceived predation risk (as measured by giving up densities: Brown 1988) or apparent survival (habitat specific return rates). Perceived predation risk appears to be reduced in urban habitats for granivores (Shochat et al 2004, Valcarcel and Fernandez Juricic 2009). Researchers have documented apparent survival to be either higher in urban habitats (Horak and Lebreton 1998), or equivalent (Leston and Rodewald 2006, Rodewald and Shustack 2008b) in urban and rural habitats. Apparent survival, however, reflects both survival and dis persal (Hoover 2003); differences in apparent survival between habitats could actually be the result of different nesting success in the two habitats. Many studies (e.g., Switzer 1993, Haas 1998, Hoover 2003) have shown that individuals are far more likely to return to a site following successful breeding than they are if they fail in the previous year. Return rates of successfully nesting birds may provide a more accurate estimate of adult survival than those that measure return rates for the entire population (Hoover 2003).
99 In this chapter, we compare nesting productivity, estimated survival, and decision rules governing site fidelity of Northern Mockingbirds in urban and rural habitats. Because the mockingbird is an urban adapter that is more abundant in urban than nonurban habitats, we predicted that (1) urban mockingbirds would begin breeding earlier than nonurban mockingbirds, (2) more young would be produced per female per year in urban habitats, (3) site fidelity within and between years would b e higher in urban habitats, (4) urban habitats would produce young at or above the sourcesink threshold (they would produce more young than necessary to compensate for estimated adult mortality), and (5) urban mockingbirds would have higher apparent survi val than nonurban mockingbirds. Alternatively, urban areas may act as ecological traps, in which case survival and reproductive success would be lower in urban habitats and reproductive success would fall below the estimated sourcesink threshold. Methods Data Collection Mockingbirds were located in four habitat types (two parking lots, two residential neighborhoods, two pastures, and one wildlife preserve) in and around Gainesville, FL during 2005 2008 (Fig. 31). These habitats differed greatly in the average proportion of ground cover consisting of pavement and buildings (Table 31). Study sites were between 2.7 km and 52.9 km apart and we never recorded banded birds moving between sites. The pastures were added in 2006 and in 2007 a third residential neighborhood was used. Data from all four years were used to estimate adult survival and juvenile returns; however, only data from 2006 2008 were used in our productivity estimates.
100 Nest searching and band sightings began at all study sites in late February to early March when the first birds started nest building and ended in late July to early August when most birds had stopped nesting. Mockingbird nests were located at each study site and adults were captured with mist nets while incubating (females only) or when feeding nestlings (males and females) and color banded. Territory boundaries were approximated from resightings on banded birds, aggressive encounters, and nest locations. When individuals were unbanded we used the location of nests and the timing of egglaying to assign nests to different females and obtain approximate territory boundaries. While these methods do not allow us to precisely define territory boundaries, we are able to use them to assess territory occupation and assign unband ed birds to a particular territory. We located mockingbird nests, monitored them every one to four days, and recorded nest contents until the nest became inactive. Nestlings were color banded in the nest in 2005 and 2006. In 2007 and 2008 nestlings were only banded with a USGS aluminum band. For each known pair we determined the ordinal date (number of days since January 1) that their first clutch was completed. We restricted this analysis to females that were present the entire breeding season (see ter ritory occupation below). For nests that we found after laying we back calculated the clutch completion date based on a 12 day incubation period and a 12 day nestling period (Derrickson and Breitwisch 1992). We did not test for endof season differences in when the birds stopped breeding because search efforts at the end of the season were not consistent across sites. For each female in each year we calculated the total number of fledglings produced across all successful nests, the number of successful nests, and the number
101 of nesting attempts. In some cases we likely did not find every nesting attempt for each pair based upon long breaks between documented nests. These pairs were excluded from our analysis of the number of nesting attempts per female. They were not excluded from our analysis of start date, total number fledged, and successful nesting attempts when there was not enough time between located nests to successfully fledge another nest. We did not exclude these pairs from our analysis of territory occupation (see below). To estimate withinseason movement, we compared the proportion of the breeding season each territory was occupied. To estimate territory occupation we first defined the length of the breeding season for each study site by subtracting the ordinal date of the first clutch at each site from the date the last clutch was completed. We then estimated when each pair appeared on the breeding territory from field notes on the presence of individuals on each territory. When there were no notes for a particular pair, we estimated their arrival/departure date based on their first/last nest of the season. We subtracted 14 days from the first nest and added 14 days to the last nest when it was depredated and 30 days when the nest succ essfully fledged. We then divided the number of days each pair was present by the total length of the breeding season at that site. We then assigned each pair into one of two categories: present all season or present less than 90% of the season. Statistical Analyses We compared the average first clutch completion date per female, total fledgling production per female, number of successful nests per female, number of nesting attempts per female, and territory occupation using generalized linear mixed model s (GLMM) with habitat and year as fixed effects and site as a random effect. We log-
102 transformed first clutch completion date to meet assumptions of normality and analyzed the data in PROC MIXED (SAS 9.2). We used a Poisson distribution to model both the n umber of successful nests per female and the number of nesting attempts per female and a negative binomial distribution to model total fledgling production per female (PROC GLIMMIX, SAS 9.2). When there was a significant interaction between year and habit at we analyzed each year separately. We performed post hoc comparisons for significant habitat effects using least squares mean differences with a Tukey adjustment. We also calculated the proportion of banded nestlings returning as breeding adults in eac h habitat to estimate juvenile survival. Full model results are presented in Appendix B. We used maximum likelihood estimation in program MARK (White and Burnham 1999) based on live encounters/recapture models to compare differences in return rates between habitat types. We used data from banded after hatchyear individuals to estimate apparent annual survival and detection probabilities of banded birds with an information theoretic approach (Burnham and Anderson 2002). Due to small sample sizes, we were unable to fit timespecific models; therefore we excluded time from our final sets of candidate models. We first tested for a difference in apparent annual survival as a function of sex (50 males and 116 females). Because males and females did not differ in apparent annual survival (males = 0.653 +/ 0.049 and females = 0.642 +/ 0.046), we combined sexes for the habitat analysis. To investigate differences in decision rules of female mockingbirds in different habitats, we used generalized linear models (PROC GENMOD, SAS) to analyze the probability of a female returning to a study site as a function of her success in the
103 previous year. We defined success in two ways: whether or not a female had at least one successful nest and the number of successful nests (we combined double and t riple brooded females into 2+ successful nests). Initially, we only used the first year for which we had banding data for each female. Because sample sizes were small, however, we then assumed that a females decision to return was independent between ye ars and added the data from additional years for females that were present in more than one year. We present results from both of these analyses. We did not investigate decision rules of males because of small sample sizes. Source/Sink Threshold We were also interested in comparing populationlevel productivity between habitats. We estimated the number of female fledglings needed to offset adult and juvenile mortality (the source/sink threshold) for each habitat based upon three different survival scenar ios. Because juvenile return rates were so low (see Results), we could not use direct estimates of juvenile survival. In the first case we assumed female survival was uniform across habitats using our estimate for overall female survival from MARK and as sumed juvenile survival was half that of adults. For the second scenario we used the habitat specific estimates of female survival and assumed the same constant juvenile survival across habitats as in the previous scenario. In the third case we assumed j uvenile survival was half of the habitat specific adult survival. For each scenario we compared the estimated source/sink threshold to the GLMM predicted estimates of number female fledglings produced per female per year to assess the viability of the population in each habitat type.
