-
Title
-
Eng
Exploring Migratory Connectivity in the Calliope Hummingbird Through Stable Isotope Analysis of Tail Feathers
-
Date
-
2012
-
Creator
-
Eng
Brown, Clare E
-
Subject
-
Eng
Environmental Studies
-
extracted text
-
EXPLORING MIGRATORY
CONNECTIVITY IN THE CALLIOPE HUMMINGBIRD
THROUGH STABLE ISOTOPE ANALYSIS OF TAIL FEATHERS
by
Clare E. Brown
A Thesis
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
August 2012
©2012 by Clare E. Brown. All rights reserved.
This Thesis for the Master of Environmental Studies Degree
by
Clare E. Brown
has been approved for
The Evergreen State College
by
________________________
Alison R. Styring, Ph.D.
Member of the Faculty
________________________
Robert H. Knapp, Jr., D.Phil.
Member of the Faculty
________________________
Jonathan A. Moran, Ph.D.
Associate Professor, School of Environment & Sustainability,
Royal Roads University
________________________
Date
ABSTRACT
Exploring migratory connectivity in the Calliope Hummingbird through stable isotope analysis
of tail feathers
Clare E. Brown
Knowing where individuals and populations of a migratory species are throughout their annual
cycle, and how those individuals and populations are connected to each other in both the
breeding and non-breeding seasons, is important for understanding the ecology, evolution, and
conservation needs of that species. Stable isotopes can be used as endogenous markers in animal
movement studies, with deuterium being particularly useful in assigning migratory origins. In
this study I use feather deuterium to explore migratory connectivity in Calliope Hummingbirds
(Stellula calliope). First, I used feathers collected from juveniles in the breeding range to
determine the species-specific relationship between feather deuterium and deuterium in
precipitation. Then, I used feathers collected in the breeding range from adult Calliope
Hummingbirds, which molt in the winter, to examine the strength of migratory connectivity in
this species, and to establish wintering locations of sampled breeding populations. Juvenile
feather deuterium values were not correlated with predicted deuterium values of capture site
precipitation, possibly due to the rugged topography of the study area and the contribution of
snowmelt to growing season water supply. Adult feather deuterium values revealed no
differences among populations from different regions, or between sexes, suggesting that sampled
populations of this species have weak or no migratory connectivity. My ability to detect
connectivity, however, was limited by the low resolution of the isotope data. Approximately
twenty percent of adult females had significantly more depleted feather deuterium values than
the remainder of adults, perhaps due to retained juvenile flight feathers in second-year birds,
elevational or latitudinal differences in molt location, or differences in molt timing. Mapping the
range of potential molt locations predicted by feather deuterium values onto deuterium isoscapes
of Mexico and North America suggests that the most likely explanation is that some Calliope
Hummingbird individuals molt their flight feathers during migration.
Table of Contents
List of figures…………………………………………………………………......v
List of tables……………………………………………………………………..vi
Acknowledgements……………………………………………………………..vii
Introduction………………………………………………………………………1
Avian migration and migratory connectivity
Studying migratory connectivity in birds
Stable isotope methods
Calliope Hummingbirds
Methods………………………………………………………………………....10
Feather collection
Isotope analysis
Statistical and isoscape analysis
Figures and tables
Results…………………………………………………………………………...17
Adults
Adult males
Adult females
Enriched cluster
Juveniles
Isoscape analysis
Alabama
Figures and tables
Discussion………………………………………………………………………37
Feather-deuterium values of juvenile hummingbirds
Migratory connectivity in Calliope Hummingbirds
Depleted cluster
Summary………………………………………………………………………..41
References………………………………………………………………………42
iv
List of Figures
Figure 1. Map of hummingbird capture sites.
Figure 2. Distribution of δ2H values of adult and juvenile Calliope Hummingbird tail feathers.
Figure 3. Distribution of δ2H values of adult Calliope Hummingbird tail feathers.
Figure 4. Distribution of δ2H values of adult Calliope Hummingbird tail feathers, divided into an
enriched and a depleted cluster by k-means clustering.
Figure 5. Distribution of δ2Hf values of juvenile Calliope Hummingbirds and depleted-cluster
adults. δ2Hf values of juvenile birds differ significantly from δ2Hf values of individuals in the
depleted cluster.
Figure 6. Distribution of δ2Hf values of enriched cluster adult female and adult male Calliope
Hummingbirds. The mean δ2Hf values of enriched cluster males and females do not differ
significantly.
Figure 7. Distribution of δ2Hf values of adult female Calliope Hummingbirds.
Figure 8. Distribution of δ2Hf values of enriched cluster Calliope Hummingbirds, divided by
region.
Figure 9. Distribution of δ2H values of juvenile Calliope Hummingbird tail feathers.
Figure 10. Mean annual δ2H values of precipitation predicted by the OIPC for feather collection
sites vs. δ2H values of tail feathers of juvenile Calliope Hummingbirds captured at those sites.
Figure 11. Mean annual δ2H values of precipitation predicted by the OIPC for feather collection
sites vs. δ2H values of tail feathers of Rufous Hummingbirds captured at those sites.
Figure 12. δ2Hf values of enriched cluster females and males mapped onto a groundwaterdeuterium isoscape of Mexico at one and two standard deviations about the mean.
Figure 13. δ2Hf values of enriched cluster individuals from British Columbia, Northern
Washington and Southern Washington mapped onto a groundwater-deuterium isoscape of
Mexico at one and two standard deviations about the mean.
Figure 14. δ2Hf values of individuals in the enriched and depleted clusters mapped onto a North
American isoscape of mean-annual deuterium in precipitation at one and two standard deviations
about the cluster means.
v
List of Tables
Table 1. Hummingbird capture sites.
Table 2. Summary of Stellula calliope captures by region, showing number of sites within each
region, period of field work, and individuals captured.
Table 3. Summary statistics.
vi
Acknowledgements
I am indebted to many people for making this project possible. I am grateful to my advisor, Alison
Styring, for both her guidance and her patience. The input of my committee members Jonathan Moran
and Rob Knapp was invaluable; Jonathan provided advice, training, and financial support, and Rob’s
insightful comments on this manuscript were extremely helpful. Dan Harville devoted an enormous
amount of time and effort to training me to handle and band hummingbirds, and he and Jan Harville were
instrumental in making my field work both successful and enjoyable. My mother, Rachel Lawson,
became my unofficial co-investigator, and spent weeks traveling to far flung corners of Washington State
with me in search of hummingbirds. I am also grateful to Mark and Colleen Myers, Scott Heinz, Don
Norman, and Alison Brown for their assistance in the field.
My hummingbird capture sites were almost all at private residences, so my project relied on the
willingness of homeowners to open their homes to me. The hospitality of many of these people, their
enthusiasm for hummingbirds, and their interest in my study made my field work a truly wonderful
experience. I am very grateful to Dianne and Rick Edmonds, Betty and Ralph Hagenbuch, Teri Pieper and
Ken Bevis, Ruthella Skagen, Libby and Victor Glick, Scott Fitkin, Tom Lamb, Cricket Webb, Jill
Streeter, Julia O’Connor, Gordon Kent, and Doris and Larry Robinson. I also appreciate the help of
Nancy Wallwork in coordinating my visits to Brooks Memorial State Park.
I was able to expand the geographic scope of feather collection through the willingness of other
hummingbird banders to collect feathers for me. In Canada, Alison Moran, Cam Finlay, and other
members of the Hummingbird Project of British Columbia collected feathers in Alberta and British
Columbia, and Ida Bacon collected feathers in Alberta. Rob Faucett at the Burke Museum helped me with
the import process. In the United States, Fred Bassett collected feathers in Alabama, and Lisa Goldberg
sent me feathers samples from California.
I also owe thanks to the people that assisted me with analysis of my feathers and of my data. Vasiliki
Demas at Lifespan Biosciences and Paul Quay and Mark Haught at the stable isotope laboratory in the
University of Washington School of Oceanography provided me with lab space and the equipment I
needed to prepare my samples for analysis, and Gabriel Bowen at Purdue University made the isotopic
analysis of my feathers possible. Once I had data in hand, Joseph Brown was an enormous help in
managing those data, determining how best to proceed with analysis, and in putting the results of that
analysis in context.
Finally, no project is possible without funding, and I am grateful for the financial support of the
Mountaineers Foundation and The Evergreen State College Foundation.
vii
INTRODUCTION
Avian migration and migratory connectivity
Avian migration is one of the world’s great natural phenomena. Every year, billions of individual birds
from thousands of species undertake a round-trip journey between their breeding and wintering ranges,
flying from breeding to wintering site in the fall and returning to the breeding site the following spring
(Berthold 1993). While the breeding and non-breeding ranges of a species as a whole may be well
described, there is often little known about the year-round geographic distribution of individuals and
populations within that species. One question in particular is how strong migratory connectivity is in a
given species. Migratory connectivity is the extent to which individuals that breed in the same geographic
area also migrate to spend the non-breeding period in the same wintering area, and can be ranked on a
continuum from strong to weak (Webster et al. 2002).
The total range of a seasonally migratory species can be roughly broken into two usually
geographically distinct areas: the breeding range and the wintering range. Within each of those areas,
individuals are grouped geographically into populations which may be quite distinct from each other. In a
species with strong migratory connectivity, most individuals from one breeding population migrate to the
same area in the non-breeding range so that population structure in the non-breeding range reflects that in
the breeding range. In a species with weak migratory connectivity, the geographic distribution of
individuals in the breeding range bears little relationship to their distribution in the non-breeding period,
and non-breeding populations will be a mixture of individuals from multiple breeding populations and
vice versa. Beyond a strict geographic definition, migratory connectivity can also be a function of age or
sex class, if males and females or birds of different ages are segregated geographically or occupy different
habitats in the winter range.
Understanding the ecology and evolution of a migratory species is impossible in the absence of
knowledge about where individuals and populations of that species are throughout their annual cycle, and
how those individuals and populations are connected to each other. As outlined by Webster and Marra
(2005), the strength of migratory connectivity in a species is important at both immediate and
evolutionary scales. Cross-seasonal interactions act at the individual or population level, with events or
conditions in one season affecting individual survival and reproduction or population dynamics in
another. Studying these interactions in migratory species is complicated in that they occur across space as
well as across time, meaning that measuring cause (e.g. wintering ground conditions) and effect (e.g.
