Reproductive ecology and habitat associations of Florida's crested caracaras

MISSING IMAGE

Material Information

Title:
Reproductive ecology and habitat associations of Florida's crested caracaras
Physical Description:
Book
Language:
English
Creator:
Morrison, Joan L
Publication Date:

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact the RDS coordinator (ufdissertations@uflib.ufl.edu) with any additional information they can provide.
Resource Identifier:
aleph - 28636940
oclc - 38855492
System ID:
AA00026804:00001


This item is only available as the following downloads:


Full Text











REPRODUCTIVE ECOLOGY AND HABITAT ASSOCIATIONS
OF FLORIDA'S CRESTED CARACARAS













By

JOAN L. MORRISON


















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

1997




























Copyright 1997

by

Joan L. Morrison
























To Clyde and all the rest.















ACKNOWLEDGMENTS

Many people contributed to this research and I am grateful to all of them. I
particularly wish to thank Paul Ebersbach and Bob Progulske at the Avon Park Air Force
Range (APAFR, DOI), Brian Millsap, Don Wood, and Dave Cook of the Florida Game
and Fresh Water Fish Commission's (FGFWFC) Non-Game Wildlife Program, and Dave
Wesley (U.S. Fish and Wildlife Service) for recognizing the need for research on this
little-known species and subsequently providing funding. This project would have never
gotten off the ground without their support and assistance. I thank the staff at APAFR for
their help, particularly Pat Walsh, Sam van Hook, Scott Penfield, Steve Hanley, Julie
Olson, and Harlan Blackburn. Nancy Douglass, Victor Echavez, and Hugh Clark of the
FGFWFC also provided field and logistic support, particularly assistance with accessing
state-owned properties. I am grateful for additional financial support received from the
South Florida Water Management District, the University of Florida's Graduate Minority
Fellowship Office, and the Ford Foundation Fellowship Program.
I gratefully acknowledge the cooperation and support provided by the personnel
and staff at the MacArthur Agro-Ecology Research Center (MAERC) and Archbold
Biological Station, particularly Drs. John Fitzpatrick, Hilary Swain, and Tom Bancroft,
and Fred Lohrer. I particularly appreciated the opportunity to live at and to operate this
project based out of MAERC. Ranch managers Dan Childs and Gene Lollis provided
many hours of interesting conversations regarding cattle ranching in Florida. I thank
them for supporting this project throughout and for sticking up for us with the Florida
Cattleman's Association. I also thank the rest of the staff at MAERC for their help and
for putting up with our activities at the ranch. I also wish to acknowledge the private
donation by Frank "Sonny" Williamson to the IFAS research effort at MAERC.



iv











Special thanks go to Steve McGehee who was a stalwart field assistant
throughout. His enthusiasm for caracaras and untiring field assistance, particularly in
finding nests and radiotracking, was invaluable. Vicky Dreitz and Danielle Todd also
proved to be excellent field assistants, during the second and third field seasons
respectively, and capably handled all aspects of the research program as well. I am
grateful to Brian Mealey for his support throughout and also for loan of Clyde, our
research caracara. Additional people who helped with field work at various times
included John Amett, Tylan Dean, Sam Gibbs, Laurie Gray, Kevin Long, Mike
McMillian, Carlos Pages, Megan Parker, Ray Poulin, Lourdas Rojas, Jim Sanderson, Jim
Solomon, and Reuven Yosef. I thank Scott Hedges, Stefanie Melvin, Tom Weber, Cathy
Ritchie, and Shannon Stull for their help at various stages of the project. Thanks also to
Lorenz Animal Foods and Brad Stith for help with the dead cows.
I am grateful to the staff at the Department of Wildlife Ecology and Conservation,
Institute of Food and Agricultural Sciences, University of Florida, whose efforts
throughout my tenure at UF kept our paperwork processed, salaries on track, and bills
paid. Sincere thanks go to Peg Margosian, Brian Sheahan, Roberta Pickett, Don
Browning, and Helen Huseman who contributed many hours of assistance with the GIS
analysis and mapping, and to Ken Portier and Jay Harrison for help with statistics. Linda
Kubitz, Cathy Henderson, Milt Putnam, and Sylvia Beauchamp from IFAS Media
Services provided excellent media coverage of the project. Thanks go to Neil Rettig and
Kim Hayes for providing additional film of our activities. I thank Mary Maltbie for
sexing the caracaras using the blood samples, Drs. Don Forrester and Marilyn Spalding,
and Garry Foster for necropsies and processing parasite data, and Dr. Reed Bowman for
providing the Mayfield nest success program. I also thank personnel at the Maitland Bird
of Prey Center, the Falcon Batchelor Bird of Prey Center at the Miami Museum of
Science, and Wrede's Wildlife Rehabilitation Center for assisting us when we had
caracara emergencies. Sincere thanks go to Mark Brady and personnel at Glades Electric
Cooperative for helping us to access the nest trees, to Jim Wyatt, Karen Dunne, Reed

v











Robbins, John McGee, and Morton Sund of Wyatt Aviation who kept us flying safely
during all the long hours of aerial radiotracking, and to the personnel at Wrenchmasters,
who kept our vehicles running.
I especially thank my graduate advisor, Dr. Steve Humphrey, who stepped in at
the beginning when everything looked tenuous. His steady and constant support
throughout helped keep up the momentum. I thank my other committee members, Drs.
Lyn Branch, John Fitzpatrick, John Smallwood, Mel Sunquist, and S. David Webb, for
their contributions, especially for reviews of this dissertation. I also thank Dr. Mike
Collopy for sparking my original interest in caracaras and for getting me to UF in the first
place.
I particularly want to acknowledge the friendship of Kim Babbitt, Patty Cramer,
and Pat Kennedy, who were there from the beginning and got me through the rough
times. Thanks also go to other friends and colleagues Eduardo Alvarez, Leslie Backus,
Tylan Dean, Nathalie Hamel, Dave Leonard, Ken Meyer, Steve Morello, Katie Sieving,
and Dale Stahlecker for their help and input at various times.
I thank my family for their continued support, particularly my parents, whose
belief and encouragement never waned. I especially thank Jim Sanderson for his love and
support throughout and for his untiring belief that this was all possible.
Most importantly, I thank landowners and cattle ranchers throughout south-central
Florida, who so generously provided access to their lands and with whom we had so
many interesting discussions regarding caracaras and cattle ranches. Without their
cooperation and support, this study could not have been accomplished at all.
This dissertation is contribution No. 33 from the MacArthur Agro-Ecology
Research Center of Archbold Biological Station.








vi















TABLE OF CONTENTS

ACKNOWLEDGMENTS ............................................ iv
LIST OF TABLES ............................. .. .... .............. ix
LIST OF FIGURES ................................................. xii
ABSTRACT ................... .............................. xiv
CHAPTERS
1 STUDY OVERVIEW ........................................ 1
Introduction .................................... ............ 1
Study Species -- History and Background ................ .......... 2
Study A rea ............ ............................... 5
Study Organization ................ ................... .... 5
2 BREEDING ECOLOGY OF FLORIDA'S CRESTED CARACARAS. 9
Introduction ................................................ 9
Study Area and Methods ...................................... 10
Nest Location and Visitation ................................ 10
Reproductive Success ..................................... 11
Productivity and Components of Variation .................... 12
Environmental Data ......... ........................... .. 13
Statistical Analyses ........................ ............ 13
Results ......... ........ .......................... 14
Territorial Occupancy and Breeding Rates..................... 14
Breeding Behavior and Site Fidelity .......................... 14
Nest Building and Incubation ................... ........ 15
Length of the Breeding Season and Timing of Broods ............ 15
Number of Broods Per Year ................................ 16
Clutch Size ............................................. 16
Broods ................................................. 16
Reproductive Success, Productivity and Components of Variation. 17
Performance and Laying Date ................. ............ 17
Nest Failure, Egg, and Chick Mortality ....................... 18
D iscussion. ........................................... 18
Weather, Timing, and Length of the Breeding Season ........... 19
Clutch Size ................... ..................... 20

vii











Timing and Number of Broods ........ ................. 20
Mortality .................................. .......... 21
Breeding Strategy and Variation in Reproductive Parameters ..... 21
Comparison With Other crested caracaras ................... 22
3 EFFECTS ON PRODUCTIVITY AND POTENTIAL COSTS OF
DOUBLE BROODING IN CRESTED CARACARAS .......... 41
Introduction ................. ............................ 41
Study Area and Methods ................. .................. 43
Results ................ ................................ 46
Extent of Occurrence of Double Brooding .................... 46
Timing of Broods .................................... 47
Productivity and Nesting Success ........................... 47
Parental Care and Post-fledging Survival .................... 48
Potential Costs to Adults ............................... 48
Effects on Population Growth and Persistence ................. 48
Influence of Factors .................................... 48
Discussion .................................... .......... 48
4 HABITAT ASSOCIATIONS AND PRODUCTIVITY OF
CARACARAS ON CATTLE RANCHES .................... 64
Introduction ................................. ... ... ....... 64
Study Area ............................................. 65
Methods ............................................... 65
Land Ownership/Land Use .............................. 65
Home Range Estimation and Spatial Dispersion of Territories ..... 66
Landscape-level Habitat Selection by Breeding Pairs ............ 68
Caracaras and Cattle Ranches ........... ........ ..... 70
Results ................................................. 70
Land Ownership/Land Use ................................ 70
Home Ranges and Spatial Dispersion of Territories ............. 70
Landscape-level Habitat Selection by Breeding Pairs............ 71
Caracaras and Cattle Ranches .......................... 72
Discussion ................ ............................. 72
5 SYNTHESIS AND CONCLUSIONS ......................... 92
LITERATURE CITED ................... ........................... 96
BIOGRAPHICAL SKETCH ................. ....................... 108





viii













LIST OF TABLES
Table page
2-1 Breeding chronology and intervals, in weeks, between broods for Florida's
crested caracara population, 1994 1996 ............................ 24
2-2 Multiple nesting in Florida's crested caracara population, 1994 1996.
Early nests are those that initiated egglaying before the median week of
egglaying for the entire sample, each year ........................... 25
2-3 Reproductive success parameters for Florida's crested caracaras,
1994 1996. Nestling survival to fledging (number of fledglings/number
of eggs laid), fledgling survival to independence (number of young alive
at end of post-fledging dependency period/number fledged), and egg
success (proportion of eggs laid that produced independent young) for eggs
from known size clutches. .................. .................... 26
2-4 Estimates of Mayfield nest success probabilities for breeding pairs of
crested caracaras in south-central Florida, 1994 1996 ................ 27
2-5 Measures of annual productivity (mean SE) and Index of Productivity
for crested caracaras in south-central Florida, 1994 1996. See text for
calculation of Index of Productivity. Productivity per successful pair
includes young from renesting attempts ............................ 28
2-6 Variability in reproductive parameters for crested caracaras during 3 years
in south-central Florida. CV = coefficient of variation (standard
deviation/mean). Productivity/successful pair includes young from
renesting efforts. n = 98 clutches .................................. 29
2-7 Seasonal decline in breeding performance (mean SE) for known-size
clutches and for all nest attempts of crested caracaras in south-central
Florida, 1994 1996. Early nests initiated egglaying before the median
week of egglaying each year, late nests initiated egglaying after the
median week. ................ ............................. 30
2-8 Fates of eggs from known-size clutches for crested caracara nests in south-
central Florida, 1994 1996. .................................... 31
2-9 Failure of crested caracara nests in south-central Florida, 1994 1996 ..... 32
3-1 Occurrence and timing of nesting attempts between years for crested
caracaras in south-central Florida. NA = not applicable for that year...... 53
3-2 Clutch size (mean SE) and nesting success (SD in parentheses) between
nest types for crested caracaras in south-central Florida, 1994 1996. The
same letter indicates significant differences between nest types. ND = no
ix











data on clutch sizes of first or second nests in 1994 because these nests
were found after hatching ............................ ........ . 54
3-3 Estimated parameters for evaluating contribution of double brooding to
population growth for crested caracaras in south-central Florida.
Parameters estimated from data collected during 1994 1996............ 55
3-4 Estimation of population size and annual growth rate for two hypothetical
populations of crested caracaras in south-central Florida, one in which all
pairs only produce a single brood each year, and one in which
approximately 10% of pairs successfully produce two broods each year.... 56
3-5 Statistics for logistic regression analysis of the influence of timing of
initiation of the first brood (INITWK) and the number fledged in the first
brood (NUMFRST) on the occurrence of double brooding in crested
caracaras in south-central Florida, 1994 1996. n = 16 double-brooded
pairs, 103 single-brooded pairs ........... ....................... 57
3-6 Statistics for logistic regression analysis of the influence of double
brooding the previous year (DBPREV), fledging date (WKPREV) and
number fledged the previous year (NUMPREV) on the occurrence of
double brooding in crested caracaras in south-central Florida, 1994 1996.
n = 11 double-brooded pairs, 34 single-brooded pairs ................. 58
4-1 Variables included in development of the logistic regression model
describing caracara home ranges. Unless noted, all variables are expressed
as proportions of the total area of the home range that was vegetated with
natural or semi-natural* vegetation. ................................ 76
4-2 Landownership and major land use for crested caracara territories in south-
central Florida, 1994 1996. n = 65 territories ....................... 77
4-3 Size and vegetation composition of crested caracara home ranges. Home
ranges are defined as the 99% kernel contour estimated using a 0.85
smoothing factor. N = number of radiolocations used to estimate home
range size ................. ................................. 78
4-4a Matrix of percent overlap of 60% contours (core areas) for 9 adjacent pairs
of breeding crested caracaras on the MacArthur Agro-Ecology Research
Center, Highlands County, south-central Florida. Birds in rows are
overlapped by birds in columns ................................... 79
4-4b Matrix of percent overlap of 60% contours (core areas) for 4 adjacent pairs
of breeding crested caracaras along the Kissimmee River, Highlands and
Okeechobee counties, south-central Florida. Birds in rows are overlapped
by birds in columns ........................................... 79











4-5a Matrix of percent overlap of 99% contours for 9 adjacent pairs of breeding
crested caracaras on the MacArthur Agro-Ecology Research Center,
Highlands County, south-central Florida. Birds in rows are overlapped by
birds in columns ................. ............................ 80
4-5b Matrix of percent overlap of 99% contours for 4 adjacent pairs of breeding
crested caracaras along the Kissimmee River, Highlands and Okeechobee
counties, south-central Florida. Birds in rows are overlapped by birds in
columns...................................................... 80
4-6 Parameter estimates for model produced using logistic regression analysis
that best distinguished between known caracara home ranges and available
units throughout the study area. Variables include pine forest (PINEFOR),
mixed forest (MIXFOR), wax myrtle (WAXMYRT), woodland (WOODL),
and contiguous pasture (CONTIGP) ............................. ... 81
4-7 Parameter estimates for model that best distinguished between known
caracara home ranges and available units throughout the study area and
that did not include forest variables. Variables include native
grassland/prairie (GRASSL), wax myrtle (WAXMYRT), improved
pasture (IMPPAST), and contiguous pasture (CONTIGP)................. 82
4-8 Predicted probabilities of use for test caracara home ranges calculated
using logistic regression models that best distinguished between caracara
home ranges and available units throughout the study area ............. 83






















xi













LIST OF FIGURES

Figure page
1-1 Distribution of the crested caracara in North and Central America ......... 7
1-2 Current known breeding range of the crested caracara in Florida. The study
area encompassed much of this range ................ ................. 8
2-1 Location of study area for crested caracaras in south-central Florida. Study
area encompassed much of the species' current known breeding range in
Florida.................................................................. 33
2-2 Mean monthly rainfall and breeding chronology of crested caracaras during
3 years in south-central Florida, 1994 (n = 33 nest attempts), 1995 (n = 62
attempts), 1996 (n = 62 attempts) ...................................
34
2-3 Relationship between month of first nest initiation and occurrence of
renesting by crested caracaras following either an early failure or a
successful first brood, south-central Florida, 1994 1996 ................ 35
2-4 Distribution of clutch sizes by year, for crested caracara nests in south-
central Florida, 1994 1996..................................... 36
2-5 Mean number of fledglings and independent young per nest for each
clutch size during each year for crested caracaras in south-central Florida,
1994 1996. n indicates number of clutches......................... 37
2-6 Relationship between mean clutch size and month of nest initiation for 98
known-size clutches of crested caracaras in south-central Florida,
1994 1996. Number of clutches is shown above each month ............ 38
2-7 Mayfield probability of nest success ( SD) by month for crested caracaras
in south-central Florida, 1994 1996 (n = 157 nest attempts) ............. 39
2-8 Relationship between productivity per nest attempt (n = 157 attempts) and
month of nest initiation for crested caracaras in Florida, 1994 1996.
Number of nest attempts is shown above each month ................... 40
3-1 Timing of nest initiation throughout the breeding season for crested caracaras
in south-central Florida, 1994 1996 ............................... 59
3-2 Length of the interbrood interval in relation to number fledged from the
first nesting attempt for double brooded crested caracaras in south-central
Florida, 1994 1996. n = 16 pairs ................................ 60


xii











3-3 Proportion of nests initiated in each month that were successful, for crested
caracaras in south central Florida, 1994 1996......................... 61
3-4 Mean number of young fledged per pair for pairs of crested caracaras
nesting in south-central Florida, 1994 1996. Number of pairs is shown
above each column ............................................... 62
3-5 Kaplan-Meier survival functions estimated for juvenile crested caracaras
from first, second, and single broods in south-central Florida, 1994 1996... 63
4-1 Study area for evaluating landscape-level habitat selection by breeding pairs
of crested caracaras in Florida, 1994-1996 ........................... 84
4-2 Relationship between occupancy, initiation of breeding attempts, and land
ownership for 37 crested caracaras territories in south-central Florida, 1994 -
1996 ......................................... ............... 85
4-3 Mean ( SE) number of radiolocations for 25 breeding adult crested
caracaras in relation to distance from the nest. n = 1283 total radiolocations. 86
4-4 Distributions of radiolocations of 25 breeding adult crested caracaras in
relation to distance from the nest during breeding and non-breeding seasons
in south-central Florida, 1994 1996 ........................... ... 87
4-5 Distribution of distances between nests of neighboring pairs of successfully
breeding crested caracaras in south-central Florida, 1994 1996. n = 30
pairings. ...................................................... 88
4-6 Relationship between proportion of improved pasture in home range and
home range size for 25 crested caracara home ranges in south-central Florida,
1994- 1996 .................................................. 89
4-7 Relationship between the amount of improved pasture in the home range
and total number of young fledged during 1994 1996, for crested caracaras
in south-central Florida ................. ........................ 90
4-8 Home ranges of the crested caracara on study area in south-central Florida... 91












xiii















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

REPRODUCTIVE ECOLOGY AND HABITAT ASSOCIATIONS
OF FLORIDA'S CRESTED CARACARAS

By

Joan L. Morrison

December, 1997

Chairman: Dr. Stephen R. Humphrey
Major Department: Wildlife Ecology and Conservation

I studied the reproductive ecology and habitat associations of the crested caracara
(Caracaraplancus) in Florida. Listed as Threatened, this population is believed to be in
decline due to habitat loss. Results from observations conducted at 155 nests during 3
years indicated that Florida's caracaras may be doing better than previously believed,
particularly on privately owned cattle ranches, where most nesting pairs were found.
Pairs nesting on ranches exhibited higher site fidelity, annual breeding rates, and nesting
success, and produced more young overall.
The breeding cycle is strongly tied to the annual precipitation cycle. Although the
modal clutch size was 2, clutch sizes of 3 produced the most independent young during
all years. Pairs can make several nesting attempts per year. Individuals that successfully
raise two broods in a year may realize higher lifetime reproductive success. Fledglings
from early-season nests had the highest probability of survival to recruitment. Home
range size and nesting density estimated using radiotelemetry were considerably smaller
and higher, respectively, than previously published estimates for this population.



xiv











Caracaras are highly territorial near the nest but apparently tolerate conspecifics in
foraging areas.
I investigated landscape-level habitat selection by breeding pairs of crested
caracaras using GIS analysis and logistic regression modeling. Results indicated a
positive association between caracara home ranges and improved pastures, particularly
with large areas of contiguous pasture. The size of individual caracara home ranges was
inversely correlated with the proportion of improved pasture in the home range.
Caracaras frequently occupy areas in close association with humans and are
generally tolerant of moderate levels of human activities. Because of the caracara's
strong site fidelity, a conservation strategy for Florida's population should be based on
protecting known nest sites and should focus at a regional scale. Management plans must
address effects of loss and degradation of nesting and foraging habitat, particularly
conversion of pastures to citrus and sugar cane. Because most suitable habitat is in
private ownership, habitat protection must involve cooperative efforts between land
management agencies and private landowners. This research has focused attention on the
importance of cattle ranches to Florida's native wildlife by using the caracara as an
example of how humans, agriculture, and native wildlife can coexist.



















xv













CHAPTER 1

STUDY OVERVIEW

Introduction
One of the most profound and widespread impacts of humans on the landscape is
alteration of natural habitats. Fragmentation and loss of natural habitats have been
implicated in the decline or extinction of many species (Askins et al 1990, Wilcox and
Murphy 1985). Consequences of these habitat changes may include reduced reproductive
success or survivorship owing to loss of suitable breeding and foraging habitats, altered
dispersal and other behavioral patterns, and increased predation pressure within
remaining habitat patches (Whitcomb et al. 1981, Wilcove et al. 1984, Hagan and
Johnston 1992, LaHaye et al. 1992, McKelvey et al. 1993, Herkert 1994, Lens and
Dhondt 1994, Dunning et al. 1995, Redpath 1995, Robinson et al. 1995, Matthysen and
Currie 1996). Formerly contiguous populations are fragmented into disjunct subgroups if
individuals are unable to disperse among isolated patches that are embedded in an
inhospitable matrix (Opdam et al. 1985, Sj6gren 1991, Fahrig and Merriam 1994).

The effects of fragmentation or loss of natural habitats on small, isolated, or
peripheral populations are of particular concern. These populations have high
conservation value because they often show some degree of genetic, behavioral, and/or
morphological divergence from central populations (Levin 1970, Lawton 1993, Lesica
and Allendorf 1995). Geographically isolated populations are subject to higher risk of
extinction owing to demographic, environmental, and genetic stochasticity (Shaffer
1981). It is reasonable to expect that effects of habitat loss and fragmentation acting in
concert with isolation and small size pose serious threats to peripheral populations.

Fragmentation and loss of natural habitats are occurring at a rapid rate throughout
peninsular Florida. Because of its geographic position and glacial history, Florida has a
rich biodiversity (Neill 1957, Myers and Ewel 1990). The peninsula is a mosaic of

1









2
habitat islands and small, often disjunct populations of species, many of which have
differentiated at least at the subspecies level from their closest relatives in the western
U.S. Continued fragmentation and conversion of natural habitats combined with
geographic isolation currently jeopardizes chances for the long-term persistence of many
species in Florida.

Study Species -- History and Background
The crested caracara (Caracara plancus) is a medium-sized tropical raptor that
occupies open habitats, typically grassland, prairie, rangeland, savanna, and desert. Its
current geographic distribution covers most of South and Central America, through parts
of northern Mexico, and into the southern U.S. (del Hoyo et al. 1994). Its current North
American distribution is limited to disjunct populations in Texas, Arizona, Cuba, and
Florida (Figure 1-1).

Fossil records and the distribution of these populations suggest that the caracara
may have been more widespread throughout southern North America during the
Pleistocene (Howard and Miller 1933, Brodkorb 1964, Chandler and Martin 1991,
Vuilleumier 1985). This epoch was characterized by extensive periods of reduced
rainfall. During these times, grassland and savanna habitats presumably reached their
maximum extent across the continent. Especially during the glacial cycles when sea
levels were lowest, an extensive subtropical savanna corridor may have extended around
the Gulf of Mexico, facilitating the exchange and movement of western biota eastward
into Florida (Neill 1957, Webb 1977, Webb and Wilkins 1984). Ancestors of the extant
caracara may have expanded northward through Central America during the biotic
interchange, followed by continued expansion throughout the Gulf region and into
peninsular Florida.

As a consequence of numerous glacial cycles, repeated fluctuations in sea level,
and associated climate and habitat changes, the geographic ranges of many North and
Central American species eventually became divided and reduced in size. Many species
having western affinities, including the caracara, were ultimately restricted to small









3
populations in Florida (Woolfenden and Fitzpatrick 1984). Today, conservation efforts
throughout Florida are increasingly focusing on these populations because of their
isolation, small size, and disjunct distributions.

The extant Florida population of the crested caracara (C.p. audubonii) occurs in
the south-central peninsula. Described by early ornithologists as inhabiting much of the
central peninsula (Scott 1892, Phelps 1912, Howell 1932, Bent 1937, Nicholson 1929,
1930), its historic distribution probably conformed closely to the original location of the
major prairie ecosystems (Davis 1967). Representatives of this ecosystem included the
Kissimmee prairie, which extended north and westward from Lake Okeechobee in
association with the Kissimmee River floodplain, and the DeSoto prairie, which extended
westward from the southernmost portion of the central ridge. These prairies constituted
vast expanses of low, open, nearly treeless grasslands, with sparse cabbage palm
hammocks and scattered patches of shrubs and saw palmettos (Harshberger 1914).

The current breeding range of the caracara in Florida encompasses parts of about
10 counties within this presumed historic range (Kale et al. 1992, Stevenson and
Anderson 1994) (Figure 1-2). The grassland and prairie ecosystems in this region are
now only limited in extent (Abrahamson and Hartnett 1990). Extensive areas of the
original prairies have been converted to improved pastures, which are dominated by non-
native grasses and are intensely managed to produce forage for cattle. What remains as
natural and semi-natural habitat in this region is rapidly being converted for agricultural
and urban development.

The status of Florida's crested caracara population has long been a subject of
concern. Apparent steep declines in numbers during the late 1960s and early 1970s
(Funderberg and Heinzman 1967, Heinzmann 1970) first focused attention on this
population. Results of surveys of historic territories in the late 1980s suggested that this
population may be undergoing a long-term decline concomitant with accelerated habitat
loss (Millsap 1989). Layne (1995), however, stated that during the period 1972 to 1991,









4
the adult population appears to have been stable. Current published estimates for this
population are approximately 400-500 individuals statewide (Layne 1995, 1996).

Because of its restricted range, apparent small population size and history of
decline, Florida's crested caracara population was listed as federally Threatened in 1987
(U.S. Fish and Wildlife Service 1987a), followed by a similar state listing (Wood 1990).
The current paucity of demographic and ecological information precludes accurate
assessment of this population's status or trends, however. Habitat fragmentation and loss
are assumed to threaten population persistence, but little data exist with which to test this
hypothesis. In 1989, a federal recovery plan was accepted for Florida's caracara
population (U.S. Fish and Wildlife Service 1989), but there was little biological
information from which to develop appropriate recovery goals.

Management of wildlife populations and development of suitable conservation
strategies require in-depth knowledge of a species' life history and population dynamics.
The caracara apparently now occurs primarily on privately-owned cattle ranches in this
region, undoubtedly contributing to the paucity of biological information. Confrontations
between land management agencies and private landowners have escalated in recent
years, particularly over issues relating to threatened and endangered species (Lewis 1995,
Bean and Wilcove 1997). Landowners are hesitant to permit access to their lands,
particularly when a threatened or endangered species is involved. This poses a major
dilemma for conservation of a species that occurs regularly on private lands.

The purpose of this research was to provide basic ecological information on
Florida's caracara population, particularly information that is relevant for management.
Although caracaras apparently persist on cattle ranches, their exact distribution
throughout the region, their habitat requirements, and their productivity on cattle ranches
are poorly known. Goals of this research included improving knowledge of the
caracara's breeding biology, determining land uses compatible with its survival, and
assessing relationships between caracaras and cattle ranches. This knowledge is critical
for developing guidelines for long-term population monitoring and management.









5
Study Area
The study area in south-central Florida encompassed much of the crested
caracara's current breeding range. Included were all or parts of Highlands, Glades,
Okeechobee, Osceola, DeSoto, Polk, Hendry, and Indian River counties (Figure 1-2), an
area totaling approximately 20,891 km2. The landscape in this region is a mosaic of
habitats, primarily grasslands and pastures interspersed with cropland, citrus groves,
dairies, pine flatwoods, freshwater marshes and lakes, oak scrub, hardwood hammocks,
and urban developments. Over half of this area is in private ownership. Primary land
uses are cattle ranching, citrus production, and other agricultural crops. The land in
public ownership is managed for a variety of uses including recreation, military training,
hunting, natural areas, and river floodplain management.

Study Organization
My research examined the reproductive ecology and habitat associations of
Florida's crested caracaras. I focused on (1) documenting reproductive strategies and
productivity during 3 breeding seasons; (2) determining spatial requirements and
dispersion of breeding pairs, (3) evaluating habitat selection of breeding pairs throughout
the landscape, and (4) assessing density and productivity of caracaras on cattle ranches.
Results and conclusions are organized into four chapters.

Chapter 2 details the breeding biology and discusses reproductive strategies of
Florida's crested caracaras. I examine territorial occupancy, breeding rates, and various
reproductive parameters for the sample population. Specifically, I test the hypothesis that
caracaras, like other large raptors, have low reproductive rates and low productivity. I
also examine relationships between breeding chronology and environmental parameters.
Finally, I examine factors that may influence overall productivity in this population.

In Chapter 3, I examine the incidence, determinants, and potential costs of double
brooding in this population, a strategy previously unknown for this species. I also discuss
potential implications of double-brooding for individual and population productivity.









6
I examine spatial requirements and evaluate habitat selection of breeding pairs
across the south-central Florida landscape in Chapter 4. I estimated home ranges using
radiotelemetry then used resource selection techniques to develop a model that describes
habitat and landscape features important to breeding pairs in selection of a territory. In
this chapter I also examine density of breeding pairs on cattle ranches and compare
productivity among caracaras nesting at sites where the major land use is or is not cattle
ranching.

Chapter 5 contains a synthesis of the conclusions and findings from this research.
I present an evaluation of the current status of this population within the context of its
historical background and my research results. I also provide recommendations for
conservation efforts, management, and monitoring of this population.







7









- p













Widely distribued year around


-''" ::: Rare and local







Figure 1-1. Distribution of the crested caracara in North and Central America.









8













































Figure 1-2. Current known breeding range of the crested caracara in Florida. The study
area encompassed much of this range.













CHAPTER 2

BREEDING ECOLOGY OF FLORIDA'S CRESTED CARACARAS

Introduction
The crested caracara (Caracaraplancus) is one of nine species in the subgroup
Caracarinae (caracaras), family Falconidae (del Hoyo et al. 1994). Although some
species are widely distributed throughout Central and South America, caracaras as a
group have been little studied. The crested caracara is the only species that reaches North
America. Breeding populations occur in Texas (Oberholser 1974), southern Arizona
(Levy 1988, Ellis et al. 1988), and Florida (Howell 1932, Stevenson and Anderson 1994).
The breeding ecology of this species is poorly known throughout its range, although
populations in Texas (Dickinson and Arnold 1996) and Baja California Sur, Mexico
(Rivera-Rodriguez and Rodriguez-Estrella, unpubl. data), have received recent attention.
Florida's crested caracara population is of conservation concern because it is
isolated from other populations, has a restricted range, and apparently is of small size
(Funderberg and Heinzmann 1967, Heinzman 1970, Millsap et al. 1990). Nesting habitat
throughout its historic range has been greatly reduced by conversion of grassland habitats
for urban and agricultural development, and this habitat loss is expected to continue.
This population is currently listed as Threatened by the U.S. Fish and Wildlife Service
and the state of Florida. Despite this listing, published information is limited to
descriptions of its natural history and distribution (Nicholson 1929, Howell 1932, Bent
1937, Sprunt 1954, Stevenson and Anderson 1994, Layne 1996).
The intent of this research was to study the breeding ecology of Florida's crested
caracaras, particularly to provide data useful for conservation and management of this
population. Specific objectives were to (1) determine breeding chronology, (2) assess
relationships between breeding and environmental variables, (3) relate breeding ecology
to annual population productivity, and (4) assess variation in reproductive parameters.

9









10
Study Area and Methods
I studied the breeding ecology of crested caracaras during 3 years in south-central
Florida (Figure 2-1). The landscape is a mosaic of habitats, primarily grasslands and
pastures interspersed with cropland, citrus groves, dairies, pine flatwoods, freshwater
marshes and lakes, oak scrub, hardwood hammocks, and urban developments. Primary
land uses are cattle ranching, citrus production, and other agricultural crops. Breeding
pairs of crested caracaras occupy open grasslands and pastures. These habitats are
characterized by large expanses of low-growing (< 0.5 m in height) ground cover of
grasses and forbs and are dotted with numerous shallow ponds, wetlands, and marshes,
and single or small clumps of live oaks (Quercus virginiana), cabbage palms (Sabal
palmetto), and cypress (Taxodium sp.).
Southern Florida has a humid, subtropical climate with a distinct annual cycle of
temperature and precipitation (Chen and Gerber 1990). During the rainy season, June
through September or October, temperatures are warm and relatively uniform, averaging
about 280C. Convective rains and thunderstorms occur almost daily, and monthly rainfall
averages more than 60 cm. Monthly rainfall during the dry season, usually from
November through May, averages around 20 cm. During the dry season, air temperatures
are cooler but rarely drop below freezing. Daily high temperatures rarely exceed 200 C,
but in May, temperatures as high as 38C have been recorded.
During the three years of this study, I monitored 33, 62, and 62 nesting attempts
in 30, 48, and 55, (65 different) territories, respectively. I color-banded 49 adult
caracaras representing 41 territories. I also color-banded and radiotagged 189 nestlings
between 7 and 8 weeks of age while they were still in the nest or just after they fledged,
when they could be captured easily on the ground.