104 Results Nesting Season Initiation In most years, mockingbirds started nesting earlier in urban areas than in rural areas. There was a significant interaction between year and habitat for the average first clutch completion dat e for pairs that were on their territories the entire breeding season (Table 51). Habitat was significant when we analyzed each year independently and females in the wildlife preserve consistently started nesting later (Fig. 51). In 2006, both pastures and wildlife preserve had later start dates (Fig. 51b). Nesting Productivity Habitat had a significant effect on the total number of fledglings produced per female per year and the number of successful nests per female per year (Table 51). The number of fledglings produced increased with urbanization; females in parking lots produced the most fledglings and females in wildlife preserves produced the fewest (Fig. 52). There was no interaction between year and habitat for the remaining variables (Table 5 1). Similarly, females in parking lots had more successful nests than females in pastures (Fig. 53). There was no effect of habitat on the number of nesting attempts per female (Table 51). Estimated Survival Very few birds banded as nestlings were r esighted as breeding adults and there did not appear to be habitat specific differences in juvenile recruitment (Table 52). Apparent adult survival of mockingbirds was affected by habitat (Fig. 54). Indeed, the model that included habitat in the estimate of survival was the only model with strong support c < 2, Table 53). Parking lots and residential neighborhoods had higher apparent survival estimates than pastures and wildlife preserves (Fig. 54).
105 Decision Rules Governing Site Fidelity Fewer pairs in the pastures remained on their territory fo r the entire breeding season than in residential neighborhoods, where most territories were occupied for the entire breeding season (Fig. 55). There was a significant interaction between a females success (successful versus unsuccessful) in a previous y ear and habitat on her decision to return to a site to breed the following year (Table 54). Females in parking lots and residential neighborhoods were more likely to return the following year if they nested successfully the previous year (Fig. 56). Suc cessful and unsuccessful females in pastures and wildlife preserves, however, showed little difference in their probability of returning (Fig. 56). The patterns were the same regardless of whether we used the first year of data for each female only or we used all years of data and assumed that a females decision was independent between years. When we examined how the number of successful nests in the previous year affected return rates, there was an interaction between success and habitat when we restri cted our analysis to a females first year of data only (Table 54). Females in residential habitats showed a stepwise increase in return rates with additional successful nests (Fig. 57a). Females in parking lots did not appear to distinguish between one and twoplus successful nests. Only one bird in the pastures and one in the wildlife preserve were doublebrooded and they both returned. The pasture and wildlife preserve females did not appear to distinguish between zero and one successful nest (Fig. 5 7a). When we use data for all the years a female was present there was no longer a significant interaction between habitat and success and only the number of successful nests significantly predicted a females return (Table 54). While not statistically significant,
106 the patterns in both analyses (one year of data per female versus all years combined) were similar (Fig. 5 7). SourceSink Analyses The number of fledglings produced per female per year (based on GLMM predictions) in pastures and wildlife preserves was below the estimated source/sink threshold regardless of the scenarios we used (constant adult and juvenile survival; habitat specific adult survival and constant juvenile survival, and habitat specific adult and juvenile survival: Table 55) The number of fledglings produced by parking lot females exceeded the source/sink threshold under all three assumptions of survival, whereas the residential females produced fledglings in excess of the threshold in all scenarios but the first (Table 55 ). Discussion The Northern Mockingbird, which we had previously defined as an urban adapter based on its greater abundance in urban than in rural habitats (Chapter 2), also appears to be an urban adapter in terms of population viability. We found no evidence for the ecological trap hypothesis urban birds produced more young per season than rural birds and their annual productivity was above the estimated sourcesink threshold regardless of the estimates for adult and juvenile survival that we used. Indeed, birds in rural habitats seemed to have productivity below the sourcesink threshold, although we may have underestimated seasonlong productivity because at least some rural mockingbirds were not present for the entire breeding season (Fig 55) and may have bred elsewhere during the remainder of the breeding season (Jackson et al. 1989). We had no indications that urban mortality rates were higher for adults or fledglings. Urban mockingbirds also seem to have decision rules regarding site fidelity that should enable
107 them to avoid ecological traps. In urban habitats (but not in rural habitats see below), mockingbirds returned mainly to sites where they nested successfully in previous years and tended to disperse away from sites where their nesting succes s was low. This decision rule should be adaptive in an environment in which nest predation rates are relatively predictable from year to year (Chapter 3) and where some sites may have predictably low levels of nest success (Hoover 2003). The extremely low return rates of juvenile mockingbirds (Table 52), however, do not allow us to provide meaningful estimates of juvenile survival, a common problem in sourcesink models (Brawn and Robinson 1996). Despite the numerous factors (e.g. cars, cats, and buildin gs) in urban areas that could decrease juvenile survival, it is far from clear that juvenile survival is actually reduced in urban habitats (Whittaker and Marzluff 2009). It is important that future studies assess post fledgling survival so that accurate estimates of population dynamics in urban and nonurban habitats can be obtained. Higher urban productivity was likely driven, in part, by differences in nest predation rates. Nest predation rates, however, did not differ between urban and nonurban habit ats in 2007 and 2008 (Chapter 3), yet there were no interactions between productivity and year. Therefore, nest predation alone can not account for higher urban productivity. Differences between fledgling productivity in urban and nonurban populations i s likely to be highly species specific. Urban avoiders, which are negatively affected by urbanization, should show the opposite pattern. The lower estimated survival of nonurban adult mockingbirds could result from higher mortality, but it may also reflect higher dispersal. Reduced winter food resources
108 could be a source of higher mortality. Urban mockingbirds remain on their breeding territories the entire year at these study sites whereas nonurban birds disappear from their breeding territories in t he winter (unpubl. data). This difference in winter territoriality is presumably driven by differences in food availability. Lower food availability and travel costs could result in higher mortality of nonurban birds. Future studies should follow indiv idual birds throughout the year to establish differences between mortality and dispersal. Urban and rural birds appear to use different decision rules regarding site fidelity in relation to nesting success. Urban birds were significantly less likely to r eturn following an unsuccessful year, whereas there were no significant differences between the return rates of successful and unsuccessful females in nonurban habitats. An individuals use of prior experience as a determinant of site fidelity should be dependent on the predictability of the habitat (Switzer 1993, Hoover 2003). Our results are consistent with the predictability of nest predation rates: nest predation rates in parking lots and residential neighborhoods are consistent from year to year; however, nest predation rates in pastures and wildlife preserves were variable (Chapter 3). Furthermore, urban habitats experience reduced temporal variation in other biotic and abiotic factors, such as temperature (Brazel et al. 2000, Shochat et al. 2006 and references therein) relative to nonurban habitats. If nonurban habitats are more variable, then the availability of suitable habitat may be a more important factor governing the decision rules of nonurban birds. Mockingbird foraging habitat, short grass and bare ground, is constantly maintained in peoples lawns and in parking lots, but is dependent upon grazing and fire in rural locations. Indeed, six pairs of unbanded mockingbirds appeared in June 2009 in
109 locations where they had never been previ ously recorded at the wildlife preserve following an extensive burn in May (pers. obs.). The low withinseason site fidelity (Fig. 5 5) may further reflect the unpredictability of mockingbird nesting habitat in rural areas where they may depend more upon both natural (fire) and unnatural (grazing) disturbances. While withinseason dispersal might provide additional nesting opportunities, withinseason dispersal has been correlated with increased mortality in grassland birds (Perlut et al. 2008) The context dependent decision rules used by mockingbirds have also been documented in Acadian Flycatchers, an urban avoider (Rodewald and Shustack 2008b). Acadian Flycatchers in forests surrounded by urban habitat were more likely to return following successful nests than unsuccessful nests, but individuals in forests surrounded by rural areas showed no difference in return rates based on their success in the previous year. While there were no differences in predation rates in relation to urbanization, nests in urban forests were more likely to be parasitized and thus experienced lower productivity than forests surrounded by rural habitat. Further s tudies are necessary to tease apart the role that environmental predictability and habitat availability play in determining whether birds return to urban and nonurban breeding sites. The success of the Northern Mockingbird in cities and towns, particularly in the southeast, is striking. Based on these data, urban (parking lot and residential) mockingbirds both produce more fledglings per female per year and have higher apparent survival than nonurban (pasture and wildlife preserve) mockingbirds, with t he number of offspring produced exceeding that required to maintain a stable population.