1
reproductive success) requires monitoring the same individuals in both the breeding and non-breeding
range. Seasonal interactions at the individual level have been most successfully demonstrated in the
American Redstart (Setophaga ruticilla). Redstart males that winter in higher quality habitat, as
determined by stable-carbon isotope measurement of claws, arrive at breeding sites earlier in the
subsequent breeding season (Marra et al. 1998) and, in turn, experience lower rates of extra-pair paternity
in their own nests and higher rates of polygyny (Reudink et al. 2009).
At the population level, survival or mortality in the non-breeding season may affect population
dynamics in the breeding season, or vice versa, by increasing or decreasing the density of breeding
populations. The importance of this effect depends on the strength of migratory connectivity. With strong
connectivity, a localized event that causes high mortality in a specific wintering population will have a
significant effect on the corresponding breeding population. When connectivity is weak, winter mortality
will be diffused across the breeding range. At the evolutionary scale, migratory connectivity may
determine a species’ degree of local adaptation to its non-breeding environment (Webster and Marra
2005). In a species with weak connectivity, gene flow among wintering populations that mix and
interbreed in the breeding range will prevent those wintering populations from adapting to their local
environments. If connectivity is strong, discrete winter populations will also be separated from each other
in the breeding season, preventing gene flow and allowing local adaption to non-breeding conditions to
occur. In practical terms, studies on migratory connectivity could inform conservation planning. When
connectivity is strong, the restricted gene flow that enables local adaptation of wintering populations may
prevent populations from adapting to habitat alteration or loss through adopting alternate migration routes
or wintering areas. Further, in a species with strong connectivity, populations that occupy stable habitats
in the breeding range may show declines due to local disturbances in the non-breeding range or vice
versa. If a species’ patterns of connectivity are understood, conservation and management plans can target
key breeding or wintering sites or migration routes in order to protect specific populations identified as in
decline, or to maintain stable populations across the breeding range (Martin et al. 2007).
Studying migratory connectivity in birds
Studies of migratory connectivity present the formidable obstacle of linking individual birds captured at
their breeding sites to their specific wintering sites, or vice versa. There are several potential approaches
to this problem, including mark-recapture methods such as leg-banding, satellite tracking, geolocator
2
devices, and endogenous markers such as genetic markers and stable isotope analysis (Webster et al.
2002). Leg-banding programs, in which captured birds are fitted with printed metal bands that serve as
unique identifiers for those individuals in subsequent encounters, began in Europe and North America at
the end of the nineteenth century, and hundreds of millions of birds have been banded in countries across
the globe (Berthold 1993). Only a very small proportion of banded birds, however, are ever encountered
in locations other than the original capture site (Webster et al. 2002). This means that the great majority of
band recoveries of migratory birds are from individuals that disappear on migration and then reappear the
next year, providing no information about where they have been during the intervening period. Enormous,
long term datasets of bird captures and recaptures from across the multiple countries often included in a
migratory species’ range are thus required to generate sufficient data points for robust analyses of
migratory connectivity from banding data. For example, Ambrosini et al. (2009) used encounter records
of banded Barn Swallows (Hirundo rustica) in Europe and Africa from 1911 to 1998 to evaluate
migratory connectivity in that species. Such datasets are not available for most bird species, meaning that
studies of migratory connectivity that are able to rely on banding data alone are unusual.
Satellite tracking was first applied to bird movements in the late 1980s (Jouventin and
Weimerskirch 1990), and it revolutionized migration studies by allowing investigators to remotely track
the movements of individual birds virtually anywhere on the planet in real time. Satellite tags are attached
to birds and transmit location information via the ARGOS satellite system. The tags can have lifespans of
up to several years, meaning that individuals can be followed over multiple migratory journeys (Webster
et al. 2002). The broad application of satellite technology to migration studies has, however, been
constrained by both high costs and the size of the transmitters themselves. The combined cost of
transmitter and satellite data acquisition is typically in the thousands of dollars for each tracked
individual, and the smallest transmitters available weigh at least five grams (Bridge et al. 2011). The high
cost has kept sample sizes in satellite tracking studies small, and the size of transmitters limits their use to
birds larger than 100 grams, which excludes many migratory passerine species. The ICARUS project
(Wikelski et al. 2007), expected to launch in 2014, plans to install an antenna in the International Space
Station that will detect radio signals from transmitters attached to small animals. First generation
transmitters are anticipated to weigh less than five grams, with future transmitters scaled down even
further. This reduction in transmitter weight relative to currently available satellite tags will greatly
increase the number of taxa that can be tracked remotely.
Geolocator devices differ from satellite tags in that they record and store data rather than
transmitting it. Solar geolocators are attached to birds and record the time of every sunrise and sunset,
3
which can then be used to determine approximate latitude and longitude (Stutchbury et al. 2009). The size
of the batteries needed to satisfy the power requirements of data transmission is the major obstacle to
miniaturizing satellite tags. Because geolocators do not transmit, they can be much smaller than satellite
transmitters. Geolocator devices have been produced that weigh less than one gram, opening up the
possibility of using them to study migratory movements of all but the smallest avian taxa (Bridge et al.
2011). Geolocators also have a much lower cost per individual than do satellite transmitters, and are
independent units that do not rely on a satellite infrastructure. However, the lack of transmission means
that a bird must be recaptured in order to download stored data from its geolocator. Research using
geolocators to study bird migration thus relies on the fact that many migratory bird species exhibit high
site fidelity, with individuals returning to the same breeding site year after year (e.g., Stutchbury et al.
2009, Bairlein et al. 2012, Beason et al. 2012). Unfortunately, not all migratory species are site-faithful,
and some species such as hummingbirds are too tiny for even the smallest geolocators.
Populations within a species show varying degrees of genetic structure across space (Avise 2000).
If populations can be distinguished from each other on the basis of population-specific genotypes at
certain loci, those loci have the potential to be used as genetic markers in migration studies. For example,
genetic variation at specific loci has been characterized across the range of some species of Pacific
salmon, and the origin of individuals caught away from their spawning sites can be identified with genetic
stock identification (e.g. Beacham et al. 2005). In migratory birds, if the geographic distribution of
genotypes in the breeding range is known, birds captured during migration or in the non-breeding range
can be assigned to a specific breeding population. Migratory bird species often have weak genetic
structure, and studies to date suggest that genetic markers may be best suited to drawing broad
conclusions about migratory connectivity at the continental scale rather than linking precise geographical
areas (Bensch et al. 1999, Wennerberg 2001, Kimura et al. 2002, Rolshausen et al. 2009, Irwin et al.
2011). The creation of a distribution-wide phylogeography is a key first step in applying this method, and
the sampling and sequencing effort required may be prohibitive.
Stable isotope methods
The use of stable isotopes emerged as a powerful tool for the study of animal movements in the mid1990s with the application of hydrogen stable isotopes to the assignment of geographic origins of
migratory birds (Chamberlain et al. 1996, Hobson and Wassenaar 1997) and butterflies (Wassenaar and
Hobson 1998, Hobson et al. 1999). Isotopes are atoms with the same number of protons and electrons, but
different numbers of neutrons, and thus different atomic masses. The most common isotope of hydrogen,
4
denoted 1H, has one proton and no neutrons, and an atomic mass of one. Deuterium, denoted as 2H or D,
has one proton and also one neutron, and an atomic mass of two. Stable isotopes are distinguished from
radioactive isotopes by their energetic stability, in that they do not decay. Isotopes of lighter elements
such as hydrogen, carbon, nitrogen and oxygen are the most commonly used in ecological studies, both
because these elements predominate in biological compounds and because the addition of a neutron to a
lighter element causes a proportionately greater and easier to detect change in mass than the addition of a
neutron to a heavier element.
Research using stable isotopes relies on biologically meaningful variation in the relative
abundance of isotopes of a given element under different conditions. That variation is generally quantified
as the deviation of the isotopic ratio (e.g. 2H:1H) of the sample of interest from the isotopic ratio of an
internationally accepted standard. In the case of hydrogen, the vast majority of hydrogen atoms in the
natural world are 1H, with deuterium atoms making up only about 0.0155% (Sulzman 2008). In
precipitation, which is the ultimate source of hydrogen in most terrestrial plant and animal tissue, the
abundance of deuterium relative to 1H varies strongly and predictably with season, climate, elevation and
latitude. Deuterium forms stronger hydrogen bonds in liquid water than does 1H, meaning that it moves
less easily to the vapor phase. This difference in bond strength and the resulting depletion of deuterium in
water vapor relative to liquid water increases at lower temperatures (Marshall et al. 2008). In clouds,
water that condenses and falls as precipitation is enriched in deuterium relative to the water vapor left
behind, and over time this drives a “rainout” effect in which clouds become more and more depleted in
deuterium (Marshall et al. 2008). Temperature and rainout effects lead to a depletion of deuterium in
precipitation as elevation and latitude increase, in winter relative to summer, towards the middle of
continents, and with overall increases in precipitation (Marshall et al. 2008). Measured geographic and
climatic trends in the relative abundance of hydrogen isotopes in precipitation can be used to generate
robust models that predict isotope ratios from environmental parameters, allowing the construction of
detailed maps (isoscapes) describing that relative abundance across continents (Meehan et al. 2004,
Bowen et al. 2005b).
Geographic patterns of deuterium abundance in precipitation are relevant to migration studies in
that the isotope ratio of hydrogen incorporated into an animal’s tissues as it eats and drinks is a function
of the local isotope ratio of hydrogen in precipitation (Chamberlain et al. 1996). The hydrogen isotope
ratio of a given tissue is not simply equivalent to the local hydrogen isotope ratio of precipitation, but the
two ratios are often tightly correlated, such that isotope ratios in animal tissues depend on and can be
predicted from isotope ratios in precipitation (Hobson 2008). As a result, an animal’s tissues will carry an
5
isotopic signature characteristic of that animal’s location. If the animal moves to a new location at which
precipitation has a different hydrogen isotope ratio, persistence of the original location’s isotopic
signature will vary among tissue types. Hydrogen in tissues with rapid turnover rates, such as blood, will
transition to a new isotopic ratio more quickly than will tissues with lower turnover rates, such as claws or
hair (Hobson and Clark 1992). Feathers, which are metabolically inert after their growth is completed,
retain the isotopic signature of the site where they were grown (i.e. the molt location) for the entire
lifetime of the feather (Chamberlain et al. 1996). Thus, if both molt strategy and the relationship between
feather hydrogen isotope ratios (δ2Hf) and precipitation hydrogen isotope ratios (δ2Hp) of feather-growth
location are known for a species, an approximate geographic location of molt can be estimated for an
individual captured far from the molt site. For example, if a bird species molts post-breeding, before
beginning fall migration, feathers from individuals of that species captured on the wintering grounds will
have hydrogen-isotopic ratios that reflect the breeding site. The hydrogen isotope ratio of those feathers
can be measured, and then used to answer questions about migratory strategies and patterns. A pattern in
which isotope ratios in feathers of individuals captured in the same area tend to be more similar to each
other than to isotope ratios of feathers of individuals captured in other regions of wintering range suggests
strong migratory connectivity, while the absence of such pattern suggests weak migratory connectivity. If
the relationship between δ2Hf and δ2Hp is known, allowing feather values to be mapped onto a hydrogen
isoscape, the distribution of individuals and populations across the breeding range can be compared to
their distribution across the wintering range. The migratory pattern of a species can then be classified as,
for example, chain migration, leapfrog migration, or telescopic migration.