Nest Location and Visitation
I located active nesting territories by visiting known historic sites (Layne unpubl
data, Millsap 1989) and by conducting observations at sites where I had seen adult
caracaras. I found nests by observing the behavior of adults and by covering entire areas









11
on foot and searching individual trees. I conducted monthly observations throughout the
year in known territories to assess site fidelity of adults and to look for nesting activity.
Although I found most nests in 1994 after hatching, I followed nest progress before
laying in most territories in 1995 and 1996.

Once found, I checked nests weekly to determine their fate. I determined clutch
size, hatching date, brood size, and chick age using a mirror attached to an extendible
pole. Each nest was assigned a week of initiation based either on the known time of egg
laying or on back-dating from known hatching or fledging dates (assuming an incubation
period of approximately 32 days and a nestling period of approximately 8 weeks). I then
calculated week of initiation from the first week of September, which I assigned as the
first week of the breeding season each year, based on the earliest known egg laying
during all 3 years. I defined breeding season length each year as the number of weeks
during which I found nests with eggs. Early nests were defined as those in which egg
laying was initiated before the median laying week each year, late nests were those in
which egg laying was initiated after the median laying week.
Following nest failures, I continued to check these territories at least monthly for
signs of renesting. I also regularly searched all territories that had successfully fledged
one brood in a breeding season, for second nests. After fledging, I monitored chicks
weekly by telemetry to determine survival.

Reproductive Success
I estimated probabilities of nest success using Mayfield's method (Mayfield 1961,
Johnson 1979) for all nest attempts during all years (n = 157 attempts). I estimated
weekly rather than daily probabilities of nest success (Mayfield 1961), because of the
length of the caracara's breeding cycle. I defined a successful nest as one in which at
least one fledgling survived to independence, because conclusions based on reproductive
output measured at fledgling independence can differ strongly from those measured at
fledging (e.g. Haukioja 1970). I defined independence as the end of the post-fledging
dependency period (PFDP), the first 8 weeks post fledging. Before reaching









12
independence, young remained near the nest and were dependent on their parents for
food. I calculated nest success for the incubation, nestling, and post-fledging periods,
using lengths of 4, 8, and 8 weeks, respectively. I compared Mayfield nest success
probabilities between periods and years using a Z-test (Hensler and Nichols 1981).

Measures of reproductive success also included clutch size, nestling survival to
fledging, fledgling survival to independence, and egg success (the proportion of eggs laid
that produced independent young). For these analyses, I used a subset of the total nests
with known clutch size (n = 98 nests representing 50 territories). I compared these
measures of reproductive success among years using Mann-Whitney and Kruskal-Wallis
nonparametric tests (Sokal and Rohlf 1995).

Along with nesting success, territorial occupancy and breeding rates are often
used as indicators of the stability and productivity of a population (Steenhof 1987,
Newton 1979). Relatively unbiased estimates of these parameters may be derived by
determining these rates over several years for a set of preselected territories (Steenhof and
Kochert 1982). I identified 37 territories for which occupancy and breeding status were
known in 1994 and tracked their status through the subsequent 2 years. I defined an
occupied territory as one in which adult caracaras were seen on 3 or more visits to the
territory. I considered a breeding attempt to be actual egg laying, not just nest
construction. To assess changes in these rates between years, I used Cochran's Q-test
(Sokal and Rohlf 1995).

Productivity and Components of Variation
I defined two measures of annual productivity: the number of independent young
per nesting attempt and the number of independent young per successful pair.
Productivity per successful pair included young from renesting attempts. To evaluate
factors that influenced overall annual productivity, I calculated an Index of Productivity
for each year using information from known clutch-size nests:

Index of Productivity = (C)(B)(S)
N









13
where C = mean annual clutch size, B = length of the breeding season each year,
S = annual egg success, and N = length of the breeding cycle (Ricklefs and Bloom 1977,
van Riper 1995). Length of the breeding season was the average number of weeks each
year when nests with eggs were found, and length of the breeding cycle was

N=b+e+y+n+f

where b = nest building (4 weeks), e = egg laying (0.5 weeks), y = length of incubation
period (4.5 weeks), n = nestling period (8 weeks), andf= fledgling dependency period
(8 weeks). All time estimates are average for the sample. Total length of the breeding
cycle for the crested caracara in Florida is approximately 25 weeks.

To evaluate seasonal trends in reproductive performance, I compared clutch size,
number of independent young per nesting attempt, and Mayfield nest success estimates
for early (initiated egg laying before the median week of egg laying) and late (initiated
egg laying after the median week of egg laying) nests for all 3 years. I assessed monthly
trends in clutch size and number of independent young per nesting attempt using
weighted regression for unequal sample sizes among categories (Sokal and Rohlf 1995).
To examine variation among reproductive parameters, I calculated the coefficient
of variation (standard deviation/mean, Ricklefs 1973) in percent, for clutch size, number
of independent young per nesting attempt, number of independent young per successful
pair, and Mayfield nest success.

Environmental Data
For monthly rainfall analysis, I used data collected from 35 stations throughout
the study area by the National Weather Service and the South Florida Water Management
District. I averaged daily precipitation values for each month to obtain monthly average
precipitation for all 35 stations. Values for all stations were then averaged to obtain
monthly precipitation averages for the overall study area.

Statistical Analyses
Because each year's sample of territories contained a subset of the previous year's
territories, data were not pooled across years, unless indicated. I conducted statistical









14
analyses using SigmaStat v. 2.0 (Jandel Scientific 1995). Statistical significance was
evaluated at the a = 0.05 level.

Results

Territorial Occupancy and Breeding Rates
Territorial occupancy rates for the subsample of 37 territories were 97% and
100% in 1994 and 1995, respectively, but were lower in 1996 (89%, Q = 6.5, P < 0.05).
One territory was only occupied during 1995, and I could not confirm occupancy in 2
additional territories during 1996. Adult caracaras were seen in these areas, but no
individuals were banded in these territories and no breeding activity was observed near
the traditional nesting site, so these territories were considered unoccupied. Among
known occupied territories, breeding rates remained high during all 3 years, at 100%,
97%, and 97%, respectively (Q = 2.0, P > 0.25).

Breeding Behavior and Site Fidelity
Pairs of breeding caracaras are apparently monogamous and remain together on
the territory throughout the year, although they are not easily observed outside the
breeding season. The beginning of the breeding season was signaled by several changes
in the adults' behavior. Pairs were often observed around the traditional nesting site
perching together, preening and allopreening, and sharing food. Adults were generally
faithful to both mate and site. Thirty-nine of 49 banded adults, including both banded
members of 5 pairs, were present in the same territory and reproduced there for at least 2
years. Although this represents an annual turnover rate of approximately 18%, I do not
know the fate of the birds that disappeared. Re-pairing within a breeding season was
observed only once, after disappearance of a banded female shortly after her single chick
fledged. During the 1995 breeding season, an entire family abandoned its breeding
territory about one month after the 2 chicks fledged, and moved to an area approximately
4 km from the nest. This new area was within the territory of another pair that had
successfully fledged two young that same year. Radiotelemetry data for the female and 2









15
chicks that had relocated indicated that they did not return to their original nesting
territory. This female nested successfully in the new territory during 1996.

Nest Building and Incubation
Both males and females build the nest. All pairs that built nests eventually
produced a clutch. Building time ranged from 2 to 4 weeks (n = 34 pairs) and often, up
to 2 months elapsed between nest completion and egg laying. Both parents developed a
brood patch and incubated the clutch. Neither mate delivered food to the other during
incubation, but one would remain on the nest while the other left the area to forage.

Length of the Breeding Season and Timing of Broods
Florida's caracaras' exhibit a prolonged breeding season (Figure 2-2). Chicks
fledged from successful nests as early as mid-December, indicating that eggs had been
laid during late September. The latest eggs were laid in July, in 1995, although these
nests failed (n = 2; both were second attempts). Most young fledged during March and
April. Chicks from late and second nests fledged as late as July.
Although yearly variation existed in onset and duration of the breeding season, the
median week of egg laying for all nesting attempts (n = 157) was in late January each
year and did not differ among years (H= 4.14, P = 0.13, Table 2-1). Secondary peaks of
egg laying in March and April (Figure 2-2) represent territories where pairs made a
second nesting attempt following a successful first brood (n = 16 pairs). The breeding
season was longest in 1995, when nests with eggs were found during 40 weeks. Breeding
season length in 1994 and 1996 was approximately 28 weeks (Table 2-1).
Although within-year variation existed in timing of nest initiation among pairs,
between-year nesting intervals for individual pairs were remarkably consistent. The
median year-to-year interval between egg laying for single brooded pairs was 50 weeks
(n = 41 pairs, Table 2-1), thus individual pairs tended to initiate nests at about the same
time each year.









16

Number of Broods Per Year
Multiple nesting within a breeding season occurred during all 3 years. Early
nesting pairs regularly renested, either by attempting a second brood or after a first
attempt failed (Figure 2-3). During 1995 and 1996, an average of 17.5% of early nests
failed in the incubation or early nestling stages. Eighty percent of these pairs renested
and half were successful (Table 2-2). Time between failure of the first nest and egg
laying in the second nest ranged from 3 to 14 weeks (n = 8 pairs, Table 2-2).
Successful double brooding was recorded during all 3 years (Table 2-2).
Although second nests failed more often and had smaller clutch sizes, pairs that
successfully raised two broods during a season had higher annual productivity than pairs
that raised only one brood. More pairs renested after early failure and attempted a second
brood in 1995 than in other years (Table 2-2).

Clutch Size
For known-size clutches (n = 98), the overall mean was 2.23 eggs (Table 2-3).
Modal clutch size was 2 eggs for all years. Mean clutch size was higher in 1994
(H = 7.61, P = 0.02, n = 16) owing to more clutches of 3 (Figure 2-4) but did not differ
between 1995 and 1996 (U = 1698, P = 0.98). For the same females, clutch size
decreased between first nests (2.18 0.20) and replacement or second brood nests
(1.91 0.09, paired m-test, t = 2.19, P = 0.03, n = 11 pairs).

Broods
Hatching was asynchronous. Eggs hatched one to two days apart. Both parents
fed and cared for the young. Brood size at approximately 4 weeks after hatching
averaged 82% of mean clutch size for all 3 years pooled (n = 81 broods, Table 2-3).
Chicks occasionally fell out of nests (pers. obs.). Many chicks simply disappeared; some
may have died from starvation. Over all years, only 38% of nests with 3-egg clutches
fledged all 3 chicks. Nestlings were in the nest for 7 to 8 weeks, although the earliest
recorded fledging occurred at age 42 days. After fledging, the young remained in the
general area around the nest and were dependent on the adults for food for at least 2









17
months. Even after independence, some young remained in the territory until the adults
begin breeding the following year.

Reproductive Success. Productivity, and Components of Variation
The overall Mayfield probability of nest success for all nest attempts (n = 157)
was 0.74 (SD = 0.03, Table 2-4). Nest success was lower in 1995 than in 1996
(Z = -4.83, P <0.0001). Nest success was also lower in 1995 than in 1994, but not
significantly so (Z = 0.73, P > 0.46, Table 2-4). Lower nest success in 1995 was caused
by higher failure rates of early nesting pairs and second brood attempts. Probabilities of
nest success during all phases of the nesting cycle were similar within years (Table 2-4)
except for the incubation period in 1994. The number of exposure weeks for the
incubation period was less in 1994 than in other years because in that year, most nests
were found after hatching. Early-season nest failures were likely missed in 1994.
Measures of productivity were lowest in 1995 (Table 2-5), owing to reduced egg
success (Table 2-3) and nest success (Table 2-4). In all years, productivity per successful
pair was enhanced because pairs successfully renested following early failure or raised
second broods. Clutch sizes of 3 produced the most fledglings and independent young
during all years (Figure 2-5). Two- and 3-egg clutches produced similar numbers of
independent young. Most clutches of 1 failed (Figure 2-5).
Measures of productivity showed the highest variability among reproductive
parameters (Table 2-6). Clutch size varied little and nesting success showed the least
variability among parameters (Table 2-6).

Performance and Laying Date
All measures of breeding performance declined as the breeding season progressed
each year (Table 2-7). Mean clutch size was negatively correlated with laying date
(r = -0.85, P < 0.01, Figure 2-6) and the regression was significant (F= 29.07, n = 98).
Approximately 76% of the variation in clutch size was due to month of nest initiation.
The probability of a nest succeeding varied throughout the breeding season (Figure 2-7).
During all years, pairs that initiated egg laying early in the breeding season had a higher









18
probability of nest success (Z = -3.06, P < 0.001, n = 76) than late nesting pairs (n = 81,
Table 2-7), although the earliest nesting pairs generally had a lower probability of nest
success (Figure 2-7). Timing of nest initiation also clearly affected productivity. The
number of independent young per nesting attempt declined as the breeding season
progressed (r = -0.95, P < 0.01, Figure 2-8) and again, the regression was significant
(F= 18.44, P < 0.005, n = 157). Approximately 67% of the variation in productivity
could be explained by month of nest initiation.

Nest Failure. Egg and Chick Mortality
Overall, 61.2% of all eggs that hatched from known-size clutches survived to
fledging and 48.4% survived to independence (Table 2-3). The fate of most eggs that did
not result in an independent young was unknown (Table 2-8). Total nest failure occurred
most often during late incubation and near the time of hatching (Table 2-9). Overall, an
average of 25% of nests failed each year (Table 2-9), but most early nesting pairs that
failed renested, and approximately 50% of renesters were successful (Table 2-2). On
average, 18% of pairs each year failed to produce any young despite a nesting attempt
(Table 2-9).

Discussion
The reproductive rate of birds presumably represents the maximum number of
young that individuals in a population can raise given the prevailing environmental
conditions during a breeding period (Amadon 1964, Lack 1968, Ricklefs 1973). Raptors
typically have lower reproductive rates than most birds (Newton 1979). Territorial
occupancy rates vary widely, and a percentage of pairs annually may not attempt to breed
or may abandon an attempt, often due to weather, food availability, or density dependent
effects (Newton and Marquiss 1976, 1986, Village 1982, 1983, Kostrzewa 1996).
My results indicate territorial occupancy and breeding rates for Florida's crested
caracaras that are higher than rates reported for most raptors. In a resident population
where pairs are monogamous and remain on the territory year-round, individuals'
familiarity with both territory and mate may facilitate successful annual breeding.









19
Although annual breeding rates were high, productivity varied considerably
among pairs and years. In all years, factors associated with productivity included
weather, timing and length of the breeding season, clutch size, timing and number of
broods, and mortality of eggs and chicks.

Weather. Timing. and Length of the Breeding Season.
Many factors interact to modify the timing of reproduction in birds, particularly
for tropical and subtropical species (Skutch 1950, Immelmann 1971, Ricklefs and Bloom
1977). Cues that function to initiate the crested caracara's breeding season in Florida are
unknown. Food supply often is a proximate factor regulating both reproductive timing
and effort (Perrins 1970, Newton and Marquiss 1981, Dijkstra et al. 1982, Daan et al.
1989). In tropical and subtropical regions, the abundance of food resources, and thus
many birds' reproductive cycles, are closely tied to rainfall cycles (Skutch 1950).
Florida's climatic cycle may provide predictable cues to resource availability for crested
caracaras. The length and timing of the caracara's breeding season in Florida was
strongly tied to the rainfall cycle (Figure 2-2). Reproduction occurred during the winter
dry season, with the peak in egg laying following periods of decreasing precipitation in
the fall. During the fall, as conditions become drier, fish and other prey become
concentrated in ephemeral wetlands, ponds, and ditches (Kushlan 1990, Babbitt 1996).
Caracaras often forage in these habitats and regularly kleptoparasitize other species
feeding there (Rivera-Rodriguez and Rodriguez-Estrella 1992, pers. obs.). Seasonal and
predictable abundance of prey during dry-down may enhance individual females' ability
to attain body condition necessary for breeding.
Rainfall cycles may also influence the length of the breeding season. Higher
precipitation during the winter of 1994-1995 (Figure 2-2) resulted in a higher water table
at the beginning of the rainy season in summer 1995. I found nests containing eggs
during a longer period in 1995 and more breeding attempts were recorded. Overall
productivity for 1995 was lower, however, despite the longer breeding season and similar









20
mean clutch size. Lower productivity was due to reduced egg and nest success during
that year (Tables 2-3 and 2-4).
Temperature cycles may also influence reproduction. High temperatures in the
spring at the end of dry season may impose additional physiological stress on breeding
adults, particularly on late nesters. In 1995, many second and late season (June-July)
nests failed, and 3 clutches were found abandoned during incubation in May, when
ambient temperatures reached > 40C for several consecutive days.

Clutch Size
Similar to many avian species, the most productive clutch size for Florida's
crested caracaras was larger than the modal clutch size, during all years. This is contrary
to Lack's hypothesis (Lack 1954, 1966) but instead supports the hypothesis that females
capable of laying a variable number of eggs lay clutches as large as they can produce
(Klomp 1970, von Haartman's 1971, H6gstedt 1980). Differences in territorial and
individual quality may exist that translate into the ability of each female to lay the largest
clutch possible.

Timing and Number of Broods.
A decline in all aspects of breeding performance with advance in laying date has
been well-documented for birds (Lack 1968, Perrins 1970, Newton and Marquiss 1984,
Daan et al. 1989, Verhulst and Tinbergen 1991). Timing of nest initiation clearly affects
the reproductive performance of Florida's crested caracaras. Clutch size, probability of
nest success, productivity, and probability of renesting all declined as the breeding season
progressed. Differences in territory or individual quality can affect the timing of egg
laying (Klomp 1970, Murphy 1978, Verhulst et al. 1995). Reduced nesting success late
in the breeding season may be indicative of seasonal environmental changes, particularly
declining food availability (Perrins 1970, Korpimaki and Lagerstr6m 1988).
Although the ability to renest is clearly advantageous for many avian species
(Smith and Roff 1980, Westmoreland et al. 1986, Guepel and DeSante 1990, den Boer-
Hazewinkel 1987) most raptors do not normally hatch more than one clutch per season









21
(Newton 1979). The ability to renest enhances annual productivity and thus lifetime
reproductive success of individual caracaras. About half of the pairs that renested
following early failure ultimately did produce independent young during the same year.
During all 3 years, pairs that successfully fledged two broods in a breeding season
incurred higher productivity than pairs that fledged only one brood.

Mortality
Loss of eggs and chicks during the breeding cycle affected annual productivity.
Less than half of all eggs from known-size clutches resulted in independent young. Most
losses occurred before fledging and the cause of most losses was unknown. Most eggs
lost during incubation disappeared at the time of hatching. With weekly nest checks, it
was difficult to determine whether eggs disappeared or whether newly hatched nestlings
were lost during the first week. Predation, possibly by crows and raccoons (Layne 1996,
pers. obs.), and weather were suspected causes of egg and chick disappearance.
Collisions with vehicles was the major cause of fledgling mortality. The loss of a
relatively large number of fledglings due to vehicle collisions in 1995 contributed heavily
to reduced overall productivity of the sample of breeding pairs during that year.

Breeding Strategy and Variation in Reproductive Parameters
Florida is characterized by both spatial and temporal environmental heterogeneity.
Florida's crested caracaras exhibit much spatial and temporal variation in the timing of
reproduction and reproductive performance by individual pairs. Clutch size and nesting
success showed the least variation among years. Since clutch size is related to the
number of young parents can raise (Lack 1954), there would be little advantage to
varying clutch size unless future conditions were predictable (Cody 1971). The low
coefficient of variation for nest success suggests that most pairs have a good chance at
producing at least one independent young during any year. High variation in productivity
may be attributable to variable food supply and a myriad of potential causes for loss of
eggs and chicks. High variation in productivity is also partly due to the fact that









22
productivity was measured after the young had become independent and mortality factors
had acted for a longer period.
Variable weather and resource conditions probably have been major factors in the
evolution of observed life history patterns for Florida's crested caracaras. Winter weather
is unpredictable in south-central Florida. Early nesting pairs, in particular, face a
tradeoff. Although reproductive success is highest for early nesting pairs, the very
earliest pairs incur a high risk of nest failure due to unpredictable winter storms and cold
fronts (Figure 2-7). The strategy of multiple nesting may offset this risk. Pairs that
successfully renest after failure or successfully double brood ultimately may have more
young recruited into the breeding population. Despite the risk of failure due to weather,
nesting as early as possible is advantageous because early nesting pairs generally fledge
more chicks per nest and have the opportunity to make a second nesting attempt.
The plastic patterns of reproduction for Florida's crested caracaras presumably
evolved because they are effective in this variable environment. Although caracaras may
breed in any month of the year, the peak of egg laying in late January during all 3 years
suggests an optimal time for reproduction. Fewer nests are initiated outside this period
and they are less successful, although some do succeed. If unpredictable, catastrophic
storms occur during the peak reproductive period, widespread nest failure may result.
During these years, pairs that initiate nesting outside the "optimal" period may exhibit
higher reproductive success. Such plasticity in spatial and temporal nesting patterns will
result in successful nesting and productivity by at least some pairs during any particular
year.

Comparison With Other Crested Caracaras
Information on the breeding biology and nesting success of other crested caracara
populations is limited. Dickinson and Arnold (1996) and Rivera-Rodriguez and
Rodriguez-Estrella (unpubl. data) have reported extended breeding seasons for crested
caracaras in Texas and Baja California Sur, Mexico, respectively. Mean clutch size of
crested caracaras in Texas was 2.14 0.14 (SD) eggs, (n = 7 clutches, Dickinson and









23
Arnold 1996). Productivity was 1.14 + 0.46 (SD) independent young per breeding pair
(n = 7 attempts) and declined as the breeding season progressed (Dickinson and Arnold
1996). Rivera-Rodriguez and Rodriguez-Estrella (unpubl. data) reported productivity of
1.91 + 0.08 (SE) fledglings per nesting attempt (n = 22 attempts) in Mexico. Double
brooding was suspected at both sites. Second clutches were observed in June in Texas
(Dickinson and Arnold 1996), and in August in Mexico (Rivera-Rodriguez and
Rodriguez-Estrella, unpubl. data).









24
Table 2-1. Breeding chronology and intervals, in weeks, between broods for Florida's
crested caracara population, 1994 1996. ND = no data.

Length of Median year-to-
breeding Median week year interval
season of initiation for single
Year (weeks) for all attempts n brooded pairs n
1994 28 21 (3rd wk Jan) 33 ND
1995 40 22 (4th wk Jan) 62 52 18
1996 28 22 (4th wk Jan) 62 48 23









Table 2-2. Multiple nesting in Florida's crested caracara population, 1994 1996. Early nests are those that initiated egglaying before
the median week of egglaying for the entire sample, each year.

Interval between
Number Attempted failed first nest
of Early Early nests renesting and replacement Attempted
Year territories nests that failed after failure Successful nest in same year second brood Successful
1994 30 19 1 0 3 (10.0%) 3 (100%)
1995 48 26 7(26.9%) 6 (85.7%) 3(50%) 4 6 weeks 8 (16.7%) 4 (50%)
1996 55 31 3 (9.7%) 2 (66.7%) 1(50%) 3, 14 weeks 5 (9.1%) 3 (60%)





t'









Table 2-3. Reproductive success parameters for Florida's crested caracaras, 1994 1996. Nestling survival to fledging (number of
fledglings/number of eggs laid), fledgling survival to independence (number of young alive at end of the post-fledging dependency
period/number fledged), and egg success (proportion of eggs laid that produced independent young) for eggs from known size
clutches.

Percent No. survived
No. Clutch Eggs Brood No. of clutch No. Nestling to end Fledgling Egg
Year clutches size Laid size broods size fledged Survival of PFDP Survival Success
1994 16 2.56(0.16) 41 2.23+0.20 13 87% 27 65.9% 22 81.5% 0.54
1995 41 2.17(0.08) 89 1.75+0.12 32 81% 51 57.3% 34 66.7% 0.38
1996 41 2.17(0.08) 89 1.75+0.12 36 81% 56 62.9% 50 89.3% 0.56
Pooled 98 2.23 (0.05) 219 1.83 + 0.07 81 82% 134 61.2% 106 79.1% 0.48


0I









Table 2-4. Estimates of Mayfield nest success probabilities for breeding pairs of crested caracaras in south-central Florida, 1994 -
1996.

Territories Nesting Incubation Nestling Fledgling Total probability
Year Followed Attempts period (SD) period (SD) period (SD) of nest success (SD)
1994 30 33 0.75 (0.06) 0.95 (0.02) 0.93 (0.02) 0.66 (0.09)
1995 48 62 0.87 (0.02) 0.82 (0.02) 0.83 (0.02) 0.59 (0.04)
1996 55 62 0.94 (0.02) 0.92 (0.01) 0.98 (0.01) 0.85 (0.03)

Overall 65 157 0.89 (0.02) 0.88 (0.01) 0.93 (0.01) 0.74 (0.03)





I,,.









28
Table 2-5. Measures of annual productivity (mean SE) and Index of Productivity for
crested caracaras in south-central Florida, 1994 1996. See text for calculation of Index
of Productivity. Productivity per successful pair includes young from renesting attempts.

Number of independent Number of independent Index of
Year young per attempt n young per successful pair n Productivity
1994 1.33 + 0.16 33 1.87 + 0.16 23 1.55
1995 0.98 + 0.11 62 1.69 + 0.10 36 1.32
1996 1.34 + 0.10 62 1.69 +0.09 49 1.36








Table 2-6. Variability in reproductive parameters for crested caracaras during 3 years in south-central Florida. CV = coefficient of
variation (standard deviation/mean). Productivity/successful pair include young from renesting efforts. n = 98 clutches.

1994 1995 1996 Total
Parameter Mean (SD) CV (%) n Mean (SD) CV (%) n Mean (SD) CV (%) n Mean (SD) CV (%) n
Clutch size 2.56(0.63) 24.6 16 2.17(0.54) 24.9 41 2.17 (0.50) 23.0 41 2.23 (0.55) 24.4 98
Productivity per 1.33 (0.89) 66.9 33 0.98(0.90) 91.8 62 1.34 (0.79) 59.0 62 1.20(0.87) 72.5 157
nesting attempt
Productivity per 1.87(0.78) 41.7 23 1.69(0.62) 36.7 36 1.69 (0.62) 36.7 49 1.73(0.65) 37.6 108
successful pair
Nesting success 0.66(0.09) 13.6 33 0.59(0.04) 6.8 62 0.85 (0.03) 3.5 62 0.74(0.03) 4.0 157
(Mayfield)









Table 2-7. Seasonal decline in breeding performance (mean + SE) for known size clutches and for all nest attempts of crested
caracaras in south-central Florida, 1994 1996. Early nests initiated egglaying before the median week of egglaying each year,
late nests initiated egglaying after the median week.

Year
1994 n 1995 n 1996 n Total n
Clutch size (n = 98)
Early 2.75+0.16 8 2.29+0.11 24 2.24+0.09 29 2.33+0.07 61a
Late 2.38+0.26 8 2.00+0.12 17 2.00+0.12 12 2.08+0.09 37a

Number independent
young per attempt (n = 157)
Early 1.53+0.19 19 1.27+0.18 26 1.39+0.16 31 1.38+0.10 76b
Late 1.07+0.25 14 0.76+0.14 36 1.29+0.12 31 1.02+0.09 81b

Nest success (n = 157)
Mayfield estimate (SD)
Early 0.95(0.03) 19 0.72(0.06) 26 0.82(0.05) 31 0.79(0.03) 76c
Late 0.55(0.11) 14 0.50(0.06) 36 0.88(0.04) 31 0.64(0.04) 81c

a t = -2.18, P = 0.03
b U= 6687.5, P = 0.02
SZ= -3.05, P < 0.001








Table 2-8. Fates of eggs from known-size clutches for crested caracara nests in south-central Florida, 1994 1996.

Fate 1994 1995 1996
Chick survived to independence 22(53.7%) 34(38.2%) 50(56.2%)
Incubation stage
Unknown egg/chicka 5(12.2%) 20 (22.5%%) 23(25.8%)
Eggs did not hatch 3 (7.3%) 1 (1.1%) 0
Died during pipping 0 1 (1.1%) 0
Eggs abandoned 0 8 (9.01%) 0
Known predation 3 (7.3%) 2 (2.2%) 0
Nestling stage
Chick died in nestb 0 0 2 (2.2%)
Chick defect/diseasec 0 2 (2.2%) 0
Chick disappeared, fate unknown 2 (4.9%) 4 (4.4%) 7 (7.9%)
Chick fell out of nest 1 (2.4%) 1 (1.1%) 1 (1.1%)
Fledgling stage
Killed by vehicle 1 (2.4%) 7 (7.9%) 6 (6.7%)
Fledgling found dead, cause unknown 2 (7.3%) 8 (9.0%) 0
Fledgling disappeared, fate unknown 2 (2.4%) 1 (1.1%) 0
Total eggs 41 89 89

a Not known whether eggs were lost during incubation or whether they hatched and chicks were lost during early nestling stage.
b Chick found dead in nest during first two weeks, from unknown causes.
CChick died of injury or defect before it fledged.









Table 2-9. Failure of crested caracara nests in south-central Florida, 1994-1996.

1994 1995 1996
Total nest attempts 32 62 62
Total territories 30 48 55
Failures (% of all nest attempts) 7 (21.9%) 24 (38.7%) 11 (17.7%)
During incubation/hatching 2 9 5
During nestling period 2 7 3
After fledging 3 8 3
Territories that did not produce any 6 (20.0%) 11 (22.9%) 5 (9.1%)
young despite nesting attempt










33


























....





















Figure 2-1. Location of study area for crested caracaras in south-central Florida. Study
area encompassed much of the species' current known breeding range in Florida.











SNumber nests initiated by month
26 I INumber of successful nests
24 1""-Rainfall + SE
S0.90
22


18 0.70

S 16 \ 0.60

a 0.50
12
1 0 0 .4 0 "0

0 8 -A
0.30

1 64 0.20
z 4
2 0.10

0 --- 0.00
Ua oa o?. o, U


Month
Figure 2-2. Mean monthly rainfall and breeding chronology of crested caracaras during 3 years in south-
central Florida. 1994 (n = 33 nest attempts), 1995 (n = 62 attempts), 1996 (n = 62 attempts).













IINumber of nests initiated, n = 157
60 Proportion of those initiated in each month that renested
60 1

0.9
50
0.8

20.7
S40
0.6

~ 304- 0.5 .
o 0
0.4 "
I 0.3 ,

0.2
10
0.1


Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul

Month of nest initiation

Figure 2-3. Relationship between month of first nest initiation and occurrence of renesting by crested
caracaras following either an early failure or a successful first brood, south-central Florida, 1994 1996.










36







01994, n = 16 clutches
80 1995, n = 41 clutches
T 1996,n = 41 clutches
701


60

50


40


10
30 30

20


10



1 2 3

Clutch size


Figure 2-4. Distribution of clutch sizes by year, for crested caracara
nests in south-central Florida, 1994 1996.









37




1994 Clutch of 1 (n= 1)
3 ...* ----Clutch of 2 (n= 5)

S-.... Clutch of 3 (n= 10)
2 A .................




0 =
I~~ ~ --.^ -- --



Eggs Number fledged Number of
independent
young


3 ... 1995 -- Clutch of 1 (n = 3)
...--- Clutch of 2 (n = 28)
2 it.... "A... *-*--Clutchof3(nO= 10)




o l ---------- -- -- -- -- ----

Eggs Number fledged Number of
independent
young



3 1i-- -1- Clutch of 1 (n = 2)
I .-...- Clutch of 2 (n = 30)
S....... -- -i- -Clutch of 3 (n = 9)



I 1=
E .....


Eggs Number fledged Number of
independent
young




Figure 2-5. Mean number of fledglings and independent young per nest for each clutch
size during each year for crested caracaras in south-central Florida, 1994 1996.
n = indicates number of clutches.












3.5

Mean clutch size

2 -.----y=2.91 -0.12x r= -0.85, P <0.01
34



S-16
2 2.5 38








1.5
Oct Nov Dec Jan Feb Mar Apr May Jun Jul

Month of nest initiation

Figure 2-6. Relationship between mean clutch size and month of nest initiation for 98 known-size
clutches of crested caracaras in south-central Florida, 1994 1996. Number of clutches is shown above
each month.











1.25



1.00



0.75 '



0.50



0.25



0.00
Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug
Month

Figure 2-7. Mayfield probability of nest success ( SD) by month for crested caracaras
in south-central Florida, 1994 1996 (n = 157 nest attempts).













2.5
- --.--y = 2.011 -0.153x r =-0.95, P <0.01
-2- Number of independent young
eo 2.0 1


c"*--. 31 54 27
1 5 3 .


1 12
4 1.0-
010


Z 0.5- "-


0.0 I
Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul

Month of nest initiation

Figure 2-8. Relationship between productivity per nest attempt (n = 157 attempts)
and month of nest initiation for crested caracaras in Florida, 1994 1996.
Number of nest attempts is shown above each month.