110 On the other hand, productivity in the nonurban habitats was not sufficient to offset lower apparent survival rates. Thus the mockingbirds status as an urban adapt er stems from its ability to successfully produce young and its high survival in urban habitats.
111 Table 51. Statistical significance of GLMMs for average start date, total number of fledglings produced per female per year, number of successful nests per female per year, number of nesting attempts per female per year, and the proportion of birds on their territory the entire breeding season. PL: parking lot, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve. Model df F P Average start d ate Habitat 3, 140 11.08 <0.0001 Year 2, 140 0.96 0.39 Habitat*year 6, 140 2.34 0.035 Total fledglings produced per female per year Habitat 3, 331 3.16 0.025 Year 2, 331 4.26 0.015 Habitat*year 6, 331 1.3 0.26 Number of successful nests per female per year Habitat 3, 348 4.9 0.0024 Year 2, 348 1.23 0.29 Habitat*year 6, 348 1.17 0.32 Number of nesting attempts per female per year Habitat 3, 355 1.59 0.19 Year 2, 355 1.62 0.20 Habitat*year 6, 355 1.04 0.40 Proportion of b irds on the territory the entire breeding season Habitat 3, 239 5.37 0.0014 Year 2, 239 4.77 0.009 Habitat*year 6, 239 0.43 0.86 Table 52. The total number of nestlings banded, the number that returned as breeding adults, and the proportion of banded nestlings that returned to breed in each habitat. Habitat Number banded Number Returned Proportion returned Parking Lot 82 3 0.037 Residential 93 2 0.022 Pasture 68 0 0.000 Wildlife preserve 44 1 0.023
112 Table 53. Model selection results of Cormack Jolly Seber models used in program MARK ranked by AICc (adjusted for small sample size) for the Northern Mockingbird. Model a K b 2Log(L) c AIC c c d w i e p (.) 5 379.497 389.749 0.000 0.745 p (g) 5 383.282 393.534 3.785 0.112 p (g) 8 377.256 393.870 4.121 0.095 p (.) 2 391.202 395.252 5.502 0.048 a p : detection probability; .: no variation. b Number of parameters in model. c 2*loglikelihood value of a model. d Scaled value of A ICc e Akaike weight.
113 Table 54. Statistical significance of GLMs for the probability of a female returning to breed based on habitat and her breeding success in the previous season. Success refers to whether or not she had at least one successful nest in the previous season. Treatment df 2 P Data from a females first banded year only Habitat 3 4.25 0.24 Success 1 4.24 0.040 Habitat*success 3 8.55 0.036 All years combined Habitat 3 3.41 0.33 Success 1 8.21 0.004 Habitat*success 3 7.58 0.056 Data from a females first banded year on ly Habitat 3 1.2 0.75 Number of successful nests 2 11.35 0.003 Habitat*number of successful nests 6 13.84 0.031 All years combined Habitat 3 1.48 0.69 Number of successful nests 2 10.59 0.005 Habitat*number of successful nests 6 8.5 0.20 Table 55 Estimation of the source/sink threshold in each habitat under three different adult and juvenile survival scenarios and the number of female fledglings per female per year predicted from the GLMM. Habitat Adult female mortality Juvenile Survival Source/ Sink Threshold Model Predicted Female Fledglings/Female/Year Scenario 1 Assuming uniform adult survival Parking Lot 0.362 0.319 1.133 1.483 Residential 0.362 0.319 1.133 1.043 Pasture 0.362 0.319 1.133 0.764 Wildlife Preserve 0.362 0.319 1.133 0.753 Scenario 2 Assuming habitat specific adult survival and uniform juvenile survival Parking Lot 0.343 0.319 1.075 1.483 Residential 0.270 0.319 0.845 1.043 Pasture 0.520 0.319 1.629 0.764 Wildlife Preserve 0.489 0.319 1.533 0.753 Scenari o 3 Assuming habitat specific adult and juvenile survival Parking Lot 0.343 0.328 1.045 1.483 Residential 0.270 0.365 0.739 1.043 Pasture 0.520 0.240 2.165 0.764 Wildlife Preserve 0.489 0.255 1.916 0.753
114 Figure 51. The average (+/ SE) ordinal date of the completion of a females first clutch in each habitat type in a) 2006, b) 2007, and c) 2008. PL: parking lot, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve. Letters represent significance at experiment
115 Figure 52. The average (+/ SE) number of total fledglings produced per female per year in each habitat type. PL: parking lot, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve. Letters represent significance at experiment
116 Figure 53. The average (+/ SE) number of successful nests per female per year in each habitat type. PL: parking lot, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve. Letters represent significance at experiment
117 Figure 54. Adult survival probability (+/ SE) in each habitat type estimated in program MARK. PL: parking lot, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve.
118 Figure 55. The proportion of pairs (+/ SE) that occupied their territory the entire breeding season by habitat type. PL: parking lot, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve. Letters represent significance at experiment
119 Figure 56. The probability (+/ SE) of a female returning to a site following her success in the previous season in each habitat type. a) Data are restricted to the first year of data for females that were present in multiple years. b) Data for females present in multiple years are pooled. PL: parking lot, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve.
120 Figure 57. The probability (+/ SE) of a female returning to a site following the number of nests that fledged young in the previous season in each habitat type. a) Data are restricted to the first year of data for females that were present in multiple years. b) Data for females present in multiple years are pooled. Sample sizes are indicated at the base of each bar. PL: parking lot, RES: residential neighborhood, PAST: pasture, WP: wildlife preserve.