The clear benefit of stable isotopic methods is that stable isotopes are endogenous markers that
can be measured in every bird captured. Stable hydrogen isotopes can be used to infer the geographic
location of parts of a bird’s life cycle that take place far away from a capture site, without the need to
mark and recapture an individual or to track individuals after capture. A substantial body of research has
applied stable isotope analysis to the study of avian migration, beginning with Hobson and Wassenaar’s
(1997) demonstration that stable hydrogen isotope ratios of songbird feathers collected in the winter in
Guatemala were consistent with stable hydrogen isotope ratios of water in the known breeding ranges of
the sampled species. Analysis of deuterium abundance in feathers has been used to explore migratory
connectivity and migratory patterns in numerous taxa, including Swainson’s Hawks (Buteo swainsoni)
(Sarasola et al. 2008), European Woodpigeons (Colmuba palumbus) (Hobson et al. 2009a), Loggerhead
Shrikes (Lanius ludovicianus) (Perez and Hobson 2009), Bicknell’s Thrushes (Catharus bicknelli)
(Hobson et al. 2001, Hobson et al. 2004a), Veerys (Catharus fuscescens) (González-Prieto et al. 2011),
Common Blackbirds (Turdus merula) (Evans et al. 2012), American Redstarts (Setophaga ruticilla)
6
(Norris et al. 2004b, Studds et al. 2008), and other wood warbler species (Parulidae) (Kelly 2006, Paxton
et al. 2007, Jones et al. 2008, Langin et al. 2009).
The relative abundances of isotopes of carbon (13C/12C) and nitrogen (15N/14N) also vary in plant
and animal tissues across landscapes in response to habitat parameters, diet, and trophic level (Marra et al.
1998, Hobson 1999, Rubenstein and Hobson 2004). Marra et al. (1998) and Norris et al. (2004a) used
stable carbon isotopes to link quality of winter habitat with migration timing and measures of
reproductive success in American Redstarts. Carbon and nitrogen isotopes in combination with each other
and with hydrogen isotopes have been used to make broad-scale inferences about migratory connectivity
in species such as Willow Warblers (Phylloscopus trochilus) (Chamberlain et al. 2000), Black-throated
Blue Warblers (Setophaga caerulescens) (Rubenstein et al. 2002), Aquatic Warblers (Acrocephalus
paludicola) (Pain et al. 2004), and Reed Warblers (Acrocephalus scirpaceus) (Prochazka et al. 2008).
Despite the potential utility of stable isotopes in studying animal migration, some caveats apply to
the interpretation of feather-isotope values. First, hydrogen-isotope ratios measured in feathers give a
range of possibilities for geographic assignment of molt site rather than a precise location, and must be
interpreted in light of what is known about the occurrence and life history of the species of interest.
Second, within-population individual variability in feather-deuterium values and within-site temporal
variability in precipitation-deuterium values may render the assignment of anything beyond a broad
geographic range for the molt site impossible (Langin et al. 2007, Farmer et al. 2008). Third, drawing
conclusions about migratory differences among individuals based on measured differences in featherisotope values assumes that sampled individuals do not vary in molt strategy, and this assumption may
not be met. For example, Norris et al. (2004b) measured feather hydrogen-isotope ratios of male
American Redstarts from a breeding site in Ontario that were known to have bred at the same location the
previous year. This species molts post-breeding, and about forty percent of sampled individuals had
feathers significantly enriched in deuterium relative to the rest, suggesting that they molted further south.
Thus, while most birds molted near their breeding sites before beginning migration, a significant
proportion of individuals employed a different strategy and molted at staging areas during migration.
Fourth, a fundamental assumption of hydrogen-isotope studies is that the ultimate source of the isotope
ratio of an animal’s tissues, i.e. the water that that animal consumes either directly by drinking or
indirectly through its food, is equivalent to precipitation sampled at the animal’s capture site. This
assumption is supported by the general correlation in stable isotope studies between feather-growth site
δ2Hp and δ2H values of feathers, but has not been directly tested.
7
Calliope Hummingbirds
The Calliope Hummingbird (Stellula calliope) is the world’s smallest long distance avian migrant. It
breeds in western North America, ranging from central California east into Utah and north into Alberta
and British Columbia, and migrates to southwestern Mexico to winter (Calder and Calder 1994). In the
1980s some individuals began wintering in the southeastern United States, where there is now a small but
reliable winter population (Dittman and Demcheck 2006). This species occurs in primarily montane
habitats in both the breeding and non-breeding ranges (DesGranges 1978, Calder and Calder 1994). In the
Pacific Northwest, breeding Calliope Hummingbirds are most common in the Ponderosa pine habitats of
the eastern slope of the Cascade mountain range and in the dry interior forests of the Okanagan
Highlands, with relatively few records from the flat, low elevation Columbia Basin and the rainy Pacific
slope (data were gathered using eBird (http://www.ebird.org)). In Mexico, the species occupies a range of
habitats in the non-breeding season (Calder and Calder 1994). On Volcán de Colima, Jalisco, DesGranges
(1978) recorded it as being most abundant in arid pine-oak forest habitat (about 1500 to 3000 meters
above sea level). The non-breeding range is split in two by the Río Balsas depression.
Very little is known about Calliope Hummingbirds outside of the breeding season, and nothing is
known about the nature of migratory connectivity in this species. Hundreds of Calliopes have been legbanded in the western United States and Canada every year since the mid-1980s, with thousands of
individuals banded annually since 2005 (North American Bird Banding Association/Western Bird
Banding Association annual reports, 1980-2010). Despite this banding effort, no encounter of a banded
Calliope Hummingbird has been ever been reported from Mexico (USGS Bird Banding Laboratory:
Summaries of banding and encounter data, retrieved online 5/19/2012). This is no doubt due to a lack of
capture effort in Mexico, and means that banding data cannot be used at this point to draw conclusions
about migratory connectivity in this species. Calliope Hummingbirds weigh, on average, less than three
grams (Calder and Calder 1994), rendering them too small to be tracked with any available transmitter or
geolocation device. Investigations into their migratory movements are thus limited to endogenous
markers. To date, there has been no phylogeographic or stable-isotopic study of this species.
This study applies stable isotope methods to an analysis of migratory connectivity in Calliope
Hummingbirds. I ask if Calliope Hummingbirds that breed in the Pacific Northwest exhibit segregation
on the wintering ground by sex or by breeding site, and use hydrogen-isotope ratios of feathers collected
from individuals breeding in Washington State, British Columbia, Alberta, and California to evaluate if
any such segregation exists. If this species has high migratory connectivity, individuals that breed in the
8
same area are expected to cluster together in the wintering range. If there is little or no migratory
connectivity, individuals breeding near each will be spread across the wintering range, distributed
panmictically with individuals from other breeding populations. These winter distribution patterns are, in
theory, represented by the hydrogen-isotope ratios of the hummingbirds’ feathers, which are molted in the
non-breeding period. If individuals from one breeding site can be distinguished from individuals from
another breeding site on the basis of the hydrogen-isotope values of their feathers, those individuals are
presumably occupying different areas on the wintering grounds, suggesting high migratory connectivity.
In addition to comparing feather isotope values among groups, I seek to determine the geographic
distribution of the sampled Calliope Hummingbirds’ wintering sites by mapping their feather isotope
values onto hydrogen isoscapes of Mexico and North America. First, I use measured deuterium values of
feathers collected from juvenile Calliope Hummingbirds at their natal sites and deuterium values of
precipitation predicted by isotope models for those sites to find the relationship between the hydrogenisotope ratio of water at a site and the hydrogen-isotope ratio of feathers grown at that site. Once this
relationship is known, measured hydrogen-isotope values of adult feathers can be converted to predicted
hydrogen-isotope values of water at the feather growth site and mapped onto a hydrogen isoscape to
produce a prediction of wintering site location.
9
METHODS
Feather collection
Feathers were collected from a total of 557 Calliope Hummingbird individuals at 23 sites in Washington
State, British Columbia, Alberta, and California (Figure 1, Table 1) from May through August of 2009.
Sites in Washington State were grouped geographically into “Northern” and “Southern” Washington (Fig.
1). Sampling at the Southern Washington sites did not begin until July, when most males had already
disappeared post-breeding, and only one male was captured at the four sites in that region. Feathers were
collected from an additional two individuals in Alabama during the subsequent winter.
All sites had established artificial hummingbird feeders. At each site, my field team and I
captured hummingbirds with either a Hall trap or a cage-wire drop-door trap (see Russel and Russel
2001). All birds were sexed and classified as either hatch-year (juvenile) or after hatch-year (adult). As
adult Calliope Hummingbirds are highly sexually dimorphic, adult males are easily identified. Adult
females and juveniles are superficially similar in plumage, but juvenile hummingbirds can be reliably
distinguished from adults by the presence of corrugations on the bill that persist for at least several
months after fledging (Pyle 1997). To sex juvenile birds, I looked for rufous color along the edges of the
central rectrices (R1). Rufous is present in juvenile males, and absent in juvenile females (Baltosser
1994). In addition to sex and age data, I measured mass, wing chord, tail length, and length of exposed
culmen, and evaluated molt, degree of tail-feather wear, and furcular fat. I also noted if adult females
were gravid – in hummingbirds, the developing egg can easily be seen through the skin of the abdomen.
All individuals were banded with bands provided by the USGS Bird Banding Laboratory, and one fourth
rectrix was pulled for isotopic analysis and placed in a paper envelope for storage. All birds were released
following processing.