CHAPTER 3

EFFECTS ON PRODUCTIVITY AND POTENTIAL COSTS OF DOUBLE
BROODING IN CRESTED CARACARAS

Introduction
Reproductive success in birds is determined by the number of broods and
productivity per brood (Drury 1961, Lack 1966). The number of successful broods
produced each breeding season by a breeding pair is an important contribution to an
individual's annual and lifetime productivity (Lack 1954, Steams 1992, Newton 1989).
Successful multiple brooding may also be important in the maintenance of populations
(Holmes et al. 1992).
In temperate latitudes, many avian species regularly raise more than one brood per
year (Smith and Roff 1980, Westmoreland et al 1986, den Boer-Hazewinkel 1987,
Guepel and DeSante 1990, Smith and Marquiss 1995). Occurrence of multiple brooding
and temporal spacing of broods are influenced by length of the breeding season, food
availability, body size, clutch size, development time, amount of parental care, and
experience of the parents (Perrins 1970, Ricklefs 1984, Bednarz 1987, Askenmo and
Unger 1986, Tinbergen and van Balen 1988, Holmes et al. 1991, 1992).
Successful production of multiple broods in a season may increase fitness only if
the strategy does not negatively influence the future fecundity or survival of either
breeders or first-brood young (Williams 1966, den Boer-Hazewinkel 1987, Nur 1988,
Smith and Marquis 1995, Verhulst and Hut 1996). Multiple-brooded individuals usually
fledge significantly more young per season than single-brooded individuals, but
reproductive success and offspring quality generally decline between first and second
broods, and the probability of independent offspring being recruited into the breeding
population is lower for young from second broods (Smith and Roff 1980, Hochachka
1990). Despite these costs, multiple brooding may be maintained in some species

41









42
because it increases an individuals' lifetime reproductive success (Perrins 1970, den
Boer-Hazewinkel 1987, Newton 1989). Under certain environmental conditions, there
may be little cost to producing additional broods (Ormerod and Tyler 1993, Smith and
Marquiss 1995), so the benefit easily outweighs the cost.
Raptors typically raise only one brood per year because of the length of time
required for the complete breeding cycle including, for many species, an extended post-
fledging dependency period (Newton 1979). Although many raptors renest after losing a
clutch (Newton 1979, Wood and Collopy 1993, Forsman et al 1995), true double
brooding (i.e., individuals attempting to rear a second brood in the same breeding season
after a successful first attempt) rarely has been confirmed. Second brood attempts,
generally in response to peaks in prey cycles, have been reported for some smaller species
with short breeding cycles such as the American Kestrel (Falco sparverius; Howell 1932,
Stahlecker and Griese 1977, Toland 1985), Eurasian Kestrel (F. tinnunculus; Fargallo et
al. 1996), Black-shouldered Kite (Elanus caerulescens; Malherbe 1963), Burrowing Owl
(Athene cunicularia; Millsap and Bear 1990), and Barn Owl (Tyto alba; Witmer and
Patrick 1987). Multiple brooding occurs in some larger raptors that live year-round in
subtropical and tropical environments where food is abundant for most of the year, such
as the Harris' Hawk (Parabuteo unicinctus; Mader 1975, Bednarz 1987), Pale Chanting
Goshawk (Melieraxpoliopterus; Smeenk and Smeenk-Enserink 1975), and Galapagos
Hawk (Buteo galapagoensis; DeVries 1975).
I documented double brooding in crested caracaras (Caracaraplancus) during
three successive years in south-central Florida. The crested caracara is a medium-sized
raptor that occurs in open grassland and savanna habitats throughout Central and South
America, in the southwestern U.S., and in Florida (del Hoyo et al 1994). Circumstantial
evidence of possible double brooding in this species has been reported in Florida (Bent
1937), Texas (Dickinson and Arnold 1996), and Mexico (Rivera-Rodriguez and
Rodriguez-Estrella unpubl. data) but the prevalence of this phenomenon remains
unknown. Recently, widespread loss of grassland habitats to expanding agriculture and
urban development has focused concern on the Florida population (Layne 1996, B.









43
Millsap pers. comm.), which is currently listed as Threatened by the U.S. Fish and
Wildlife Service and by the state of Florida. Because successful double brooding may
affect individuals' lifetime reproductive success and annual population productivity, I
was interested in evaluating (1) the extent to which double brooding occurs, (2) the
contribution of second broods to an individual's annual productivity, (3) potential costs in
terms of reduced survivorship of first- or second-brood young, (4) potential costs to
double brooded adults in terms of reduced survival or fecundity, or delayed nesting in the
following year, and (5) factors that may contribute to the occurrence of double brooding.

Study Area and Methods
The study area in south-central Florida included all or parts of Highlands, Glades,
Okeechobee, Osceola, DeSoto, Polk, Hendry, and Indian River counties and was centered
on the MacArthur Agro-Ecology Research Center (2710'N, 81012'W), Highlands
County. This area encompasses much of the species' currently known breeding range in
Florida. Breeding pairs of crested caracaras occupy open grassland and pasture habitats
containing single or small clumps of live oaks (Quercus virginiana), cypress (Taxodium
sp.), and cabbage palms (Sabal palmetto). Breeding pairs are monogamous and remain
on territories year-round (Morrison 1996).
Beginning in January 1994, I conducted monthly surveys in known territories to
look for signs of nesting activity. Once found, nests were checked weekly until fates
were determined. I considered a nesting attempt to be actual egg-laying, not just nest
construction. I determined clutch size, hatching date, brood size, and chick age at each
nest using a mirror attached to an extendible pole. Age was determined by assessing the
chicks' size and feather development.
I assigned each nest a week of initiation based either on known time of egglaying
or on back-dating from known hatching or fledging dates (assuming an incubation period
of approximately 32 days and a nestling period of approximately 8 weeks). I then
calculated week of initiation from the first week of September, which I assigned as the









44
first week of the breeding season based on the earliest known egglaying during all three
years.
I estimated annual nesting success using Mayfield's method (Mayfield 1961). I
defined a successful nest as one that fledged at least one young. I compared Mayfield
estimates of nest success among nest types using a Z-test (Hensler and Nichols 1981).
I measured annual productivity as the number of young fledged per successful
pair. I assigned the classification of "single brood" to pairs that made only one nesting
attempt during an annual breeding season. Double brooded pairs were defined as those
that made a second nesting attempt following a successful first effort, in the same
breeding season. Territories in which pairs produced replacement clutches after nest
failures were not included. Clutch size is reported only for known-size clutches. I
defined interbrood interval as the number of weeks between fledging of the first brood
and initiation of egglaying in the second nest.
To evaluate whether laying a second clutch affected the duration of post-fledging
care or survival of first-brood fledglings, I used radiotelemetry to determine times of
departure of first and second-brood young from their natal territory and survival past
independence. Nestling caracaras were color banded and radiotagged (one per nest)
while still in the nest at 7-8 weeks of age or just after fledging, when they could be
captured easily on the ground. Young continue to beg and be fed by their parents for
several months after fledging (Morrison unpubl. data), so I measured the duration of
parental care as the period during which young remained in their natal territory. The
natal territory was defined as the area within a radius of 3 km from the nest, which
approximates a breeding pair's home range (see Chapter 4). I defined permanent
departure from the natal territory as the first time a juvenile was located more than 3 km
from the nest followed by a minimum of 3 consecutive weeks at distances > 3 km from
the nest.
I obtained survival data for 61 radiotagged young through their first year via
weekly aerial radiotracking surveys. I estimated probabilities of survival to recruitment
for young from single, first, and second nests using these survival data along with the









45
Kaplan-Meier procedure modified for staggered entry of animals (Pollock et al. 1989).
Age at first reproduction is not known for this species but is estimated to be 3 to 4 years
(Voous 1983, Layne 1996), so I used 1200 days post fledging as an estimate of
recruitment age. I compared survival probability functions between groups using the log-
rank test (Cox and Oakes 1984, Pollock et al. 1989).
To evaluate potential costs to double brooded adults, I followed the fate of color-
banded individuals. I used the likelihood ratio test in PROC LIFETEST in SAS (SAS
Institute 1985) to compare predicted survival estimates for 16 known double-brooded and
38 known single-brooded adults. Survival through at least 2 years was known for 11
double-brooded and 27 single-brooded adults. Survival was defined as remaining and
breeding on the same territory in subsequent breeding seasons. To assess the potential
cost of reduced fecundity in the year following a double brood event, for 1995 and 1996,
I compared clutch sizes between the group of known adults that had double brooded the
previous year and the group of known adults that produced only one brood. I also
examined whether double brooding had any influence on timing of the first nest attempt
of these pairs in the following year. For 1995 and 1996, I1 compared timing of first nest
initiation between pairs that had double brooded the previous year and pairs that
produced only one brood.
I evaluated potential effects of double brooding on population growth rates and
likelihood of persistence. For this analysis, I used values previously calculated for
survival of single- and double-brooded adults through 2 years, survival of first-, second-,
and single-brood young to 2 years post-fledging, and mean numbers of young fledged for
single- and double-brooded pairs. I calculated population size and population growth rate
during 3 consecutive years for two hypothetical populations, one in which double
brooding occurred in approximately 10% of the population each year, and one with no
double brooding. I assumed each initial population size to be 100 breeding pairs (50
females) and age of first breeding to be 4 years.
I used logistic regression to evaluate the effects of several factors on the
occurrence of double brooding in this population. I examined whether initiation date of









46
the first clutch (INITWK) and numbers of fledglings in the first brood (NUMFRST)
influenced the occurrence of double brooding during the same year. For pairs whose
reproductive history was known for at least 2 years, I examined whether the occurrence of
double brooding during any year was influenced by whether or not the pair had double
brooded during the previous year (DBPREV), the number of fledglings from the previous
year (NUMPREV), and the fledging date from the previous year (WKPREV).
Statistical tests were conducted using SigmaStat (Jandel Scientific 1995).
Significance for all tests was assumed at the a = 0.05 level. Comparisons of fecundity,
timing of nest initiation, and productivity were made using Wilcoxon rank sum and
Kruskal-Wallis nonparametric tests (Sokal and Rohlf 1995). Post-hoc power estimates
for two-sided Wilcoxon tests were obtained using the program Resampling Stats
(Resampling Stats Inc. 1995). I first simulated 10,000 values of the rank sum statistic
under the null hypothesis of no differences between nest types. The 95th percentile of
these values was used as the critical value for determining significance of the difference
between nest types, given the two corresponding sample sizes. Real data values were
resampled to approximate the alternate distribution of the rank sum statistic as suggested
by the data. The power of the Wilcoxon test at the 0.05 level was then estimated as the
proportion of times out of 10,000 iterations that the resampled test statistic was at least as
large as the critical value.

Results

Extent of Occurrence of Double Brooding
I monitored breeding activity on 60 different territories throughout the study,
although only a subset of those each year (Table 3-1). A second brood attempt was made
on 13 territories (21.7%, Table 3-1). In three territories, the pair initiated a second
nesting attempt during two consecutive years, although none of these pairs successfully
fledged two broods in more than one year. Both adults were color banded in one territory
where double brooding occurred. In all other territories where double brooding occurred,
one adult was banded and participated in both nesting efforts.









47

Timing of Broods
Pairs that attempted a second brood initiated their first nests significantly earlier
(U= 1504.5, P = 0.00, n = 16) than single-brooded pairs (n = 103) (Table 3-1). Second
nests of double-brooded pairs were initiated 9 13 weeks later (U = 26.5, P = 0.00,
n = 16) than nests of single-brooded pairs. Median week of nest initiation did not differ
among years for each nest type (single brood, H = 0.79, P = 0.67, n = 103; first of two,
H= 1.88, P = 0.39, n = 16; second of two, H= 0.65, P = 0.72, n = 16, Table 3-1).
Only early breeders attempted to double brood. The number of pairs that
attempted second broods decreased with increasing laying date (Figure 3-1). Of pairs that
attempted a second brood (n = 16), 72% fledged young from their first brood before 1
March, 94% before 1 April. Of all pairs that fledged a brood before 1 March (n = 16),
69% made a second nesting attempt. The mean interbrood interval was 7.6 weeks (range
2 13 weeks, n = 16), and interval length was positively correlated with the number of
young fledged from the first nest (r = 0.50, P < 0.05, Figure 3-2).

Productivity and Nesting Success
Mean clutch size declined between the first and second nests of double-brooded
pairs (Table 3-2). Laying times of second clutches and the first clutches of late breeders
overlapped during March and early April (Figure 3-1). Mayfield success estimates for
second nests were significantly lower than those for first nests or nests of single-brooded
pairs (Table 3-2). Of the pairs that attempted two broods, those that had an early first
nest produced the earliest second broods; these early second broods had the highest
probability of success (Figure 3-3). Taking into account all nesting attempts, double
brooded pairs did not necessarily fledge more young than single-brooded pairs
(T= 1697.50, P = 0.21, Figure 3-4), primarily owing to high failure rates of late second
broods. Pairs that nested earliest however, had the highest chance of successfully
fledging both broods (Figure 3-3), thus also had the highest chance of fledging more
young each year than single brooded pairs.









48

Parental Care and Post-fledging Survival
First-brood young incurred shorter periods of post-fledging parental care. Mean
time to departure from the natal territory of first-brood fledglings (11 wk 2.56 SE,
n =14 broods) was significantly shorter (U = 290.5, P = 0.00) than that of fledglings from
single-brood pairs (33 wk + 1.61 SE, n = 32 broods) or of second-brood fledglings
(29.7 wk + 1.73 SE, n = 6 broods, U= 33.0, P = 0.01). Post-fledging care of first-brood
young overlapped incubation of the second clutch in all cases. Young from first broods
typically left the natal territory during the early nestling stage of the second brood, but
whether the parents enforced this departure is not known. The period of post-fledging
parental care for fledglings from single-brood pairs and second-brood fledglings did not
differ (U= 54.4, P = 0.40). In many cases, single- and second-brood fledglings remained
in the natal territory until the parents began breeding the following year.
Estimates of survival probability to recruitment differed between young from first
and second broods (Kaplan-Meier survival function, X = 8.51, P = 0.004, Figure 3-5).
Survival estimates for young from first broods were also higher than estimates for young
from single broods (X2 = 75.95, P = 0.00, Figure 3-5).

Potential Costs to Adults
Survival estimates to subsequent breeding seasons did not differ (X2 = 0.01,
P = 0.92) between known double brooded (n = 16) and single-brooded (n = 38) adults
(Table 3-3). Clutch sizes did not differ (U = 95.0, P = 1.0) between known adults that
had successfully double brooded in the previous year (2.67 + 0.33, n = 10) and known
adults that produced only one brood (2.22 + 0.22, n = 19), but power of this test was low
(I-p = 0.05) because of small sample sizes each year. I did not detect a delay in nesting
by double brooded adults in the year following a double brood event. The median week
of egglaying did not differ between pairs that had successfully double-brooded the
previous year and pairs that fledged only one brood in either 1995 (U= 80.0, P = 0.13;
median week = 19, n = 3 double brooded pairs; median week = 22, n = 35 single-brooded
pairs) or 1996 (U = 105.0, P = 0.46; median week = 20, n = 4 double brooded pairs;









49
median week = 22, n = 43 single-brooded pairs). Power of these tests was low, however,
at 0.23 and 0.12, respectively.

Effects on Population Growth and Persistence
The hypothetical population that included double brooding by approximately 10%
of pairs annually grew at a slightly faster rate than the population comprised of only
single-brooded pairs. After 3 years, differences in size were negligible between the two
populations, however (Table 3-4).

Influence of Factors
Timing of initiation of the first brood (INITWK) was the only factor important in
predicting the occurrence of double brooding that same year by caracara pairs
(Table 3-5). For pairs whose reproductive status was known for the previous year
(n = 45), the occurrence of double brooding in any year was not influenced by double
brooding the previous year, the number fledged during the previous year, or the date of
fledging the previous year (Table 3-6).

Discussion
Life history theory predicts that a breeding attempt will occur when the likely
benefits to lifetime reproductive success exceed expected costs (Williams 1966, Steams
1992). In Florida's crested caracara population, only a small percentage of pairs attempt
a second brood in any year. This strategy may be maintained, however, because double
brooded individuals may incur higher lifetime reproductive success (Perrins 1970, den
Boer-Hazewinkel 1987). Few second nests are successful and the probability of
recruitment is lower for second-brood young. Individuals that successfully fledge two
broods in a season are more productive, however, so they potentially have more young
recruited into the breeding population. In addition, costs of double brooding to adults and
young seem minimal.
Second clutches may reduce the reproductive value of the first brood because of
reduced post-fledging care of first-brood fledglings (Verhulst and Hut 1996). I found no
evidence that first-brood young incurred reduced post-fledging survival despite









50
substantially reduced post-fledging parental care. Indeed, first-brood fledglings actually
had the highest probability of survival to recruitment. This suggests that juvenile
caracaras can become independent foragers in a period as short as 7-8 weeks.
I found little evidence of costs to double brooded adults, though these conclusions
are preliminary because of the short time frame of the study, small sample sizes, and low
power of statistical tests. Caracaras are long-lived (Klimkiewicz and Fletcher 1989,
Morrison 1996), so true costs may not be evident in the short term. Additional years of
tracking the survival and reproductive efforts of known adults would be needed to
properly evaluate these potential costs.
The occurrence of double brooding in Florida's crested caracara population may
be related to both length of the breeding season and to food availability, two factors that
determine the number and the timing of broods produced by a breeding pair (Perrins
1970, Newton 1979, Tinbergen and van Balen 1988). Crested caracaras in Florida exhibit
a long breeding season, generally beginning in September and continuing throughout the
winter and into June or early July (Morrison 1996). Pairs that nest early may have
sufficient time to raise two broods.
Food supply is known to influence the occurrence and timing of breeding in
raptors (Cavd 1968, Dijkstra et al. 1982, Newton and Marquiss 1984). Food supply has
been shown to influence the number of young fledged annually by a breeding pair
primarily by influencing the occurrence of multiple brooding (Malherbe 1963, Bednarz
1987). The crested caracara has a varied diet and feeds on live vertebrate and invertebrate
prey as well as carrion (Richmond 1976, Ellis et al. 1988, Morrison 1996), therefore
breeding need not be synchronized with availability of particular foods. The caracara
typically inhabits areas associated with human activity such as cattle ranches, agricultural
lands, and urban dumps (Rodriguez-Estrella 1996, pers. obs.). These habitats may
provide a diversity of predictable food resources, perhaps owing to regular schedules of
management activities. The combination of a relatively mild climate and the perpetual
availability of a variety of food resources apparently allows early initiation of breeding
and occasionally, double brooding by some pairs. Whether individuals actually attempt a









51
second brood or even attempt to nest early in the season may depend on the age and
quality of individuals, experience of pairs, and territory quality. The occurrence of
double-brooding in any year was not influenced by the previous year's reproductive
effort, suggesting that double brooded individuals may be the most fit.
Year-round residency and persistence of the pair bond between nesting seasons
may facilitate double brooding in crested caracaras because of individuals' familiarity
with habitats and resources. Males and females have similar parental roles, so pairs may
have little difficulty feeding first-brood fledglings and incubating a second clutch.
Caracaras also frequently reuse old nests for first and second breeding attempts. Reuse of
old nests reduces the time required for nest-building (Mader 1975, Toland 1985,
Westmoreland et al. 1986).
Double brooding appears to be a regular but uncommon occurrence in Florida's
population of crested caracaras and may occur in other populations (Dickinson and
Arnold 1996, Rodriguez-Estrella and Rivera-Rodriguez, unpubl. data). An extended
breeding season and post-fledging dependency period may have contributed to a previous
lack of its confirmation. Some juveniles remain on their natal territory for several months
after fledging (Morrison 1996). Therefore, determining whether juveniles seen with
adults during any month of the year are from a single or second brood is not possible,
unless individual birds are marked and have a known history.
The ultimate value of double brooding must be that double-brooded individuals
produce more young with some probability of establishing themselves in the breeding
population. Whether or not double-brooding contributes to overall population growth
rates is of lesser importance; the overriding goal for individuals is to leave more offspring
than the average member of the population. Combined with an extended breeding
season, environmental and food conditions during some years may be favorable for some
caracara pairs to attempt a second nest. In addition, costs may be small, at worst. Of
greater interest is why more individuals do not attempt second broods. For many avian
species, the occurrence and timing of a breeding effort is influenced by age and
experience of the parents or energetic constraints related to territory or individual quality









52
(e.g., Smith and Roff 1980, Evans Ogden and Stutchbury 1996). The absence of double
brooding in some early nesting pairs of crested caracaras and the lack of earlier nesting by
many single-brooded pairs may be related to these factors. Further studies of conditions
that influence the occurrence and timing of breeding by individuals and a more accurate
assessment of recruitment rates of first and second-brood young are needed to better
understand the contribution of double brooding to individuals' lifetime reproductive
success and to the overall dynamics of Florida's crested caracara population.









Table 3-1. Occurrence and timing of nesting attempts between years for crested caracaras in south-central Florida. NA = not
applicable for that year.

No. of 2nd nest Median week Median week Median week
Total territories attempts found (%) of initiation of initiation of initiation
examined/new /Different from of single of first of two of second of
Year territories added previous year brood nests n nest attempts n two nest attempts n
1994 28/28 3 (10.7)/NA 21 (3rd wk Jan) 25 14 (5th wk Nov) 3 32 (2nd wk Apr) 3
1995 43/17 8 (18.6)/7 22 (4th wk Jan) 35 12 (4th wk Nov) 8 32 (2nd wk Apr) 8
1996 48/15 5 (10.4)/3 22 (4th wk Jan) 43 16 (3rd wk Dec) 5 35 (4th wk Apr) 5
Overall 60 territories 16 attempts/13 territories 103 16 16









54
Table 3-2. Clutch size (mean SE) and nesting success (SD in parentheses) between nest
types for crested caracaras in south-central Florida, 1994-1996. The same letter indicates
significant differences between nest types. ND = no data on clutch sizes of first or
second nests in 1994 because these nests were found after hatching.

Clutch Mayfield nest
Nest type Year size n success n
Single broods 1994 2.60 + 0.16 15 0.67(0.09) 25
1995 2.28 + 0.11 25 0.72(0.05) 35
1996 2.22 + 0.09 32 0.92(0.03) 43
Pooled 2.32 + 0.07 72 0.81 (0.03)a 103


First of two 1994 ND 1.00 (0.00) 3
1995 2.66 + 0.32 3 1.00 (0.00) 8
1996 2.33 + 0.33 3 1.00 (0.00) 5
Pooled 2.50 + 0.22b 6 1.00 (0.00)c 16


Second of two 1994 ND 1.00 (0.00) 3
1995 1.80+ 0.02 5 0.35 (0.15) 8
1996 2.00 + 0.00 2 0.25 (0.20) 5
Pooled 1.86_+ 0.14b 7 0.35 (0.12)a'c 16

a Z = 3.72, P < 0.0002
b U = 33.0, P = 0.03
c Z= 4.19,. P < 0.0001









55

Table 3-3. Estimated parameters for evaluating contribution of double brooding to
population growth for Florida's crested caracaras. Parameters estimated from data
collected during 1994 1996.

Probability of Probability of
Mean number adult survival fledgling survival
Parameter fledged per year to subsequent year to subsequent year
Year 1
Single 1.80 0.82 0.71
Double 3.00 0.88 0.75a

Year 2
Single 1.60 0.70 0.46
Double 1.19 0.78 0.43

Year 3
Single 1.53 0.70 0.43
Double 1.30 0.78 0.36

a Survival estimates were calculated separately for young from first and second broods. I assumed half of
all young produced by double-brooded adults were from single broods and half were from second
broods.
Year 1 (0.5*1.00+0.5*0.50)
Year 2 (0.5*0.73 + 0.5*0.13)
Year 3 (0.5*0.58 + 0.5*0.13)









56
Table 3-4. Estimation of population size and annual growth rate for two hypothetical
populations of crested caracaras in south-central Florida, one in which all pairs only
produce a single brood each year, and one in which approximately 10% of pairs
successfully produce two broods each year.

Single broods only
Total number Total number Total number
Year Adults young fledged 1 year olds 2 year olds Total N X
1 50.00 90.00 140.00 2.80
2 41.00 65.60 63.90 170.50 1.22
3 28.70 43.91 46.58 29.39 148.58 0.87
Mean 1.63


Single and double broods
Total number Total number Total number
Year Adults young fledged 1 year olds 2 year olds Total N X

single 45.00 81.00 126.00 2.92
double 5.00 15.00 20.00
2
single 37.17 59.47 57.51 154.15 1.14
double 4.13 4.91 3.69 12.73
3
single 26.32 40.26 42.23 26.45 135.26 0.88
double 2.92 3.80 2.85 1.23 10.80
Mean 1.65









57
Table 3-5. Statistics for logistic regression analysis of the influence of timing of
initiation of the first brood (INITWK) and the number fledged in the first brood
(NUMFRST) on the occurrence of double brooding in crested caracaras in south-central
Florida, 1994 1996.
n = 16 double-brooded pairs, 103 single-brooded pairs.

Variable Coefficient SE P
Constant 7.50 2.22 0.001
INITWK -0.52 0.12 <0.001
NUMFRST 0.01 0.43 0.97









58
Table 3-6. Statistics for logistic regression analysis of the influence of double brooding
the previous year (DBPREV), fledging date (WKPREV) and number fledged the previous
year (NUMPREV) on the occurrence of double brooding in crested caracaras in south-
central Florida, 1994 1996. n = 11 double-brooded pairs, 34 single-brooded pairs.

Variable Coefficient SE P
Constant 2.40 0.28 0.39
DBPREV 0.72 0.92 0.44
WKPREV -0.12 0.86 0.15
NUMPREV 0.27 0.55 0.63















14 I Number of single nests initiated, n = 103
Number of pairs that attempted a second nest, n = 16
12 D Number of second nests initiated, n = 16
Medians

10 -

10
8


6

4






-, 0 '.0 0q O 1-

Oct Nov Dec Jan Feb Mar Apr May Jun Jul
Initiation week

Figure 3-1. Timing of nest initiation throughout the breeding season for crested caracaras in south-
central Florida, 1994 1996.









60









14


12


'" 10i S


a 8 0


g 6
o
e 0 r = 0.50, P < 0.05
s 4 0


2 0


0 -- i i i-
0.5 1.0 1.5 2.0 2.5 3.0 3.5

Number fledged from first brood


Figure 3-2. Length of the interbrood interval in relation to number
fledged from the first nesting attempt for double brooded crested
caracaras in south-central Florida, 1994 1996. n = 16 pairs.












1.00 ----------- - - ---
\ --- Single broods
-0 4 -- Second of two broods
U First of two broods
0.75





S I 0.50 \ -



\ \






0.00 [
Sept Oct Nov Dec Jan Feb Mar Apr May Jun July

Month

Figure 3-3. Proportion of nests initiated in each month that were successful, for crested caracaras in south
central Florida, 1994 1996.









62









4.5 -
Pairs that only attempted one brood
4--
3 0 Pairs that attempted two broods
3.5 --


o

I 2.5

2 35
43 5
8
S1.5




0.5

0
1994 1995 1996

Year
Figure 3-4. Mean number of young fledged per pair for pairs of
crested caracaras nesting in south-central Florida, 1994 -1996.
Number of pairs is shown above each column.















Bis"",









1.0 -- Fledglings from first broods, n = 14 broods
Si| Fledglings from second broods, n = 6 broods
0.9 Fledglings from single broods, n = 32 broods
--

0.8 -

0.7


. 0.3 -


0.2 -
0.3
0.1 ----m- ; ---u-ii-l---u--Im
0.0
0 200 400 600 800 1000 1200

Days since fledging


Figure 3-5. Kaplan-Meier survival functions estimated for juvenile crested caracaras
from first, second, and single broods in south-central Florida, 1994 1996.













CHAPTER 4

HABITAT ASSOCIATIONS AND PRODUCTIVITY OF CARACARAS ON CATTLE
RANCHES

Introduction
Wildlife managers face the increasingly difficult task of providing suitable areas
and habitats for wildlife in landscapes that are continually changing due to a complex
array of human-induced processes. Management of wide-ranging species can be
particularly challenging because individuals' activity areas may cross socio-political or
landownership boundaries.
The crested caracara is a gregarious raptor with a limited distribution in North
America. In Florida, a small population occurs in the south-central peninsula, a region
historically dominated by native prairie. Much of this ecosystem, in which the caracara
historically occurred (Phelps 1912, Nicholson 1929, 1930, Howell 1932), has been
replaced by large acreages of privately-owned cattle ranches. Although these ranches
consist of open grasslands with scattered trees, vegetation cover types and land
management practices differ considerably from those occurring on native prairies. Most
ranches include large areas of improved pasture, grasslands dominated by non-native
vegetation species that are intensively managed to produce forage for cattle. The natural
hydrologic regime in improved pastures also has been altered by extensive ditching and
draining of wetlands.
The current landscape in south-central Florida reflects complex patterns of land
ownership and land uses. Within recent decades, urban and agricultural development has
escalated in this region. On private lands, remnant prairie and, increasingly, pasture
habitats are being converted to citrus and sugar cane. The crested caracara is likely to be
affected by these land use changes. This raptor is now often found on the cattle ranches
(Layne 1978, Millsap 1989), but the degree of dependence on ranch habitats is not well

64









65
understood. Because the spatial and habitat requirements of the crested caracara and land
management practices compatible with its survival remain poorly known, consequences
of land use changes on this Threatened population are difficult to predict.
Understanding relationships between caracaras and land uses is critical to
conservation planning for Florida's population. The objective of this study was to
examine these relationships by considering habitat and spatial use of breeding adults and
productivity of pairs nesting on the cattle ranches. Specifically, I identify patterns of land
ownership and land use associated with active nest sites, examine home ranges of
breeding adults and spatial relationships among neighboring pairs, evaluate landscape-
level habitat selection by breeding pairs, and examine productivity by pairs in relation to
habitat variables characteristic of cattle ranches.

Study Area
The study area encompassed approximately 21,000 km2 and included all or parts
of Highlands, Glades, Okeechobee, DeSoto, Osceola, Polk, Hendry, and Indian River
Counties (Figure 4-1). The landscape is a mosaic of habitats, primarily grasslands and
pastures interspersed with cropland, citrus groves, dairies, pine flatwoods, freshwater
marshes and lakes, oak scrub, hardwood hammocks, and urban developments. Over half
of this area is in private ownership. Principal land uses are cattle ranching, citrus
production, and other agricultural crops. The land in public ownership is managed for a
variety of uses including recreation, military training, hunting, natural areas, and river
floodplain management.

Methods

Land Ownership/Land Use
I located active caracara nests by visiting known historic territories (B. Millsap
pers. comm., J. Layne pers. comm.) and areas where caracaras had been observed within
the past 5 years, and by searching other areas of potentially suitable habitat on both public
and private lands. The searching methodology involved conducting observations near
potential nest sites. I also covered entire areas on foot to look into individual cabbage









66
palm (Sabal palmetto) and live oak (Quercus virginiana) trees. When caracaras were
observed on private land, the landowner was contacted for access permission. For each
territory located, I classified landownership as private or public and primary land use in
the vicinity of the nest tree as ranching, other (including urban or other agriculture), or
natural.
To evaluate whether occupancy and breeding rates in caracara territories were
related to patterns of land use and land ownership, I used a set of 37 territories that were
monitored during all 3 years. For each year, I recorded whether or not each territory was
occupied and if so, whether or not the pair made a breeding attempt. Using a Z-test with
Yates' correction (Sokal and Rohlf 1995), I tested the null hypothesis of no differences in
the proportions of territories that were occupied or where the pair made a breeding
attempt during all 3 years, between territories on privately owned (n = 30) and publicly-
owned lands (n = 7).

Home Range Estimation and Spatial Dispersion of Territories
I trapped and radiotagged 39 (16 females, 23 males) breeding adult caracaras
during the nesting seasons in 1994 through 1996 (Morrison and McGehee 1996). Each
captured adult was banded with a standard U.S. Fish and Wildlife Service aluminum leg
band and a colored aluminum band with a unique alphanumeric character that permitted
identification of individuals. Radiotransmitters (Holohil, Ltd.) were attached using a
standard backpack harness constructed of Teflon ribbon and secured at the bird's chest.
Because caracaras are monomorphic, gender of captured adults was determined using
DNA from blood samples collected in the field (Morrison and Maltbie in review).
I monitored radiotagged adult caracaras throughout the year. I located all
radiotagged adults (Mech 1983) and obtained GPS coordinates of their locations during
weekly aerial surveys from a fixed-wing Cessna 172 aircraft. I ascertained additional
locations for adult caracaras by triangulation (White and Garrott 1990). Determining
ground locations was facilitated by access to ranch interiors granted by local ranchers.
Most locations were taken from a distance of< 500m. I also recorded visual observations









67
of radiotagged adults during scheduled visits to territories. Triangulation and visual
locations used for home range estimation comprised approximately 18% and 15%,
respectively, of all locations for adults.
The 95% confidence circle around aerial location estimates (Marzluff et al. 1994)
was approximately 42 ha. Error polygons around triangulated locations (White and
Garrott 1990, Kennedy et al. 1994) were all < 50 ha; most were < 20 ha. Because all
locations used for home range estimation were taken at least 24 hours apart, I did not
conduct tests for independence of observations (Swihart and Slade 1985).
I delineated home ranges using kernel methods (Worton 1989, 1995, Seamen and
Powell 1996) and the software RANGESV (Kenward and Hodder 1995). The boundary
of each caracara's home range was defined as the 99% contour estimated with the fixed
kernel method and a 0.85 smoothing factor (Kenward and Hodder 1995, Seamen and
Powell 1996). Core areas (Wray et al. 1992) were identified as the 60% kernel contour
using cluster analysis in RANGESV.
I compared home range estimates between males and females and between
breeding and non-breeding seasons using a two-tailed Mann-Whitney test (Sokal and
Rolf 1995). The breeding season was defined as the period from the beginning of
incubation to the end of the post-fledging dependency period, 8 weeks after the young
fledged from the nest. After this time, young were considered independent, and parents
and young were together less often as a family unit.
To evaluate nesting density and spatial relationships among territories, I
calculated distances between adjacent caracara nests. Using RANGESV, I evaluated
overlap of 99% contours and core areas among 9 neighboring, successfully breeding pairs
of caracaras inhabiting the MacArthur Agro-Ecology Research Center (MAERC) and an
adjoining ranch, and among 4 neighboring, successfully breeding pairs at a site along the
Kissimmee River.









68

Landscape-level Habitat Selection by Breeding Pairs
Habitat data
Data on vegetation cover types throughout the study area were obtained from the
Florida GAP Analysis project at the University of Florida. Analyses were performed in
ArcINFO GRID (Environmental Sciences Research Institute 1996) The study area map,
initially classified from 1994 LANDSAT satellite imagery, was imported into the GIS in
grid-cell (raster) format. The map was 1909 X 1571 cells at a grid cell resolution of 90m.
I recognized 13 natural and semi-natural vegetation cover types (Table 4-1). Other cover
types reflected land use conversion by humans such as citrus, croplands, sod fields,
quarries, and urban areas.