121 CHAPTER 6 CONCLUSIONS The Northern Mockingbird is a classic urban adapter it is a native species that is quite abundant in urban areas, but which also occurs at lower abundances in nonurban habitats (Chapter 2). Abundance alone however, is not an accurate assessment of habitat quality (Van Horne 1983) and urban areas could represent sinks (Pulliam 1988) or even ecological traps (Gates and Gysel 1978) for urban adapters (but see Leston and Rodewald 2006). Based on this research urban areas do not represent sink or trap habitats for mockingbirds. Productivity of urban pairs exceeded that of nonurban pairs and more than offset estimated adult mortality (Chapter 5). Apparent adult survival was also higher in urban habitats than in nonurban habitats, although this could be driven by dispersal and not solely mortality (Chapter 5). Therefore, I conclude that urban landscapes represent good mockingbird habitat. Additional food resources and reduced nest predation rates are reasons often cited for the success of urban adapters (Marzluff et al. 2001, Shochat et al. 2004a, Faeth et al. 2005, Shochat et al. 2006). I found no evidence of increased food availability on a per capita basis (Chapter 4). Nestling quality and quantity were similar per nesting attempt in urban and nonurban habitats. There were some indications (s maller prey delivered at urban nests) that nestling food quality might be reduced in urban habitats; mockingbirds, however, appeared to compensate for this reduction by increasing the number of feeding trips in residential neighborhoods. While individual mockingbirds did not appear to have access to increased food resources, on a population level there did appear to be enough food to sustain the increased densities
122 of mockingbirds in urban habitats and total fledgling production per unit area was higher in urban areas. The role of nest predation in urban bird communities is far from clear and researchers have been unable to reach general conclusions as to whether nest predation is reduced, increased, or equivalent in urban relative to nonurban habitats. T hese data on mockingbird nest predation rates (Chapters 2 and 3) indicate that urban habitats provide a refuge from nest predation in some years, but in other years there are no differences between habitats. Avian nest predators, however, are more abundant in urban habitats, leading to the urban nest predator paradox. Numerous factors contribute to urban nest predation rates, from changes in nest predator communities to habitat dependent changes in the diet of predators (Table 33). Using video cameras o n nests, I identified cats as the dominant urban nest predator and Coopers Hawks ( Accipiter cooperii ) as the dominant nonurban nest predator. The most abundant avian nest predators in urban habitats accounted for none of our recorded predation events. I conclude that changes in nest predator community composition, as well as dietary shifts of predators and nest defense behavior of prey, are responsible for the urban nest predator paradox. Further support for the nest defense hypothesis is provided in Chapter 2. Lifehistory traits of birds differed considerably between urban and nonurban habitats. Smallbodied individuals were more abundant in nonurban habitats, whereas largebodied individuals many of which are facultative nest predators and most of which mob avian nest predators, were more abundant in urban habitats. Most (90%) of the small bodied (<40g) birds that were detected in urban habitats nest in enclosed sites. Small -
123 bodied opencup nesters were almost entirely absent from urban areas, but made up about 50% of the detections in nonurban habitats. My results are therefore consistent with the hypothesis that nest predation at least partly determines which species are able survive in urban habitats and which species cannot invade or sustai n populations in urban habitats. Understanding the factors that contribute to the success of urban adapters is necessary if we are to promote urban biodiversity. As the worlds urban population continues to increase dramatically, ecologists can no longer ignore the contribution that humandominated landscapes make to regional patterns of biodiversity and to human wellbeing. Furthermore, urban ecology offers opportunities for scientists to address peoples disconnect with nature and with scientists. Based on our studies, we can predict that urban bird communities will change dramatically as new predator populations arise and new invasive species change the trophic structure of communities.
124 APPENDIX A COMMUNITY CENSUS DAT A Table A 1. Total number of indi viduals detected in each habitat during 100m, fixedradius point counts of each species along with their mass (Dunning 1993) and nesting guild (Ehrlich et al. 1988). The number of points per habitat is indicated in parentheses. Nest: o open cup nest, c enclosed nest. PL: parking lot, RES: residential neighborhood, UFF: urban forest fragment, PAST: pasture, NO: nonurban scrub, NF: nonurban forest. Species Mass (g) Nest PL (43) RES (16) UFF (11) PAST (15) NO (52) NF (48) Grand Total (185) Laughing Gull 325 o 5 1 1 0 0 0 7 Cattle Egret 338 o 0 0 0 1 0 0 1 Eurasian Collared Dove 149 o 67 13 6 2 0 0 88 Mourning Dove 119 o 33 26 4 6 14 2 85 White Winged Dove 153 o 0 0 0 0 1 0 1 Common Ground Dove 30.1 o 0 0 0 0 6 0 6 Rock Pigeon 354.5 o 13 1 1 0 0 0 1 5 Turkey Vulture 1467 o 0 0 0 0 1 0 1 Mississippi Kite 278 o 0 3 0 0 0 0 3 Wild Turkey 5811 o 0 0 0 1 0 1 2 Northern Bobwhite 178 o 0 0 0 4 2 1 7 Coopers Hawk 439 o 0 1 0 0 0 0 1 Yellow throated Warbler 10.02 o 0 1 0 0 0 0 1 Red tailed Hawk 1126 o 0 0 1 0 1 0 2 Red shouldered Hawk 559 o 2 2 0 0 2 4 10 Osprey 1485.5 o 1 1 2 0 0 0 4 Barred Owl 731 c 0 0 0 0 0 1 1 Burrowing Owl 155 c 0 0 0 1 0 0 1 Black hooded Parakeet 110 c 2 1 0 0 0 0 3 Monk Parakeet 101 c 13 7 0 0 0 0 20 Yellow billed Cuckoo 6 4 o 0 0 0 0 2 3 5 Hairy Woodpecker 66.25 c 0 0 0 0 2 0 2 Downy Woodpecker 27 c 14 5 0 5 18 20 62 Pileated Woodpecker 287 c 0 2 0 0 4 4 10