Field work in Washington State was carried out by John D. Harville and me. Members of the
Hummingbird Project of British Columbia collected feathers in Canada. Fred Bassett collected feathers
from hummingbirds wintering in Alabama. The feathers from hummingbirds captured in California were
collected as part of a project at the University of California, Davis, and were breast feathers rather than
tail feathers. Hummingbirds were handled in accordance with the guidelines laid out in the North
American Bird Banding Manual (Gustafson et al. 1997) and the North American Banders’ Manual for
Banding Hummingbirds (Russel and Russel 2001).
10
Isotope analysis
Feather samples from 236 of the 559 hummingbirds captured were prepared and sent to the Purdue Stable
Isotope Facility (PSIF) for isotopic analysis. I included all Calliope Hummingbird individuals from
Canada, California and Alabama and a subset of individuals from Washington State in the isotopic
analysis. For each site in Washington State, I analyzed a maximum of ten individuals each of juveniles,
adult males, and adult females. Where possible for a site, I analyzed individuals that were captured on the
same day. If more than ten individuals in one of those three categories had been captured at a site on the
same day, I randomly selected ten for analysis. To reduce bias during analysis, I randomized the order in
which feather samples were prepared and analyzed. Due to delays in the import process, however, all
feathers from Canada were prepared and analyzed together.
Feathers were cleaned with a 2:1 chloroform: methanol wash in order to remove dirt and surface
oils before isotopic analysis. This cleaning method is widely used in hydrogen-isotope studies (e.g.,
Hobson et al. 2009b, Langin et al. 2009, Hardesty and Fraser 2010, González-Prieto et al. 2011). I
followed the protocols provided by the PSIF for cleaning and weighing feather samples, modified due to
the small size of hummingbird feather. I placed each feather in a microcentrifuge tube, filled the tube with
enough 2:1 chloroform:methanol solution to cover the feather, and left the feather in the solution for at
least five minutes, agitating the tube at least once during the soak. I then decanted the solution and
repeated the soaking and agitation step two more times. Finally, I rinsed both sides of the feathers with
the chloroform:methanol solution and dried them on a sheet of aluminum foil under a fume hood.
Feathers were stored in clean paper envelopes between cleaning and weighing. To prepare the samples for
isotopic analysis, I weighed 0.15 ± 0.01 mg of material cut from the distal part of the cleaned feathers into
individual 3.5 x 5 mm silver capsules and crushed the capsules.
Measurement of hydrogen-isotope ratios in organic materials such as feather-keratin is
complicated by the presence of exchangeable hydrogen in the sample. While most hydrogen in keratin is
bound to carbon and is non-exchangeable, a significant proportion of hydrogen atoms in feather-keratin
are free to exchange with hydrogen atoms in ambient water vapor (Wassenaar and Hobson 2000). If the
hydrogen-isotope ratio of ambient vapor differs from that of the sample, the replacement of exchangeable
hydrogen atoms in the sample with atoms from the ambient environment will alter the sample’s total
hydrogen-isotope ratio. Feathers are metabolically inert, meaning that once feather growth is complete,
the isotope ratio of non-exchangeable hydrogen in the feather does not change (see Hobson and Clark
1992). The isotope ratio of exchangeable hydrogen, however, and thus of the total hydrogen content in
11
feathers of migratory birds, will change as the feathers move to different locations, as when the bird
leaves the site of feather growth and when feathers are collected and moved into a laboratory for analysis.
Only isotope ratios of total hydrogen content can be measured directly, meaning that uncorrected sample
isotope ratios measured at different facilities, or at different times within the same facility, are not
comparable. The isotope ratio of non-exchangeable hydrogen alone can be determined using the
comparative equilibrium method proposed by Wassenaar and Hobson (2000, 2003, also see Bowen et al.
2005a). In comparative equilibrium, keratin reference standards with known hydrogen isotope ratios and
the samples of interest are simultaneously allowed to equilibrate with ambient laboratory conditions. The
samples and the standards are then analyzed together, and a formula to correct for exchangeable hydrogen
is determined using linear regression.
After cleaning and preparation, hummingbird feather samples were sent to the PSIF for analysis
on an isotope ratio mass spectrometer. Samples were subjected to comparative equilibrium, so that
reported δ2H values can be interpreted as the isotope ratio of non-exchangeable hydrogen in the feathers.
Analytical precision was determined by repeated analysis of two reference keratins. Hydrogen isotope
ratios (2H/1H) are reported in terms of their deviation, in parts per thousand (‰), from the isotope ratio in
Vienna standard mean ocean water-Vienna standard light Antarctic precipitation (VSMOW-VSLAP), a
widely used international standard:
where Rsample is the measured ratio of deuterium:hydrogen in the sample, and Rstandard is the known ratio of
deuterium:hydrogen in VSMOW-VSLAP.
Statistical and isoscape analysis
All statistical analyses other than reduced major axis regressions were conducted in R version 2.11.1 (R
Development Core Team 2011). δ2Hf values of adult Calliope Hummingbirds samples are bimodally
distributed. I segregated them into two groups based on their feather δ2H values with k-means clustering,
using the k-means clustering algorithm developed by Hartigan and Wong (1979), the default option in R.
This algorithm groups data points into an a priori specified number of clusters such that the within-cluster
sum of squares is minimized.
12
Calliope Hummingbirds molt in the winter after completing southward migration to their
wintering sites (Calder and Calder 1994, Dittman and Demcheck 2006). The hydrogen isotope ratio
measured in feathers (δ2Hf) collected from an adult Calliope in the breeding range will thus reflect the
hydrogen isotope ratio of precipitation (δ2Hp) of that individual’s wintering site. As hydrogen is
incorporated into plant and animal tissues, however, isotope fractionation alters the hydrogen isotope ratio
so that δ2Hf will not simply equal δ2Hp (Chamberlain et al. 1996). A key step, then, in discovering
migratory origins is to translate measured δ2Hf values into δ2Hp values that can be mapped onto a
hydrogen isoscape. This translation is straightforward - deuterium values of feathers and of feather
growth site precipitation tend to have a strong linear correlation (e.g. Bowen et al. 2005b) that can be
determined by analyzing feathers of known origin, i.e., feathers collected from adult birds at the site of
feather growth, or from juvenile birds near the nesting site. Regressing δ2Hp of the collection sites against
δ2H of these feathers will produce a linear equation that can then be used to convert measured δ2H values
of feathers of unknown origin into predicted δ2Hp values for the sites where those feathers were grown.
As there is error in both the model-based prediction of δ2Hp (the dependent variable) and in the
measurement of δ2Hf (the independent variable), reduced major axis regression (RMA) is more
appropriate in this case than simple least squares regression (Hobson et al. 2004b).
To determine the relationship between δ2Hf and δ2Hp of feather-growth location for Calliope
Hummingbirds, I analyzed feathers collected from juvenile birds in Washington State and British
Columbia. Mapping breeding Calliope Hummingbirds onto their putative wintering sites requires
converting measured δ2Hf into a predicted δ2Hp value, so I performed RMA regressions with juvenile δ2Hf
as the independent variable and δ2Hp of feather collection site as the dependent variable, using RMA for
Java v1.21 (Bohonak and van der Linde 2004). I regressed both mean annual and June δ2Hp values
predicted by the Online Isotopes in Precipitation Calculator (OIPC) (Bowen et al. 2005b, Bowen 2011b)
against juvenile δ2Hf values. Juvenile δ2Hf correlated poorly, however, with δ2Hp of collection site in my
data set (see Discussion), so I was unable to use juvenile Calliope feathers to predict the relationship
between δ2Hp and δ2Hf in adults. Instead, I used data provided by Dr. Jonathan Moran for Rufous
Hummingbird (Selasphorus rufus) feathers collected in Canada, the south-eastern United States, and
Mexico. Feathers were collected from the individuals in this data set at the sites where the feathers were
presumed to have been grown (i.e., juveniles in the breeding range, and adults or first-year birds in the
wintering range following molt). Calliope and Rufous Hummingbirds are closely related (Stellula actually
falls within the Selasphorus clade (McGuire et al. 2009)), and occur in the same habitats in both the
breeding and non-breeding season. I assume that, due to their phylogenetic relatedness and ecological
similarity, the effect of isotope fractionation on the isotope ratios of hydrogen incorporated into tissue will
13
be similar between the two species. It follows from this assumption that the relationship observed
between Rufous Hummingbird δ2Hf and collection site δ2Hp should be a reasonable proxy for that
relationship in Calliope Hummingbirds, and I use the equation obtained with the Rufous Hummingbird
dataset to convert Calliope Hummingbird δ2Hf to δ2Hp.
I performed an RMA regression on the Rufous Hummingbird data set with δ2Hf as the
independent variable and mean annual δ2Hp estimated for feather collection sites with the OIPC as the
dependent variable. Mean annual deuterium values are more appropriate for this regression than are
monthly values, as the relationship obtained from the RMA regression is used to map adult Calliopes onto
their wintering range with a hydrogen-isoscape map of Mexico constructed by Wassenaar et al. (2009)
using deuterium in groundwater. Using groundwater effectively averages precipitation throughout the
year, smoothing out the intra-and inter-annual variation of δ2Hp values (Wassenaar et al. 2009). This
means that groundwater-deuterium values are more comparable to annual than to monthly precipitationdeuterium values. I used the linear relationship produced with the RMA regression to convert measured
δ2Hf values of adult Calliope Hummingbirds to predicted δ2Hp values of their feather-growth sites.
These δ2Hp values fall into two distinct clusters, one with more enriched hydrogen-isotope ratios
and the other with more depleted isotope ratios. I divided the enriched cluster by sex and by region (Fig.
1), and mapped both one standard deviation and two standard deviations about the mean δ2Hp values of
each group onto the hydrogen-isoscape of groundwater in Mexico (Wassenaar et al. 2009). The δ2Hp
values of the depleted cluster fall outside the range ofδ2Hp values that occur in Mexico, so I used a global
hydrogen-isoscape of precipitation (Bowen and Revenaugh 2003, Bowen 2011a) to map one standard
deviation and two standard deviations about the mean δ2Hp values of the enriched and depleted clusters.
Both the global and Mexican isoscapes are available as ArcGIS rasters, and I mapped Calliope
Hummingbird δ2Hp values onto the rasters with ArcGIS 10.