Model development
I used resource selection techniques (Thomas and Taylor 1990, Manly et al. 1993)
to identify habitats and landscape features that distinguished between known caracara
home ranges and potentially "available" home ranges throughout the study area. These
"available" units were 1296 ha in size, comparable to the average caracara's home range,

and were sampled across the entire study area using a sliding window technique and the
GIS. Because of the finite 90m grid cell size, square units 40 cells (3600m) on a side
were used.
Resource categories used for model development consisted of variables describing
proportions of the 13 natural and semi-natural vegetation cover types (Table 4-1). Cover
types that reflected habitat conversion, such as citrus groves, row crops, and urban
development were not included, although some caracara home ranges contained them. I
assumed that conversion occurred after a caracara pair had selected the territory, so these
cover types were associated by default with cover types actually selected by caracaras.
Other variables included (1) total proportion vegetated, (2) total proportion of open
habitat, (3) total proportion of shrubs, (4) total amount of improved pasture habitat, and
(5) largest contiguous patch of improved pasture habitat (Table 4-1).









69
Using the GIS, I determined that caracara home ranges were at least 53%
vegetated with natural or semi-natural vegetation. For model development, I only used
available units that contained more than 53% natural or semi-natural vegetation (n = 925)
and the 25 known caracara home ranges. I used the GIS to calculate values for all
variables for all home ranges and available units.
I used logistic regression procedures in SAS (SAS Institute 1995) to calculate
coefficients and standard errors for all variables (Manly et al. 1993). These procedures fit
linear logistic regression models for binary data by the method of maximum likelihood
(Menard 1995). The response probability produced by the models was the probability of
a sample unit being a territory, or "used." I assumed that variables that distinguished
between the home ranges and the available units indicated selection by caracaras.
Significance of variables was assumed at the a = 0.10 level.
Because use of multiple variables results in many alternative models, selection of
the "best" or "correct" model is an important step in data analysis (Burnham and
Anderson 1992). To choose among the models produced in SAS, I used AIC (Akaike
1985, Burnham and Anderson 1992) and Goodman's criterion (Goodman 1984), both of
which consider the principle of parsimony in model selection. To evaluate model fit, I
conducted Hosmer and Lemeshow's goodness of fit test (Hosmer and Lemeshow 1989),
which tests the null hypothesis that there are no differences between the observed and
expected number of observations in each of several groups based on the percentiles of the
estimated probabilities.

Model validation
A model used to make inferences about important habitats for a species is only as
good as its ability to make correct predictions about new observations. I tested the ability
of the "best" logistic regression model to predict use of 29 other known caracara home
ranges. Because I did not have telemetry data for these test ranges, they were delineated
using the GIS. For each test range, I created an area 1296 ha in size (square units 3600m









70
on a side) centered on the nest. I then applied the model to each test range to calculate
probability of use.

Caracaras and Cattle Ranches
To evaluate the caracara's use of cattle ranches, I investigated relationships
between home range size, productivity, and the proportion of improved pasture in each
home range. I defined productivity as the total number of young produced during all 3
years of the study by each pair. For this analysis, I only used home ranges for which I
had these data (n = 18). Statistical tests were conducted using SigmaStat (Jandel
Scientific 1995).

Results

Land Ownership/Land Use
Land ownership and land use patterns were determined for 65 active caracara
territories. Most territories found were on privately-owned land (Table 4-2). The major
land use in the vicinity of caracara nests at all sites was cattle ranching (Table 4-2).
Lands under public ownership where I found caracara territories were primarily managed
for native habitats, including prairie, although moderate grazing occurred in some areas.
The proportions of the 37 preselected territories that were occupied and where adults
attempted breeding during all 3 years were higher (z = 2.64, P = 0.008, 1-P = 0.71 and
z = 2.02, P = 0.04, 1-p = 0.52, respectively) for territories on privately-owned lands
(Figure 4-2).

Home Ranges and Spatial Dispersion of Territories
Breeding adult caracaras remained on territory throughout the year. I obtained
usable data for home range estimation from 25 individuals (16 males and 9 females).
Home range sizes varied from approximately 551 to 2778 ha (Table 4-3). Most home
ranges were more than 80% vegetated with natural or semi-natural vegetation. All ranges
contained improved pasture, ranging from approximately 11% to 94% of the naturally
vegetated area (Table 4-3).









71
Approximately 90% of locations for all adults were within 2.5 km of the nest
(Figure 4-3). Home range sizes did not differ between males and females (t = 1.13,
P = 0.27, df= 23), although power of this test was low (1-P = 0.10). Adult caracaras
ranged farther from the nest, on average, during the non-breeding season (A12 = 56.17,
P < 0.001, df= 7, Figure 4-4). For 11 adult caracaras (7M, 4F) having approximately
equal numbers of radiolocations for each season, breeding season home ranges were
smaller than non-breeding season home ranges (paired M-tests, t = -3.45, P = 0.006,
df= 10, 1-1 = 0.86).
Internest distances for neighboring successfully breeding pairs averaged 3.08 km
0.18 SE (range 1.01 5.08 km, n = 30 pairs, Figure 4-5). Core areas rarely overlapped
(Table 4-4), although 99% contours overlapped on average 14.7 % 2.82 SE at MAERC
and 10.08% + 5.39 SE at the Kissimmee River site (Table 4-5). At MAERC, 6 caracara
home ranges overlapped home ranges of at least 2 other pairs of caracaras.

Landscape-level Habitat Selection by Breeding Pairs
The model that best distinguished between known caracara home ranges and
available units throughout the study area indicated a negative association of caracara
home ranges with the variables pine forest, mixed forest, and woodland, and a positive
association with the variables improved pasture, and contiguous pasture. All variables
were significant at the 0.10 level (Table 4-6). Hosmer and Lemeshow's test indicated
good fit of this model (G = 10.16, P = 0.25).
Since caracaras are known to be associated with open rather than forested habitats,
I reran logistic regression analysis with the forest and woodland variables omitted.
Results of this analysis indicated a positive association between caracara home ranges
and improved pasture, native grassland/prairie, and wax myrtle, a shrub commonly
associated with grassland habitats. Again, all variables were significant at the 0.10 level
(Table 4-7). Hosmer and Lemeshow's goodness of fit test indicated a poorer fit of this
model, however (G = 22.86, P = 0.003).









72
Although these 2 models were biologically meaningful and statistically fit the
data, their predictive capability was poor. For the model that included the forest and
woodland variables, the mean predicted probability of use for the 29 test ranges was 0.10
(range = 0.00 to 0.31, Table 4-8). The model that did not contain the forest variables was
only slightly better at predicting probability of use for the test ranges (mean = 0.13, range
0.00 to 0.42, Table 4-8).

Caracaras and Cattle Ranches
I found a strong inverse relationship between the proportion of improved pasture
and home range size (r= -0.57, P = 0.003, df = 23, n = 25, Figure 4-6). For pairs for
which I had telemetry data and reproductive information for all 3 years (n = 18), the total
number of young fledged during all 3 years was positively but not significantly correlated
with the proportion of improved pasture in the home range (r= 0.20, P = 0.42, df = 16,
Figure 4-7). Home ranges on the study area landscape are shown in Figure 4-8.

Discussion
In south-central Florida, breeding pairs of crested caracaras are distributed across
the landscape in well-defined, year-round territories. Caracara home ranges are strongly
associated with improved pasture habitats on cattle ranches. Little overlap of core areas
among neighboring pairs suggests that breeding pairs are highly territorial near the nest.
Food resources may be shared in areas distant from the nest, however, as indicated by
overlapping 99% contours among neighboring pairs. This territorial dispersion pattern is
typical of raptors. Many species are highly territorial around the nest site. In habitats
containing adequate food resources, nests may occur close together and feeding areas
may overlap (Schoener 1968, Newton 1979, Village 1982).
Estimates of home range size varied considerably among breeding individuals.
Despite this variation, results indicate substantial differences in overall estimates of home
range size and nesting density from those reported by Layne (1996). Layne (1996)
estimated home ranges of 1800-7400 ha based on maximum observed distances of 4.8 km
to 9.7 km between sightings of unmarked individuals presumed to be members of the









73
same pair. He also recorded maximum local density of 1 pair/1400 ha, but suggested that
average density was much less. In this study, home ranges varied from approximately
500 to 2800 ha. Local density was 9 successfully breeding pairs within about 5000 ha, or
approximately 1 pair/555 ha on MAERC and an adjacent ranch in Highlands County.
This density is more than twice the maximum local density estimated by Layne (1996).
The availability of food and nest sites generally limits breeding density (Village 1982,
1983). Typically, large raptors breed at lower densities than small raptors, which utilize a
wider variety of prey species (Newton 1979). High local density of crested caracaras on
MAERC may reflect habitat quality (availability of both food and nest sites) in that area.
Although the density of caracaras at MAERC may not occur throughout the population's
entire Florida range, results of this study suggest that at least in some areas, nesting
densities may be substantially higher than previously thought for this population.
A positive association between caracaras and improved pastures suggests that
caracaras may benefit from land management activities routinely conducted on cattle
ranches such as grazing, burning, mowing, and disking. Territory occupancy,
reproductive activity, and nesting success were more consistent at territories on ranches.
Pairs nesting on cattle ranches apparently produce more fledglings than pairs nesting on
lands managed as natural areas. In general, however, comparisons between territories
associated with different land uses were difficult because I had trouble locating caracaras
on lands other than cattle ranches. Despite regular searching on public lands or other
lands managed as natural areas, few territories on entirely native habitat were found.
Movement patterns of crested caracaras reflects breeding commitments and
temporal and spatial variation in food availability, particularly the amount and
configuration of suitable foraging habitat within each home range. The availability of
nest sites also influences the size and shape of the home range. In this study, few
caracara ranges were circular or centered on the nest. On most cattle ranches, suitable
nest trees typically are few in number and occur in clumps.. Home range size and shape
primarily reflected the relationship between the nest site and improved pasture habitat
(see map).









74
Caracaras exhibit strong fidelity to a nest site and year-round home range.
Management recommendations for raptors typically focus on protecting the nest site and
a circular buffer around the nest in which development or management activities are
limited (U.S. Fish and Wildlife Service 1987b). Suitable caracara nest trees are often
clumped and associated with hammocks that do not constitute suitable foraging habitat.
A circular buffer around a nest may not encompass areas required for all activities
associated with reproduction and year-round survival of a breeding pair of caracaras.
Management plans for this species must consider the arrangement as well as the amount
of both nest sites and suitable foraging habitat.
A landscape approach to identifying important habitats and assessing a species'
distribution is particularly appropriate for raptors (Austin et al. 1996, Johansson et al
1994, Carey et al. 1992). Many species require large areas of land to support viable
populations and are particularly sensitive to changes in land use (Marquiss et al. 1985,
Carey et al. 1992). Conservation of such species requires consideration of land uses and
management practices throughout entire regions. Landscape-level models that evaluate
resource selection by animals can provide information about important habitats and about
land uses compatible with species' survival. These models ultimately can be used to
evaluate a species' distribution, delineate key protection areas, predict effects of future
land use changes, and identify potential areas for habitat improvement or restoration.
The models developed here represent an important first step in identifying habitat
and landscape features important to crested caracaras in Florida. These models
corroborate the observed association between breeding crested caracaras and improved
pasture on cattle ranches. These models must be used with caution, however. While they
showed a good fit statistically, their biological usefulness for management is currently
limited, as indicated by the poor ability of the statistically "best" models to predict use of
the 29 test ranges.
Future model development would benefit from consideration of two current
constraints. First, agricultural and grassland habitats were poorly classified on the study
area map obtained from the Florida GAP Analysis Project. The resolution of the









75
available GIS data did not permit detailed differentiation among these cover types or
delineation of habitats within them such as patches of hammocks, wetlands, and ditches,
which might be important to caracaras in selection of a territory. Second, these models
did not consider context of habitats either within or around caracara home ranges.
Habitat quality is determined not only by content but by the context of suitable habitats
(Kushlan 1981, Janzen 1986, Fahrig and Merriam 1994, Harris et al 1996). Further
research directed towards understanding relationships between caracaras and habitats
should focus on refining cover types that represent open habitats and on considering
habitat context both within the home range and at a larger scale.









76
Table 4-1. Variables included in development of the logistic regression model
describing caracara home ranges. Unless noted, all variables are expressed as
proportions of the total area of the home range that was vegetated with natural or semi-
natural* vegetation.


Variable Symbol
oak forest OAKFOR
pine forest PINEFOR
cypress CYPRESS
mixed forest MIXFOR
saw palmetto SAWP
oak shrub OAKSHRB
hypericum shrub HYPSHRB
woodland WOODL
willow WILL
wax myrtle WAXMYRT
seasonally flooded native grassland or prairie GRASSL
marsh MARSH
improved pasture* IMPPAST
proportion vegetated with natural or semi-natural vegetation PROPVEG
proportion of open habitat (pasture + grassland) OPEN
proportion of vegetated area in shrubs ALLSHRB
total area of improved pasture* AMTP
area of largest patch of contiguous improved pasture* CONTIGP









77

Table 4-2. Landownership and major land use for crested caracara territories in south-
central Florida, 1994 1996. n = 65 territories.

Major land use
Landownership Ranching Other Natural Total
Private 56(98.0%) 1 (2%) 0 57(87.7%)
Public 3(37.5%) 2 (2.5) 3(37.5%) 8(12.3%)
Total 59(90.8%) 3 (4.6%) 3 (4.6%) 65









78

Table 4-3. Size and vegetation composition of crested caracara home ranges. Home
ranges are defined as the 99% kernel contour estimated using a 0.85 smoothing factor.
N = number of radiolocations used to estimate home range size.

% of total home proportion of largest
Home range in natural vegetated area area of
Bird range or semi-natural in improved contiguous
ID Sex N size (ha) vegetation pasture pasture (ha)
120 F 48 1265.273 0.858 0.597 492.342
121 M 26 2190.443 0.933 0.493 373.879
200 M 28 1251.969 0.973 0.406 350.991
280 M 68 551.039 0.531 0.831 227.695
360 F 68 1363.798 0.979 0.732 653.605
458 M 96 1879.477 0.836 0.663 722.800
479 M 43 2777.219 0.976 0.109 53.460
498 M 91 848.295 0.905 0.535 188.930
518 F 27 571.288 0.899 0.849 410.240
538 F 27 1014.189 0.917 0.287 125.636
609 F 67 994.957 0.843 0.741 596.455
689 F 37 736.403 0.986 0.759 526.300
969 M 72 1609.962 0.960 0.690 919.980
1110 M 81 1176.541 0.958 0.581 498.884
1379 F 25 1301.698 0.900 0.571 323.155
1469 M 71 1454.535 0.990 0.642 578.182
1490 M 42 659.221 0.989 0.765 400.091
1558 M 39 1129.249 0.937 0.632 522.410
1568 M 30 1083.789 0.942 0.522 213.750
1589 M 97 1604.085 0.924 0.577 323.356
1600 F 25 1685.152 0.865 0.717 803.179
1608 M 45 681.562 0.790 0.707 325.852
1711 M 38 859.928 0.720 0.691 279.356
1969 F 77 1072.958 0.946 0.937 788.290
1989 M 33 1537.786 0.821 0.709 817.595









79
Table 4-4a. Matrix of percent overlap of 60% contours (core areas) for 9 adjacent pairs
of breeding crested caracaras on MacArthur Agro-Ecology Research Center, Highlands
County, south-central Florida. Birds in rows are overlapped by birds in columns.

Bird ID 360 689 969 1469 1490 1558 1600 1608 1711
360 0 2.6 0 0 0 0 0 0
689 0 0 0 0 4.4 0 0 0
969 2.9 0 18.8 0 0 0 0 0
1469 0 0 21.9 0 0 0 0 0
1490 0 0 0 0 2.8 0 0 0
1558 0 3.7 0 0 1.5 0 0 0
1600 0 0 0 0 0 0 0 0
1608 0 0 0 0 0 0 0 0
1711 0 0 0 0 0 0 0 0 -


Table 4-4b. Matrix of percent overlap of 60% contours (core areas) for 4 adjacent pairs
of breeding crested caracaras along the Kissimmee River, Highlands and Okeechobee
counties, south-central Florida. Birds in rows are overlapped by birds in columns.

Bird 498 120 1379 1568
498 0 0 0
120 0 0 0
1379 0 0 8.5
1568 0 0 7.5 -









80
Table 4-5a. Matrix of percent overlap of 99% contours for 9 adjacent pairs of breeding
crested caracaras on MacArthur Agro-Ecology Research Center, Highlands County,
south-central Florida. Birds in rows are overlapped by birds in columns.

Bird 360 689 969 1469 1490 1558 1600 1608 1711
360 0 33.1 12.4 6.2 0 9.7 0 0
689 0 0 0 25.9 21.4 3.1 1.9 14.6
969 27.2 0 38.8 7.6 4.1 0 0 0
1469 11.0 0 41.7 0 0 0 0 0
1490 13.6 31.5 19.8 0 24.8 23.0 0 0
1558 0 14.1 5.9 0 13.4 0 0 0
1600 7.7 1.4 0 0 8.3 0 0 1.5
1608 0 2.4 0 0 0 0 0 32.9
1711 0 12.6 0 0 0 0 2.9 22.7 -


Table 4-5b. Matrix of percent overlap of 99% contours for 4 adjacent pairs of breeding
crested caracaras along the Kissimmee River, Highlands and Okeechobee counties, south-
central Florida. Birds in rows are overlapped by birds in columns.

Bird 498 120 1379 1568
498 3.9 11.2 5.7
120 2.7 0 0
1379 7.3 0 22.4
1568 4.3 0 26.0 -









81
Table 4-6. Parameter estimates for model produced using logistic regression analysis that
best distinguished between known caracara home ranges and available units throughout
the study area. Variables include pine forest (PINEFOR), mixed forest (MIXFOR), wax
myrtle (WAXMYRT), woodland (WOODL), and contiguous pasture (CONTIGP).

Parameter DF Coefficient Std. Err. P
Constant 1 -3.04 0.69 0.000
PINEFOR 1 -0.06 0.04 0.087
MIXFOR 1 -0.22 0.12 0.062
WAXMYRT 1 0.10 0.05 0.053
WOODL 1 -0.17 0.07 0.012
CONTIGP 1 0.003 0.001 0.003









82

Table 4-7. Parameter estimates for model that best distinguished between known
caracara home ranges and available units throughout the study area and that did not
include forest variables. Variables include native grassland/prairie (GRASSL), wax
myrtle (WAXMYRT), improved pasture (IMPPAST), and contiguous pasture
(CONTIGP).

Parameter DF Coefficient Std. Err. P
Constant 1 -9.18 1.31 0.000
GRASSL 1 0.04 0.02 0.069
WAXMYRT 1 0.19 0.06 0.000
IMPPAST 1 0.08 0.02 0.000
CONTIGP 1 0.002 0.001 0.090









83
Table 4-8. Predicted probabilities of use for test caracara home ranges calculated using
logistic regression models that best distinguished between caracara home ranges and
available units throughout the study area.

Model that included forest Model with forest and
Nest No. and woodland variables woodland variables omitted
1 0.000 0.019
2 0.021 0.010
3 0.047 0.052
4 0.256 0.349
5 0.211 0.391
6 0.189 0.172
7 0.096 0.061
8 0.270 0.356
9 0.132 0.351
10 0.015 0.019
11 0.049 0.067
12 0.006 0.001
13 0.151 0.116
14 0.161 0.142
15 0.017 0.013
16 0.042 0.015
17 0.194 0.292
18 0.314 0.418
19 0.056 0.128
20 0.113 0.125
21 0.009 0.033
22 0.000 0.001
23 0.011 0.001
24 0.175 0.154
25 0.122 0.130
26 0.068 0.123
27 0.021 0.021
28 0.045 0.087
29 0.036 0.243

Mean 0.098 0.134









84































^-'"
















Figure 4-1. Study area for evaluating landscape-level habitat selection by breeding
pairs of crested caracaras in Florida, 1994 -1996.













100
90 0 Private ownership, n = 30
U Public ownership, n = 7
80
S70
60

8 50
40
30 30 00



10
0

Occupied Initiated
breeding
attempt

Figure 4-2. Relationship between occupancy, initiation of breeding attempts, and land ownership for 37
crested caracara territories in south-central Florida, 1994 1996.




Full Text

PAGE 1

REPRODUCTIVE ECOLOGY AND HABITAT ASSOCIATIONS OF FLORIDA'S CRESTED CARACARAS By JOAN L. MORRISON 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 1997

PAGE 2

Copyright 1997 by Joan L. Morrison

PAGE 3

To Clyde and all the rest.

PAGE 4

ACKNOWLEDGMENTS Many people contributed to this research and I am grateful to all of them. I particularly wish to thank Paul Ebersbach and Bob Progulske at the Avon Park Air Force Range (APAFR, DOI), Brian Millsap, Don Wood, and Dave Cook of the Florida Game and Fresh Water Fish Commission's (FGFWFC) Non-Game Wildlife Program, and Dave Wesley (U.S. Fish and Wildlife Service) for recognizing the need for research on this little-known species and subsequently providing funding. This project would have never gotten off the ground without their support and assistance. I thank the staff at APAFR for their help, particularly Pat Walsh, Sam van Hook, Scott Penfield, Steve Hanley, Julie Olson, and Harlan Blackburn. Nancy Douglass, Victor Echavez, and Hugh Clark of the FGFWFC also provided field and logistic support, particularly assistance with accessing state-owned properties. I am grateful for additional financial support received from the South Florida Water Management District, the University of Florida's Graduate Minority Fellowship Office, and the Ford Foundation Fellowship Program. I gratefully acknowledge the cooperation and support provided by the personnel and staff at the MacArthur Agro-Ecology Research Center (MAERC) and Archbold Biological Station, particularly Drs. John Fitzpatrick, Hilary Swain, and Tom Bancroft, and Fred Lohrer. I particularly appreciated the opportunity to live at and to operate this project based out of MAERC. Ranch managers Dan Childs and Gene Lollis provided many hours of interesting conversations regarding cattle ranching in Florida. I thank them for supporting this project throughout and for sticking up for us with the Florida Cattleman's Association. I also thank the rest of the staff at MAERC for their help and for putting up with our activities at the ranch. I also wish to acknowledge the private donation by Frank "Sonny" Williamson to the IFAS research effort at MAERC. iv

PAGE 5

Special thanks go to Steve McGehee who was a stalwart field assistant throughout. His enthusiasm for caracaras and untiring field assistance, particularly in finding nests and radiotracking, was invaluable. Vicky Dreitz and Danielle Todd also proved to be excellent field assistants, during the second and third field seasons respectively, and capably handled all aspects of the research program as well. I am grateful to Brian Mealey for his support throughout and also for loan of Clyde, our research caracara. Additional people who helped with field work at various times included John Arnett, Tylan Dean, Sam Gibbs, Laurie Gray, Kevin Long, Mike McMillian, Carlos Pages, Megan Parker, Ray Poulin, Lourdas Rojas, Jim Sanderson, Jim Solomon, and Reuven Yosef. I thank Scott Hedges, Stefanie Melvin, Tom Weber, Cathy Ritchie, and Shannon Stull for their help at various stages of the project. Thanks also to Lorenz Animal Foods and Brad Stith for help with the dead cows. I am grateful to the staff at the Department of Wildlife Ecology and Conservation, Institute of Food and Agricultural Sciences, University of Florida, whose efforts throughout my tenure at UF kept our paperwork processed, salaries on track, and bills paid. Sincere thanks go to Peg Margosian, Brian Sheahan, Roberta Pickett, Don Browning, and Helen Huseman who contributed many hours of assistance with the GIS analysis and mapping, and to Ken Portier and Jay Harrison for help with statistics. Linda Kubitz, Cathy Henderson, Milt Putnam, and Sylvia Beauchamp from IFAS Media Services provided excellent media coverage of the project. Thanks go to Neil Rettig and Kim Hayes for providing additional film of our activities. I thank Mary Maltbie for sexing the caracaras using the blood samples, Drs. Don Forrester and Marilyn Spalding, and Garry Foster for necropsies and processing parasite data, and Dr. Reed Bowman for providing the Mayfield nest success program. I also thank personnel at the Maitland Bird of Prey Center, the Falcon Batchelor Bird of Prey Center at the Miami Museum of Science, and Wrede's Wildlife Rehabilitation Center for assisting us when we had caracara emergencies. Sincere thanks go to Mark Brady and personnel at Glades Electric Cooperative for helping us to access the nest trees, to Jim Wyatt, Karen Dunne, Reed

PAGE 6

Robbins, John McGee, and Morton Sund of Wyatt Aviation who kept us flying safely during all the long hours of aerial radiotracking, and to the personnel at Wrenchmasters, who kept our vehicles running. I especially thank my graduate advisor, Dr. Steve Humphrey, who stepped in at the beginning when everything looked tenuous. His steady and constant support throughout helped keep up the momentum. I thank my other committee members, Drs. Lyn Branch, John Fitzpatrick, John Smallwood, Mel Sunquist, and S. David Webb, for their contributions, especially for reviews of this dissertation. I also thank Dr. Mike Collopy for sparking my original interest in caracaras and for getting me to UF in the first place. I particularly want to acknowledge the friendship of Kim Babbitt, Patty Cramer, and Pat Kennedy, who were there from the beginning and got me through the rough times. Thanks also go to other friends and colleagues Eduardo Alvarez, Leslie Backus, Tylan Dean, Nathalie Hamel, Dave Leonard, Ken Meyer, Steve Morello, Katie Sieving, and Dale Stahlecker for their help and input at various times. I thank my family for their continued support, particularly my parents, whose belief and encouragement never waned. I especially thank Jim Sanderson for his love and support throughout and for his untiring belief that this was all possible. Most importantly, I thank landowners and cattle ranchers throughout south-central Florida, who so generously provided access to their lands and with whom we had so many interesting discussions regarding caracaras and cattle ranches. Without their cooperation and support, this study could not have been accomplished at all. This dissertation is contribution No. 33 from the MacArthur Agro-Ecology Research Center of Archbold Biological Station. vi

PAGE 7

TABLE OF CONTENTS page ACKNOWLEDGMENTS iv LIST OF TABLES ix LIST OF FIGURES xii ABSTRACT xiv CHAPTERS 1 STUDY OVERVIEW 1 Introduction 1 Study Species — History and Background 2 Study Area 5 Study Organization 5 2 BREEDING ECOLOGY OF FLORIDA'S CRESTED CARACARAS 9 Introduction 9 Study Area and Methods 10 Nest Location and Visitation 10 Reproductive Success 11 Productivity and Components of Variation 12 Environmental Data 13 Statistical Analyses 13 Results 14 Territorial Occupancy and Breeding Rates 14 Breeding Behavior and Site Fidelity 14 Nest Building and Incubation 15 Length of the Breeding Season and Timing of Broods 15 Number of Broods Per Year 16 Clutch Size 16 Broods 16 Reproductive Success, Productivity and Components of Variation 17 Performance and Laying Date 17 Nest Failure, Egg, and Chick Mortality 18 Discussion 18 Weather, Timing, and Length of the Breeding Season 19 Clutch Size 20 vii

PAGE 8

Timing and Number of Broods 20 Mortality 21 Breeding Strategy and Variation in Reproductive Parameters 21 Comparison With Other crested caracaras 22 3 EFFECTS ON PRODUCTIVITY AND POTENTIAL COSTS OF DOUBLE BROODING IN CRESTED CARACARAS 41 Introduction 41 Study Area and Methods 43 Results 46 Extent of Occurrence of Double Brooding 46 Timing of Broods 47 Productivity and Nesting Success 47 Parental Care and Post-fledging Survival 48 Potential Costs to Adults 48 Effects on Population Growth and Persistence 48 Influence of Factors 48 Discussion 48 4 HABITAT ASSOCIATIONS AND PRODUCTIVITY OF CARACARAS ON CATTLE RANCHES 64 Introduction 64 Study Area 65 Methods 65 Land Ownership/Land Use 65 Home Range Estimation and Spatial Dispersion of Territories 66 Landscape-level Habitat Selection by Breeding Pairs 68 Caracaras and Cattle Ranches 70 Results 70 Land Ownership/Land Use 70 Home Ranges and Spatial Dispersion of Territories 70 Landscape-level Habitat Selection by Breeding Pairs 71 Caracaras and Cattle Ranches 72 Discussion 72 5 SYNTHESIS AND CONCLUSIONS 92 LITERATURE CITED 96 BIOGRAPHICAL SKETCH 108 viii

PAGE 9

LIST OF TABLES Table page 2-1 Breeding chronology and intervals, in weeks, between broods for Florida's crested caracara population, 1994 1996 24 2-2 Multiple nesting in Florida's crested caracara population, 1994 1996. Early nests are those that initiated egglaying before the median week of egglaying for the entire sample, each year 25 2-3 Reproductive success parameters for Florida's crested caracaras, 1994 1996. Nestling survival to fledging (number of fledglings/number of eggs laid), fledgling survival to independence (number of young alive at end of post-fledging dependency period/number fledged), and egg success (proportion of eggs laid that produced independent young) for eggs from known size clutches 26 2-4 Estimates of Mayfield nest success probabilities for breeding pairs of crested caracaras in south-central Florida, 1994 1996 27 2-5 Measures of annual productivity (mean + SE) and Index of Productivity for crested caracaras in south-central Florida, 1994 1996. See text for calculation of Index of Productivity. Productivity per successful pair includes young from renesting attempts 28 2-6 Variability in reproductive parameters for crested caracaras during 3 years in south-central Florida. CV = coefficient of variation (standard deviation/mean). Productivity/successful pair includes young from renesting efforts, n — 98 clutches 29 2-7 Seasonal decline in breeding performance (mean SE) for known-size clutches and for all nest attempts of crested caracaras in south-central Florida, 1994 1996. Early nests initiated egglaying before the median week of egglaying each year, late nests initiated egglaying after the median week 30 2-8 Fates of eggs from known-size clutches for crested caracara nests in southcentral Florida, 1 994 1 996 31 29 Failure of crested caracara nests in south-central Florida, 1994 1996 32 31 Occurrence and timing of nesting attempts between years for crested caracaras in south-central Florida. NA = not applicable for that year 53 3-2 Clutch size (mean SE) and nesting success (SD in parentheses) between nest types for crested caracaras in south-central Florida, 1994 1996. The same letter indicates significant differences between nest types. ND = no ix

PAGE 10

data on clutch sizes of first or second nests in 1 994 because these nests were found after hatching 54 3-3 Estimated parameters for evaluating contribution of double brooding to population growth for crested caracaras in south-central Florida. Parameters estimated from data collected during 1994 1996 55 3-4 Estimation of population size and annual growth rate for two hypothetical populations of crested caracaras in south-central Florida, one in which all pairs only produce a single brood each year, and one in which approximately 10% of pairs successfully produce two broods each year. ... 56 3-5 Statistics for logistic regression analysis of the influence of timing of initiation of the first brood (INITWK) and the number fledged in the first brood (NUMFRST) on the occurrence of double brooding in crested caracaras in south-central Florida, 1994 1996. n = 16 double-brooded pairs, 103 single-brooded pairs 57 36 Statistics for logistic regression analysis of the influence of double brooding the previous year (DBPREV), fledging date (WKPREV) and number fledged the previous year (NUMPREV) on the occurrence of double brooding in crested caracaras in south-central Florida, 1994 1996. n = 1 1 double-brooded pairs, 34 single-brooded pairs 58 41 Variables included in development of the logistic regression model describing caracara home ranges. Unless noted, all variables are expressed as proportions of the total area of the home range that was vegetated with natural or semi-natural* vegetation 76 4-2 Landownership and major land use for crested caracara territories in southcentral Florida, 1994 1996. n = 65 territories 77 4-3 Size and vegetation composition of crested caracara home ranges. Home ranges are defined as the 99% kernel contour estimated using a 0.85 smoothing factor. N = number of radiolocations used to estimate home range size 78 4-4a Matrix of percent overlap of 60% contours (core areas) for 9 adjacent pairs of breeding crested caracaras on the MacArthur Agro-Ecology Research Center, Highlands County, south-central Florida. Birds in rows are overlapped by birds in columns 79 4-4b Matrix of percent overlap of 60% contours (core areas) for 4 adjacent pairs of breeding crested caracaras along the Kissimmee River, Highlands and Okeechobee counties, south-central Florida. Birds in rows are overlapped by birds in columns 79 X

PAGE 11

4-5a Matrix of percent overlap of 99% contours for 9 adjacent pairs of breeding crested caracaras on the MacArthur Agro-Ecology Research Center, Highlands County, south-central Florida. Birds in rows are overlapped by birds in columns 80 45b Matrix of percent overlap of 99% contours for 4 adjacent pairs of breeding crested caracaras along the Kissimmee River, Highlands and Okeechobee counties, south-central Florida. Birds in rows are overlapped by birds in columns 80 4-6 Parameter estimates for model produced using logistic regression analysis that best distinguished between known caracara home ranges and available units throughout the study area. Variables include pine forest (PINEFOR), mixed forest (MIXFOR), wax myrtle (WAXMYRT), woodland (WOODL), and contiguous pasture (CONTIGP) 81 4-7 Parameter estimates for model that best distinguished between known caracara home ranges and available units throughout the study area and that did not include forest variables. Variables include native grassland/prairie (GRASSL), wax myrtle (WAXMYRT), improved pasture (IMPPAST), and contiguous pasture (CONTIGP) 82 4-8 Predicted probabilities of use for test caracara home ranges calculated using logistic regression models that best distinguished between caracara home ranges and available units throughout the study area 83 xi