125 Table A 1. Continued Species Mass (g) Nest PL (43) RES (16) UFF (11) PAST (15) NO (52) NF (48) Grand Total (185) Red headed Woodpecker 71.6 c 0 0 0 0 2 0 2 Red bellied Woodpecker 61.7 c 19 7 14 10 24 27 101 Yellow shafted Flicker 132 c 4 2 0 1 2 0 9 Common Nighthawk 61.5 o 0 0 0 0 1 0 1 Chimney Swift 23.6 c 1 0 0 0 0 0 1 Ruby throated Hummingbird 3.15 o 0 1 1 0 1 0 3 Eastern Kingbird 43.6 o 0 0 0 3 0 0 3 Gray Kingbird 43.8 o 3 4 0 0 0 0 7 Great crested Flycatcher 33.5 c 10 6 2 6 10 16 50 Eastern Wood Pewee 14.1 o 0 0 0 0 0 1 1 Acadian Flycatcher 12.9 o 0 0 0 0 0 8 8 Blue Jay 86.8 o 22 6 2 10 17 8 65 Flori da Scrub Jay 80.2 o 0 0 0 0 7 0 7 American Crow 448 o 12 3 0 6 4 4 29 Fish Crow 285 o 38 16 2 0 0 0 56 European Starling 82.3 c 26 6 3 0 0 0 35 Brown headed Cowbird 43.9 2 1 0 2 8 12 25 Red -winged Blackbird 52.55 o 3 8 1 14 2 2 30 Eastern Meadowlar k 89 o 0 0 0 16 1 0 17 Orchard Oriole 19.6 o 0 0 0 3 0 0 3 Common Grackle 113.5 o 35 12 4 0 5 0 56 Boat tailed Grackle 166.5 o 23 15 0 1 0 0 39 House Finch 21.4 c 18 7 0 0 0 0 25 Bachmans Sparrow 19.65 o 0 0 0 0 4 0 4 Eastern Towhee 40.5 o 0 0 0 0 1 05 21 126 Northern Cardinal 44.64 o 24 12 11 15 70 78 210 Blue Grosbeak 34.4 o 0 1 0 5 5 0 11 Indigo Bunting 14.5 o 0 0 0 1 3 1 5 Summer Tanager 28.2 o 2 0 1 1 4 20 28
126 Table A 1. Continued Species Mass (g) Nest PL (43) RES (16) UFF (11) PAST (15) NO (52) NF (48) Grand Total (185) Purple Martin 49.4 c 0 0 0 0 1 0 1 Loggerhead Shrike 47.4 o 6 4 0 1 0 0 11 Red eyed Vireo 16.7 o 0 0 1 0 1 27 29 Yellow throated Vireo 18 o 2 0 1 1 2 6 12 White eyed Vireo 11.4 o 0 0 3 3 39 32 77 Northern Parula 8.6 o 1 2 2 2 6 36 49 Pine Warbler 11.9 o 3 1 0 0 25 11 40 Common Yellowthroat 10.1 o 0 0 0 3 39 7 49 Yellow -breasted Chat 25.3 o 0 0 0 0 1 0 1 Hooded Warbler 10.45 o 0 0 0 0 0 8 8 Painted Bunting 15.55 o 2 0 0 0 21 9 32 House Sparrow 27.7 c 74 9 5 0 0 0 88 Northern Mockingbird 48.5 o 93 33 5 14 30 1 176 Brown Thrasher 68.8 o 1 2 1 6 13 3 26 Carolina Wren 18.7 c 19 10 14 14 33 89 179 Brown headed Nutchatch 10.2 c 0 0 0 0 3 0 3 Eastern Tufted Titmouse 21.6 c 9 6 3 5 12 35 70 Carolina Chickadee 10.15 c 5 3 0 0 6 12 26 Blue gray Gnatcatcher 6 o 2 0 0 1 14 15 32 Eastern Bluebird 31.6 c 1 0 0 3 7 0 11 Grand Total 610 241 91 167 581 525 2215
127 APPENDIX B MODEL RESULTS Table B 1. Statistical significance of models testing reproductive parameters related to food availability from Chapter 4. Habitat refers to parking lot, residential, pasture, and wildlife preserve. Landuse refers to urban and nonurban. Model df F P Density 2005 Habitat 2, 4 27.64 0.005 2006 2008 Habitat 3, 10 45.32 < 0.0001 Year 2, 10 0.70 0.52 Habitat*Year 6, 10 1.88 0.18 Number of fledglings per successful nest 2005 Habitat 2, 43 0.33 0.72 Date 1, 43 1.49 0.23 2006 2008 Habitat 3, 78 1.25 0.30 Year 2, 78 3.71 0.029 Date 1, 78 4.58 0.036 Habitat*Year 6 78 1.85 0.10 Number of fledglings per hectare 2005 Habitat 2, 4 8.65 0.035 2006 2008 Habitat 3, 10 15.42 0.0004 Year 2, 10 0.53 0.60 Habitat*Year 6, 10 0.85 0.56 Clutch size 2005 Habitat 2, 98 0.73 0.49 Date 1, 98 3.02 0.09 2006 2008 Habitat 3, 281 5.81 0.0007 Year 2, 281 1.25 0.29 Date 1, 281 1.47 0.23 Habitat*year 6, 281 1.80 0.10
128 Table B 1. Continued Model df F P Proportion of eggs laid that hatched 2005 Habitat 2, 67 0.81 0.45 Date 1, 67 0.01 0 .91 2006 2008 Habitat 3, 140 0.75 0.52 Year 2, 140 1.15 0.32 Date 1, 140 0.08 0.78 Habitat*year 6, 140 0.64 0.70 Proportion of nestlings that fledged 2005 Habitat 2, 43 1.64 0.21 Date 1, 43 1.02 0.32 2006 2008 Habitat 3, 79 1.41 0. 25 Year 2, 79 0.36 0.70 Date 1, 79 0.86 0.36 Habitat*year 6, 79 0.49 0.81 Nestling Mass 2007 2008 Habitat 3, 10 0.49 0.70 Year 1, 10 0.20 0.66 Date 1, 10 2.47 0.15 Habitat*Year 3, 10 0.64 0.60 Variance in nestling mass 2008 habitat 3, 77 0.74 0.53 Number of feeding trips per nestling per hour 2008, Age 3 Land use 1, 3 4.81 0.12 Date 1, 3 1.67 0.29 2008, Age 6 Habitat 3, 13 4.02 0.032 Date 1, 13 1.02 0.33 Land use 1, 13 1.21 0.29 Date 1, 13 0.30 0.60 2008, Age 9 Category 1, 17 0.53 0.48 Date 1, 17 0.12 0.74
129 Table B 1. Continued Model df F P Average food size relative to bill 2008, Age 6 Category 1, 12 5.74 0.033 Date 1, 12 7.55 0.018 Adult Mass (Female) 2006 2008 Habitat 3, 100 0.28 0.84 Year 3, 100 0.76 0.52 Date 1, 100 3.28 0.073 Habitat*year 8, 100 1.40 0.21 Adult Mass (Male) 2006 2008 Habitat 3, 31 1.35 0.28 Year 3, 31 1.45 0.25 Date 1, 31 5.64 0.024 Habitat*year 5, 31 0.86 0.52 Varia nce in adult mass (Female) 2006 Habitat 3, 57 1.00 0.40 Variance in adult mass (Male) 2006 Land use 1, 26 3.57 0.07
130 LIST OF REFERENCES ANDERIES, J. M., M. KATTI, AND E. SHOCHAT. 2007. Living in the city: Resource availability, predati on, and bird population dynamics in urban areas. Journal of Theoretical Biology 247: 3649. BAKER, P. J., S. E. MOLONY, E. STONE, I. C. STONE, AND S. HARRIS. 2008. Cats about town: is predation by freeranging pet cats likely to affect urban bird populations? Ibis 150: 86 99. BECK, N. R., AND R. HEINSOHN. 2006. Group composition and reproductive success of cooperatively breeding whitewinged choughs ( Corcorax melanorhamphos ) in urban and nonurban habitat. Austral Ecology 31:588 596 BLAIR, R. B. 1996. Land use and avian species diversity along an urban gradient. Ecological Applications 6: 506 519. BLAND, R. L., J. TULLY, AND J. J. D. GREENWOOD. 2004. Birds breeding in British gardens: an underestimated population? Bird Study 51: 97 106. BORGMANN, K. L. AND A D. RODEWALD. 2004. Nest predation in an urbanizing landscape: the role of exotic shrubs. Ecological Applications 14: 1757 1765. BOUTIN, S 1990. Food supplementation experiments with terrestrial vertebrates: pattern, problems, and the future. Canadian Journal of Zoology 68: 371394. Brawn, J. D. and S. K. Robinson. 1996. Sourcesink population dynamics may complicate the interpretation of longterm census data. Ecology 77: 3 12. Brazel, A., N. Selover, R. Vose and G. Heisler. 2000. The tale of two climates Baltimore and Phoenix urban LTER sites. Climate Research 15: 123 135. BREITWISCH, R 1988. Sex differences in defense of eggs and nestlings by Northern Mockingbirds, Mimus polyglottos Animal Behaviour 36: 62 72. BROWN, J. L. 1964. The evolution of diver sity in avian territorial systems. Wilson Bulletin 76: 160 169. BROWN, W. P. AND P. J. SULLIVAN. 2005. Avian community composition in isolated forest fragments: A conceptual revision. Oikos 111: 1 8. BURGER, J. AND M. GOCHFELD. 1992. Vulnerability and mort ality of young Australian magpies on roads. Wilson Bulletin 104: 365 367. BURNHAM, K. P. AND D. R. ANDERSON. 2002. Model selection and multimodel inferences: a practical informationtheoretic approach, 2nd ed. Springer, New York.