14
TABLE 1 – Hummingbird capture sites
Latitude Longitude
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
49.47
50.38
49.45
49.54
49.54
50.19
50.20
50.20
50.29
48.08
48.37
48.37
48.38
48.42
48.47
48.5
49.45
45.95
46.12
46.84
47.15
47.89
37.37
30.36
30.38
-114.19
-114.65
-120.36
-120.51
-120.45
-119.42
-119.48
-119.28
-118.97
-120.01
-119.58
-119.72
-119.72
-120.16
-120.22
-120.26
-120.36
-120.67
-118.14
-120.96
-120.87
-120.21
-119.83
-87.49
-87.46
Elevation
(m asl)
1400
1510
600
720
920
400
375
450
683
332
347
259
927
588
597
558
948
817
524
621
802
335
590
8
13
Region
Alberta
British Columbia
Northern
Washington
Southern
Washington
California
Alabama
15
FIGURE 1 – Map of hummingbird capture sites. Site groupings used in the analysis are circled and are, from south to north, Southern
Washington, Northern Washington, and British Columbia.
16
RESULTS
The overall precision in the isotopic analysis was 2‰ for one of two independent reference keratins and
3‰ for the other. The PSIF reported imprecision in the bracketing reference keratins for 12 samples, and
I excluded those samples from further analysis. I also excluded three hatch year birds with feathers
extremely enriched in deuterium (δ2Hf = -71.5, -77.6, -80.3‰) relative to other juveniles, as my notes on
either the bird or on the feather itself indicated that in all three cases the individuals were misidentified
adult females. The two individuals from Alabama were not included in the general analysis. With the 15
questionable samples excluded, the following analysis includes a total of 219 Calliope Hummingbird
individuals from the western United States and Canada, comprising 172 adults (128 females, 44 males)
and 47 juveniles (14 females, 33 males) (Table 2).
The δ2Hf values for all individuals range from -36.6 to -169.4‰ with a clearly bimodal
distribution (Fig. 2). As expected, juveniles, which grow their feathers in the nest in the breeding range,
occupy the more depleted (more negative) peak of the distribution, while adults, which grow their
feathers much further south in the wintering range, occupy the more enriched (less negative) peak.
Adults
The δ2H values of adult feathers (N = 172) range from -148.1 to -36.6‰, and are bimodally distributed
(Table 3, Fig. 3). A k-means cluster analysis with an initial assignment of two clusters groups the data in
apparent agreement with the two peaks of the distribution (Table 3, Fig. 4). The cluster of individuals
with feathers more depleted in deuterium ranges from -148.1 to -100.7‰ and comprises 1 male and 29
females. The remaining 43 males and 99 females make up the cluster of individuals with feathers more
enriched in deuterium, with δ2H values that range from -96.7 to -36.6‰. The means of the two clusters
differ significantly from each other (t = 21.29, df = 42.38, P < 0.001), while the variances do not (F =
1.02, df = 29,141, P = 0.99).
One simple explanation for the presence of the depleted cluster is that these individuals are
juvenile birds mistakenly identified as adult females. However, the δ2Hf values of the birds in the depleted
cluster differ significantly from the δ2Hf values of juvenile birds (t = 7.83, df = 54.68, P < 0.001), and
about half of the juveniles have more negative δ2Hf values than the most depleted individual in the
depleted cluster (Fig. 5). Additionally, if adult females with very negative δ2Hf values are actually
juveniles, cluster membership should depend on capture date, with birds in the depleted cluster tending to
17
be captured after juveniles began to visit feeders in early July. This is not the case, as cluster membership
and capture date (defined as before July 1 or after July 1) of adult females are independent (X2 = 0.30, df
= 1, P = 0.58). These results indicate that the individuals in the depleted cluster are in fact adults, and not
misidentified juvenile birds.
Adult males
All but one of the 44 adult males (δ2Hf = -116.5‰, from British Columbia) were grouped into the more
enriched k-means cluster. The δ2Hf values of these enriched adult males appear normally distributed, and
range from -36.6‰ to -96.7‰ (Table 3, Fig. 6). With the depleted-cluster male excluded, there is no
significant difference in feather δ2H between males from Northern Washington and males from British
Columbia (t = 0.94, df = 18.58, P = 0.36), the only regions for which sample size allowed comparison.
While the only male in the depleted cluster was captured in British Columbia, within the enriched cluster
males from Northern Washington are slightly and non-significantly more variable than males from British
Columbia (F = 0.52, df = 8,28, P = 0.33) (Table 3). Male feather δ2H is uncorrelated with the elevation of
collection sites, both across all regions and within regions.
Adult females
The feather δ2H values of adult females are bimodally distributed (Fig. 7), and drive the overall
distribution of sampled adults – all but one of the 30 individuals in the depleted cluster are female. A kmeans cluster analysis of adult females alone produced results similar to the cluster analysis of all adults,
with all females assigned to the same cluster as before. As with adult males, adult female feather δ2H is
uncorrelated with collection site elevation, both across all regions and clusters and within regions and
clusters.
As the sample sizes for Alberta and California are small, I focused on adult females from British
Columbia, Northern Washington, and Southern Washington. For females, cluster membership is not
independent of sampling locality (Fisher’s Exact Test, P < 0.001), with more females from British
Columbia and fewer females from Northern Washington in the depleted cluster than would be expected if
cluster membership and sampling region were independent.
While Northern Washington has the lowest proportion of individuals in the depleted cluster, and
thus the lowest overall variability, females from this region have the greatest variability within the
18
enriched cluster. The variance of females from Northern Washington is significantly greater than that of
females from Southern Washington (F = 2.13, df = 28,53, P = 0.03), but is not significantly greater than
that of females from British Columbia (F = 2.31, df = 7,53, P = 0.25). The variances of females from
Southern Washington and British Columbia are virtually identical (Table 3). As the assumption of
homoscedasticity for Fisher’s analysis of variance is not met (Fligner-Killeen test: X2 = 6.49, df = 2, P =
0.04) and the disparity in sample size among groups is substantial, I compared the means of the three
regions using a one-way Welch’s analysis of variance (Welch 1951). The mean δ2Hf values of adult
females do not vary significantly among the three regions (F = 2.48, df = 2.00,21.65, P = 0.11).
Only one male falls within the depleted cluster, and other than one individual from California
(δ Hf = -107.1‰), all individuals in the depleted cluster are from British Columbia, Northern
2
Washington, or Southern Washington. The variances of the δ2Hf values for the depleted-cluster females
from these three regions are homogeneous (Fligner-Killeen test: X2 = 2.90, df = 2, P = 0.23), while the
means are marginally significantly different (one-way analysis of variance, F = 3.18, df = 2,25, P = 0.06).
This difference is driven by the more enriched mean δ2Hf value of depleted-cluster females from Northern
Washington (see Table 3).
Enriched cluster
Within the cluster of hummingbirds with feathers more enriched in deuterium, the mean (t = 0.78, df =
72.73, P = 0.44) and variance (F = 1.24, df = 42,98, P = 0.38) of δ2Hf values do not differ significantly
between males and females (Table 3, Fig. 6). Looking within regions, mean (t = 0.38, df = 54.41, P =
0.70) and variance (F = 1.13, df = 28,53, P = 0.68) of δ2Hf do not differ between males and females from
Northern Washington. I only captured one male from Southern Washington, so a comparison between
sexes in that region is not possible. In British Columbia, the variance of δ2Hf does not differ between
sexes (F = 1.35, df = 7,8, P = 0.71), but females have a significantly more depleted mean δ2Hf than do
males (t = 2.49, df = 14.99, P = 0.02, see Table 3). There is one male from British Columbia with very
enriched feather δ2H (-41.3‰) relative to the other males from the region, but the difference between
males and females is significant even with that male removed (t = 2.22, df = 13.47, P = 0.04).
As above, the following focuses on British Columbia, Northern Washington, and Southern
Washington. For males and females together, as with females alone, the variance of δ2Hf appears greatest
in Northern Washington (Fig. 8). The difference in variance is significant between Northern and Southern
Washington (F = 2.25, df = 29,82, P = 0.02), and is not significant between Northern Washington and
19
British Columbia (F = 1.53, df = 16,82, P = 0.34). The greater variance could be a factor of Northern
Washington’s large sample size relative to the other two regions. As for females alone, due to the unequal
variances (Fligner-Killeen test: X2 = 8.04, df = 2, P = 0.02) and the differences in sample size among
groups, I compared the mean δ2Hf of enriched-cluster adults among the three regions with a one-way
Welch’s anova. The means do not vary significantly (F = 0.20, df = 2.00,43.00, P = 0.82).
Juveniles
δ2H values of juvenile feathers (N = 47) range from -169.4 to -121.8‰, and appear to be normally
distributed, although with a slight positive skew (Table 3, Fig. 9). It is possible that juveniles with the
most enriched δ2H values are misidentified adult females. Feather deuterium values of juvenile males and
of juvenile females do not significantly differ in either mean (t = -1.52, df = 29.17, P = 0.14) or variance
(F = 1.43, df = 13,32, P = 0.50). Contrary to expectation, there is no correlation between juvenile δ2Hf
and either the mean annual or June precipitation deuterium (δ2Hp) values predicted by the Online Isotopes
in Precipitation calculator (Bowen et al. 2005b, Bowen 2011b) for the sites where the juvenile birds were
captured (RMA regressions, N = 47: mean annual, R2 = 0.04, P = 0.18; June, R2 = 0.02, P = 0.31).
Additionally, the difference between predicted δ2Hp and measured juvenile δ2Hf is greater than expected.
The juvenile feather values tend to be depleted by about 50‰ or more relative to predicted precipitation
values (Fig. 10), while the “standard” fractionation value expected from the literature is -25‰ (Wassenaar
and Hobson 2001, Bowen et al. 2005b).
Isoscape analysis
In the Selasphorus rufus data set, which comprises 43 individuals, measured δ2H feather values are tightly
correlated with capture site mean annual δ2H precipitation values predicted by the Online Isotopes in
Precipitation Calculator (Fig. 11; δ2Hp = 29.006 + 1.151 δ2Hf, R2 = 0.92, P < 0.001). I used this
fractionation equation to convert adult Calliope Hummingbird δ2H feather values to the δ2Hp values of the
locations where adult birds grew those feathers, i.e., their putative wintering sites. In order to compare the
geographic distributions of these putative wintering sites, I mapped one and two standard deviations about
the mean delta2Hp values of the enriched and depleted clusters and of sexes and regions within the
enriched cluster onto global annual precipitation and Mexican groundwater hydrogen-isoscapes,
respectively.