PAGE 12

LIST OF FIGURES Figure page 1-1 Distribution of the crested caracara in North and Central America 7 12 Current known breeding range of the crested caracara in Florida. The study area encompassed much of this range 8 21 Location of study area for crested caracaras in south-central Florida. Study area encompassed much of the species' current known breeding range in Florida 33 2-2 Mean monthly rainfall and breeding chronology of crested caracaras during 3 years in south-central Florida, 1994 (n = 33 nest attempts), 1995 (n = 62 attempts), 1996 (n = 62 attempts) 34 2-3 Relationship between month of first nest initiation and occurrence of renesting by crested caracaras following either an early failure or a successful first brood, south-central Florida, 1994 1996 35 2-4 Distribution of clutch sizes by year, for crested caracara nests in southcentral Florida, 1994 1996 36 2-5 Mean number of fledglings and independent young per nest for each clutch size during each year for crested caracaras in south-central Florida, 1994 1996. n indicates number of clutches 37 2-6 Relationship between mean clutch size and month of nest initiation for 98 known-size clutches of crested caracaras in south-central Florida, 1994 1996. Number of clutches is shown above each month 38 2-7 Mayfield probability of nest success (+ SD) by month for crested caracaras in south-central Florida, 1994 1996 ( = 157 nest attempts) 39 28 Relationship between productivity per nest attempt (n= 157 attempts) and month of nest initiation for crested caracaras in Florida, 1994 1996. Number of nest attempts is shown above each month 40 31 Timing of nest initiation throughout the breeding season for crested caracaras in south-central Florida, 1994 1996 59 3-2 Length of the interbrood interval in relation to number fledged from the first nesting attempt for double brooded crested caracaras in south-central Florida, 1994 1996. n = 16 pairs 60 xii

PAGE 13

3-3 Proportion of nests initiated in each month that were successful, for crested caracaras in south central Florida, 1994 1996 61 3-4 Mean number of young fledged per pair for pairs of crested caracaras nesting in south-central Florida, 1994 1996. Number of pairs is shown above each column 62 35 Kaplan-Meier survival functions estimated for juvenile crested caracaras from first, second, and single broods in south-central Florida, 1994 1996. 63 41 Study area for evaluating landscape-level habitat selection by breeding pairs of crested caracaras in Florida, 1 9941 996 84 4-2 Relationship between occupancy, initiation of breeding attempts, and land ownership for 37 crested caracaras territories in south-central Florida, 1994 1996 85 4-3 Mean ( SE) number of radiolocations for 25 breeding adult crested caracaras in relation to distance from the nest, n = 1283 total radiolocations. 86 4-4 Distributions of radiolocations of 25 breeding adult crested caracaras in relation to distance from the nest during breeding and non-breeding seasons in south-central Florida, 1 994 1 996 87 4-5 Distribution of distances between nests of neighboring pairs of successfully breeding crested caracaras in south-central Florida, 1994 1996. n = 30 pairings 88 4-6 Relationship between proportion of improved pasture in home range and home range size for 25 crested caracara home ranges in south-central Florida, 19941996 89 4-7 Relationship between the amount of improved pasture in the home range and total number of young fledged during 1994 1996, for crested caracaras in south-central Florida 90 4-8 Home ranges of the crested caracara on study area in south-central Florida. 91 xiii

PAGE 14

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 REPRODUCTIVE ECOLOGY AND HABITAT ASSOCIATIONS OF FLORIDA'S CRESTED CARACARAS By Joan L. Morrison December, 1997 Chairman: Dr. Stephen R. Humphrey Major Department: Wildlife Ecology and Conservation I studied the reproductive ecology and habitat associations of the crested caracara (Caracara plancus) in Florida. Listed as Threatened, this population is believed to be in decline due to habitat loss. Results from observations conducted at 155 nests during 3 years indicated that Florida's caracaras may be doing better than previously believed, particularly on privately owned cattle ranches, where most nesting pairs were found. Pairs nesting on ranches exhibited higher site fidelity, annual breeding rates, and nesting success, and produced more young overall. The breeding cycle is strongly tied to the annual precipitation cycle. Although the modal clutch size was 2, clutch sizes of 3 produced the most independent young during all years. Pairs can make several nesting attempts per year. Individuals that successfully raise two broods in a year may realize higher lifetime reproductive success. Fledglings from early-season nests had the highest probability of survival to recruitment. Home range size and nesting density estimated using radiotelemetry were considerably smaller and higher, respectively, than previously published estimates for this population. xiv

PAGE 15

Caracaras are highly territorial near the nest but apparently tolerate conspecifics in foraging areas. I investigated landscape-level habitat selection by breeding pairs of crested caracaras using GIS analysis and logistic regression modeling. Results indicated a positive association between caracara home ranges and improved pastures, particularly with large areas of contiguous pasture. The size of individual caracara home ranges was inversely correlated with the proportion of improved pasture in the home range. Caracaras frequently occupy areas in close association with humans and are generally tolerant of moderate levels of human activities. Because of the caracara's strong site fidelity, a conservation strategy for Florida's population should be based on protecting known nest sites and should focus at a regional scale. Management plans must address effects of loss and degradation of nesting and foraging habitat, particularly conversion of pastures to citrus and sugar cane. Because most suitable habitat is in private ownership, habitat protection must involve cooperative efforts between land management agencies and private landowners. This research has focused attention on the importance of cattle ranches to Florida's native wildlife by using the caracara as an example of how humans, agriculture, and native wildlife can coexist. XV

PAGE 16

CHAPTER 1 STUDY OVERVIEW Introduction One of the most profound and widespread impacts of humans on the landscape is alteration of natural habitats. Fragmentation and loss of natural habitats have been implicated in the decline or extinction of many species (Askins et al 1 990, Wilcox and Murphy 1985). Consequences of these habitat changes may include reduced reproductive success or survivorship owing to loss of suitable breeding and foraging habitats, altered dispersal and other behavioral patterns, and increased predation pressure within remaining habitat patches (Whitcomb et al. 1981, Wilcove et al. 1984, Hagan and Johnston 1992, LaHaye et al. 1992, McKelvey et al. 1993, Herkert 1994, Lens and Dhondt 1994, Dunning et al. 1995, Redpath 1995, Robinson et al. 1995, Matthysen and Currie 1996). Formerly contiguous populations are fragmented into disjunct subgroups if individuals are unable to disperse among isolated patches that are embedded in an inhospitable matrix (Opdam et al. 1985, Sjogren 1991, Fahrig and Merriam 1994). The effects of fragmentation or loss of natural habitats on small, isolated, or peripheral populations are of particular concern. These populations have high conservation value because they often show some degree of genetic, behavioral, and/or morphological divergence from central populations (Levin 1970, Lawton 1993, Lesica and Allendorf 1995). Geographically isolated populations are subject to higher risk of extinction owing to demographic, environmental, and genetic stochasticity (Shaffer 1981). It is reasonable to expect that effects of habitat loss and fragmentation acting in concert with isolation and small size pose serious threats to peripheral populations. Fragmentation and loss of natural habitats are occurring at a rapid rate throughout peninsular Florida. Because of its geographic position and glacial history, Florida has a rich biodiversity (Neill 1957, Myers and Ewel 1990). The peninsula is a mosaic of 1

PAGE 17

2 habitat islands and small, often disjunct populations of species, many of which have differentiated at least at the subspecies level from their closest relatives in the western U.S. Continued fragmentation and conversion of natural habitats combined with geographic isolation currently jeopardizes chances for the long-term persistence of many species in Florida. Study Species History and Background The crested caracara {Caracara plancus) is a medium-sized tropical raptor that occupies open habitats, typically grassland, prairie, rangeland, savanna, and desert. Its current geographic distribution covers most of South and Central America, through parts of northern Mexico, and into the southern U.S. (del Hoyo et al. 1994). Its current North American distribution is limited to disjunct populations in Texas, Arizona, Cuba, and Florida (Figure 1-1). Fossil records and the distribution of these populations suggest that the caracara may have been more widespread throughout southern North America during the Pleistocene (Howard and Miller 1933, Brodkorb 1964, Chandler and Martin 1991, Vuilleumier 1985). This epoch was characterized by extensive periods of reduced rainfall. During these times, grassland and savanna habitats presumably reached their maximum extent across the continent. Especially during the glacial cycles when sea levels were lowest, an extensive subtropical savanna corridor may have extended around the Gulf of Mexico, facilitating the exchange and movement of western biota eastward into Florida (Neill 1957, Webb 1977, Webb and Wilkins 1984). Ancestors of the extant caracara may have expanded northward through Central America during the biotic interchange, followed by continued expansion throughout the Gulf region and into peninsular Florida. As a consequence of numerous glacial cycles, repeated fluctuations in sea level, and associated climate and habitat changes, the geographic ranges of many North and Central American species eventually became divided and reduced in size. Many species having western affinities, including the caracara, were ultimately restricted to small

PAGE 18

populations in Florida (Woolfenden and Fitzpatrick 1984). Today, conservation efforts throughout Florida are increasingly focusing on these populations because of their isolation, small size, and disjunct distributions. The extant Florida population of the crested caracara (C.p. audubonii) occurs in the south-central peninsula. Described by early ornithologists as inhabiting much of the central peninsula (Scott 1892, Phelps 1912, Howell 1932, Bent 1937, Nicholson 1929, 1930), its historic distribution probably conformed closely to the original location of the major prairie ecosystems (Davis 1967). Representatives of this ecosystem included the Kissimmee prairie, which extended north and westward from Lake Okeechobee in association with the Kissimmee River floodplain, and the DeSoto prairie, which extended westward from the southernmost portion of the central ridge. These prairies constituted vast expanses of low, open, nearly treeless grasslands, with sparse cabbage palm hammocks and scattered patches of shrubs and saw palmettos (Harshberger 1914). The current breeding range of the caracara in Florida encompasses parts of about 10 counties within this presumed historic range (Kale et al. 1992, Stevenson and Anderson 1994) (Figure 1-2). The grassland and prairie ecosystems in this region are now only limited in extent (Abrahamson and Hartnett 1990). Extensive areas of the original prairies have been converted to improved pastures, which are dominated by nonnative grasses and are intensely managed to produce forage for cattle. What remains as natural and semi-natural habitat in this region is rapidly being converted for agricultural and urban development. The status of Florida's crested caracara population has long been a subject of concern. Apparent steep declines in numbers during the late 1960s and early 1970s (Funderberg and Heinzman 1967, Heinzmann 1970) first focused attention on this population. Results of surveys of historic territories in the late 1980s suggested that this population may be undergoing a long-term decline concomitant with accelerated habitat loss (Millsap 1989). Layne (1995), however, stated that during the period 1972 to 1991,

PAGE 19

4 the adult population appears to have been stable. Current published estimates for this population are approximately 400-500 individuals statewide (Layne 1995, 1996). Because of its restricted range, apparent small population size and history of decline, Florida's crested caracara population was listed as federally Threatened in 1987 (U.S. Fish and Wildlife Service 1987a), followed by a similar state listing (Wood 1990). The current paucity of demographic and ecological information precludes accurate assessment of this population's status or trends, however. Habitat fragmentation and loss are assumed to threaten population persistence, but little data exist with which to test this hypothesis. In 1989, a federal recovery plan was accepted for Florida's caracara population (U.S. Fish and Wildlife Service 1989), but there was little biological information from which to develop appropriate recovery goals. Management of wildlife populations and development of suitable conservation strategies require in-depth knowledge of a species' life history and population dynamics. The caracara apparently now occurs primarily on privately-owned cattle ranches in this region, undoubtedly contributing to the paucity of biological information. Confrontations between land management agencies and private landowners have escalated in recent years, particularly over issues relating to threatened and endangered species (Lewis 1995, Bean and Wilcove 1997). Landowners are hesitant to permit access to their lands, particularly when a threatened or endangered species is involved. This poses a major dilemma for conservation of a species that occurs regularly on private lands. The purpose of this research was to provide basic ecological information on Florida's caracara population, particularly information that is relevant for management. Although caracaras apparently persist on cattle ranches, their exact distribution throughout the region, their habitat requirements, and their productivity on cattle ranches are poorly known. Goals of this research included improving knowledge of the caracara' s breeding biology, determining land uses compatible with its survival, and assessing relationships between caracaras and cattle ranches. This knowledge is critical for developing guidelines for long-term population monitoring and management.

PAGE 20

5 Study Area The study area in south-central Florida encompassed much of the crested caracara's current breeding range. Included were all or parts of Highlands, Glades, Okeechobee, Osceola, DeSoto, Polk, Hendry, and Indian River counties (Figure 1 -2), an area totaling approximately 20,891 km The landscape in this region is a mosaic of habitats, primarily grasslands and pastures interspersed with cropland, citrus groves, dairies, pine flatwoods, freshwater marshes and lakes, oak scrub, hardwood hammocks, and urban developments. Over half of this area is in private ownership. Primary land uses are cattle ranching, citrus production, and other agricultural crops. The land in public ownership is managed for a variety of uses including recreation, military training, hunting, natural areas, and river floodplain management. Study Organization My research examined the reproductive ecology and habitat associations of Florida's crested caracaras. I focused on (1) documenting reproductive strategies and productivity during 3 breeding seasons; (2) determining spatial requirements and dispersion of breeding pairs, (3) evaluating habitat selection of breeding pairs throughout the landscape, and (4) assessing density and productivity of caracaras on cattle ranches. Results and conclusions are organized into four chapters. Chapter 2 details the breeding biology and discusses reproductive strategies of Florida's crested caracaras. I examine territorial occupancy, breeding rates, and various reproductive parameters for the sample population. Specifically, I test the hypothesis that caracaras, like other large raptors, have low reproductive rates and low productivity. I also examine relationships between breeding chronology and environmental parameters. Finally, I examine factors that may influence overall productivity in this population. In Chapter 3, 1 examine the incidence, determinants, and potential costs of double brooding in this population, a strategy previously unknown for this species. I also discuss potential implications of double-brooding for individual and population productivity.

PAGE 21

6 I examine spatial requirements and evaluate habitat selection of breeding pairs across the south-central Florida landscape in Chapter 4. I estimated home ranges using radiotelemetry then used resource selection techniques to develop a model that describes habitat and landscape features important to breeding pairs in selection of a territory. In this chapter I also examine density of breeding pairs on cattle ranches and compare productivity among caracaras nesting at sites where the major land use is or is not cattle ranching. Chapter 5 contains a synthesis of the conclusions and findings from this research. I present an evaluation of the current status of this population within the context of its historical background and my research results. I also provide recommendations for conservation efforts, management, and monitoring of this population.

PAGE 22

Figure 1-1. Distribution of the crested caracara in North and Central America.

PAGE 23

8 Figure 1 -2. Current known breeding range of the crested caracara in Florida. The study area encompassed much of this range.

PAGE 24

CHAPTER 2 BREEDING ECOLOGY OF FLORIDA'S CRESTED CARACARAS Introduction The crested caracara (Caracara plancus) is one of nine species in the subgroup Caracarinae (caracaras), family Falconidae (del Hoyo et al. 1994). Although some species are widely distributed throughout Central and South America, caracaras as a group have been little studied. The crested caracara is the only species that reaches North America. Breeding populations occur in Texas (Oberholser 1974), southern Arizona (Levy 1988, Ellis et al. 1988), and Florida (Howell 1932, Stevenson and Anderson 1994). The breeding ecology of this species is poorly known throughout its range, although populations in Texas (Dickinson and Arnold 1996) and Baja California Sur, Mexico (Rivera-Rodriguez and Rodriguez-Estrella, unpubl. data), have received recent attention. Florida's crested caracara population is of conservation concern because it is isolated from other populations, has a restricted range, and apparently is of small size (Funderberg and Heinzmann 1967, Heinzman 1970, Millsap et al. 1990). Nesting habitat throughout its historic range has been greatly reduced by conversion of grassland habitats for urban and agricultural development, and this habitat loss is expected to continue. This population is currently listed as Threatened by the U.S. Fish and Wildlife Service and the state of Florida. Despite this listing, published information is limited to descriptions of its natural history and distribution (Nicholson 1929, Howell 1932, Bent 1937, Sprunt 1954, Stevenson and Anderson 1994, Layne 1996). The intent of this research was to study the breeding ecology of Florida's crested caracaras, particularly to provide data useful for conservation and management of this population. Specific objectives were to (1) determine breeding chronology, (2) assess relationships between breeding and environmental variables, (3) relate breeding ecology to annual population productivity, and (4) assess variation in reproductive parameters.

PAGE 25

10 Study Area and Methods I studied the breeding ecology of crested caracaras during 3 years in south-central Florida (Figure 2-1). The landscape is a mosaic of habitats, primarily grasslands and pastures interspersed with cropland, citrus groves, dairies, pine flatwoods, freshwater marshes and lakes, oak scrub, hardwood hammocks, and urban developments. Primary land uses are cattle ranching, citrus production, and other agricultural crops. Breeding pairs of crested caracaras occupy open grasslands and pastures. These habitats are characterized by large expanses of low-growing (< 0.5 m in height) ground cover of grasses and forbs and are dotted with numerous shallow ponds, wetlands, and marshes, and single or small clumps of live oaks (Quercus virginiana), cabbage palms (Sabal palmetto), and cypress (Taxodium sp.). Southern Florida has a humid, subtropical climate with a distinct annual cycle of temperature and precipitation (Chen and Gerber 1990). During the rainy season, June through September or October, temperatures are warm and relatively uniform, averaging about 28C. Convective rains and thunderstorms occur almost daily, and monthly rainfall averages more than 60 cm. Monthly rainfall during the dry season, usually from November through May, averages around 20 cm. During the dry season, air temperatures are cooler but rarely drop below freezing. Daily high temperatures rarely exceed 20 C, but in May, temperatures as high as 38C have been recorded. During the three years of this study, I monitored 33, 62, and 62 nesting attempts in 30, 48, and 55, (65 different) territories, respectively. I color-banded 49 adult caracaras representing 41 territories. I also color-banded and radiotagged 189 nestlings between 7 and 8 weeks of age while they were still in the nest or just after they fledged, when they could be captured easily on the ground. Nest Location and Visitation I located active nesting territories by visiting known historic sites (Layne unpubl data, Millsap 1989) and by conducting observations at sites where I had seen adult caracaras. I found nests by observing the behavior of adults and by covering entire areas

PAGE 26

11 on foot and searching individual trees. I conducted monthly observations throughout the year in known territories to assess site fidelity of adults and to look for nesting activity. Although I found most nests in 1994 after hatching, I followed nest progress before laying in most territories in 1995 and 1996. Once found, I checked nests weekly to determine their fate. I determined clutch size, hatching date, brood size, and chick age using a mirror attached to an extendible pole. Each nest was assigned a week of initiation based either on the known time of egg laying or on back-dating from known hatching or fledging dates (assuming an incubation period of approximately 32 days and a nestling period of approximately 8 weeks). I then calculated week of initiation from the first week of September, which I assigned as the first week of the breeding season each year, based on the earliest known egg laying during all 3 years. I defined breeding season length each year as the number of weeks during which I found nests with eggs. Early nests were defined as those in which egg laying was initiated before the median laying week each year, late nests were those in which egg laying was initiated after the median laying week. Following nest failures, I continued to check these territories at least monthly for signs of renesting. I also regularly searched all territories that had successfully fledged one brood in a breeding season, for second nests. After fledging, I monitored chicks weekly by telemetry to determine survival. Reproductive Success I estimated probabilities of nest success using Mayfield's method (Mayfield 1961, Johnson 1979) for all nest attempts during all years ( = 157 attempts). I estimated weekly rather than daily probabilities of nest success (Mayfield 1961), because of the length of the caracara's breeding cycle. I defined a successful nest as one in which at least one fledgling survived to independence, because conclusions based on reproductive output measured at fledgling independence can differ strongly from those measured at fledging (e.g. Haukioja 1970). I defined independence as the end of the post-fledging dependency period (PFDP), the first 8 weeks post fledging. Before reaching

PAGE 27

12 independence, young remained near the nest and were dependent on their parents for food. I calculated nest success for the incubation, nestling, and post-fledging periods, using lengths of 4, 8, and 8 weeks, respectively. I compared Mayfield nest success probabilities between periods and years using a Z-test (Hensler and Nichols 1981). Measures of reproductive success also included clutch size, nestling survival to fledging, fledgling survival to independence, and egg success (the proportion of eggs laid that produced independent young). For these analyses, I used a subset of the total nests with known clutch size (n = 98 nests representing 50 territories). I compared these measures of reproductive success among years using MannWhitney and Kruskal-Wallis nonparametric tests (Sokal and Rohlf 1995). Along with nesting success, territorial occupancy and breeding rates are often used as indicators of the stability and productivity of a population (Steenhof 1987, Newton 1979). Relatively unbiased estimates of these parameters may be derived by determining these rates over several years for a set of preselected territories (Steenhof and Kochert 1982). I identified 37 territories for which occupancy and breeding status were known in 1994 and tracked their status through the subsequent 2 years. I defined an occupied territory as one in which adult caracaras were seen on 3 or more visits to the territory. I considered a breeding attempt to be actual egg laying, not just nest construction. To assess changes in these rates between years, I used Cochran's Q-test (Sokal and Rohlf 1995). Productivity and Components of Variation I defined two measures of annual productivity: the number of independent young per nesting attempt and the number of independent young per successful pair. Productivity per successful pair included young from renesting attempts. To evaluate factors that influenced overall annual productivity, I calculated an Index of Productivity for each year using information from known clutch-size nests: Index of Productivity = (£MM>

PAGE 28

13 where C = mean annual clutch size, B = length of the breeding season each year, S = annual egg success, and N = length of the breeding cycle (Ricklefs and Bloom 1 977, van Riper 1995). Length of the breeding season was the average number of weeks each year when nests with eggs were found, and length of the breeding cycle was N = b + e + y + n + f where b = nest building (4 weeks), e = egg laying (0.5 weeks), y = length of incubation period (4.5 weeks), n = nestling period (8 weeks ), and / = fledgling dependency period (8 weeks). All time estimates are average for the sample. Total length of the breeding cycle for the crested caracara in Florida is approximately 25 weeks. To evaluate seasonal trends in reproductive performance, I compared clutch size, number of independent young per nesting attempt, and Mayfield nest success estimates for early (initiated egg laying before the median week of egg laying) and late (initiated egg laying after the median week of egg laying) nests for all 3 years. I assessed monthly trends in clutch size and number of independent young per nesting attempt using weighted regression for unequal sample sizes among categories (Sokal and Rohlf 1995). To examine variation among reproductive parameters, I calculated the coefficient of variation (standard deviation/mean, Ricklefs 1973) in percent, for clutch size, number of independent young per nesting attempt, number of independent young per successful pair, and Mayfield nest success. Environmental Data For monthly rainfall analysis, I used data collected from 35 stations throughout the study area by the National Weather Service and the South Florida Water Management District. I averaged daily precipitation values for each month to obtain monthly average precipitation for all 35 stations. Values for all stations were then averaged to obtain monthly precipitation averages for the overall study area. Statistical Analyses Because each year's sample of territories contained a subset of the previous year's territories, data were not pooled across years, unless indicated. I conducted statistical

PAGE 29

14 analyses using SigmaStat v. 2.0 (Jandel Scientific 1995). Statistical significance was evaluated at the a = 0.05 level. Results Territorial Occupancy and Breeding Rates Territorial occupancy rates for the subsample of 37 territories were 97% and 100% in 1994 and 1995, respectively, but were lower in 1996 (89%, Q = 6.5, P < 0.05). One territory was only occupied during 1995, and I could not confirm occupancy in 2 additional territories during 1996. Adult caracaras were seen in these areas, but no individuals were banded in these territories and no breeding activity was observed near the traditional nesting site, so these territories were considered unoccupied. Among known occupied territories, breeding rates remained high during all 3 years, at 100%, 97%, and 97%, respectively (Q 2.0, P > 0.25). Breeding Behavior and Site Fidelity Pairs of breeding caracaras are apparently monogamous and remain together on the territory throughout the year, although they are not easily observed outside the breeding season. The beginning of the breeding season was signaled by several changes in the adults' behavior. Pairs were often observed around the traditional nesting site perching together, preening and allopreening, and sharing food. Adults were generally faithful to both mate and site. Thirty-nine of 49 banded adults, including both banded members of 5 pairs, were present in the same territory and reproduced there for at least 2 years. Although this represents an annual turnover rate of approximately 1 8%, I do not know the fate of the birds that disappeared. Re-pairing within a breeding season was observed only once, after disappearance of a banded female shortly after her single chick fledged. During the 1 995 breeding season, an entire family abandoned its breeding territory about one month after the 2 chicks fledged, and moved to an area approximately 4 km from the nest. This new area was within the territory of another pair that had successfully fledged two young that same year. Radiotelemetry data for the female and 2

PAGE 30

15 chicks that had relocated indicated that they did not return to their original nesting territory. This female nested successfully in the new territory during 1996. Nest Building and Incubation Both males and females build the nest. All pairs that built nests eventually produced a clutch. Building time ranged from 2 to 4 weeks {n = 34 pairs) and often, up to 2 months elapsed between nest completion and egg laying. Both parents developed a brood patch and incubated the clutch. Neither mate delivered food to the other during incubation, but one would remain on the nest while the other left the area to forage. Length of the Breeding Season and Timing of Broods Florida's caracaras' exhibit a prolonged breeding season (Figure 2-2). Chicks fledged from successful nests as early as mid-December, indicating that eggs had been laid during late September. The latest eggs were laid in July, in 1995, although these nests failed (n = 2; both were second attempts). Most young fledged during March and April. Chicks from late and second nests fledged as late as July. Although yearly variation existed in onset and duration of the breeding season, the median week of egg laying for all nesting attempts (n = 157) was in late January each year and did not differ among years (H= 4.14, P = 0.13, Table 2-1). Secondary peaks of egg laying in March and April (Figure 2-2) represent territories where pairs made a second nesting attempt following a successful first brood ( = 16 pairs). The breeding season was longest in 1995, when nests with eggs were found during 40 weeks. Breeding season length in 1994 and 1996 was approximately 28 weeks (Table 2-1). Although within-year variation existed in timing of nest initiation among pairs, between-year nesting intervals for individual pairs were remarkably consistent. The median year-to-year interval between egg laying for single brooded pairs was 50 weeks (n = 41 pairs, Table 2-1), thus individual pairs tended to initiate nests at about the same time each year.

PAGE 31

16 Number of Broods Per Year Multiple nesting within a breeding season occurred during all 3 years. Early nesting pairs regularly renested, either by attempting a second brood or after a first attempt failed (Figure 2-3). During 1995 and 1996, an average of 17.5% of early nests failed in the incubation or early nestling stages. Eighty percent of these pairs renested and half were successful (Table 2-2). Time between failure of the first nest and egg laying in the second nest ranged from 3 to 14 weeks ( = 8 pairs, Table 2-2). Successful double brooding was recorded during all 3 years (Table 2-2). Although second nests failed more often and had smaller clutch sizes, pairs that successfully raised two broods during a season had higher annual productivity than pairs that raised only one brood. More pairs renested after early failure and attempted a second brood in 1995 than in other years (Table 2-2). Clutch Size For known-size clutches (n = 98), the overall mean was 2.23 eggs (Table 2-3). Modal clutch size was 2 eggs for all years. Mean clutch size was higher in 1994 (H = 7.61, P = 0.02, n = 16) owing to more clutches of 3 (Figure 2-4) but did not differ between 1995 and 1996 (U = 1698, P = 0.98). For the same females, clutch size decreased between first nests (2.18 0.20) and replacement or second brood nests (1.91 0.09, paired Mest, t = 2.19, P = 0.03, n = 1 1 pairs). Broods Hatching was asynchronous. Eggs hatched one to two days apart. Both parents fed and cared for the young. Brood size at approximately 4 weeks after hatching averaged 82% of mean clutch size for all 3 years pooled (n = 81 broods, Table 2-3). Chicks occasionally fell out of nests (pers. obs.). Many chicks simply disappeared; some may have died from starvation. Over all years, only 38% of nests with 3 -egg clutches fledged all 3 chicks. Nestlings were in the nest for 7 to 8 weeks, although the earliest recorded fledging occurred at age 42 days. After fledging, the young remained in the general area around the nest and were dependent on the adults for food for at least 2

PAGE 32

17 months. Even after independence, some young remained in the territory until the adults begin breeding the following year. Reproductive Success. Productivity, and Components of Variation The overall Mayfield probability of nest success for all nest attempts (n = 1 57) was 0.74 (SD = 0.03, Table 2-4). Nest success was lower in 1995 than in 1996 (Z = -4.83, P O.0001). Nest success was also lower in 1995 than in 1994, but not significantly so (Z = 0.73, P > 0.46, Table 2-4). Lower nest success in 1995 was caused by higher failure rates of early nesting pairs and second brood attempts. Probabilities of nest success during all phases of the nesting cycle were similar within years (Table 2-4) except for the incubation period in 1994. The number of exposure weeks for the incubation period was less in 1994 than in other years because in that year, most nests were found after hatching. Early-season nest failures were likely missed in 1994. Measures of productivity were lowest in 1995 (Table 2-5), owing to reduced egg success (Table 2-3) and nest success (Table 2-4). In all years, productivity per successful pair was enhanced because pairs successfully renested following early failure or raised second broods. Clutch sizes of 3 produced the most fledglings and independent young during all years (Figure 2-5). Twoand 3-egg clutches produced similar numbers of independent young. Most clutches of 1 failed (Figure 2-5). Measures of productivity showed the highest variability among reproductive parameters (Table 2-6). Clutch size varied little and nesting success showed the least variability among parameters (Table 2-6). Performance and Laying Date All measures of breeding performance declined as the breeding season progressed each year (Table 2-7). Mean clutch size was negatively correlated with laying date (r = -0.85, P < 0.01, Figure 2-6) and the regression was significant (F '=29.07, n = 98). Approximately 76% of the variation in clutch size was due to month of nest initiation. The probability of a nest succeeding varied throughout the breeding season (Figure 2-7). During all years, pairs that initiated egg laying early in the breeding season had a higher

PAGE 33

18 probability of nest success (Z = -3.06, P < 0.001, n = 76) than late nesting pairs (n = 81, Table 2-7), although the earliest nesting pairs generally had a lower probability of nest success (Figure 2-7). Timing of nest initiation also clearly affected productivity. The number of independent young per nesting attempt declined as the breeding season progressed (r = -0.95, P < 0.01, Figure 2-8) and again, the regression was significant (F= 18.44, P < 0.005, n = 157). Approximately 67% of the variation in productivity could be explained by month of nest initiation. Nest Failure. Egg and Chick Mortality Overall, 61.2% of all eggs that hatched from known-size clutches survived to fledging and 48.4% survived to independence (Table 2-3). The fate of most eggs that did not result in an independent young was unknown (Table 2-8). Total nest failure occurred most often during late incubation and near the time of hatching (Table 2-9). Overall, an average of 25% of nests failed each year (Table 2-9), but most early nesting pairs that failed renested, and approximately 50% of renesters were successful (Table 2-2). On average, 1 8% of pairs each year failed to produce any young despite a nesting attempt (Table 2-9). Discussion The reproductive rate of birds presumably represents the maximum number of young that individuals in a population can raise given the prevailing environmental conditions during a breeding period (Amadon 1964, Lack 1968, Ricklefs 1973). Raptors typically have lower reproductive rates than most birds (Newton 1979). Territorial occupancy rates vary widely, and a percentage of pairs annually may not attempt to breed or may abandon an attempt, often due to weather, food availability, or density dependent effects (Newton and Marquiss 1976, 1986, Village 1982, 1983, Kostrzewa 1996). My results indicate territorial occupancy and breeding rates for Florida's crested caracaras that are higher than rates reported for most raptors. In a resident population where pairs are monogamous and remain on the territory year-round, individuals' familiarity with both territory and mate may facilitate successful annual breeding.