131 CHALFOUN, A. D., M. J. RAT NASWAMY, AND F.R. THOMPSON. 2002. Songbird nest predators in forest pasture edge and forest interior in a fragmented landscape. Ecological Applications 12: 858 867. CHAMBERLAIN, D. E., A. R. CANNON, M. P. TOMS, D. I. LEECH, B. J. HATCHWELL, AND J. K. GASTO N. 2009. Avian productivity in urban landscapes: a review and metaanalysis. Ibis 151: 1 18. CHASE, J.F., AND J. J. WALSH. 2006. Urban effects on native avifauna: a review. Landscape and Urban Planning 74: 46 69. CROOKS, K. R 2002. Relative sensitivities of mammalian carnivores to habitat fragmentation. Conservation Biology 16: 488 502. CROOKS, K. R. AND M. E. SOULE. 1999. Mesopredator release and avifaunal extinctions in a fragmented system. Nature 400: 563 566. CZECH, B. AND P. R. KRAUSMAN. 1997. Distrib ution and causation of species endangerment in the United States. Science 277: 1116 1117. DAILY, G. C. AND P. R. EHRLICH. 1994. Influence of social status on individual foraging and community structure in a bird guild. Oecologia 100: 153 165. DERRICKSON, K .C. AND R. BREITWISCH. 1992. Northern Mockingbird ( Mimus polyglottos ). In The Birds of North America, No. 7. A. Poole, P. Stettenheim and F. Gill (Eds). The Academy of Natural Sciences, Philadelphia, PA; and the American Ornithologists' Union, Washington, D.C. DUNCAN RASTOGI, A., L. ZANETTE, AND M. CLINCHY. 2006. Food availability affects diurnal nest predation and adult antipredator behavior in song sparrows, Melospiza melodia Animal Behaviour 72: 933 940. DUNNING, J. B. 1993. CRC handbook of avian body m asses. CRC Press, Boca Raton, FL. DURNER, G. M. AND J. E. GATES. 1993. Spatial ecology of Black Rat Snakes on Remington Farms, Maryland. Journal of Wildlife Management 57:812 826. EHRLICH, P. R., D. WHEYE, AND D. DOBKIN. 1988. Birders Handbook. Simon and Schuster, New York, NY. ESTES, W. A. AND W. MANNAN. 2003. Feeding behavior of Coopers Hawks at urban an rural nests in Southeastern Arizona. The Condor 105: 107 116. FAETH, S. H., P. S. WARREN, E. SHOCHAT, AND W. A. MARUSSICH. 2005. Trophic dynamics in ur ban communities. BioScience 55: 399 407.
132 FRENCH, A. G. AND T. B. SMITH. 2005. Importance of body size in determining dominance hierarchies among diverse tropical frugivores. Biotropica 37: 96 101. FRETWELL, D., AND H. L. LUCAS, JR. 1969. On territorial behavior and other factors influencing habitat distribution in birds. I. Theoretical development. Acta Biotheoretica 19: 16 36. FULLER, R. A., K. N. IRVINE, P. DEVINEWRIGHT, P. H. WARREN, AND K. J. GASTON. 2007. Psychological benefits of greenspace increase with biodiversity. Biology Letters 3: 390394. FULLER, R. A., J. TRATALOS, AND K. J. GASTON. 2009. How many birds are there in a city of half a million people? Diversity and Distributions 15: 328337. GATES, J.E., AND L. W. GYSEL. 1978. Avian nest dispersi on and fledging success in forest field ecotones. Ecology 58: 871883. GERING, J. C. AND R. B. BLAIR. 1999. Predation on artificial bird nests along an urban gradient: predatory risk or relaxation in urban environments? Ecography 22: 532 541. GREGORY, R. D. AND J. H. MARCHANT. 1996. Population trends of Jays, Magpies, Jackdaws, and Carrion Crows in the United Kingdom. Bird Study 43: 28 37. GRIMM, N. B., S. H. FAETH, N. E. GOLUBIEWSKI, C. L. REDMAN, J. WU, X. BAI, AND J. M. BRIGGS. 2008. Global change and t he ecology of cities. Science 319: 756 760. HAAS, C. A. 1998. Effects of prior nesting success on site fidelity and breeding dispersal: An experimental approach. Auk 115:929 936. HAGER, S. B., H. TRUDELL, K. J. MCKAY, S. M. CRANDALL, AND L. MAYER. 2008. Bi rd density and mortality at windows. Wilson Journal of Ornithology 120:550 564. HARTLEY, D. J., M. J. KOIVULA, J. R. SPENCE, R. PELLETIER, AND G. E. BALL. 2007. Effects of urbanization on ground beetle assemblages (Coleoptera, Carabidae) of grassland habit ats in western Canada. Ecography 30: 673 684. HASKELL, D. G., A. M. KNUPP, AND M. C. SCHNEIDER. 2001. Nest predator abundance and urbanization. Pp. 243258 in J.M. Marzluff, R. Bowman, and R. Donnelly (Editors). Avian ecology and conservation in an urbaniz ing world. Kluwer Academic Publishers, Boston, MA. HEDRICK, L.D., AND A. D. WOODY. 1983. Northern Mockingbird kills Cedar Waxwing. Wilson Bulletin 95: 157 158. HESKE, E. J., S. K. ROBINSON, AND J. D. BRAWN. 1999. Predator activity and predation on songbird nests on forest edges in east central Illinois. Landscape Ecology 14: 345 354.