Within the enriched cluster, there is no differentiation between the predicted wintering site of
20
Calliope Hummingbird males and females (Fig. 12) or among the wintering site of individuals breeding in
different regions (Fig. 13). There is high variability in hummingbird feather-isotope values relative to the
range of groundwater-isotope values modeled for Mexico by Wassenaar et al. (2009). The 95 percent
confidence intervals (two standard deviations about the mean) of δ2Hp of the groups in the enriched
cluster each cover close to the entire range of modeled groundwater δ2H in Mexico.
K-means clustering split adult Calliope Hummingbirds into a more enriched and a more depleted
cluster, with hydrogen-isotope ratios greater than and less than -100‰, respectively (Fig. 4, Table 3).
When mapped onto a global hydrogen isoscape, the two clusters show different and non-overlapping
geographic distributions (Fig. 14). The known Calliope Hummingbird wintering range falls completely
within the predicted geographic distribution of the enriched cluster at one standard deviation about the
cluster mean, whereas one standard deviation about the mean of the depleted cluster almost entirely
contains the known breeding range.
Alabama
Both birds captured in Alabama hatched in 2009, and were in juvenile plumage at the time of capture. The
first, a female, was captured on 12/16/2009, while the second, a male, was captured a month later, on
1/21/2010. Their feather δ2H values are -142.5 and 130.9‰, respectively. The first value falls squarely
into the range of juveniles captured in Washington State and British Columbia, while the second falls
towards the more enriched end of the range (Table 3).
21
TABLE 2 – Summary of Stellula calliope captures by region, showing number of sites within each region, period of field work, and
individuals captured. The numbers in parentheses are the numbers of sites or individuals which were used in the final analysis. With
the exception of sites in Alberta and Alabama, locations were visited multiple times during the indicated date range.
Adults
Region
Number
of sites
Date range of captures
Alberta
British Columbia
Northern Washington
Southern Washington
California
Alabama
2 (2)
7 (5)
8 (8)
4 (4)
1 (1)
2 (2)
5/28/2009-5/31/2009, 8/11/2009
6/31/2009-7/11/2009
6/20/2009-7/25/2009
7/8/2009-8/12/2009
5/3/2009
12/16/2009, 1/21/2010
Juveniles
Males
Females
Males
Females
2 (2)
13 (10)
47 (29)
1 (1)
2 (2)
1 (1)
5 (4)
27 (19)
307 (60)
70 (40)
5 (5)
0
0
4 (3)
46 (28)
3 (2)
0
0
0
8 (5)
14 (8)
1 (1)
0
1 (1)
22
TABLE 3 – Summary statistics
mean
(‰)
standard
deviation (‰)
N
-149.9
11.4
47
-151.4
-146.3
11.8
9.9
33
14
NA
NA
172
-69.6
13.5
142
Males
Females
-68.2
-70.2
13.0
14.5
43
99
British Columbia
Males
Females
Northern Washington
Males
Females
Southern Washington
Males
Females
-69.6
-63.7
-76.2
-68.9
-68.1
-69.4
-67.7
-73.8
-67.5
12.0
11.1
9.6
14.8
15.5
14.6
9.9
NA
10.0
17
9
8
83
29
54
30
1
29
-126.8
13.3
30
Males
Females
-116.5
-127.1
NA
13.4
1
29
British Columbia (females)
Northern Washington
Southern Washington
-132.6
-117.2
-128.9
7.7
17.5
12.5
11
6
11
JUVENILES (range -169.4 to -121.8‰)
Males
Females
ADULTS (range -148.1 to -36.6‰)
Enriched cluster (range -96.7 to -36.6‰)
Depleted cluster (range -148.1 to -100.7‰)
23
30
20
10
0
Frequency
40
50
FIGURE 2 – Distribution of δ2H values of adult and juvenile Calliope Hummingbird tail
feathers. Juveniles occupy the more depleted peak to the left, adults the more enriched peak to
the right.
-200
-150
-100
-50
0
2
Hf
24
30
20
10
0
Frequency
40
50
FIGURE 3 – Distribution of δ2H values of adult Calliope Hummingbird tail feathers. The smaller
peak to the left comprises the 30 individuals in the depleted cluster, while the larger peak to the
right is the enriched cluster.
-200
-150
-100
-50
0
2
Hf
25
FIGURE 4 – Distribution of δ2H values of adult Calliope Hummingbird tail feathers, divided into
an enriched (upper panel) and a depleted cluster (lower panel) by k-means clustering with an
initial assignment of two clusters.
30
0 10
Frequency
50
Enriched cluster, n = 142
-200
-150
-100
-50
0
-50
0
30
0 10
Frequency
50
Depleted cluster, n = 30
-200
-150
-100
2
Hf
26
FIGURE 5 – Distribution of δ2Hf values of juvenile Calliope Hummingbirds (upper panel) and
depleted-cluster adults (lower panel). δ2Hf values of juvenile birds differ significantly (t = 7.83,
df = 54.68, P < 0.001) from δ2Hf values of individuals in the depleted cluster.
10
5
0
Frequency
15
Juveniles
-180
-160
-140
-120
-100
-120
-100
10
5
0
Frequency
15
Depleted cluster
-180
-160
-140
2
Hf
27
FIGURE 6 – Distribution of δ2Hf values of enriched cluster adult female (upper panel) and adult
male (lower panel) Calliope Hummingbirds. The mean δ2Hf values of enriched cluster males and
females do not differ significantly (t = 0.78, df = 72.73, P = 0.44).
15
10
5
0
Frequency
20
Females, n = 99
-100
-90
-80
-70
-60
-50
-40
-30
-50
-40
-30
15
10
5
0
Frequency
20
Males, n = 43
-100
-90
-80
-70
-60
2
Hf
28
30
20
10
0
Frequency
40
50
FIGURE 7 – Distribution of δ2Hf values of adult female Calliope Hummingbirds. The peak on
the left falls within the depleted cluster, while the peak on the right falls within the enriched
cluster.
-200
-150
-100
-50
0
2
Hf
29
FIGURE 8 – Distribution of δ2Hf values of enriched cluster Calliope Hummingbirds, divided by
region.
10
5
0
Frequency
15
British Columbia, n = 17
-100
-90
-80
-70
-60
-50
-40
-30
-40
-30
-40
-30
10
5
0
Frequency
15
Northern Washington State, n = 83
-100
-90
-80
-70
-60
-50
10
5
0
Frequency
15
Southern Washington State, n = 30
-100
-90
-80
-70
-60
-50
2
Hf
30
10
5
0
Frequency
15
FIGURE 9 – Distribution of δ2H values of juvenile Calliope Hummingbird tail feathers.
-200
-180
-160
-140
-120
-100
2
Hf
31
FIGURE 10 – Mean annual δ2H values of precipitation predicted by the OIPC for feather
collection sites vs. δ2H values of tail feathers of juvenile Calliope Hummingbirds captured at
those sites.
2
Hp vs. juvenile Calliope Hummingbird
2
Hf
-100
-102
-104
2
2
Hp
2
R
140.03 0.26 Hf
0.04, d.f.=43, P=0.18
-106
Mean annual
2
Hp
-98
Predicted mean annual
-170
-160
-150
-140
-130
-120
2
Hf
32
FIGURE 11 – Mean annual δ2H values of precipitation predicted by the OIPC for feather
collection sites vs. δ2H values of tail feathers of Rufous Hummingbirds captured at those sites.
2
Hp vs. Rufous Hummingbird
2
Hf
-60
-70
-80
2
2
Hp 29.006 1.151 Hf
2
R 0.92, d.f.=41, P<0.001
-90
Mean annual
2
Hp
-50
-40
Predicted mean annual
-110
-100
-90
-80
-70
-60
2
Hf
33
FIGURE 12 – δ2Hf values of enriched cluster females (A, C) and males (B, D) mapped onto a
groundwater-deuterium isoscape of Mexico. A, B – one standard deviation about the mean; C, D
- two standard deviations about the mean.
34
FIGURE 13 - δ2Hf values of enriched cluster individuals from British Columbia (A, D), Northern Washington (B, E) and Southern
Washington (C, F) mapped onto a groundwater-deuterium isoscape of Mexico. A-C – one standard deviation about the mean; D-F two standard deviations about the mean.
35
FIGURE 14 – δ2Hf values of individuals in the enriched (darker grey) and depleted (lighter grey)
clusters mapped onto a North American isoscape of mean-annual deuterium in precipitation. A one standard deviation about the cluster means; B – two standard deviations about the cluster
means.
A
B
36
DISCUSSION
Feather-deuterium values of juvenile hummingbirds
Hydrogen-isotope ratios of juvenile Calliope Hummingbird feathers are unexpected first in their lack of
correlation with predicted hydrogen-isotope ratios of local precipitation and second in their distance from
predicted precipitation values. The absence of a correlation is likely a product of both the limited
geographic extent of feather sampling and the complex topography of the study area. Virtually all
juveniles were captured at sites in the mountainous regions of northeastern Washington State and eastern
British Columbia, spanning only a few degrees of latitude. There is little predicted variation in δ2Hp
across these sites; the range of OIPC-predicted mean annual δ2Hp values for all sites at which juveniles
were captured (-106‰ to -97‰) is much smaller than the range of δ2Hf values measured in juvenile
feathers (-169.4‰ to -121.8‰). This region is characterized by rugged topography, and elevation changes
of thousands of feet can occur within a few miles. Hydrogen-isotope ratios in precipitation vary with
elevation, so the region’s high degree of variation in elevation across space is matched by a
correspondingly variable hydrogen isoscape. For example, an increase in elevation of approximately 500
meters between two points about 5 km apart in the Methow Valley is matched by a depletion of
approximately 8‰ in OIPC-predicted mean annual δ2Hp. This change in δ2Hp is as great as the total range
of δ2Hp values predicted for juvenile capture sites. The situation is further complicated by the rain shadow
effect of the coastal ranges. Hummingbirds are very mobile, and females in the region may forage across
a broad elevational range while raising nestlings. The resulting isotopic variation in nestling diet would
create noise in the hydrogen-isotopic signal of juvenile feathers, preventing juveniles from sites without
significant latitudinal separation from being distinguished on the basis of their δ2Hf values and weakening
the correlation between δ2Hf and nesting site δ2Hp. The mountainous nature of the study area thus enables
variation in the elevational source of nestling diet, and this combined with the limited geographic extent
of feather sampling likely contributes to the poor correlation between predicted δ2Hp and measured δ2Hf
of juvenile Calliope Hummingbirds.