PAGE 34

19 Although annual breeding rates were high, productivity varied considerably among pairs and years. In all years, factors associated with productivity included weather, timing and length of the breeding season, clutch size, timing and number of broods, and mortality of eggs and chicks. Weather. Timing, and Length of the Breeding Season. Many factors interact to modify the timing of reproduction in birds, particularly for tropical and subtropical species (Skutch 1950, Immelmann 1971, Ricklefs and Bloom 1977). Cues that function to initiate the crested caracara's breeding season in Florida are unknown. Food supply often is a proximate factor regulating both reproductive timing and effort (Perrins 1970, Newton and Marquiss 1981, Dijkstra et al. 1982, Daan et al. 1989). In tropical and subtropical regions, the abundance of food resources, and thus many birds' reproductive cycles, are closely tied to rainfall cycles (Skutch 1950). Florida's climatic cycle may provide predictable cues to resource availability for crested caracaras. The length and timing of the caracara's breeding season in Florida was strongly tied to the rainfall cycle (Figure 2-2). Reproduction occurred during the winter dry season, with the peak in egg laying following periods of decreasing precipitation in the fall. During the fall, as conditions become drier, fish and other prey become concentrated in ephemeral wetlands, ponds, and ditches (Kushlan 1990, Babbitt 1996). Caracaras often forage in these habitats and regularly kleptoparasitize other species feeding there (Rivera-Rodriguez and Rodriguez-Estrella 1992, pers. obs.). Seasonal and predictable abundance of prey during dry-down may enhance individual females' ability to attain body condition necessary for breeding. Rainfall cycles may also influence the length of the breeding season. Higher precipitation during the winter of 1994-1995 (Figure 2-2) resulted in a higher water table at the beginning of the rainy season in summer 1995. I found nests containing eggs during a longer period in 1995 and more breeding attempts were recorded. Overall productivity for 1995 was lower, however, despite the longer breeding season and similar

PAGE 35

20 mean clutch size. Lower productivity was due to reduced egg and nest success during that year (Tables 2-3 and 2-4). Temperature cycles may also influence reproduction. High temperatures in the spring at the end of dry season may impose additional physiological stress on breeding adults, particularly on late nesters. In 1995, many second and late season (JuneJuly) nests failed, and 3 clutches were found abandoned during incubation in May, when ambient temperatures reached > 40C for several consecutive days. Clutch Size Similar to many avian species, the most productive clutch size for Florida's crested caracaras was larger than the modal clutch size, during all years. This is contrary to Lack's hypothesis (Lack 1954, 1966) but instead supports the hypothesis that females capable of laying a variable number of eggs lay clutches as large as they can produce (Klomp 1970, von Haartman's 1971, Hogstedt 1980). Differences in territorial and individual quality may exist that translate into the ability of each female to lay the largest clutch possible. Timing and Number of Broods. A decline in all aspects of breeding performance with advance in laying date has been well-documented for birds (Lack 1968, Perrins 1970, Newton and Marquiss 1984, Daan et al. 1989, Verhulst and Tinbergen 1991). Timing of nest initiation clearly affects the reproductive performance of Florida's crested caracaras. Clutch size, probability of nest success, productivity, and probability of renesting all declined as the breeding season progressed. Differences in territory or individual quality can affect the timing of egg laying (Klomp 1970, Murphy 1978, Verhulst et al. 1995). Reduced nesting success late in the breeding season may be indicative of seasonal environmental changes, particularly declining food availability (Perrins 1970, Korpimaki and Lagerstrom 1988). Although the ability to renest is clearly advantageous for many avian species (Smith and Roff 1980, Westmoreland et al. 1986, Guepel and DeSante 1990, den BoerHazewinkel 1987) most raptors do not normally hatch more than one clutch per season

PAGE 36

21 (Newton 1979). The ability to renest enhances annual productivity and thus lifetime reproductive success of individual caracaras. About half of the pairs that renested following early failure ultimately did produce independent young during the same year. During all 3 years, pairs that successfully fledged two broods in a breeding season incurred higher productivity than pairs that fledged only one brood. Mortality Loss of eggs and chicks during the breeding cycle affected annual productivity. Less than half of all eggs from known-size clutches resulted in independent young. Most losses occurred before fledging and the cause of most losses was unknown. Most eggs lost during incubation disappeared at the time of hatching. With weekly nest checks, it was difficult to determine whether eggs disappeared or whether newly hatched nestlings were lost during the first week. Predation, possibly by crows and raccoons (Layne 1996, pers. obs.), and weather were suspected causes of egg and chick disappearance. Collisions with vehicles was the major cause of fledgling mortality. The loss of a relatively large number of fledglings due to vehicle collisions in 1995 contributed heavily to reduced overall productivity of the sample of breeding pairs during that year. Breeding Strategy and Variation in Reproductive Parameters Florida is characterized by both spatial and temporal environmental heterogeneity. Florida's crested caracaras exhibit much spatial and temporal variation in the timing of reproduction and reproductive performance by individual pairs. Clutch size and nesting success showed the least variation among years. Since clutch size is related to the number of young parents can raise (Lack 1954), there would be little advantage to varying clutch size unless future conditions were predictable (Cody 1971). The low coefficient of variation for nest success suggests that most pairs have a good chance at producing at least one independent young during any year. High variation in productivity may be attributable to variable food supply and a myriad of potential causes for loss of eggs and chicks. High variation in productivity is also partly due to the fact that

PAGE 37

22 productivity was measured after the young had become independent and mortality factors had acted for a longer period. Variable weather and resource conditions probably have been major factors in the evolution of observed life history patterns for Florida's crested caracaras. Winter weather is unpredictable in south-central Florida. Early nesting pairs, in particular, face a tradeoff. Although reproductive success is highest for early nesting pairs, the very earliest pairs incur a high risk of nest failure due to unpredictable winter storms and cold fronts (Figure 2-7). The strategy of multiple nesting may offset this risk. Pairs that successfully renest after failure or successfully double brood ultimately may have more young recruited into the breeding population. Despite the risk of failure due to weather, nesting as early as possible is advantageous because early nesting pairs generally fledge more chicks per nest and have the opportunity to make a second nesting attempt. The plastic patterns of reproduction for Florida's crested caracaras presumably evolved because they are effective in this variable environment. Although caracaras may breed in any month of the year, the peak of egg laying in late January during all 3 years suggests an optimal time for reproduction. Fewer nests are initiated outside this period and they are less successful, although some do succeed. If unpredictable, catastrophic storms occur during the peak reproductive period, widespread nest failure may result. During these years, pairs that initiate nesting outside the "optimal" period may exhibit higher reproductive success. Such plasticity in spatial and temporal nesting patterns will result in successful nesting and productivity by at least some pairs during any particular year. Comparison With Other Crested Caracaras Information on the breeding biology and nesting success of other crested caracara populations is limited. Dickinson and Arnold (1996) and Rivera-Rodriguez and Rodriguez-Estrella (unpubl. data) have reported extended breeding seasons for crested caracaras in Texas and Baja California Sur, Mexico, respectively. Mean clutch size of crested caracaras in Texas was 2.14 0.14 (SD) eggs, (n = 7 clutches, Dickinson and

PAGE 38

23 Arnold 1996). Productivity was 1.14 0.46 (SD) independent young per breeding pair (n = 7 attempts) and declined as the breeding season progressed (Dickinson and Arnold 1996). Rivera-Rodriguez and Rodriguez-Estrella (unpubl. data) reported productivity of 1.91 0.08 (SE) fledglings per nesting attempt (n = 22 attempts) in Mexico. Double brooding was suspected at both sites. Second clutches were observed in June in Texas (Dickinson and Arnold 1996), and in August in Mexico (Rivera-Rodriguez and Rodriguez-Estrella, unpubl. data).

PAGE 39

24 Table 2-1. Breeding chronology and intervals, in weeks, between broods for Florida's crested caracara population, 1994 1996. ND = no data. Length of Median year-tobreeding Median week year interval season of initiation for single Year (weeks) for all attempts n brooded pairs n 1994 28 21 (3rd wk Jan) 33 ND 1995 40 22 (4th wk Jan) 62 52 18 1996 28 22 (4th wk Jan) 62 48 23

PAGE 40

in U U 3 T3 O •a 2 §"2 >/-> r-„ ifl vo O 00 sD NO (N on no on i-m O VO -h w< Tfr IT) 't >n On On On On On On

PAGE 41

1 1 I a 3 O) o > 3 CO CO o P-, O 1 S o 00 o Dp .S c — a> 2 u H > £ Hi? ON DO 3 i 3 On 2 O — to E 13 I> U 3 CO 5 | op & £ co *" as <-> !2 O c3 3 — 1 co co U 00 > 00 • il 1) 3 O O 3 3 ^ a CM 3 4J B t— 1 eg 09 00 00 IU <+-, o s 5 '5 o Ml c to to U u 'J 3 to co u T3 6 00 -a u a u X CO 00 "3 O 1) O N CO a to _, oo 72 00 aj W -J X u 5 "55 to X o o Z 3 oo no 00 in m in tjo o o m r-m ^ ^NO ON 00 NO OO tN O tN m n On On iri K (N NO m NO h rH ^) tN m m T-H oo 00 00 O ri o + O + o + m ir> in tN rrtN >— < — ON ON tJ00 00 NO — i >n no On ON On ON NO o (N NO 00 m tN vo -i m oo o o + oo ON tN NO 0O 0O O O o o o no r~rtN tN tN tN O tN oo ON •3 (D "o O

PAGE 42

C3 JO O CO to O O 3 1/1 3 i S3 Q 00 a O CO •rc o 1 u d Q 'C u & 00 e '-*-> CO Z a. S u S < CO ^3 2 2 o S3 o m o o o o o no On ir> v >/-> 00 o o (N (N "1 O O O do d m m oo On 00 On ''' (N N — o V~i (N (N ON 0O On NO CN (N o o o P, r00 ON o o m CM (N m no no o oo m m ^ m "
PAGE 43

28 Table 2-5. Measures of annual productivity (mean SE) and Index of Productivity for crested caracaras in south-central Florida, 1994 1996. See text for calculation of Index of Productivity. Productivity per successful pair includes young from renesting attempts. Number of independent Number of independent Index of Year young per attempt n young per successful pair n Productivity 1994 1.33 0.16 33 1.87 0.16 23 1.55 1995 0.98 0.11 62 1.69 0.10 36 1.32 1996 1.34 0.10 62 1.69 0.09 49 1.36

PAGE 44

<4H o a CD (U 8 s ii | > 3 U O oo eg on 1 ii p., CO II V DC 5 J § i .3 — i CO g c oo 3 .9 m >> -a -a 2 .3 11 o 2 b 0 a, 1 I o o s I E ^ ^ 0 9 6 O .9 .3 1 "S s 1 > I CO (N O •£> "C ca H > no ON a* > U a CD in ON ON •3ON CD I 0 oo r00 ON m o in i — 1 — 1 — 1 IT) NO o ri "t (N rm in cn m oc o o o d d o (N cs t d — ri o> nC O O in NO n NO oo ri d r ) NO v N NO OC OO NO rl m in ON NO o o roc ON Os ON Vr"N V 1 (N o o NO CO CO NO ON o (N NO t 00 c> oc ro ^ d NO rNO in ro oc NO £ o .S O ru 3 u "2 3 2 S to in cu o o 3 U Z 2

PAGE 45

30 s o NO ON S3 u ON On ON ON no m r* on o o o o + + m oo m o (N (N ON fN CN ~ On fN O — I o o + + ^to fN O fN CN O + ON fN O + o fN fN NO NO — < fN O + + ITi OO fN fN* oo ON S3 "S3 B W _) u no o + oo ON o o + 00 fN CO O NO fN o o + + ON ON on fN NO NO — r 1 oo CO 3 CO CO IS CO CD T3 "5 C3 S3 "S fS S d § d 11 <=> II "i. v a, p a, "o oo £ — \o NO O N

PAGE 46

31 o -T touts o in o x rs od ro oo tN m tN o o o o U-> —< m O tN (N '-<'—< On (N (N) w w w w O 1— I i-H 00 r*> tN x> > — tN O r> so O x 0 s o x tN Tt tN <* —I O CN Tf OS r? O O tN -i CO C eq Si i s =4 -22 u CO 03 U 09 CJ -a U •4-* CO CO O & o CO •s a O O OS oo xo s so o x o x o x OS O r-^ On i-! 00 fi NO SO SO o x o x o x m tN r-' cn (N (N -h u cj o cj c U U u u Oil 5 u -£3 > Cfl CO o CO -o u 3 <+-( i CD & •a Oil C — co 00 00 T3 00 M b b 73 +- c t— 1 M .3 i i M I n u IS T3 -a u J= i ID u B _o '•*- C8 JO 3 U c B o c IS c o 50 M 6 O o I s •a to e 6 3 -a I o C3 -J T3 -a c =!
PAGE 47

no 10 NO /"> NO oo 3>n m m On •- on r~ oo < ON O o CM s o S E 00 Dh ci O w> •a c J3 T3 u f— 1 00 d 3 O

PAGE 48

33 Figure 2-1. Location of study area for crested caracaras in south-central Florida. Study area encompassed much of the species' current known breeding range in Florida.

PAGE 49

34 £6-po £6-§ny 2P 6 C u 1 i 1 l o — Jl Os T3 _i
PAGE 51

36 Figure 2-4. Distribution of clutch sizes by year, for crested caracara nests in south-central Florida, 1994 1996.

PAGE 52

37 u Ej C OJ 2 1994 •Clutch of 1 (n = 1) i — Clutch of2(n = 5) r • Clutch of 3 (n = 10) Eggs Number fledged Number of independent young u s C 3 1995 -•Clutch of 1 (n = 3) •—Clutch of2(n = 28) tr-Clutch of3(n = 10) Eggs Number fledged Number of independent young •j Jo g B ca u 3 1996 1 j 0 Eggs Clutch of 1 (n = 2) Clutch of 2 (n = 30) Clutch of 3 (n = 9) Number fledged Number of independent young Figure 2-5. Mean number of fledglings and independent young per nest for each clutch size during each year for crested caracaras in south-central Florida, 1994 1996. n = indicates number of clutches.

PAGE 53

38

PAGE 54

39 sssoDns jssu jo XjijiqBqojj

PAGE 55

SunoX juapusdapui jo J9quinjv[

PAGE 56

CHAPTER 3 EFFECTS ON PRODUCTIVITY AND POTENTIAL COSTS OF DOUBLE BROODING IN CRESTED CARACARAS Introduction Reproductive success in birds is determined by the number of broods and productivity per brood (Drury 1961, Lack 1966). The number of successful broods produced each breeding season by a breeding pair is an important contribution to an individual's annual and lifetime productivity (Lack 1954, Stearns 1992, Newton 1989). Successful multiple brooding may also be important in the maintenance of populations (Holmes etal. 1992). In temperate latitudes, many avian species regularly raise more than one brood per year (Smith and Roff 1980, Westmoreland et al 1986, den Boer-Hazewinkel 1987, Guepel and DeSante 1990, Smith and Marquiss 1995). Occurrence of multiple brooding and temporal spacing of broods are influenced by length of the breeding season, food availability, body size, clutch size, development time, amount of parental care, and experience of the parents (Perrins 1970, Ricklefs 1984, Bednarz 1987, Askenmo and Unger 1986, Tinbergen and van Balen 1988, Holmes et al. 1991, 1992). Successful production of multiple broods in a season may increase fitness only if the strategy does not negatively influence the future fecundity or survival of either breeders or first-brood young (Williams 1966, den Boer-Hazewinkel 1987, Nur 1988, Smith and Marquis 1995, Verhulst and Hut 1996). Multiple-brooded individuals usually fledge significantly more young per season than single-brooded individuals, but reproductive success and offspring quality generally decline between first and second broods, and the probability of independent offspring being recruited into the breeding population is lower for young from second broods (Smith and Roff 1980, Hochachka 1990). Despite these costs, multiple brooding may be maintained in some species 41

PAGE 57

42 because it increases an individuals' lifetime reproductive success (Perrins 1970, den Boer-Hazewinkel 1987, Newton 1989). Under certain environmental conditions, there may be little cost to producing additional broods (Ormerod and Tyler 1993, Smith and Marquiss 1995), so the benefit easily outweighs the cost. Raptors typically raise only one brood per year because of the length of time required for the complete breeding cycle including, for many species, an extended postfledging dependency period (Newton 1979). Although many raptors renest after losing a clutch (Newton 1979, Wood and Collopy 1993, Forsman et al 1995), true double brooding (i.e., individuals attempting to rear a second brood in the same breeding season after a successful first attempt) rarely has been confirmed. Second brood attempts, generally in response to peaks in prey cycles, have been reported for some smaller species with short breeding cycles such as the American Kestrel (Falco sparverius; Howell 1932, Stahlecker and Griese 1977, Toland 1985), Eurasian Kestrel [F. tinnunculus; Fargallo et al. 1996), Black-shouldered Kite (Elanus caerulescens; Malherbe 1963), Burrowing Owl (Athene cunicularia; Millsap and Bear 1990), and Barn Owl (Tyto alba; Witmer and Patrick 1987). Multiple brooding occurs in some larger raptors that live year-round in subtropical and tropical environments where food is abundant for most of the year, such as the Harris' Hawk (Parabuteo unicinctus; Mader 1975, Bednarz 1987), Pale Chanting Goshawk (Melierax poliopterus; Smeenk and Smeenk-Enserink 1975), and Galapagos Hawk (Buteo galapagoensis; DeVries 1975). I documented double brooding in crested caracaras (Caracara plancus) during three successive years in south-central Florida. The crested caracara is a medium-sized raptor that occurs in open grassland and savanna habitats throughout Central and South America, in the southwestern U.S., and in Florida (del Hoyo et al 1994). Circumstantial evidence of possible double brooding in this species has been reported in Florida (Bent 1937), Texas (Dickinson and Arnold 1996), and Mexico (Rivera-Rodriguez and Rodriguez-Estrella unpubl. data) but the prevalence of this phenomenon remains unknown. Recently, widespread loss of grassland habitats to expanding agriculture and urban development has focused concern on the Florida population (Layne 1996, B.

PAGE 58

43 Millsap pers. comm.), which is currently listed as Threatened by the U.S. Fish and Wildlife Service and by the state of Florida. Because successful double brooding may affect individuals' lifetime reproductive success and annual population productivity, I was interested in evaluating (1) the extent to which double brooding occurs, (2) the contribution of second broods to an individual's annual productivity, (3) potential costs in terms of reduced survivorship of firstor second-brood young, (4) potential costs to double brooded adults in terms of reduced survival or fecundity, or delayed nesting in the following year, and (5) factors that may contribute to the occurrence of double brooding. Study Area and Methods The study area in south-central Florida included all or parts of Highlands, Glades, Okeechobee, Osceola, DeSoto, Polk, Hendry, and Indian River counties and was centered on the MacArthur Agro-Ecology Research Center (2710'N, 8112'W), Highlands County. This area encompasses much of the species' currently known breeding range in Florida. Breeding pairs of crested caracaras occupy open grassland and pasture habitats containing single or small clumps of live oaks (Quercus virginiana), cypress {Taxodium sp.), and cabbage palms (Sabal palmetto). Breeding pairs are monogamous and remain on territories year-round (Morrison 1996). Beginning in January 1994, 1 conducted monthly surveys in known territories to look for signs of nesting activity. Once found, nests were checked weekly until fates were determined. I considered a nesting attempt to be actual egg-laying, not just nest construction. I determined clutch size, hatching date, brood size, and chick age at each nest using a mirror attached to an extendible pole. Age was determined by assessing the chicks' size and feather development. I assigned each nest a week of initiation based either on known time of egglaying or on back-dating from known hatching or fledging dates (assuming an incubation period of approximately 32 days and a nestling period of approximately 8 weeks). I then calculated week of initiation from the first week of September, which I assigned as the

PAGE 59

44 first week of the breeding season based on the earliest known egglaying during all three years. I estimated annual nesting success using Mayfield's method (Mayfield 1961). I defined a successful nest as one that fledged at least one young. I compared Mayfield estimates of nest success among nest types using a Z-test (Hensler and Nichols 1981). I measured annual productivity as the number of young fledged per successful pair. I assigned the classification of "single brood" to pairs that made only one nesting attempt during an annual breeding season. Double brooded pairs were defined as those that made a second nesting attempt following a successful first effort, in the same breeding season. Territories in which pairs produced replacement clutches after nest failures were not included. Clutch size is reported only for known-size clutches. I defined interbrood interval as the number of weeks between fledging of the first brood and initiation of egglaying in the second nest. To evaluate whether laying a second clutch affected the duration of post-fledging care or survival of first-brood fledglings, I used radiotelemetry to determine times of departure of first and second-brood young from their natal territory and survival past independence. Nestling caracaras were color banded and radiotagged (one per nest) while still in the nest at 7-8 weeks of age or just after fledging, when they could be captured easily on the ground. Young continue to beg and be fed by their parents for several months after fledging (Morrison unpubl. data), so I measured the duration of parental care as the period during which young remained in their natal territory. The natal territory was defined as the area within a radius of 3 km from the nest, which approximates a breeding pair's home range (see Chapter 4). I defined permanent departure from the natal territory as the first time a juvenile was located more than 3 km from the nest followed by a minimum of 3 consecutive weeks at distances > 3 km from the nest. I obtained survival data for 61 radiotagged young through their first year via weekly aerial radiotracking surveys. I estimated probabilities of survival to recruitment for young from single, first, and second nests using these survival data along with the

PAGE 60

45 Kaplan-Meier procedure modified for staggered entry of animals (Pollock et al. 1989). Age at first reproduction is not known for this species but is estimated to be 3 to 4 years (Voous 1983, Layne 1996), so I used 1200 days post fledging as an estimate of recruitment age. I compared survival probability functions between groups using the logrank test (Cox and Oakes 1984, Pollock et al. 1989). To evaluate potential costs to double brooded adults, I followed the fate of colorbanded individuals. I used the likelihood ratio test in PROC LIFETEST in SAS (SAS Institute 1985) to compare predicted survival estimates for 16 known double-brooded and 38 known single-brooded adults. Survival through at least 2 years was known for 1 1 double-brooded and 27 single-brooded adults. Survival was defined as remaining and breeding on the same territory in subsequent breeding seasons. To assess the potential cost of reduced fecundity in the year following a double brood event, for 1995 and 1996, I compared clutch sizes between the group of known adults that had double brooded the previous year and the group of known adults that produced only one brood. I also examined whether double brooding had any influence on timing of the first nest attempt of these pairs in the following year. For 1995 and 1996, 1 compared timing of first nest initiation between pairs that had double brooded the previous year and pairs that produced only one brood. I evaluated potential effects of double brooding on population growth rates and likelihood of persistence. For this analysis, I used values previously calculated for survival of singleand double-brooded adults through 2 years, survival of first-, second-, and single-brood young to 2 years post-fledging, and mean numbers of young fledged for singleand double-brooded pairs. I calculated population size and population growth rate during 3 consecutive years for two hypothetical populations, one in which double brooding occurred in approximately 10% of the population each year, and one with no double brooding. I assumed each initial population size to be 100 breeding pairs (50 females) and age of first breeding to be 4 years. I used logistic regression to evaluate the effects of several factors on the occurrence of double brooding in this population. I examined whether initiation date of

PAGE 61

46 the first clutch (INITWK) and numbers of fledglings in the first brood (NUMFRST) influenced the occurrence of double brooding during the same year. For pairs whose reproductive history was known for at least 2 years, I examined whether the occurrence of double brooding during any year was influenced by whether or not the pair had double brooded during the previous year (DBPREV), the number of fledglings from the previous year (NUMPREV), and the fledging date from the previous year (WKPREV). Statistical tests were conducted using SigmaStat (Jandel Scientific 1995). Significance for all tests was assumed at the a = 0.05 level. Comparisons of fecundity, timing of nest initiation, and productivity were made using Wilcoxon rank sum and Kruskal-Wallis nonparametric tests (Sokal and Rohlf 1995). Post-hoc power estimates for two-sided Wilcoxon tests were obtained using the program Resampling Stats (Resampling Stats Inc. 1995). I first simulated 10,000 values of the rank sum statistic under the null hypothesis of no differences between nest types. The 95th percentile of these values was used as the critical value for determining significance of the difference between nest types, given the two corresponding sample sizes. Real data values were resampled to approximate the alternate distribution of the rank sum statistic as suggested by the data. The power of the Wilcoxon test at the 0.05 level was then estimated as the proportion of times out of 1 0,000 iterations that the resampled test statistic was at least as large as the critical value. Results Extent of Occurrence of Double Brooding I monitored breeding activity on 60 different territories throughout the study, although only a subset of those each year (Table 3-1). A second brood attempt was made on 13 territories (21.7%, Table 3-1). In three territories, the pair initiated a second nesting attempt during two consecutive years, although none of these pairs successfully fledged two broods in more than one year. Both adults were color banded in one territory where double brooding occurred. In all other territories where double brooding occurred, one adult was banded and participated in both nesting efforts.

PAGE 62

47 Timing of Broods Pairs that attempted a second brood initiated their first nests significantly earlier (U= 1504.5, P = 0.00, n = 16) than single-brooded pairs (n = 103) (Table 3-1). Second nests of double-brooded pairs were initiated 9-13 weeks later (U = 26.5, P = 0.00, n = 16) than nests of single-brooded pairs. Median week of nest initiation did not differ among years for each nest type (single brood, H= 0.79, P = 0.67, n = 103; first of two, H= 1.88, P = 0.39, n= 16; second of two, H= 0.65, P = 0.72, n= 16, Table 3-1). Only early breeders attempted to double brood. The number of pairs that attempted second broods decreased with increasing laying date (Figure 3-1). Of pairs that attempted a second brood ( = 16), 72% fledged young from their first brood before 1 March, 94% before 1 April. Of all pairs that fledged a brood before 1 March (n = 16), 69% made a second nesting attempt. The mean interbrood interval was 7.6 weeks (range 2-13 weeks, n = 1 6), and interval length was positively correlated with the number of young fledged from the first nest (r= 0.50, P < 0.05, Figure 3-2). Productivity and Nesting Success Mean clutch size declined between the first and second nests of double-brooded pairs (Table 3-2). Laying times of second clutches and the first clutches of late breeders overlapped during March and early April (Figure 3-1). Mayfield success estimates for second nests were significantly lower than those for first nests or nests of single-brooded pairs (Table 3-2). Of the pairs that attempted two broods, those that had an early first nest produced the earliest second broods; these early second broods had the highest probability of success (Figure 3-3). Taking into account all nesting attempts, double brooded pairs did not necessarily fledge more young than single-brooded pairs (T= 1697.50, P = 0.21, Figure 3-4), primarily owing to high failure rates of late second broods. Pairs that nested earliest however, had the highest chance of successfully fledging both broods (Figure 3-3), thus also had the highest chance of fledging more young each year than single brooded pairs.

PAGE 63

48 Parental Care and Post-fledging Survival First-brood young incurred shorter periods of post-fledging parental care. Mean time to departure from the natal territory of first-brood fledglings (1 1 wk 2.56 SE, n =14 broods) was significantly shorter (U = 290.5, P = 0.00) than that of fledglings from single-brood pairs (33 wk 1 .61 SE, n = 32 broods) or of second-brood fledglings (29.7 wk 1.73 SE, n = 6 broods, U= 33.0, P = 0.01). Post-fledging care of first-brood young overlapped incubation of the second clutch in all cases. Young from first broods typically left the natal territory during the early nestling stage of the second brood, but whether the parents enforced this departure is not known. The period of post-fledging parental care for fledglings from single-brood pairs and second-brood fledglings did not differ (U= 54.4, P = 0.40). In many cases, singleand second-brood fledglings remained in the natal territory until the parents began breeding the following year. Estimates of survival probability to recruitment differed between young from first and second broods (Kaplan-Meier survival function, X 2 = 8.51, P = 0.004, Figure 3-5). Survival estimates for young from first broods were also higher than estimates for young from single broods (X 2 = 75.95, P = 0.00, Figure 3-5). Potential Costs to Adults Survival estimates to subsequent breeding seasons did not differ (X 2 = 0.01, P = 0.92) between known double brooded (n = 16) and single-brooded (n = 38) adults (Table 3-3). Clutch sizes did not differ (U = 95.0, P = 1.0) between known adults that had successfully double brooded in the previous year (2.67 0.33, n = 10) and known adults that produced only one brood (2.22 0.22, n = 19), but power of this test was low (1-P = 0.05) because of small sample sizes each year. I did not detect a delay in nesting by double brooded adults in the year following a double brood event. The median week of egglaying did not differ between pairs that had successfully double-brooded the previous year and pairs that fledged only one brood in either 1995 (U= 80.0, P = 0.13; median week = 19, n = 3 double brooded pairs; median week = 22, n = 35 single-brooded pairs) or 1996 (U= 105.0, P = 0.46; median week = 20, n = 4 double brooded pairs;

PAGE 64

49 median week -22, n = 43 single-brooded pairs). Power of these tests was low, however, at 0.23 and 0.12, respectively. Effects on Population Growth and Persistence The hypothetical population that included double brooding by approximately 1 0% of pairs annually grew at a slightly faster rate than the population comprised of only single-brooded pairs. After 3 years, differences in size were negligible between the two populations, however (Table 3-4). Influence of Factors Timing of initiation of the first brood (INITWK) was the only factor important in predicting the occurrence of double brooding that same year by caracara pairs (Table 3-5). For pairs whose reproductive status was known for the previous year (n = 45), the occurrence of double brooding in any year was not influenced by double brooding the previous year, the number fledged during the previous year, or the date of fledging the previous year (Table 3-6). Discussion Life history theory predicts that a breeding attempt will occur when the likely benefits to lifetime reproductive success exceed expected costs (Williams 1 966, Stearns 1992). In Florida's crested caracara population, only a small percentage of pairs attempt a second brood in any year. This strategy may be maintained, however, because double brooded individuals may incur higher lifetime reproductive success (Perrins 1970, den Boer-Hazewinkel 1987). Few second nests are successful and the probability of recruitment is lower for second-brood young. Individuals that successfully fledge two broods in a season are more productive, however, so they potentially have more young recruited into the breeding population. In addition, costs of double brooding to adults and young seem minimal. Second clutches may reduce the reproductive value of the first brood because of reduced post-fledging care of first-brood fledglings (Verhulst and Hut 1996). I found no evidence that first-brood young incurred reduced post-fledging survival despite

PAGE 65

50 substantially reduced post-fledging parental care. Indeed, first-brood fledglings actually had the highest probability of survival to recruitment. This suggests that juvenile caracaras can become independent foragers in a period as short as 7-8 weeks. I found little evidence of costs to double brooded adults, though these conclusions are preliminary because of the short time frame of the study, small sample sizes, and low power of statistical tests. Caracaras are long-lived (Klimkiewicz and Fletcher 1989, Morrison 1996), so true costs may not be evident in the short term. Additional years of tracking the survival and reproductive efforts of known adults would be needed to properly evaluate these potential costs. The occurrence of double brooding in Florida's crested caracara population may be related to both length of the breeding season and to food availability, two factors that determine the number and the timing of broods produced by a breeding pair (Perrins 1970, Newton 1979, Tinbergen and van Balen 1988). Crested caracaras in Florida exhibit a long breeding season, generally beginning in September and continuing throughout the winter and into June or early July (Morrison 1996). Pairs that nest early may have sufficient time to raise two broods. Food supply is known to influence the occurrence and timing of breeding in raptors (Cave 1968, Dijkstra et al. 1982, Newton and Marquiss 1984). Food supply has been shown to influence the number of young fledged annually by a breeding pair primarily by influencing the occurrence of multiple brooding (Malherbe 1963, Bednarz 1987). The crested caracara has a varied diet and feeds on live vertebrate and invertebrate prey as well as carrion (Richmond 1976, Ellis et al. 1988, Morrison 1996), therefore breeding need not be synchronized with availability of particular foods. The caracara typically inhabits areas associated with human activity such as cattle ranches, agricultural lands, and urban dumps (Rodriguez-Estrella 1996, pers. obs.). These habitats may provide a diversity of predictable food resources, perhaps owing to regular schedules of management activities. The combination of a relatively mild climate and the perpetual availability of a variety of food resources apparently allows early initiation of breeding and occasionally, double brooding by some pairs. Whether individuals actually attempt a

PAGE 66

51 second brood or even attempt to nest early in the season may depend on the age and quality of individuals, experience of pairs, and territory quality. The occurrence of double-brooding in any year was not influenced by the previous year's reproductive effort, suggesting that double brooded individuals may be the most fit. Year-round residency and persistence of the pair bond between nesting seasons may facilitate double brooding in crested caracaras because of individuals' familiarity with habitats and resources. Males and females have similar parental roles, so pairs may have little difficulty feeding first-brood fledglings and incubating a second clutch. Caracaras also frequently reuse old nests for first and second breeding attempts. Reuse of old nests reduces the time required for nest-building (Mader 1975, Toland 1985, Westmoreland et al. 1986). Double brooding appears to be a regular but uncommon occurrence in Florida's population of crested caracaras and may occur in other populations (Dickinson and Arnold 1996, Rodriguez-Estrella and Rivera-Rodriguez, unpubl. data). An extended breeding season and post-fledging dependency period may have contributed to a previous lack of its confirmation. Some juveniles remain on their natal territory for several months after fledging (Morrison 1996). Therefore, determining whether juveniles seen with adults during any month of the year are from a single or second brood is not possible, unless individual birds are marked and have a known history. The ultimate value of double brooding must be that double-brooded individuals produce more young with some probability of establishing themselves in the breeding population. Whether or not double-brooding contributes to overall population growth rates is of lesser importance; the overriding goal for individuals is to leave more offspring than the average member of the population. Combined with an extended breeding season, environmental and food conditions during some years may be favorable for some caracara pairs to attempt a second nest. In addition, costs may be small, at worst. Of greater interest is why more individuals do not attempt second broods. For many avian species, the occurrence and timing of a breeding effort is influenced by age and experience of the parents or energetic constraints related to territory or individual quality

PAGE 67

52 (e.g., Smith and Roff 1980, Evans Ogden and Stutchbury 1996). The absence of double brooding in some early nesting pairs of crested caracaras and the lack of earlier nesting by many single-brooded pairs may be related to these factors. Further studies of conditions that influence the occurrence and timing of breeding by individuals and a more accurate assessment of recruitment rates of first and second-brood young are needed to better understand the contribution of double brooding to individuals' lifetime reproductive success and to the overall dynamics of Florida's crested caracara population.