133 HOOVER, J. P. 2003. Decision rules for site fidelity in a migratory bird, the prothonotary warbler. Ecology 84:416 430. INGOLD, D. J. 1994. Influence of nest site competition between European Starlings and woodpeckers. Wilson Bulletin 106: 227 241. ISHITANI, M., D. J. KOTZE, AND J. NIEMELA. 2003. Changes in carabid beetle assemblages across an urbanrural gradient in Japan. Ecography 26: 481 489. JACKSON, W. M. AND S. ROHWER. 1989. Withinseason breeding dispersal in Prairie Warblers and other species. Condor 91: 233 241. JOKIMKI, J. AND E. HUHTA. 2000. Artificial nest predation and abundance of birds along an urban gradient. Condor 102: 838 847. KARK, S., A. IWANIUK, A. SCHALIM TZEK, AND E. BANKER. 2007. Living in the city: can anyone become and urban exploiter? Journal of Biogeography 34: 638 651. KLEM, D 1990. Collisions between birds and windows mortality and prevention. Journal of Field Ornithology 61:120 128. KOMDEUR, J. AND R. KATS. 1999. Predation risk affects tradeoff between nest guarding and foraging in Seychelles Warblers. Behavioral Ecology 10: 648 658. KO, Z. 2001. Effects of urbanization on nest site selection and nesting success of the Greenfinch Carduelis chloris in Krotoszyn, Poland. Ornis Fennica 78: 175 183. LEPCZYK, C. A., A. G. MERTIG, AND J. LIU. 2003. Landowners and cat predation across rur al to urban landscapes. Biological Conservation 115: 191 201. LESTON, L. F. V. AND A. D. RODEWALD. 2006. Are urban forests ecological traps for understory birds? An examination using Northern Cardinals. Biological Conservation 131: 566 574. LEVEY, D. J., G A. L, J. UNGVARIMARTIN, M. R. HIERSOUX, J. E. JANKOWSKI, J. R. POULSEN, C. M. STRACEY, AND S. K. ROBINSON. 2009. Urban mockingbirds quickly learn to identify individual humans. Proceedings of the National Academy of Sciences 106: 8959 8962. MAC NA LLY, R. AND C. A. R. TIMEWELL. 2005. Resource availability controls birdassemblage composition through interspecific aggression. Auk 122: 1097 1111. MARTIN, T. E 1993. Nest predation and nest sites, new perspectives on old patterns. Bioscience 43: 523 53 2.
134 MARZLUFF, J. M. AND E. NEATHERLIN. 2006. Corvid response to human settlements and campgrounds: Causes, consequences, and challenges for conservation. Biological Conservation 130: 301 314. MARZLUFF, J. M., R. BOWMAN, AND R. DONNELLY. 2001. Avian ecology and conservation in an urbanizing world. Kluwer Academic Publishers, Boston, MA. MASON, C. F. 2000. Thrushes now largely restricted to the built environment in eastern England. Diversity and Distributions 6: 189194. MCINTYRE, N.E. 2000. Ecology of urban arthropods: a review and a call to action. Annals of the Entomological Society of America 93: 825835. MILLER, J. R. 2005. Biodiversity conservation and the extinction of experience. Trends in Ecology and Evolution 20: 430 434. MILLER, J. R. AND R. J. HOBBS. 2002. Conservation where people live and work. Conservation Biology 16: 330 337. MELAMPY, M. N., E. L. KERSHNER, AND M. A. JONES. 1999. Nest predation in suburban and rural woodlots of northern Ohio. American Midland Naturalist 141: 284 292. MENNECHEZ, G., AND P. CLERGEAU. 2006. Effects of urbanisation on habitat generalists: starlings not so flexible? Acta Oecologica 30: 182 191. MOORE, F.R. 1978. Interspecific aggression: toward whom should a mockingbird be aggressive? Behavioral. Ecology and Sociobiol ogy 3: 173 176. MUMME, R. L., S. J. SCHOECH, G. E. WOOLFENDEN, AND J. W. FITZPATRICK. 2000. Life and death in the fast lane: Demographic consequences of road mortality in the Florida ScrubJay. Conservation Biology 14:501 512. NAVARRETE, S. A. AND B. A. ME NGE. 1996. Keystone predation and interaction strength: Interactive effects of predators on their main prey. Ecological Monographs 66: 409 429. NEWHOUSE, M.J., P. P. MARRA, AND L. S. JOHNSON. 2008. Reproductive success of House Wrens in suburban and rural landscapes. The Wilson Journal of Ornithology 120: 99 104. NEWTON, I. 1998. Population Limitation in Birds. Academic Press, London. PATTEN, M. A., AND D. T. BOLGER. 2003. Variation in topdown control of avian reproductive success across a fragmentation gr adient. Oikos 101: 479 488. PICMAN, J., S. PRIBIL, AND A. ISABELLE. 2002. Antipredation value of colonial nesting in Yellowheaded Blackbirds. Auk 119: 461 472.
135 PERLUT, N. G., A. M. STRONG, T. M. DONOVAN, AND N. J. BUCKLEY. 2008. Grassland songbird surv ival and recruitment in agricultural landscapes: Implications for sourcesink demography. Ecology 89:1941 1952. PIERCE, A. J., AND K. POBPRASERT. 2007. A portable system for continuous monitoring of bird nests using digital video recorders. Jounral of Fiel d Ornithology 78: 322 328. PIPER, S. D. AND C. P. CATTERALL. 2003. A particular case and a general pattern: Hyperaggressive behaviour by one species may mediate avifaunal decreases in fragmented Australian forests. Oikos 101: 602 614. PRANGE, S., S. D. GEH RT AND E. P. WIGGERS. 2003. Demographic factors contributing to high raccoon densities in urban landscapes. Journal of Wildlife Management 67: 324 333. PULLIAM, H. R 1988. Sources, sinks, and population regulation. American Naturalist 132: 652 661. RALPH, C. J., S. DROEGE, AND J. R. SAUER. 1995. Managing and monitoring birds using point counts: standards and applications. Pp. 161169 in C.J. Ralph, S. Droege, and J.R. Sauer (editors). USDA Forest Service General Technical Report PSW GTR 149. USDA Forest Service, Pacific Southwest Research Station, Albany, CA. RANDA, L. A. AND J. A. YUNGER. 2006. Carnivore occurrence along an urbanrural gradient: A landscapelevel analysis. Journal of Mammalogy 87: 1154 1164. ROBERTSON, W. B., AND G. E. WOOLFENDEN. 1992. Florida bird species: an annotated list. Florida Ornithological Society, Gainesville, FL. ROBINSON, S. K 1985. Coloniality in the Yellow rumped Cacique as a defense against nest predators. Auk 102: 506 519. ROBINSON, S. K., F. R. THOMPSON, T. M. DONOVAN, D. R. WHITEHEAD, AND J. FAABORG. 1995. Regional forest fragmentation and the nesting success of migratory birds. Science 267: 1987 1990. RODEWALD, A.D., AND D. P. SHUSTACK. 2008a. Consumer resource matching in urbanizing landscapes: are synanthropic species over matching? Ecology 89: 515 521. RODEWALD, A. D., AND D. P. SHUSTACK. 2008b. Urban flight: understanding individual and populationlevel responses of Nearctic Neotropical migratory birds to urbanization. Journal of Animal Ecology 77:83 91. ROLLINSON, D. J. AND D. N. JONES. 2002. Variation in breeding parameters of the Australian magpie Gymnorhina tibicen in suburban and rural environments. Urban Ecosystems 6: 257 269.