Beyond the lack of correlation, juvenile Calliope Hummingbird δ2Hf values are more depleted
relative to predicted capture site δ2Hp values than expected. Measured δ2Hf of the majority of juveniles in
this study is at least 50‰ more depleted than capture site δ2Hp, whereas correlations reported in the
literature typically predict that δ2Hf will be depleted by only about 25‰ relative to δ2Hp (Wassenaar and
Hobson 2001, Bowen et al. 2005b). The Selasphorus rufus correlation reported here is close to the
expected, with a predicted depletion of about 30‰. One explanation for the greater depletion found in
37
juvenile feathers is that the source of water taken up by plants during the growing season may be melting
winter snowpack from higher elevations rather than rainfall. The region where most juvenile Calliope
Hummingbirds were captured is arid, with little rain in the summer months but with significant snow
accumulation at higher elevations during the winter. OIPC predicted δ2Hp of winter precipitation higher in
the mountains near capture sites is negative enough to reach the expected value of 25‰ more enriched in
deuterium than measured juvenile δ2Hf. If winter snowpack is, in fact, a major source of water available to
plants when juvenile hummingbirds are in the nest and growing their feathers, the gap between when and
where precipitation falls and capture sites could be another factor contributing to the lack of correlation
between capture site δ2Hp and juvenile Calliope Hummingbird δ2Hf.
Migratory connectivity in Calliope Hummingbirds
Hydrogen-isotope ratios of tail feathers collected from adult Calliope Hummingbirds in Washington State
and British Columbia suggest that individuals from this part of the breeding range have weak or no
migratory connectivity. Variation in feather-deuterium values of adults in the enriched cluster is high
relative to the predicted variation in precipitation-deuterium in the species’ winter range. Ninety-five
percent confidence intervals of enriched cluster δ2Hf values for all three regions and both sexes cover the
virtually the entire known Calliope Hummingbird non-breeding range when converted to δ2Hp values and
mapped onto a hydrogen isoscape of Mexico (Figs. 12,13). There is no evidence that individuals are
segregating by sex or by breeding site in the non-breeding season. However, while there is no observed
difference in δ2Hf either among regions or between sexes, high within-group variation may be masking
among-group differences, and the presence of strong migratory connectivity cannot be ruled out.
Variation in groundwater δ2H across Mexico is tied to elevation, with water becoming
increasingly depleted in deuterium as elevation increases away from the coasts (Wassenaar et al. 2009).
As with juvenile Calliope Hummingbird δ2Hf values, complex topography combined with individual
mobility may have resulted in the high variation found in hydrogen-isotope ratios of adult feathers. Other
deuterium studies have had similar results. House Sparrows (Passer domesticus) sampled across Mexico
had high within-site variability in δ2Hf, leading to a weak relationship between δ2Hf and interpolated δ2H
of groundwater (Hobson et al. 2009b). Bird species sampled along an elevation gradient in Ecuador
showed an overall trend of decrease in δ2Hf with increased elevation, but high within-site variability
resulted in weak and non-significant correlations between δ2Hf and elevation for all but one species
(Hardesty and Fraser 2010). These studies and the present research suggest that for species that occur in
mountainous areas where deuterium abundance in water is mainly a function of elevation, work with
38
feather hydrogen-isotope ratios may limited to general trends across broad geographic extents. The nonbreeding range of Calliope Hummingbirds is quite small, and it is unlikely that further work with
hydrogen isotopes will shed additional light on the strength of migratory connectivity in the species.
Depleted cluster
About twenty percent of adult female Calliope Hummingbirds have tail feathers significantly depleted in
deuterium relative to the rest of the adults (Fig. 4). The two most plausible explanations for the presence
of the depleted cluster are (1) that these individuals are second-year birds with retained tail feathers from
the previous year, or (2) that these individuals molt at a more northern latitude than most other Calliope
Hummingbirds.
Retained tail feathers
The δ2Hf values of individuals in the depleted cluster map to the Calliope Hummingbird breeding range
(Fig. 14). As such, the sampled feathers with depleted δ2Hf values could be rectrices grown in the nest by
juvenile birds that did not molt completely during their first winter. The individuals in the depleted cluster
tend to have feathers more enriched in deuterium than juveniles sampled in the Pacific Northwest (Fig. 5).
If the depleted cluster is made up of second-year birds with retained rectrices, these individuals have
dispersed northward from natal sites in more southern parts of the breeding range. However, Pyle et al.’s
(1997) study of molt in North American hummingbirds recorded only one percent of sampled individuals
as having retained flight feathers. This renders it improbable that the depleted deuterium values of such a
substantial proportion of individuals as in this study could be due to retained feathers. Additionally, only
one male falls into the depleted cluster, with 29 females, and there is no reason to believe that a strong sex
bias would be present in the retention of juvenile feathers.
Molt latitude
Precipitation becomes increasingly depleted in deuterium with increasing latitude, so the significantly
more negative δ2Hf values of depleted cluster individuals can be interpreted as meaning that these birds
are molting at higher latitudes than the majority of sampled individuals. Calliope Hummingbirds winter in
Mexico and, in recent decades, in the southeastern United States. The δ2Hf values of the depleted cluster
places the location of tail-feather growth for these birds far outside of the species’ known wintering range
(Fig. 14). Thus, these individuals are molting during migration, pausing for at least a partial molt either
39
during the southward migration after leaving their breeding sites or on their way north the following year.
There is only one male in the depleted cluster, so females are more likely to adopt this alternate molt
strategy. Cluster membership and sampling region are not independent, suggesting that a female’s
molting strategy is related to where she breeds. Females from British Columbia have a tendency to molt
during migration, females from Northern Washington tend to molt at their wintering site, and females
from Southern Washington do not appear to be biased in either direction.
Calliope Hummingbirds have been recorded as subordinate to virtually all other hummingbird
species in the non-breeding season, but they do exhibit intraspecific territoriality (DesGranges 1978).
Male hummingbirds often defend higher quality territories, leaving females to occupy less desirable
resource patches (Lyon 1976, Kodric-Brown and Brown 1978, Kuban and Neill 1980). Molt is
energetically costly, and females that are subordinate to both males and other females may be unable to
obtain sufficient resources to complete molt before leaving their wintering sites to migrate northwards.
Instead, they may devote energy to preparing for migration, and then pause en route to complete their
molt. These females would make up the depleted cluster.
40
SUMMARY
This project used stable hydrogen isotopes to examine migratory connectivity in Calliope Hummingbirds.
Study design was based on two predictions. First, I predicted that a correlation would exist between
measured hydrogen-isotope ratios of juvenile Calliope feathers (δ2Hf) and predicted hydrogen-isotope
ratios of capture-site precipitation (δ2Hp), such that measured δ2Hf of adult birds could be transformed into
δ2Hp values of putative wintering sites. This prediction was not borne out. Such a correlation does not
exist in this data set, and juvenile δ2Hf values proved to be both more variable and more depleted relative
to δ2Hp than expected. The lack of correlation is likely due to the complex topography of the study area,
the contribution of snowpack to water available to plants during the growing season, and the mobility of
female hummingbirds foraging for nestlings. Research into animal movement that relies on isotopic
analyses should take these factors into consideration. Second, I predicted that hydrogen-isotope analysis
of tail feathers from adult birds would allow an evaluation of migratory connectivity in Calliope
Hummingbirds. Variation in δ2Hf values of adult birds, however, is too great to allow potentially distinct
wintering populations to be distinguished; even if strong connectivity were present, it would be obscured
by variability in feather-isotope values. This high variation is likely due to the topographic complexity
and small size of the non-breeding range of this species. This result suggests that stable isotopic methods
are better suited to broad-scale conclusions, and may be most appropriate for species that occupy large
breeding and wintering ranges with ample water-isotopic variation.
Adult Calliope Hummingbird δ2Hf values fell into two groups, an enriched and a depleted cluster. These
two clusters likely reflect different molting strategies. Birds in the depleted cluster (about 20 percent of
adult females) appear to grow their tail feathers well outside of the known wintering range for this
species, probably molting during migration. Isotope studies typically rely on knowledge of the molt
schedule of the species of interest, and this result suggests that homogeneity of molt strategy within a
species cannot be assumed, and that potential differences in molt timing and location must be taken into
account when interpreting feather isotope values.
41
REFERENCES
Ambrosini, R., A. P. Moller, and N. Saino. 2009. A quantitative measure of migratory
connectivity. Journal of Theoretical Biology 257:203-211.
Avise, J. C. 2000. Phylogeography. The history and formation of species. Harvard University
Press, Cambridge, MA.
Bairlein, F., D. R. Norris, R. Nagel, M. Bulte, C. C. Voigt, J. W. Fox, D. J. T. Hussell, and H.
Schmaljohann. 2012. Cross-hemisphere migration of a 25 g songbird. Biology Letters.
Baltosser, W. H. 1994. Age and sex determination in the Calliope Hummingbird. Western Birds
25:104-109.
Beacham, T. D., J. R. Candy, B. McIntosh, C. MacConnachie, A. Tabata, K. Kaukinen, L. Deng,
K. M. Miller, R. E. Withler, and N. Varnavskaya. 2005. Estimation of Stock Composition
and Individual Identification of Sockeye Salmon on a Pacific Rim Basis Using
Microsatellite and Major Histocompatibility Complex Variation. Transactions of the
American Fisheries Society 134:1124-1146.
Beason, J. P., C. Gunn, K. M. Potter, R. A. Sparks, and J. W. Fox. 2012. The Northern Black
Swift: Migration Path And Wintering Area Revealed. The Wilson Journal of Ornithology
124:1-8.
Bensch, S., T. Andersson, and S. Akesson. 1999. Morphological and Molecular Variation across
a Migratory Divide in Willow Warblers, Phylloscopus trochilus. Evolution 53:19251935.
Berthold, P. 1993. Bird Migration. A general survey. Oxford University Press, Oxford.
Bohonak, A. J. and K. van der Linde. 2004. RMA: Software for reduced major axis regression,
Java version. Version 1.21 http://www.kimvdlinde.com/professional/rma.html.