PAGE 68

M U u -a o u a e C o c C3 T3 U o S .2 as 1 o i cm o § ^ S <4— ( <4— i O O 5 .is w •a 5 •= o o o x> ^3 ii^ ~ *r< T3 a ,o E 2 5 to 1 1 is 2 £ .2 > E £3 U m oo m 5PH S a < (S (N ^ m m f> cn oo >r> > > o o S5 55 -*- -*- U u Q M % S3 +3 1_ •— > •— > M M M % £ £ "p J=l m "t — i (N CM ON ON ON On On On

PAGE 69

54 Table 3-2. Clutch size (mean SE) and nesting success (SD in parentheses) between nest types for crested caracaras in south-central Florida, 1994-1996. The same letter indicates significant differences between nest types. ND = no data on clutch sizes of first or second nests in 1994 because these nests were found after hatching. Clutch Mayfield nest Nest type Year size n success n Single broods 1994 2.60+ 0.16 15 0.67(0.09) 25 1995 2.28+ 0.11 25 0.72(0.05) 35 1996 2.22 + 0.09 32 0.92 (0.03) 43 Pooled 2.32 + 0.07 72 0.81 (0.03) a 103 First of two 1994 ND 1.00 (0.00) 3 1995 2.66 + 0.32 3 1.00 (0.00) 8 1996 2.33 + 0.33 3 1.00 (0.00) 5 Pooled 2.50 0.22 b 6 1.00 (0.00) c 16 Second of two 1994 ND 1.00 (0.00) 3 1995 1.80+ 0.02 5 0.35 (0.15) 8 1996 2.00 + 0.00 2 0.25 (0.20) 5 Pooled 1.86 0.14 b 7 0.35 (0.12)^ 16 Z= 3.72, P< 0.0002 b U= 33.0, /> = 0.03 Z = 4.19, P< 0.0001

PAGE 70

55 Table 3-3. Estimated parameters for evaluating contribution of double brooding to population growth for Florida's crested caracaras. Parameters estimated from data collected during 1994 1996. Parameter Year 1 Single Double Mean number fledged per year 1.80 3.00 Probability of adult survival to subsequent year 0.82 0.88 Probability of fledgling survival to subsequent year 0.71 0.75 a Year 2 Single Double 1.60 1.19 0.70 0.78 0.46 0.43 Year 3 Single Double 1.53 1.30 0.70 0.78 0.43 0.36 Survival estimates were calculated separately for young from first and second broods. I assumed half of all young produced by double-brooded adults were from single broods and half were from second broods. Yearl (0.5*1.00 + 0.5*0.50) Year 2 (0.5*0.73 + 0.5*0.13) Year 3 (0.5*0.58 + 0.5*0.13)

PAGE 71

56 Table 3-4. Estimation of population size and annual growth rate for two hypothetical populations of crested caracaras in south-central Florida, one in which all pairs only produce a single brood each year, and one in which approximately 1 0% of pairs successfully produce two broods each year. Single broods only Year Adults Total number young fledged Total number 1 year olds Total number 2 year olds Total N X 1 2 3 Mean 50.00 41.00 28.70 90.00 65.60 43.91 63.90 46.58 29.39 140.00 170.50 148.58 2.80 1.22 0.87 1.63 Single and double broods Total number Year Adults young fledged Total number 1 year olds Total number 2 year olds Total N X 1 single double 45.00 5.00 81.00 15.00 126.00 20.00 2.92 2 single double 37.17 4.13 59.47 4.91 57.51 3.69 154.15 12.73 1.14 3 single double Mean 26.32 2.92 40.26 3.80 42.23 2.85 26.45 1.23 135.26 10.80 0.88 1.65

PAGE 72

57 Table 3-5. Statistics for logistic regression analysis of the influence of timing of initiation of the first brood (INITWK) and the number fledged in the first brood (NUMFRST) on the occurrence of double brooding in crested caracaras in south-central Florida, 19941996. n = 16 double-brooded pairs, 103 single-brooded pairs. Variable Coefficient SE P Constant 7.50 2.22 0.001 INITWK -0.52 0.12 <0.001 NUMFRST 0.01 0.43 0.97

PAGE 73

58 Table 3-6. Statistics for logistic regression analysis of the influence of double brooding the previous year (DBPREV), fledging date (WKPREV) and number fledged the previous year (NUMPREV) on the occurrence of double brooding in crested caracaras in southcentral Florida, 1994 1996. n = 1 1 double-brooded pairs, 34 single-brooded pairs. Variable Coefficient SE P Constant 2.40 0.28 0.39 DBPREV 0.72 0.92 0.44 WKPREV -0.12 0.86 0.15 NUMPREV 0.27 0.55 0.63

PAGE 74

59 a .s II a j a -a c o o u -a s i c £ M in uO iu pC E 3 z £ e u B C o 0 It? 0 E 3 B B T3 3£ 6z; 9Z 3 a, < 2 — a a o ZZ oz u u Q > o o O S 1 o U --• tz> U o M CO C3 CD bp § u 25 3 o Jq 00 3 o I -4— I I vi On U ON C — 00 on a on 'Eh o i ^ 00 c 6 jo asquint

PAGE 75

60 14 T 12 10 13 8 4 2 r = 0.50, P < 0.05 0.5 1.0 1.5 2.0 2.5 3.0 3.5 Number fledged from first brood Figure 3-2. Length of the interbrood interval in relation to number fledged from the first nesting attempt for double brooded crested caracaras in south-central Florida, 1994 1996. n = 16 pairs.

PAGE 76

61 1 o s lis 00 1 .5 oo -*- on P ON d • .2 +-> ON 5 2 (X m o = a jnjssaoons 3J3M ipUOUI 1{0E3 UI p3JBi;iUI S}S3U JO 3§BJU30J3J

PAGE 77

62 Figure 3-4. Mean number of young fledged per pair for pairs of crested caracaras nesting in south-central Florida, 1994 -1996. Number of pairs is shown above each column.

PAGE 78

Ajijiqeqcud |baiajtis

PAGE 79

CHAPTER 4 HABITAT ASSOCIATIONS AND PRODUCTIVITY OF CARACARAS ON CATTLE RANCHES Introduction Wildlife managers face the increasingly difficult task of providing suitable areas and habitats for wildlife in landscapes that are continually changing due to a complex array of human-induced processes. Management of wide-ranging species can be particularly challenging because individuals' activity areas may cross socio-political or landownership boundaries. The crested caracara is a gregarious raptor with a limited distribution in North America. In Florida, a small population occurs in the south-central peninsula, a region historically dominated by native prairie. Much of this ecosystem, in which the caracara historically occurred (Phelps 1912, Nicholson 1929, 1930, Howell 1932), has been replaced by large acreages of privately-owned cattle ranches. Although these ranches consist of open grasslands with scattered trees, vegetation cover types and land management practices differ considerably from those occurring on native prairies. Most ranches include large areas of improved pasture, grasslands dominated by non-native vegetation species that are intensively managed to produce forage for cattle. The natural hydrologic regime in improved pastures also has been altered by extensive ditching and draining of wetlands. The current landscape in south-central Florida reflects complex patterns of land ownership and land uses. Within recent decades, urban and agricultural development has escalated in this region. On private lands, remnant prairie and, increasingly, pasture habitats are being converted to citrus and sugar cane. The crested caracara is likely to be affected by these land use changes. This raptor is now often found on the cattle ranches (Layne 1978, Millsap 1989), but the degree of dependence on ranch habitats is not well 64

PAGE 80

65 understood. Because the spatial and habitat requirements of the crested caracara and land management practices compatible with its survival remain poorly known, consequences of land use changes on this Threatened population are difficult to predict. Understanding relationships between caracaras and land uses is critical to conservation planning for Florida's population. The objective of this study was to examine these relationships by considering habitat and spatial use of breeding adults and productivity of pairs nesting on the cattle ranches. Specifically, I identify patterns of land ownership and land use associated with active nest sites, examine home ranges of breeding adults and spatial relationships among neighboring pairs, evaluate landscapelevel habitat selection by breeding pairs, and examine productivity by pairs in relation to habitat variables characteristic of cattle ranches. Study Area The study area encompassed approximately 21 ,000 km and included all or parts of Highlands, Glades, Okeechobee, DeSoto, Osceola, Polk, Hendry, and Indian River Counties (Figure 4-1). The landscape is a mosaic of habitats, primarily grasslands and pastures interspersed with cropland, citrus groves, dairies, pine flatwoods, freshwater marshes and lakes, oak scrub, hardwood hammocks, and urban developments. Over half of this area is in private ownership. Principal land uses are cattle ranching, citrus production, and other agricultural crops. The land in public ownership is managed for a variety of uses including recreation, military training, hunting, natural areas, and river floodplain management. Methods Land Ownership/Land Use I located active caracara nests by visiting known historic territories (B. Millsap pers. comm., J. Layne pers. comm.) and areas where caracaras had been observed within the past 5 years, and by searching other areas of potentially suitable habitat on both public and private lands. The searching methodology involved conducting observations near potential nest sites. I also covered entire areas on foot to look into individual cabbage

PAGE 81

66 palm (Sabal palmetto) and live oak (Quercus virginiana) trees. When caracaras were observed on private land, the landowner was contacted for access permission. For each territory located, I classified landownership as private or public and primary land use in the vicinity of the nest tree as ranching, other (including urban or other agriculture), or natural. To evaluate whether occupancy and breeding rates in caracara territories were related to patterns of land use and land ownership, I used a set of 37 territories that were monitored during all 3 years. For each year, I recorded whether or not each territory was occupied and if so, whether or not the pair made a breeding attempt. Using a Z-test with Yates' correction (Sokal and Rohlf 1995), I tested the null hypothesis of no differences in the proportions of territories that were occupied or where the pair made a breeding attempt during all 3 years, between territories on privately owned (n = 30) and publiclyowned lands (n = 7). Home Range Estimation and Spatial Dispersion of Territories I trapped and radiotagged 39 (16 females, 23 males) breeding adult caracaras during the nesting seasons in 1994 through 1996 (Morrison and McGehee 1996). Each captured adult was banded with a standard U.S. Fish and Wildlife Service aluminum leg band and a colored aluminum band with a unique alphanumeric character that permitted identification of individuals. Radiotransmitters (Holohil, Ltd.) were attached using a standard backpack harness constructed of Teflon ribbon and secured at the bird's chest. Because caracaras are monomorphic, gender of captured adults was determined using DNA from blood samples collected in the field (Morrison and Maltbie in review). I monitored radiotagged adult caracaras throughout the year. I located all radiotagged adults (Mech 1983) and obtained GPS coordinates of their locations during weekly aerial surveys from a fixed-wing Cessna 1 72 aircraft. I ascertained additional locations for adult caracaras by triangulation (White and Garrott 1990). Determining ground locations was facilitated by access to ranch interiors granted by local ranchers. Most locations were taken from a distance of < 500m. I also recorded visual observations

PAGE 82

67 of radiotagged adults during scheduled visits to territories. Triangulation and visual locations used for home range estimation comprised approximately 1 8% and 1 5%, respectively, of all locations for adults. The 95% confidence circle around aerial location estimates (Marzluff et al. 1994) was approximately 42 ha. Error polygons around triangulated locations (White and Garrott 1990, Kennedy et al. 1994) were all < 50 ha; most were < 20 ha. Because all locations used for home range estimation were taken at least 24 hours apart, I did not conduct tests for independence of observations (Swihart and Slade 1985). I delineated home ranges using kernel methods (Worton 1989, 1995, Seamen and Powell 1996) and the software RANGESV (Kenward and Hodder 1995). The boundary of each caracara's home range was defined as the 99% contour estimated with the fixed kernel method and a 0.85 smoothing factor (Kenward and Hodder 1995, Seamen and Powell 1996). Core areas (Wray et al. 1992) were identified as the 60% kernel contour using cluster analysis in RANGESV. I compared home range estimates between males and females and between breeding and non-breeding seasons using a two-tailed MannWhitney test (Sokal and Rolf 1995). The breeding season was defined as the period from the beginning of incubation to the end of the post-fledging dependency period, 8 weeks after the young fledged from the nest. After this time, young were considered independent, and parents and young were together less often as a family unit. To evaluate nesting density and spatial relationships among territories, I calculated distances between adjacent caracara nests. Using RANGESV, I evaluated overlap of 99% contours and core areas among 9 neighboring, successfully breeding pairs of caracaras inhabiting the MacArthur Agro-Ecology Research Center (MAERC) and an adjoining ranch, and among 4 neighboring, successfully breeding pairs at a site along the Kissimmee River.

PAGE 83

68 Landscape-level Habitat Selection by Breeding Pairs Habitat data Data on vegetation cover types throughout the study area were obtained from the Florida GAP Analysis project at the University of Florida. Analyses were performed in ArcINFO GRID (Environmental Sciences Research Institute 1996) The study area map, initially classified from 1994 LANDSAT satellite imagery, was imported into the GIS in grid-cell (raster) format. The map was 1909 X 1571 cells at a grid cell resolution of 90m. I recognized 13 natural and semi-natural vegetation cover types (Table 4-1). Other cover types reflected land use conversion by humans such as citrus, croplands, sod fields, quarries, and urban areas. Model development I used resource selection techniques (Thomas and Taylor 1990, Manly et al. 1993) to identify habitats and landscape features that distinguished between known caracara home ranges and potentially "available" home ranges throughout the study area. These "available" units were 1296 ha in size, comparable to the average caracara's home range, and were sampled across the entire study area using a sliding window technique and the GIS. Because of the finite 90m grid cell size, square units 40 cells (3600m) on a side were used. Resource categories used for model development consisted of variables describing proportions of the 13 natural and semi-natural vegetation cover types (Table 4-1). Cover types that reflected habitat conversion, such as citrus groves, row crops, and urban development were not included, although some caracara home ranges contained them. I assumed that conversion occurred after a caracara pair had selected the territory, so these cover types were associated by default with cover types actually selected by caracaras. Other variables included (1) total proportion vegetated, (2) total proportion of open habitat, (3) total proportion of shrubs, (4) total amount of improved pasture habitat, and (5) largest contiguous patch of improved pasture habitat (Table 4-1).

PAGE 84

69 Using the GIS, I determined that caracara home ranges were at least 53% vegetated with natural or semi-natural vegetation. For model development, I only used available units that contained more than 53% natural or semi-natural vegetation (n = 925) and the 25 known caracara home ranges. I used the GIS to calculate values for all variables for all home ranges and available units. I used logistic regression procedures in SAS (SAS Institute 1995) to calculate coefficients and standard errors for all variables (Manly et al. 1993). These procedures fit linear logistic regression models for binary data by the method of maximum likelihood (Menard 1995). The response probability produced by the models was the probability of a sample unit being a territory, or "used." I assumed that variables that distinguished between the home ranges and the available units indicated selection by caracaras. Significance of variables was assumed at the a = 0.10 level. Because use of multiple variables results in many alternative models, selection of the "best" or "correct" model is an important step in data analysis (Burnham and Anderson 1992). To choose among the models produced in SAS, I used AIC (Akaike 1985, Burnham and Anderson 1992) and Goodman's criterion (Goodman 1984), both of which consider the principle of parsimony in model selection. To evaluate model fit, I conducted Hosmer and Lemeshow's goodness of fit test (Hosmer and Lemeshow 1989), which tests the null hypothesis that there are no differences between the observed and expected number of observations in each of several groups based on the percentiles of the estimated probabilities. Model validation A model used to make inferences about important habitats for a species is only as good as its ability to make correct predictions about new observations. I tested the ability of the "best" logistic regression model to predict use of 29 other known caracara home ranges. Because I did not have telemetry data for these test ranges, they were delineated using the GIS. For each test range, I created an area 1296 ha in size (square units 3600m

PAGE 85

70 on a side) centered on the nest. I then applied the model to each test range to calculate probability of use. Caracaras and Cattle Ranches To evaluate the caracara's use of cattle ranches, I investigated relationships between home range size, productivity, and the proportion of improved pasture in each home range. I defined productivity as the total number of young produced during all 3 years of the study by each pair. For this analysis, I only used home ranges for which I had these data (n = 1 8). Statistical tests were conducted using SigmaStat (Jandel Scientific 1995). Results Land Ownership/Land Use Land ownership and land use patterns were determined for 65 active caracara territories. Most territories found were on privately-owned land (Table 4-2). The major land use in the vicinity of caracara nests at all sites was cattle ranching (Table 4-2). Lands under public ownership where I found caracara territories were primarily managed for native habitats, including prairie, although moderate grazing occurred in some areas. The proportions of the 37 preselected territories that were occupied and where adults attempted breeding during all 3 years were higher (z = 2.64, P = 0.008, 1-P = 0.71 and z = 2.02, P = 0.04, 1-P = 0.52, respectively) for territories on privately-owned lands (Figure 4-2). Home Ranges and Spatial Dispersion of Territories Breeding adult caracaras remained on territory throughout the year. I obtained usable data for home range estimation from 25 individuals (16 males and 9 females). Home range sizes varied from approximately 551 to 2778 ha (Table 4-3). Most home ranges were more than 80% vegetated with natural or semi-natural vegetation. All ranges contained improved pasture, ranging from approximately 1 1% to 94% of the naturally vegetated area (Table 4-3).

PAGE 86

71 Approximately 90% of locations for all adults were within 2.5 km of the nest (Figure 4-3). Home range sizes did not differ between males and females (t = 1.13, P = 0.27, df = 23), although power of this test was low (1-P = 0.10). Adult caracaras ranged farther from the nest, on average, during the non-breeding season (A" 2 = 56.17, P < 0.001, df = 7, Figure 4-4). For 1 1 adult caracaras (7M, 4F) having approximately equal numbers of radiolocations for each season, breeding season home ranges were smaller than non-breeding season home ranges (paired /-tests, t = -3.45, P = 0.006, df= 10, 1-0 = 0.86). Internest distances for neighboring successfully breeding pairs averaged 3.08 km 0.18 SE (range 1.01 5.08 km, n = 30 pairs, Figure 4-5). Core areas rarely overlapped (Table 4-4), although 99% contours overlapped on average 14.7 % 2.82 SE at MAERC and 10.08% 5.39 SE at the Kissimmee River site (Table 4-5). At MAERC, 6 caracara home ranges overlapped home ranges of at least 2 other pairs of caracaras. Landscape-level Habitat Selection by Breeding Pairs The model that best distinguished between known caracara home ranges and available units throughout the study area indicated a negative association of caracara home ranges with the variables pine forest, mixed forest, and woodland, and a positive association with the variables improved pasture, and contiguous pasture. All variables were significant at the 0. 1 0 level (Table 4-6). Hosmer and Lemeshow's test indicated good fit of this model (G = 10.16, P = 0.25). Since caracaras are known to be associated with open rather than forested habitats, I reran logistic regression analysis with the forest and woodland variables omitted. Results of this analysis indicated a positive association between caracara home ranges and improved pasture, native grassland/prairie, and wax myrtle, a shrub commonly associated with grassland habitats. Again, all variables were significant at the 0.10 level (Table 4-7). Hosmer and Lemeshow's goodness of fit test indicated a poorer fit of this model, however (G = 22.86, P = 0.003).

PAGE 87

72 Although these 2 models were biologically meaningful and statistically fit the data, their predictive capability was poor. For the model that included the forest and woodland variables, the mean predicted probability of use for the 29 test ranges was 0.10 (range = 0.00 to 0.31, Table 4-8). The model that did not contain the forest variables was only slightly better at predicting probability of use for the test ranges (mean = 0.13, range 0.00 to 0.42, Table 4-8). Caracaras and Cattle Ranches I found a strong inverse relationship between the proportion of improved pasture and home range size (r= -0.57, P = 0.003, df = 23, n= 25, Figure 4-6). For pairs for which I had telemetry data and reproductive information for all 3 years (n = 1 8), the total number of young fledged during all 3 years was positively but not significantly correlated with the proportion of improved pasture in the home range (r= 0.20, P = 0.42, df = 16, Figure 4-7). Home ranges on the study area landscape are shown in Figure 4-8. Discussion In south-central Florida, breeding pairs of crested caracaras are distributed across the landscape in well-defined, year-round territories. Caracara home ranges are strongly associated with improved pasture habitats on cattle ranches. Little overlap of core areas among neighboring pairs suggests that breeding pairs are highly territorial near the nest. Food resources may be shared in areas distant from the nest, however, as indicated by overlapping 99% contours among neighboring pairs. This territorial dispersion pattern is typical of raptors. Many species are highly territorial around the nest site. In habitats containing adequate food resources, nests may occur close together and feeding areas may overlap (Schoener 1968, Newton 1979, Village 1982). Estimates of home range size varied considerably among breeding individuals. Despite this variation, results indicate substantial differences in overall estimates of home range size and nesting density from those reported by Layne (1996). Layne (1996) estimated home ranges of 1800-7400 ha based on maximum observed distances of 4.8 km to 9.7 km between sightings of unmarked individuals presumed to be members of the

PAGE 88

73 same pair. He also recorded maximum local density of 1 pair/ 1400 ha, but suggested that average density was much less. In this study, home ranges varied from approximately 500 to 2800 ha. Local density was 9 successfully breeding pairs within about 5000 ha, or approximately 1 pair/555 ha on MAERC and an adjacent ranch in Highlands County. This density is more than twice the maximum local density estimated by Layne (1996). The availability of food and nest sites generally limits breeding density (Village 1982, 1983). Typically, large raptors breed at lower densities than small raptors, which utilize a wider variety of prey species (Newton 1979). High local density of crested caracaras on MAERC may reflect habitat quality (availability of both food and nest sites) in that area. Although the density of caracaras at MAERC may not occur throughout the population's entire Florida range, results of this study suggest that at least in some areas, nesting densities may be substantially higher than previously thought for this population. A positive association between caracaras and improved pastures suggests that caracaras may benefit from land management activities routinely conducted on cattle ranches such as grazing, burning, mowing, and disking. Territory occupancy, reproductive activity, and nesting success were more consistent at territories on ranches. Pairs nesting on cattle ranches apparently produce more fledglings than pairs nesting on lands managed as natural areas. In general, however, comparisons between territories associated with different land uses were difficult because I had trouble locating caracaras on lands other than cattle ranches. Despite regular searching on public lands or other lands managed as natural areas, few territories on entirely native habitat were found. Movement patterns of crested caracaras reflects breeding commitments and temporal and spatial variation in food availability, particularly the amount and configuration of suitable foraging habitat within each home range. The availability of nest sites also influences the size and shape of the home range. In this study, few caracara ranges were circular or centered on the nest. On most cattle ranches, suitable nest trees typically are few in number and occur in clumps.. Home range size and shape primarily reflected the relationship between the nest site and improved pasture habitat (see map).

PAGE 89

74 Caracaras exhibit strong fidelity to a nest site and year-round home range. Management recommendations for raptors typically focus on protecting the nest site and a circular buffer around the nest in which development or management activities are limited (U.S. Fish and Wildlife Service 1987b). Suitable caracara nest trees are often clumped and associated with hammocks that do not constitute suitable foraging habitat. A circular buffer around a nest may not encompass areas required for all activities associated with reproduction and year-round survival of a breeding pair of caracaras. Management plans for this species must consider the arrangement as well as the amount of both nest sites and suitable foraging habitat. A landscape approach to identifying important habitats and assessing a species' distribution is particularly appropriate for raptors (Austin et al. 1996, Johansson et al 1994, Carey et al. 1992). Many species require large areas of land to support viable populations and are particularly sensitive to changes in land use (Marquiss et al. 1985, Carey et al. 1992). Conservation of such species requires consideration of land uses and management practices throughout entire regions. Landscape-level models that evaluate resource selection by animals can provide information about important habitats and about land uses compatible with species' survival. These models ultimately can be used to evaluate a species' distribution, delineate key protection areas, predict effects of future land use changes, and identify potential areas for habitat improvement or restoration. The models developed here represent an important first step in identifying habitat and landscape features important to crested caracaras in Florida. These models corroborate the observed association between breeding crested caracaras and improved pasture on cattle ranches. These models must be used with caution, however. While they showed a good fit statistically, their biological usefulness for management is currently limited, as indicated by the poor ability of the statistically "best" models to predict use of the 29 test ranges. Future model development would benefit from consideration of two current constraints. First, agricultural and grassland habitats were poorly classified on the study area map obtained from the Florida GAP Analysis Project. The resolution of the

PAGE 90

75 available GIS data did not permit detailed differentiation among these cover types or delineation of habitats within them such as patches of hammocks, wetlands, and ditches, which might be important to caracaras in selection of a territory. Second, these models did not consider context of habitats either within or around caracara home ranges. Habitat quality is determined not only by content but by the context of suitable habitats (Kushlan 1981, Janzen 1986, Fahrig and Merriam 1994, Harris et al 1996). Further research directed towards understanding relationships between caracaras and habitats should focus on refining cover types that represent open habitats and on considering habitat context both within the home range and at a larger scale.

PAGE 91

76 Table 4-1 Variables included in development of the logistic regression model describing caracara home ranges. Unless noted, all variables are expressed as proportions of the total area of the home range that was vegetated with natural or seminatural* vegetation. Variable Symbol oak forest OAKFOR pine forest PINEFOR cypress CYPRESS mixed forest MIXFOR saw palmetto SAWP oak shrub OAKSHRB hypericum shrub HYPSHRB woodland WOODL willow WILL wax myrtle WAXMYRT seasonally flooded native grassland or prairie GRASSL marsh MARSH improved pasture* IMPPAST proportion vegetated with natural or semi-natural vegetation PROPVEG proportion of open habitat (pasture + grassland) OPEN proportion of vegetated area in shrubs ALLSHRB total area of improved pasture* AMTP area of largest patch of contiguous improved pasture* CONTIGP

PAGE 92

77 Table 4-2. Landownership and major land use for crested caracara territories in southcentral Florida, 1994 1996. n = 65 territories. Landownership Major land use Ranching Other Natural Total Private 56(98.0%) 1 (2%) 0 57(87.7%) Public 3 (37.5%) 2 (2.5) 3 (37.5%) 8(12.3%) Total 59(90.8%) 3 (4.6%) 3 (4.6%) 65

PAGE 93

78 Table 4-3. Size and vegetation composition of crested caracara home ranges. Home ranges are defined as the 99% kernel contour estimated using a 0.85 smoothing factor. N = number of radiolocations used to estimate home range size. % of total home proportion of largest Home range in natural vegetated area area of Bird range or semi-natural in improved contiguous ID Sex N size (ha) vegetation pasture pasture (ha) 120 F 48 1265.273 0.858 0.597 492.342 121 M 26 2190.443 0.933 0.493 373.879 200 M 28 1251.969 0.973 0.406 350.991 280 M 68 551.039 0.531 0.831 227.695 360 F 68 1363.798 0.979 0.732 653.605 458 M 96 1879.477 0.836 0.663 722.800 479 M 43 2777.219 0.976 0.109 53.460 498 M 91 848.295 0.905 0.535 188.930 518 F 27 571.288 0.899 0.849 410.240 538 F 27 1014.189 0.917 0.287 125.636 609 F 67 994.957 0.843 0.741 596.455 689 F 37 736.403 0.986 0.759 526.300 969 M 72 1609.962 0.960 0.690 919.980 1110 M 81 1176.541 0.958 0.581 498.884 1379 F 25 1301.698 0.900 0.571 323.155 1469 M 71 1454.535 0.990 0.642 578.182 1490 M 42 659.221 0.989 0.765 400.091 1558 M 39 1129.249 0.937 0.632 522.410 1568 M 30 1083.789 0.942 0.522 213.750 1589 M 97 1604.085 0.924 0.577 323.356 1600 F 25 1685.152 0.865 0.717 803.179 1608 M 45 681.562 0.790 0.707 325.852 1711 M 38 859.928 0.720 0.691 279.356 1969 F 77 1072.958 0.946 0.937 788.290 1989 M 33 1537.786 0.821 0.709 817.595

PAGE 94

79 Table 4-4a. Matrix of percent overlap of 60% contours (core areas) for 9 adjacent pairs of breeding crested caracaras on MacArthur Agro-Ecology Research Center, Highlands County, south-central Florida. Birds in rows are overlapped by birds in columns. Bird ID 360 689 969 1469 1490 1558 1600 1608 1711 360 0 2.6 0 0 0 0 0 0 689 0 0 0 0 4.4 0 0 0 969 2.9 0 18.8 0 0 0 0 0 1469 0 0 21.9 0 0 0 0 0 1490 0 0 0 0 2.8 0 0 0 1558 0 3.7 0 0 1.5 0 0 0 1600 0 0 0 0 0 0 0 0 1608 0 0 0 0 0 0 0 0 1711 0 0 0 0 0 0 0 0 Table 4-4b. Matrix of percent overlap of 60% contours (core areas) for 4 adjacent pairs of breeding crested caracaras along the Kissimmee River, Highlands and Okeechobee counties, south-central Florida. Birds in rows are overlapped by birds in columns. Bird 498 120 1379 1568 498 0 0 0 120 0 0 0 1379 0 0 8.5 1568 0 0 7.5

PAGE 95

80 Table 4-5a. Matrix of percent overlap of 99% contours for 9 adjacent pairs of breeding crested caracaras on MacArthur Agro-Ecology Research Center, Highlands County, south-central Florida. Birds in rows are overlapped by birds in columns. Bird 360 689 969 1469 1490 1558 1600 1608 1711 360 0 33.1 12.4 6.2 0 9.7 0 0 689 0 0 0 25.9 21.4 3.1 1.9 14.6 969 27.2 0 38.8 7.6 4.1 0 0 0 1469 11.0 0 41.7 0 0 0 0 0 1490 13.6 31.5 19.8 0 24.8 23.0 0 0 1558 0 14.1 5.9 0 13.4 0 0 0 1600 7.7 1.4 0 0 8.3 0 0 1.5 1608 0 2.4 0 0 0 0 0 32.9 1711 0 12.6 0 0 0 0 2.9 22.7 Table 4-5b. Matrix of percent overlap of 99% contours for 4 adjacent pairs of breeding crested caracaras along the Kissimmee River, Highlands and Okeechobee counties, southcentral Florida. Birds in rows are overlapped by birds in columns. Bird 498 120 1379 1568 498 3.9 11.2 5.7 120 2.7 0 0 1379 7.3 0 22.4 1568 4.3 0 26.0

PAGE 96

81 Table 4-6. Parameter estimates for model produced using logistic regression analysis that best distinguished between known caracara home ranges and available units throughout the study area. Variables include pine forest (PINEFOR), mixed forest (MIXFOR), wax myrtle (WAXMYRT), woodland (WOODL), and contiguous pasture (CONTIGP). Parameter DF Coefficient Std. Err. p Constant 1 -3.04 0.69 0.000 PINEFOR 1 -0.06 0.04 0.087 MIXFOR 1 -0.22 0.12 0.062 WAXMYRT 1 0.10 0.05 0.053 WOODL 1 -0.17 0.07 0.012 CONTIGP 1 0.003 0.001 0.003

PAGE 97

82 Table 4-7. Parameter estimates for model that best distinguished between known caracara home ranges and available units throughout the study area and that did not include forest variables. Variables include native grassland/prairie (GRASSL), wax myrtle (WAXMYRT), improved pasture (IMPPAST), and contiguous pasture (CONTIGP). Parameter DF Coefficient Std. Err. P Constant 1 -9.18 1.31 0.000 GRASSL 1 0.04 0.02 0.069 WAXMYRT 1 0.19 0.06 0.000 IMPPAST 1 0.08 0.02 0.000 CONTIGP 1 0.002 0.001 0.090

PAGE 98

83 Table 4-8. Predicted probabilities of use for test caracara home ranges calculated using logistic regression models that best distinguished between caracara home ranges and available units throughout the study area. Model that included forest Model with forest and Nest No. and woodland variables woodland variables omitted 1 0.000 0.019 2 0.021 0.010 3 0.047 0.052 4 0.256 0.349 5 0.211 0.391 6 0.189 0.172 7 0.096 0.061 8 0.270 0.356 9 0.132 0.351 10 0.015 0.019 11 0.049 0.067 12 0.006 0.001 13 0.151 0.116 14 0.161 0.142 15 0.017 0.013 16 0.042 0.015 17 0.194 0.292 1 8 0 314 0 418 19 0.056 0.128 20 0.113 0.125 21 0.009 0.033 22 0.000 0.001 23 0.011 0.001 24 0.175 0.154 25 0.122 0.130 26 0.068 0.123 27 0.021 0.021 28 0.045 0.087 29 0.036 0.243 Mean 0.098 0.134

PAGE 99

84 Figure 4-1. Study area for evaluating landscape-level habitat selection by breeding pairs of crested caracaras in Florida, 1994 -1996.

PAGE 100

85 •a o S & •a u J i l c/3 a i 2 o c o ^11-1 1 1 3 g o V 0 js u o 1 1 o .2
PAGE 101

86 o\ \C rn suoijBooioipBJ jo jaquin^

PAGE 102

87 o>oonomoino ooooooooo suoijbooi jo uoiyodojj

PAGE 103

88 s§uured jo uoiyodoij

PAGE 104

89 o o rd i II k o o o o O o NC o 4o o o m o o o o o n o o o o o o o c rn o o o o o u £ o u M cd O, T3 > O E | O c o o p N o E o -o E o G NO On 0) ON Cl, ON i ? < .2 S fa 1 O t: _c 0 -5 9I 2 & d c £ bo ex 1 ja 1 g § o NO i3 _L *T *> I? I I 00 m (ell) azis aSirej auioH

PAGE 105

90 rs II ft, so r<-) ci o o o OS o oo o ro SO o o o o O o o o o I a 1 c a c3 *o u _o £ E 0 J3 cd g V fi i 1 I o* n C3 > o o CO .1 § O £ 4-> v. G U d Q <2 vo" OS os a, os 00 C S3 I I c2 T3 00 T3 U oa 0X) .SP § sjbsX £ §uunp paSpsg SunoX jo jaqumu pjox

PAGE 106

91

PAGE 107

CHAPTER 5 SYNTHESIS AND CONCLUSIONS The Florida population of the crested caracara is isolated and remote from other caracara populations. Because it is relatively small (estimated at 400 500 individuals) and yet is a characteristic element of the Florida prairie ecosystem, the species is of both conservation and evolutionary interest. Conversion of grassland and pasture habitats is occurring rapidly throughout the region of Florida that constitutes much of the species' current breeding range in that state. Florida's caracara population is listed as Threatened by the U.S. Fish and Wildlife Service and the Florida Game and Fresh Water Fish Commission. Currently, no habitat protection guidelines or management activities focus on conservation of this population. This is largely due to the paucity of ecological information on the species in Florida or elsewhere. I undertook this research with the intent of expanding the basic biological knowledge of this raptor and evaluating the status of the Florida population. This study constitutes the first in-depth examination of the ecology of the crested caracara in Florida, or anywhere throughout its geographic range. I examined various aspects of the caracara' s biology that might be pertinent to management of Florida's population and attempted to understand relationships between caracaras and privately-owned cattle ranches, where they now mostly persist. I found that Florida's crested caracaras differ markedly from other, better studied raptors. Crested caracaras in Florida have relatively high nest success and productivity, both compared to most large raptors and in contrast to previously reported estimates for this population. The reproductive strategy of Florida's crested caracaras is typical of other sedentary tropical bird populations. The breeding cycle is closely tied to the annual precipitation cycle. A fairly stable climate, the absence of migration, and, perhaps, an abundance of food in human-influenced habitats facilitate a long breeding season. Spatial 92

PAGE 108

93 and temporal variability in breeding activity exhibited by individual pairs may be a response to variable environmental conditions both within and between breeding seasons. High nest success and productivity is indicated by high territorial occupancy and breeding rates coupled with the ability of certain pairs to make multi-nesting attempts. Pairs that successfully fledge two broods in one breeding season may realize higher lifetime reproductive success than single-brooded pairs. Although little historic information is available, reproductive rates for the sample population seem adequate for long-term population maintenance. Florida's caracara population appears to be stable at present. Caution is warranted in evaluating trends, however. Difficulties with access to many areas of suitable habitat, particularly away from major roads, preclude a completely systematic search of the species' Florida range. Counts are hampered by the effects of the birds' terrestrial behavior on their detectability. Because caracaras are long-lived and site-faithful, a significant time lag could occur before any decline is detected. Results of my research indicate a strong association between breeding pairs of crested caracaras and cattle ranches. The crested caracara is a generalist predator that seems to be tolerant of some human activity and moderate habitat changes as long as they are not harassed and the nest site is left intact. Certain pairs are quite tolerant of smallscale burning, clearing, removal of sod, building construction, and general human activity, even during the nesting season. Normal levels of ranching activities are welltolerated. Indeed, these activities may have aided in maintaining suitable caracara habitat on private ranchlands. Pasture management programs include controlled fires set during winter and spring. Fires often sweep through active nest stands with little effect. Estimates of spatial requirements and nesting density for Florida's crested caracaras differed considerably from previously published estimates. I found caracaras nesting as close together as 0.5 km on the cattle ranches. Local nesting density on a cattle ranch in Highlands County was over twice that reported by Layne (1996). These results support the hypothesis that cattle ranches provide good quality habitat for caracaras. Similar densities of breeding pairs may occur on other large ranches where I did not

PAGE 109

94 survey for territories. These findings suggest that (1) it is unlikely that the Florida population was reduced to < 100 birds between 1967 and 1970 (Heinzmann 1970); (2) breeding densities of Florida's crested caracaras may be higher than previously suspected in some areas of their current breeding range; and (3) total population numbers may be higher than published estimates. Caracara territories were not randomly distributed throughout the study area. Logistic regression analysis revealed a strong association between nesting territories and improved pasture and particularly, with large areas of contiguous pasture. Management activities on ranches such as grazing, burning, plowing, and mowing influence vegetation structure. The short ground cover vegetation on these ranches may provide suitable conditions for this mainly terrestrial forager. Development of a reliable predictive model for suitable caracara habitat was constrained by the resolution of the available GIS data, however, particularly by the inability to identify ditches and wetland habitats within the pastures. Future model development should consider improving differentiation among open grassland habitats as well as habitat context both within caracara home ranges and at a larger, landscape scale. My research indicates that Florida's caracaras may be doing better than previously believed, especially on privately-owned cattle ranches. If the population is in a long-term decline, factors other than reproduction are likely involved. The greatest threat continues to be loss and degradation of nesting habitat. Observations made during this study suggest that most suitable habitat may be occupied. I found caracara pairs in most areas where I looked for them, particularly on private ranches. Urban development continues to expand inland from the coast and off Florida's central ridge. The accelerating economic value of citrus and sugar cane favors conversion of remaining prairies and grasslands and these pressures are expected to continue. On some ranches, removal of palms has precluded the use of these areas as nesting sites. Although the short-term outlook for Florida's caracaras appears positive, the long-term outlook is less clear. A decline seems inevitable if habitat conversion continues at the present rate throughout the caracara' s limited range.