136 ROSENZWEIG, M. L. 2003. Win Win Ecology. Oxford University Press, New York. ROTH, T. C., AND S. L. LIMA. 2003. Hunting behavior and diet of Coopers Hawks: an urban view of the small bird in winter paradigm. Condor 105: 474 483. SCHMIDT, K. A., J. R. GOHEEN, R. NAUMANN, R. S. OSTFELD, E. M. SCHAUBER, AND A. BERKOWITZ. 2001. Experimental removal of strong and weak predators: Mice and chipmunks preying on songbird nests. Ecology 82: 2927 2936. SCHMIDT, K. A., R. S. OSTFELD, AND K. N. SMYTH. 2006. Spatial heterogeneity in predator activity, nest survivorship, and nest site selection in two for est thrushes. Oecologia 148: 22 29. SCHMIDT, K. A. AND C. J. WHELAN. 2001. Effects of exotic Lonicera and Rhamnus on songbird nest predation. Conservation Biology 13: 1502 1506. SCHOECH, S. J., R. BOWMAN AND S. J. REYNOLDS. 2004. Food supplementation and p ossible mechanisms underlying early breeding in the Florida ScrubJay ( Aphelocoma coerulescens ). Hormones and Behavior 46: 565 573. SHAWKEY, M. D., R. BOWMAN, AND G. E. WOOLFENDEN. 2004. Why is brood reduction in Florida scrubjays higher in suburban than in wildland habitats? Canadian Jounal of Zoology 82: 1427 1435. SHAFFER, T. L 2004. A unified approach to analyzing nest success. Auk 121: 26 540. SHELLEY, E. L., M. Y. U. TANAKA, A. R. RATNATHICAM, AND D. T. BLUMSTEIN. 2004. Can Lanchesters laws help ex plain interspecific dominance in birds? Condor 106: 395 400. SHOCHAT, E 2004. Credit or debit? Resource input changes population dynamics of city slicker birds. Oikos 106: 622 626. SHOCHAT, E., S. B. LERMAN, M. KATTI, AND D. B. LEWIS. 2004a. Linking optim al foraging behavior to bird community structure in an urbandesert landscape: field experiments with artificial food patches. American Naturalist 164: 232 243. SHOCHAT, E., W. L. STEFANOV, M. E. A. WHITEHOUSE, AND S. H. FAETH. 2004b. Spider diversity in t he greater Phoenix area: the influence of human modification of habitat structure and productivity. Ecological Applications 14: 268 280. SHOCHAT, E., P. S. WARREN, S. H. FAETH, N. E. MCINTYRE, AND D. HOPE. 2006. From patterns to emerging processes in mechanistic urban ecology. Trends in Ecology and Evolution 21: 186 191. SMITH, T. M., AND H. H. SHUGART. 1987. Territory size variation in the Ovenbird: the role of habitat structure. Ecology 68: 695 704.
137 SOL, D., D. M. SANTOS, J. GARCIA, M. CUADRADO. 1998. Competition for food in urban pigeons: The cost of being juvenile. Condor 100: 298 304. SORACE, A 2002. High density of bird and pest species in urban habitats and the role of predator abundance. Ornis Fennica 79: 60 71. SOULE, M. E., D. T. BOLGER, A. C. AL BERTS, J. WRIGHT, M. SORICE, AND S. HILL. 1988. Reconstructed dynamics of rapid extinctions of chaparral requiring birds in urban habitat islands. Conservation Biology 2: 75 92. STEVENSON, H. M., AND B. H. ANDERSON. 1994. The birdlife of Florida. Universit y Press of Florida, Gainesville, FL. SWITZER, P. V. 1993. Site fidelity in predictable and unpredictable habitats. Evolutionary Ecology 7:533 555. TAYLOR, A. F. AND F. E. KUO. 2009. Children with attention deficits concentrate better after walk in the park Journal of Attention Disorders 12: 402 409. THORINGTON, K. K. AND R. BOWMAN. 2003. Predation rate on artificial nests increases with human housing density in suburban habitats. Ecography 26: 188 196. ULRICH, R. S. 1984. View through a window may influenc e recovery from surgery. Science 224: 420 422. UNITED NATIONS POPULATION DIVISION. 2 006. World Urbanization Prospects: the 2005 Revision. United Nations, New York. VALCARCEL, A. AND E. FERNANDEZJURICIC. 2009. Antipredator strategies of house finches: are urban habitats safe spots from predators even when humans are around? Behavioral Ecology and Sociobiology 63:673685. VAN HEEZIK, Y., A. SMYTH, A. ADAMS, AND J. GORDON. 2010. Do domestic cats impose an unsustainable harvest on urban bird populations? Biolo gical COnservation 143: 121 130. VAN HORNE, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47: 893 901. VICKERY, P. D., M. L. HUNTER, AND J. V. WELLS. 1992. Evidence of incidental nest predation and its effect s on nests of threatened grassland birds. Oikos 63: 281 288. WEATHERHEAD, P. J. AND G. BLOUINDEMERS. 2004. Understanding avian nest predation: why ornithologists should study snakes. Journal of Avian Biology 35: 185 190. WEATHERHEAD, P. J. AND M. B. CHARL AND. 1985. Habitat selection in an Ontario population of the snake, Elaphe obsoleta. Journal of Herpetology 19: 12 19.
138 WEINSTEIN, N., A. K. PRZYBYLSKI, AND R. M. RYAN. 2009. Can nature make us more caring? Effects of immersion in nature on intrinsic aspirations and generosity. Personality and Social Psychology Bulletin 35: 1315 1329. WELDON, A. J. AND N. M. HADDAD. 2005. The effects of patch shape on Indigo Buntings: Evidence for an ecological trap. Ecology 86: 1422 1431. WHITCOMB, R. F., C. S. ROBBINS, J. F. LYNCH, B. L. WHITCOMB, M. K. KLIMKIEWICZ, AND D. BYSTRAK. 1981. Effects of forest fragmentation on avifauna of the eastern deciduous forest. Pp. 125205 in R.L. Burgess and D.M. Sharpe (editors). Forest Island Dynamics in ManDominated Landscapes. Sprin ger Verlag, New York, NY. WHITE, G. C. AND K. P. BURNHAM. 1999. Program MARK: survival estimation from populations of marked animals. Bird Study 46:120 139. WHITTAKER, K. A., AND J. M. MARZLUFF. 2009. Species specific survival and relative habitat use in an urban landscape during the postfledging period. Auk 126:288 299. WIKLUND, C. G. AND M. ANDERSON. 1994. Natural selection of colony size in a passerine bird. Journal of Animal Ecology 63: 765 774. WOODS, M., R. A. MCDONALD AND S. HARRIS. 2003. Predation of wildlife by domestic cats ( Felis catus ) in Great Britain. Mammal Review 33: 174 188.
139 BIOGRAPHICAL SKETCH Christine M. Stracey was born in New Jersey. She grew up in Springfield, NJ with her older brother, Tom, and graduated from Jonathan Dayton Regional High School in 1996. She attended The College of New Jersey and graduated with honors obtaining a Bachelor of Science in biology, with a minor in chemistry. While at the College of New Jersey, Christine was in the Honors Program and a member and captain of the Womens Swimming and Diving Team. Upon completing her undergraduate degree, she attended Columbia University, where she received her Master of Arts under the guidance of Dr. Stuart Pimm in conservation biology. Her thesis was entitled Testing the Equilibrium Theory of Island Biogeography: immigration rates of birds on British islands. Christine began her Ph.D. work at the University of Illinois under the guidance of Dr. Scott Robinson. After her first year there, Dr. Robinson moved to the University of Florida and Christine followed. Upon completion of her Ph.D., Christine will become an Assistant Professor at Westminster College in Salt Lake City. She has been married to Tim Richard for eight years and has two Beagles, Mango and Addie.