PAGE 110

95 A conservation strategy for this population should be based on protecting known nest sites and large areas of open grassland habitats, particularly pasture. Such a strategy should focus at a regional scale, and include nesting sites on both public and private lands. Because of the variation in context and levels of human disturbance among caracara nest sites, each site may have to be treated separately in management planning. Private ownership of most suitable habitat will complicate conservation efforts. Because caracaras seem to be doing well on cattle ranches, efforts being made in some areas to restore native prairies may target species that require these native habitats. While beneficial for prairie-dependent species, these decisions may preclude management for caracaras on public lands. Habitat conservation must incorporate cooperative efforts among land management agencies and private landowners. Developing means to preserve the large cattle ranches that now apparently support the bulk of this population is critical to conservation of Florida's crested caracaras. These efforts will of necessity include creative alternatives to direct land purchase. A regional focus to conservation mandates obtaining better information about the distribution of nesting territories throughout the region. Future research should focus on development of better predictive models of the species' distribution and improving the understanding of relationships between caracaras and cattle ranches. Management efforts must also incorporate a strong education component, to reduce the widespread negative impression of scavengers and to stress the value of the crested caracara as an example of the potential for coexistence of sustainable economic activities and native wildlife. This study represents an important beginning towards understanding the population dynamics, demographics, and habitat ecology of Florida's crested caracara population. As with any species that is long-lived and ranges widely across habitats, however, additional study is required to fully understand the role of the caracara in the Florida landscape.

PAGE 111

LITERATURE CITED Abrahamson, W.G. and D.C. Hartnett. 1990. Pine flatwoods and dry prairies. Pages 103149 in Ecosystems of Florida, R.L. Myers and J.J. Ewel, ed. Univ. Central Florida Press, Orlando. Akaike, H. 1985. Prediction and Entropy. Pages 1-24 in A.C. Atkinson and S.E. Fienberg, eds. A Celebration of Statistics. The ISI Centenary volume. Springer Verlag, New York, NY. Amadon, D. 1964. The evolution of low reproductive rates in birds. Evolution 18:105110. Askenmo, C, and U. Unger. 1986. How to be double-brooded: trends and timing of breeding performance in the rock pipit. Ornis Scandinavica 17: 237-244. Askins, R.A., J.F. Lynch, and R. Greenburg. 1990. Population declines in migratory birds in eastern North America. Curr. Ornithol. 7:1-57. Austin, G.E., C.J. Thomas, D.C. Houston, and D.B. Thompson. 1996. Predicting the spatial distribution of buzzard Buteo buteo nesting areas using a Geographical Information System and remote sensing. J. Appl. Ecol. 33:1541-1550. Babbitt, K. 1996. Tadpoles and predators: patterns in space and time and the influence of habitat complexity on their interactions. Ph.D. diss., University of Florida, Gainesville. Bean, M.J. and D.S. Wilcove. 1997. The private land problem. Conserv. Biol. 11:1-2. Bednarz, J.C. 1987. Successive nesting and autumnal breeding in Harris' Hawks. Auk 104: 85-96. Bent, A.C. 1938. Life histories of North American birds of prey. Part 2. U.S. Natl. Mus. Bull. No. 167. Brodkorb, P. 1964. Catalogue of Fossil Birds: Pt. 2 (Anseriformes through Galliformes). Bull. Florida State Mus. 8: 195-335. Burnham, K.P. and D.R. Anderson. 1992. Data-based selection of an appropriate biological model: the key to modern data analysis. Pages 16-30 in D.R. McCullough and R.H. Barrett, eds. Wildlife 2001: Populations. Elsevier, London, UK. 96

PAGE 112

97 Carey, A.B., S.P. Horton, and B.L. Biswell. 1992. Northern spotted owls: influence of prey base and landscape character. Ecol. Monogr. 62:223-250. Cave, A.J. 1968. The breeding of the Kestrel, Falco tinnunculus, in the reclaimed area Oostelijk Flevoland. Netherlands J. Zool. 18:313-407. Chandler, R.M., and L.D. Martin. 1991 A record for caracara from the late Pleistocene of Nebraska. Cur. Res. Pleistocene 8: 87-88. Chen, E., and J.F. Gerber. 1990. Climate. Pages 1 1-34 in R.L. Myers and J.J. Ewel, eds. Ecosystems of Florida. Univ. Central Florida Press, Orlando. Cody, M.L. 1971. Ecological aspects of reproduction. Pages 462-512 in D.S. Farner and J.R. King, eds. Avian Biology Vol. 1. Academic Press, New York, NY. Cox, D.R., and D. Oakes. 1984. Analysis of Survival Data. Chapman and Hall, New York, NY. Daan, S., C. Dijkstra, R. Drent, and T. Meijer. 1989. Food supply and the annual timing of avian reproduction. Pages 392-407 in H. Ouellet, ed. Acta XIX Congr. Inter. Ornithol. Vol. I. Univ. Ottawa Press, Ottawa, Canada. Davis, J.H. 1967. General map of natural vegetation of Florida. Agric. Exp. Sta., Inst. Food Agric. Sci., Univ. Florida, Circular S-178, Gainesville. De Vries, T.J. 1975. The breeding biology of the Galapagos Hawk, Buteo galapagoensis. Gerfaut 65: 29-57. del Hoyo, J., A. Elliot, and J. Sargatal (eds.). 1994. Handbook of Birds of the World, Vol. 2. New World Vultures to Guineafowl. Lynx Edicions, Barcelona, Spain. den Boer Hazewinkel, J. 1987. On the costs of reproduction: parental survival and production of second clutches in the Great Tit. Ardea 75:99-1 10. Dickinson, V.M., and K.A. Arnold. 1996. Breeding biology of the Crested Caracara in south Texas. Wilson Bull. 108: 516-523. Dijkstra, C, L. Vuursteen, S. Daan, and D. Masman. 1982. Clutch size and laying date in the Kestrel Falco tinnunculus: effect of supplementary food. Ibis 124:210-213. Drury, W.H. 1961. Studies of the breeding biology of the Horned Lark, Water Pipit, Lapland Longspur, and Snow Bunting on Bylot Island, Northwest Territories, Canada. Bird-Banding 32: 1-44. Dunning, J.G. Jr., R. Borgella, Jr., K. Clements, and G.K. Meffe. 1995. Patch isolation, corridor effects, and colonization by a resident sparrow in a managed pine woodlands. Conserv. Biol. 9:542-550.

PAGE 113

98 Environmental Sciences Research Institute. 1996. ArcINFO Geographic Information System software, Version 7.0.4. Redlands, CA. Ellis, D.H., D.G. Smith, W.H. Waley, and C.H. Ellis. 1988. Crested Caracara. Pages 119126 in Southwest Raptor Management Symposium and Workshop, Natl. Wildl. Fed. Sci. Tech. Ser. No. 1 1, Washington DC. Evans Ogden, L.J., and B.J. Stutchbury. 1996. Constraints on double brooding in a Neotropical migrant, the Hooded Warbler. Condor 98:736-744. Fahrig, L. and G. Merriam. 1994. Conservation of fragmented populations. Conserv. Biol. 8(l):50-59. Fargallo, J.A., G. Blanco, and E. Soto-Largo. 1996. Possible second clutches in a Mediterranean montane population of the Eurasian Kestrel {Falco tinnunculus). J. Raptor Res. 30: 70-73. Forsman, E.D., A. Giese, D. Manson, S. Sovern, and D.R. Herter. 1995. Renesting by spotted owls. Condor 97:1078-1080. Funderberg, J.B. and G. Heinzman. 1967. Status of the caracara in Florida. Florida Nat. 40(4): 150-151. Goodman, L.A. 1984. The Analysis of Cross-Classified Data Having Ordered Categories. Harvard Univ. Press, Cambridge, MA. Guepel. G.R., and D.F. DeSante. 1990. Incidence and determinants of double brooding in Wrentits. Condor 92: 67-75. Hagan, J.M. Ill, and D.W. Johnston, editors. 1992. Ecology and Conservation of Neotropical Migrant Landbirds. Smithsonian Inst. Press, Washington, DC. Harris, L.D., T.S. Hoctor, and S.E. Gergel. 1996. Landscape processes and their significance to biodiversity conservation. Pages 319-347 in O.E. Rhodes, Jr., R.K. Chesser, and M.H. Smith. Population Dynamics in Ecological Space and Time. Univ. Chicago Press, Chicago, IL Harshberger, J.W. 1914. The vegetation of south Florida. Trans. Wagner Free Inst. Sci. Philos. 3:51-189. Haukioja, E. 1970. Clutch size of the Reed Bunting Emberiza schoeniclus. Ornis Fennica 47:101-135. Heinzman, G. 1970. The caracara survey: A four year report. Florida Nat. 43:149. Hensler, G.L., and J.D. Nichols. 1981. The Mayfield method of estimating nesting success: A model, estimators, and simulation results. Wilson Bull. 93: 42-53.

PAGE 114

99 Herkert, J.R. 1994. The effects of habitat fragmentation on midwestern grassland bird communities. Ecological Applications 4(3):461-471. Hochachka, W. 1990. Seasonal decline in reproductive performance of Song Sparrows. Ecology 71: 1279-1288. Hogstedt, G. 1980. Evolution of clutch size in birds: Adaptive variation in relation to territory quality. Science 210:1 148-1 150. Holmes, R.T., T.W. Sherry, and F.W. Sturges. 1991. Numerical and demographic responses of temperate forest birds to annual fluctuations in their food resources. Pages 1559-1567 in H. Ouellet, ed. Acta XX Congr. Int. Ornithol. Vol. II. Univ. Ottawa Press, Ottawa, Canada. Holmes, R.T., T.W. Sherry, P.P. Marra, and K.E. Petit. 1992. Multiple brooding and productivity of a Neotropical migrant, the Black-throated Blue Warbler (Dendroica caerulescens), in an unfragmented temperate forest. Auk 109: 321-333. Hosmer, D.W. and S. Lemeshow. 1989. Applied Logistic Regression. Wiley and Sons, New York, NY. Howard, H., and A.H. Miller. 1933. Bird remains from cave deposits in New Mexico. Condor 35(1): 15-18. Howell, A. 1932. Florida Birdlife. J.J. Little and Ives. Co., New York, NY. Immelmann, K. 1971. Ecological aspects of periodic reproduction. Pages 342-391 in D.S. Farner and J.R. King, eds. Avian Biology, Vol. 1 Academic Press, New York, NY. Jandel Scientific. 1995. SigmaStat version 2.0. Jandel Scientific, San Rafael, CA. Janzen, D. 1986. The eternal external threat. Pages 286-303 in M.E. Soule, ed. Conservation Biology. Sinauer Associates, Inc., Sunderland, MA. Johansson, C, P.J. Hardin, and CM. White. 1994. Large-area goshawk habitat modeling in Dixie National Forest using vegetation and elevation data. Stud. Avian Biol. 16:50-57. Johnson, D.H. 1979. Estimating nest success: The Mayfield method and an alternative. Auk 96:651-661. Kale, H.W. II, B. Pranty, B. Stith, and C.W. Biggs. 1992. The atlas of the breeding birds of Florida. Final report to the Florida Game and Fresh Water Fish Comm., Tallahassee. Kennedy, P.L., J.M. Ward, G.A. Rinker, and J. A. Gessamen. 1994. Post-fledging areas in Northern Goshawk home ranges. Stud. Avian Biol. 16: 85-82.

PAGE 115

100 Kenward, R.E. and K.H. Hodder. 1996. RANGESV. An Analysis System for Biological Location Data. Inst. Terr. Ecol.. Furzebrook Res. Sta., Wareham, Dorset, UK. Klimkiewicz, M.K., and A.G. Fletcher. 1989. Longevity records of North American birds. J. Field Ornithol. 60, Supplement 1 : 477. Klomp, H. 1970. The determination of clutch size in birds: A review. Ardea 58:1-124. Korpimaki, E. and M. Lagerstrom. 1988. Survival and natal dispersal of fledglings of Tengmalm's Owl in relation to fluctuating food conditions and hatching date. J. Anim. Ecol. 57:433-441. Kostrzewa, A. 1996. A comparative study of nest-site occupancy and breeding performance as indicators for nesting habitat quality in three European raptor species. Ethol. Ecol. Evol. 8:1-18. Kushlan, J. A. 1981. Design and management of continental wildlife reserves: Lessons from the Everglades. Biol. Conserv. 15:281-290. Kushlan, J. A. 1990. Freshwater Marshes. Pages 324-363 in R. L. Myers and J. J. Ewel, eds.. Ecosystems of Florida. Univ. Central Florida Press, Orlando. Lack, D. 1954. The natural regulation of animal numbers. Oxford Univ. Press, London, UK Lack, D. 1966. Population studies of birds. Oxford Univ. Press, London, UK. Lack, D. 1968. Ecological adaptations for breeding in birds. Methuen and Co., Ltd., London, UK. LaHaye, W., R.J. Gutierrez, and D.R. Call. 1992. Demography of an insular population of spotted owls (Strix occidentalis occidentalis). Pages 803-814 in D. R. McCullough and R. H Barrett, eds. Wildlife 2001: Populations. Elsevier Publishers, London, UK. Lawton, J.H. 1993. Range, population abundance, and conservation. Trends Ecol. Evol. 8:409-413. Layne, J.N. 1978. Audubon's Caracara. Pages 34-35 in Rare and endangered biota of Florida. Vol. 2: Birds, H.W. Kale, II, ed., Univ. Press of Florida, Gainesville. Layne, J.N. 1995. Audubon's Crested Caracara in Florida. Pages 82-83 in E. T. LaRoe, G. S. Farris, C. E. Puckett, P. D. Doran, and M. J. Mac, eds. Our Living Resources: A Report to the Nation on the Distribution, Abundance, and Health of U.S. Plants, Animals, and Ecosystems. U.S. Dep. Int. Nat. Biol. Serv., Washington DC.

PAGE 116

101 Layne, J.N. 1996. Crested Caracara. Pages 197-210 in J. Rogers, Jr., H.W. Kale II, and H. Smith, eds. Rare and Endangered Biota of Florida, Vol. 5: Birds. Univ. Press of Florida, Gainesville. Lens, L. and A. A. Dhondt. 1994. Effects of habitat fragmentation on the timing of Crested Tit Parus cristatus natal dispersal. Ibis 136:147-152. Lesica, P. and F.W. Allendorf. 1995. When are peripheral populations valuable for conservation? Conserv. Biol. 9(4):753-760. Levin, D.A. 1970. Development instability and evolution in peripheral isolates. Am. Nat. 104:343-353. Levy, S.H. 1988. Status and distribution of the Crested Caracara in Arizona. Unpubl. Report to Ariz. Game and Fish Dept., Phoenix. Lewis, G. 1995. Private property rights. The conflict and the movement. J. For. 93:25-26. Mader, W.J. 1975. Biology of the Harris' Hawk in southern Arizona. Living Bird 14: 5984. Malherbe, A. P. 1963. Notes on birds of prey and some others at Boshoek north of Rustenburg during a rodent plague. Ostrich 34:95-96. Manly, B.F., L.L. McDonald, and D.L. Thomas. 1993. Resource selection by animals. Chapman and Hall, London, UK. Marquiss, M., D.A. Ratcliffe, and R. Roxburgh. 1985. The numbers, breeding success and diet of golden eagles in southern Scotland in relation to changes in land use. Biol. Conserv. 34:121-140. Marzluff, J.M., M.S. Vekasy, and C. Coody. 1994. Comparative accuracy of aerial and ground telemetry locations of foraging raptors. Condor 96:447-454. Matthysen, E. and D. Currie. 1996. Habitat fragmentation reduces disperser success in juvenile nuthatches Sitta europaea: evidence from patterns of establishment. Ecography 19:67-72. Mayfield, H.F. 1961. Nesting success calculated from exposure. Wilson Bull. 73: 255261. McKelvey, K., B.R. Noon, and R.H. Lamberson. 1993. Conservation planning for species occupying fragmented landscapes: The case of the Northern spotted owl. Pages 424-450 in P. M. Kareiva, J. G. Kingsolver, and R. B Huey, eds. Biotic Interactions and Global Change. Sinauer Associates, Sunderland, MA. Mech, L.D. 1983. Handbook of Animal Radiotracking. Univ. of Minnesota Press, Minneapolis.

PAGE 117

102 Menard, S. 1995. Applied Logistic Regression Analysis. Sage Publications, Inc. Thousand Oaks, CA. Millsap, B.A. 1989. Final progress report, Crested Caracara population survey. NonGame Wildlife Program, Florida Game and Fresh Water Fish Comm. Tallahassee. 8 pp. Millsap, B.A., and C. Bear. 1990. Double brooding by Florida burrowing owls. Wilson Bull. 102: 313-317. Millsap, B.A., J.A. Gore, D.E. Runde, and S.I. Cerulean. 1990. Setting priorities for the conservation of fish and wildlife species in Florida. Wildl. Monogr. No. 1 1 1 57 pp Morrison, J.L. 1996. Crested Caracara {Caracara plancus). No. 249. in The Birds of North America. A. Poole and F. Gill, eds. Acad. Nat. Sci., Philadelphia, and Am. Ornithol. Union, Washington, DC. Morrison, J.L. and S.M. McGehee. 1996. Capture methods for Audubon's Crested Caracara. J. Field Ornithol. 67(4):630-636. Morrison, J. L. and M. Maltbie. in review. Methods for gender identification of Crested Caracaras. Auk. Murphy, E.G. 1978. Seasonal variation in reproductive output of House Sparrows: the determination of clutch size. Ecology 59(6): 1 189-1 199. Myers, R.L. and J.J. Ewel, editors. 1990. Ecosystems of Florida. Univ. Central Florida Press. Orlando. Neill, W.T. 1957. Historical biogeography of present day Florida. Bull. Florida State Mus. 2:175-222. Newton, I. 1979. Population Ecology of Raptors. Buteo Books, Vermillion, SD. Newton, I., editor. 1989. Lifetime reproduction in birds. Academic Press, London, UK. Newton, I., and M. Marquiss. 1976. Occupancy and success of nesting territories in the European Sparrowhawk. Raptor Res. 10:65-71. Newton, I. and M. Marquiss. 1981. Effect of additional food on laying dates and clutch sizes of sparrowhawks. Ornis Scandinavica 12:224-229. Newton, I., and M. Marquiss. 1984. Seasonal trend in the breeding performance of Sparrowhawks. J. Anim. Ecol. 53:809-829. Newton, I., and M. Marquiss. 1986. Population regulation in Sparrowhawks. J. Anim. Ecol. 55:463-480.

PAGE 118

103 Nicholson, D.J. 1929. The Audubon Caracara, a little known bird in Florida. Florida Nat. 2:67-69 Nicholson, D.J. 1930. Notes from Florida. Auk 57:267. Nur, N. 1988. The cost of reproduction in birds: An examination of the evidence. Ardea 76: 155-168. Oberholser, H.C. 1974. The Bird Life of Texas. Univ. Texas Press, Austin. Opdam, P., G. Rijsdijk, and F. Hustings. 1985. Bird communities in small woods in an agricultural landscape: Effects of area and isolation. Biol. Conserv. 34:333-352. Ormerod, S.J., and S.J. Tyler. 1993. The adaptive significance of brood size and time of breeding in the Dipper Cinclus cinclus (Aves: Passeriformes) as seen from postfledging survival. J. Zool. London 231:371-381. Perrins, CM. 1970. The timing of birds' breeding seasons. Ibis 1 12:242-255. Phelps, F.M. 1912. A March list from the Caloosahatchee River and Lake Okeechobee. Wilson Bull. 24: 117-125. Pollock, K.H., S.R. Winterstein, CM. Bunck, and P.D. Curtis. 1989. Survival analysis in telemetry studies: The staggered entry design. J. Wildl. Manage. 53: 7-15. Redpath, S.M. 1995. Impact of habitat fragmentation on activity and hunting behavior in the tawny owl, Strix aluco. Behav. Ecol. 6(4):410-415. Resampling Stats Inc. 1995. Resampling Stats Version 4.0.1. Resampling Stats Inc., Arlington, Virginia. Richmond, A.R. 1976. Feeding of nestlings by the Caracara in Costa Rica. Wilson Bull. 88: 667. Ricklefs, R.E. 1973. Fecundity, mortality, and avian demography. Pages 366-435 in D.S. Farner, ed. Breeding Biology of Birds. Natl. Acad. Sci., Washington DC. Ricklefs, R.E. 1984. The optimization of growth rate in altricial birds. Ecology 65: 16021616. Ricklefs, R.E. and G. Bloom. 1977. Components of avian breeding productivity. Auk 94:86-96. Rivera-Rodriguez, L.B., and R. Rodriguez-Estrella. 1992. Kleptoparasitism and other interactions of Crested Caracara in the Cape region, Baja California, Mexico. J. Field Ornithol. 63(2): 177-180.

PAGE 119

104 Robinson, S.K., F.R. Thompson III, T.M. Donovan, D.R. Whitehead, and J. Faaborg. 1995. Regional forest fragmentation and the nesting success of migratory birds. Science 267:1987-1990. Rodriguez-Estrella, R. 1996. Response of Common Black Hawks and Crested Caracara to human activities in Mexico. Pages 355-363 in D. Bird, D. Varland, and J.J. Negro, eds. Raptors in Human Landscapes: Adaptations to Built and Cultivated Environments. Academic Press, London, UK. SAS Institute, 1995. SAS User's Guide. SAS Inst. Cary, NC. Schoener, T.W. 1968. Sizes of feeding territories among birds. Ecology 49:123-141. Scott, W.E. 1886. On the avifauna of Pinal County, with remarks of some birds of Pima and Gila Counties, Ariz. Auk 3: 421-432. Seamen, D.E. and R.A. Powell. 1996. An evaluation of the accuracy of kernel density estimators for home range analysis. Ecology 77:2075-2085. Shaffer, M.L. 1981. Minimum population sizes for species conservation. Bioscience 31(2):131-134. Sjogren, P. 1991. Extinction and isolation gradients in metapopulations: the case of the pool frog (Rana lessonae). Biol. J. Linn. Soc. 42:135-147. Skutch, A.F. 1950. The nesting seasons of Central American birds in relation to climate and food supply. Ibis 92:185-222. Smeenk, C, and N. Smeenk-Enserink. 1975. Observations on the Pale Chanting Goshawk Melierax poliopterus, with comparative notes on the Gabar Goshawk Micronisus gabar. Ardea 63:93-1 15. Smith, J.N., and D.A. Roff. 1980. Temporal spacing of broods, brood size, and parental care in Song Sparrows {Melospiza melodia). Can. J. Zool. 58: 1007-1015. Smith, R.D., and M. Marquiss. 1995. Production and costs of nesting attempts in Snow Buntings Plectrophenax nivalis: Why do they attempt second broods? Ibis 137: 469-476. Sokal, R.R., and F.J. Rohlf. 1995. Biometry, 3rd ed. W. H. Freeman and Co., New York, NY. Sprunt, A., Jr. 1954. Audubon's caracara. Florida Nat. 27: 99-101, 119. Stahlecker, D.W., and H.J. Griese. 1977. Evidence of double brooding by American Kestrels in the Colorado high plains. Wilson Bull. 89: 618-619. Stearns, S.C. 1992. The Evolution of Life Histories. Oxford Univ. Press. Oxford, UK.

PAGE 120

105 Steenhof, K. 1987. Assessing raptor reproductive success and productivity. Pages 157170 in B.A. Pendleton, B.A. Millsap, K.W. Cline, D.M. Bird, eds. Raptor Management Techniques Manual. Natl. Wildl. Fed. Inst. Wildl. Res., Sci. Tech. Ser. No. 10. Washington DC Steenhof, K. and M.N. Kochert. 1982. An evaluation of methods used to estimate raptor nesting success. J. Wildl. Manage. 46(4):885-893. Stevenson, H.M., and B.H. Anderson. 1994. The birdlife of Florida. Univ. Press of Florida, Gainesville. Swihart, R.K. and N.A. Slade. 1985. Testing for independence of observations in animal movements. Ecology 66(4): 1 1 761 1 84. Thomas, D.L. and E.J. Taylor. 1990. Study designs and tests for comparing resource use and availability. J. Wildl. Manage. 54:322-330. Tinbergen, J.M., and J.H. van Balen. 1988. Food and multiple breeding. Pages 380-391 in XIX Congr. Inter. Ornithol. H. Ouellet, ed. Ottawa, Ontario, Canada. 1986. Natl. Mus. of Nat. Sci., Ottawa. Toland, B.R. 1985. Double brooding by American Kestrels in central Missouri. Condor 87: 434-436. U.S. Fish and Wildlife Service. 1987a. Endangered and threatened wildlife and plants; threatened status for the Florida population of the Audubon's crested caracara. Final Rule. Fed. Reg. 52:25229-25231. U.S. Fish and Wildlife Service. 1987b. Habitat management guidelines for the bald eagle in the southeast region. 3rd revision. Atlanta, GA. U.S. Fish and Wildlife Service. 1989. Recovery plan for the Florida population of Audubon's crested caracara. Prepared for Southeast Region, Jacksonville, FL. van Riper, C. III. 1995. Ecology and breeding biology of the Hawaiian Elepaio (Chasiempsis sandwichensis bryani). Condor 97:512-527. Verhulst, S. and J.M. Tinbergen. 1991. Experimental evidence for causal relationship between timing and success of reproduction in the Great Tit, Parus m. major. J. Anim. Ecol. 60:269-282. Verhulst, S., and R. Hut. 1996. Post-fledging care, multiple breeding and the costs of reproduction in the Great Tit. Anim. Behav. 5 1 : 957-966. Verhulst, S., J.H. van Balen, and J.M. Tinbergen. 1995. Seasonal decline in reproductive success of the Great Tit: Variation in time or quality? Ecology 76(8):2392-2403.

PAGE 121

106 Village, A. 1982. The home range and density of kestrels in relation to vole abundance. J. Anim. Ecol. 51:413-428. Village, A. 1983. The role of nest-site availability and territorial behaviour in limiting the breeding density of kestrels. J. Anim. Ecol. 52:635-645. von Haartman, L. 1971. Population dynamics. Pages 391-459 in D.S. Farner and J.R. King, eds. Avian Biology, Vol. 1 Academic Press, New York, NY. Voous, K.H. 1983. Birds of the Netherlands Antilles. De Walberg Pers. Vuilleumier, F. 1985. Fossil and recent avifaunas and the Interamerican Interchange. Pages 387-424 in F.G. Stehli and S.D. Webb, eds. The Great American Biotic Interchange. Plenum Press, New York, NY. Webb, S.D. 1977. A history of savanna vertebrates in the New World, Part 1 : North America. Ann. Rev. Ecol. Syst. 8:355-380. Webb, S.D. and K.T. Wilkins. 1984. Historical biogeography of Florida Pleistocene mammals. Spec. Publ. Carnegie Mus. 8:370-383. Westmoreland, D., L.B. Best, and D.E. Blockstein. 1986. Multiple brooding as a reproductive strategy: Time-conserving adaptations in Mourning Doves. Auk 103: 196-203. 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 forests. Pages 125-205 in R.L. Burgess and D.M. Sharpe, eds. Forest Island Dynamics in Man-Dominated Landscapes. SpringerVerlag, New York, NY. White, G.C. and R.A. Garrott. 1990. Analysis of wildlife radiotracking data. Academic Press, New York, NY. 383 pp. Wilcove, D.S., C.H. McLellan, and A.P. Dobson. 1984. Habitat fragmentation in the temperate zone. Pages 237-256 in M.E. Soule, ed. Conservation Biology: the Science of Scarcity and Diversity. Sinauer Associates, Sunderland, MA. Wilcox, B.A. and D.D. Murphy. 1985. Conservation strategy: The effects of fragmentation on extinction. Am. Nat. 125:879-888. Williams, G.C. 1966. Natural selection, the costs of reproduction, and a refinement of Lack's principle. Am. Nat. 100: 687-690. Witmer, M.C., and R.E. Patrick. 1987. Continuous breeding at a Barn Owl nest in Texas. Southwest. Nat. 32: 402-403. Wood, D.A. 1990. Official lists of endangered and potentially endangered fauna and flora in Florida. Florida Game and Freshwater Fish Comm. Tallahassee. 19 pp.

PAGE 122

107 Wood, P.B., and M.W. Collopy. 1993. Effects of egg removal on Bald Eagle productivity in northern Florida. J. Wildl. Manage. 57:1-9. Woolfenden, G.E. and J.W. Fitzpatrick. 1984. The Florida Scrub Jay. Demography of a Cooperative-Breeding Bird. Princeton Univ. Press, Princeton, NJ. Worton, B. J. 1989. Kernel methods for estimating the utilization distribution in homerange studies. Ecology 70:164-168. Worton, B.J. 1995. Using Monte Carlo simulation to evaluate kernel-based home range estimators. J. Wildl. Manage. 59:794-800. Wray, S., W.J. Cresswell, P.C. White, and S. Harris. 1992. What, if anything, is a core area? An analysis of the problems of describing internal range configurations. Pages 256-271 in I. G. Priede and S.M. Swift, eds. Wildlife Telemetry: Remote Monitoring and Tracking of Animals. Ellis Harwood, New York, NY

PAGE 123

BIOGRAPHICAL SKETCH Joan L. Morrison received a Bachelor of Arts degree in biology from the College of Wooster, Wooster, Ohio, in 1975. After working for a year in a research laboratory at the University of Michigan, she attended the School of Natural Resources at the University of Michigan to study the development and pathogenicity of a blood parasite in Canada geese. Subsequent to receiving a Master of Science degree in resource ecology from the University of Michigan in 1979, Joan worked as a biologist for the U.S. Fish and Wildlife Service in Michigan and Colorado and for the U.S. Forest Service in Washington, Alaska, and New Mexico. Beginning in 1985, Joan began working as a contract research biologist for the New Mexico Department of Game and Fish. She studied the life history and distribution of the southwestern subspecies of the meadow jumping mouse {Zapus hudsonius luteus) and collaborated on studies of the reproductive and foraging ecology of Accipiter hawks in the mountains of New Mexico. In 1991, Joan entered the doctoral program at the University of Florida. She studied the reproductive ecology, population biology, and habitat associations of the crested caracara, a Threatened bird of prey in Florida. Joan completed her Ph.D. in fall 1997. Currently, she is continuing research on Florida's caracaras and has expanded her research program to include studies of seasonal movements of the Cape Sable seaside sparrow (Ammodramus maritimus mirabilis) in Everglades National Park and the nesting and foraging ecology of the chimango caracara (Milvago chimango) in Chile. 108

PAGE 124

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. Stephen R. Humphrey, Chair Professor of Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of PhUosot Lyn C. Branch Associate Professor of Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^rofessor of Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. John A. Smallwood Affiliate Professor of Wildlife Ecology and Conservation

PAGE 125

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doct Melvin E. Sunquist U Associate Professor of Wildlife Ecology and Conservation I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope-and quality, as a dissertation for the degree of Doctor S. David Webb Distinguished Research Professor of Geology This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate School and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. December 1997 Dean, College of Agriculture Dean, Graduate School