Bowen, G. J. 2011a. Gridded maps of the isotopic composition of meteoric waters.
http://www.waterisotopes.org.
Bowen, G. J. 2011b. The Online Isotopes in Precipitation Calculator. Version 2.2
http://www.waterisotopes.org.
Bowen, G. J., L. Chesson, K. Nielson, T. E. Cerling, and J. R. Ehleringer. 2005a. Treatment
methods for the determination of delta H-2 and delta O-18 of hair keratin by continuousflow isotope-ratio mass spectrometry. Rapid Communications in Mass Spectrometry
19:2371-2378.
Bowen, G. J. and J. Revenaugh. 2003. Interpolating the isotopic composition of modern meteoric
precipitation. Water Resources Research 39:1299.
Bowen, G. J., L. I. Wassenaar, and K. A. Hobson. 2005b. Global application of stable hydrogen
and oxygen isotopes to wildlife forensics. Oecologia 143:337-348.
Bridge, E. S., K. Thorup, M. S. Bowlin, P. B. Chilson, R. H. Diehl, R. W. Fléron, P. Hartl, K.
Roland, J. F. Kelly, W. D. Robinson, and M. Wikelski. 2011. Technology on the Move:
Recent and Forthcoming Innovations for Tracking Migratory Birds. Bioscience 61:689698.
Calder, W. A. and L. L. Calder. 1994. Calliope Hummingbird (Stellula calliope).in A. Poole,
editor. The Birds of North America Online. Cornell Lab of Ornithology, Ithaca.
Chamberlain, C. P., S. Bensch, X. Feng, S. Åkesson, and T. Andersson. 2000. Stable isotopes
examined across a migratory divide in Scandinavian willow warblers (Phylloscopus
trochilus trochilus and Phylloscopus trochilus acredula) reflect their African winter
42
quarters. Proceedings of the Royal Society of London. Series B: Biological Sciences
267:43-48.
Chamberlain, C. P., J. D. Blum, R. T. Holmes, X. Feng, T. W. Sherry, and G. R. Graves. 1996.
The use of isotope tracers for identifying populations of migratory birds. Oecologia
109:132-141.
DesGranges, J.-L. 1978. Organization of a tropical nectar feeding bird guild in a variable
environment. Living Bird 17:199-236.
Dittman, D. L. and D. K. Demcheck. 2006. Contributions to our knowledge of molt in the
Calliope Hummingbird. Birding 38.
Evans, K. L., J. Newton, K. J. Gaston, S. P. Sharp, A. McGowan, and B. J. Hatchwell. 2012.
Colonisation of urban environments is associated with reduced migratory behaviour,
facilitating divergence from ancestral populations. Oikos 121:634-640.
Farmer, A., B. S. Cade, and J. Torres-Dowdall. 2008. Fundamental limits to the accuracy of
deuterium isotopes for identifying the spatial origin of migratory animals. Oecologia
158:183-192.
González-Prieto, A. M., K. A. Hobson, N. J. Bayly, and C. Gómez. 2011. Geographic Origins
and Timing of Fall Migration of the Veery in Northern Colombia. The Condor 113:860868.
Gustafson, M. E., J. Hildenbrand, and L. Metras. 1997. The North American Bird Banding
Manual (Electronic Version). Version 1.0.
Hardesty, J. L. and K. C. Fraser. 2010. Using deuterium to examine altitudinal migration by
Andean birds. Journal of Field Ornithology 81:83-91.
Hartigan, J. A. and M. A. Wong. 1979. A K-means clustering algorithm. Journal of the Royal
Statistical Society: Series C (Applied Statistics) 28:100-108.
Hobson, K. A. 1999. Stable-carbon and nitrogen isotope ratios of songbird feathers grown in two
terrestrial biomes: implications for evaluating trophic relationships and breeding origins.
The Condor 101:799-805.
Hobson, K. A. 2008. Applying Isotopic Methods to Tracking Animal Movements. Pages 45-78
in K. A. Hobson and L. I. Wassenaar, editors. Tracking Animal Migration with Stable
Isotopes, Terrestrial Ecology Vol. 2. Academic Press, London.
Hobson, K. A., Y. Aubry, and L. I. Wassenaar. 2004a. Migratory connectivity in Bicknell's
Thrush: Locating missing populations with hydrogen isotopes. Condor 106:905-909.
Hobson, K. A., G. J. Bowen, L. I. Wassenaar, Y. Ferrand, and H. Lormee. 2004b. Using stable
hydrogen and oxygen isotope measurements of feathers to infer geographical origins of
migrating European birds. Oecologia 141:477-488.
Hobson, K. A. and R. G. Clark. 1992. Assessing avian diets using stable isotopes. I. Turnover of
C-13 in tissues. Condor 94:181-188.
Hobson, K. A., H. Lormee, S. L. Van Wilgenburg, L. I. Wassenaar, and J. M. Boutin. 2009a.
Stable isotopes (delta D) delineate the origins and migratory connectivity of harvested
animals: the case of European woodpigeons. Journal of Applied Ecology 46:572-581.
Hobson, K. A., K. P. McFarland, L. I. Wassenaar, C. C. Rimmer, and J. E. Goetz. 2001. Linking
breeding and wintering grounds of Bicknell's Thrushes using stable isotope analysis of
feathers. The Auk 118:16-23.
Hobson, K. A., S. L. Van Wilgenburg, K. Larson, and L. I. Wassenaar. 2009b. A feather
hydrogen isoscape for Mexico. Journal of Geochemical Exploration 102:167-174.
43
Hobson, K. A. and L. I. Wassenaar. 1997. Linking breeding and wintering grounds of
neotropical migrant songbirds using stable hydrogen isotopic analysis of feathers.
Oecologia 109:142-148.
Hobson, K. A., L. I. Wassenaar, and O. R. Taylor. 1999. Stable isotopes (δD and δ 13C) are
geographic indicators of natal origins of monarch butterflies in eastern North America.
Oecologia 120:397-404.
Irwin, D. E., J. H. Irwin, and T. B. Smith. 2011. Genetic variation and seasonal migratory
connectivity in Wilson’s warblers (Wilsonia pusilla): species-level differences in nuclear
DNA between western and eastern populations. Molecular Ecology 20:3102-3115.
Jones, J., D. R. Norris, M. K. Girvan, J. J. Barg, T. K. Kyser, and R. J. Robertson. 2008.
Migratory connectivity and rate of population decline in a vulnerable songbird. Condor
110:538-544.
Jouventin, P. and H. Weimerskirch. 1990. Satellite tracking of Wandering albatrosses. Nature
343:746-748.
Kelly, J. F. 2006. Stable isotope evidence links breeding geography and migration timing in
wood warblers (Parulidae). Auk 123:431-437.
Kimura, M., S. M. Clegg, I. J. Lovette, K. R. Holder, D. J. Girman, B. Milá, P. Wade, and T. B.
Smith. 2002. Phylogeographical approaches to assessing demographic connectivity
between breeding and overwintering regions in a Nearctic−Neotropical warbler (Wilsonia
pusilla). Molecular Ecology 11:1605-1616.
Kodric-Brown, A. and J. H. Brown. 1978. Influence of economics, interspecific competition, and
sexual dimorphism on territoriality of migrant Rufous Hummingbirds. Ecology 59:285296.
Kuban, J. F. and R. L. Neill. 1980. Feeding ecology of hummingbirds in the highlands of the
Chisos Mountains, Texas. Condor 82:180-185.
Langin, K. M., P. P. Marra, Z. Nemeth, F. R. Moore, T. K. Kyser, and L. M. Ratcliffe. 2009.
Breeding latitude and timing of spring migration in songbirds crossing the Gulf of
Mexico. Journal of Avian Biology 40:309-316.
Langin, K. M., M. W. Reudink, P. P. Marra, D. R. Norris, T. K. Kyser, and L. M. Ratcliffe.
2007. Hydrogen isotopic variation in migratory bird tissues of known origin: implications
for geographic assignment. Oecologia 152:449-457.
Lyon, D. L. 1976. A montane hummingbird territorial system in Oaxaca, Mexico. The Wilson
Bulletin 88:280-299.
Marra, P. P., K. A. Hobson, and R. T. Holmes. 1998. Linking winter and summer events in a
migratory bird by using stable-carbon isotopes. Science 282:1884-1886.
Marshall, J. D., J. R. Brooks, and K. Lajtha. 2008. Sources of Variation in the Stable Isotopic
Composition of Plants. Pages 22-60 Stable Isotopes in Ecology and Environmental
Science. Blackwell Publishing Ltd.
Martin, T. G., I. Chadès, P. Arcese, P. P. Marra, H. P. Possingham, and D. R. Norris. 2007.
Optimal conservation of migratory species. Plos One 2:e751.
McGuire, J. A., C. C. Witt, J. V. Remsen, R. Dudley, and D. L. Altshuler. 2009. A higher-level
taxonomy for hummingbirds. Journal of Ornithology 150:155-165.
Meehan, T. D., J. T. Giermakowski, and P. M. Cryan. 2004. GIS-based model of stable hydrogen
isotope ratios in North American growing-season precipitation for use in animal
movement studies. Isotopes in Environmental and Health Studies 40:291-300.
44
Norris, D. R., P. P. Marra, T. K. Kyser, T. W. Sherry, and L. M. Ratcliffe. 2004a. Tropical
winter habitat limits reproductive success on the temperate breeding grounds in a
migratory bird. Proceedings of the Royal Society of London Series B-Biological Sciences
271:59-64.
Norris, D. R., P. P. Marra, R. Montgomerie, T. K. Kyser, and L. M. Ratcliffe. 2004b.
Reproductive effort molting latitude, and feather color in a migratory songbird. Science
306:2249-2250.
Pain, D. J., R. E. Green, B. Giessing, A. Kozulin, A. Poluda, U. Ottosson, M. Flade, and G. M.
Hilton. 2004. Using stable isotopes to investigate migratory connectivity of the globally
threatened aquatic warbler Acrocephalus paludicola. Oecologia 138:168-174.
Paxton, K. L., C. Van Riper, T. C. Theimer, and E. H. Paxton. 2007. Spatial and temporal
migration patterns of Wilson's Warbler (Wilsonia pusilla) in the southwest as revealed by
stable isotopes. Auk 124:162-175.
Perez, G. E. and K. A. Hobson. 2009. Winter habitat use by Loggerhead Shrikes (Lanius
Ludovicianus) in Mexico: separating migrants from residents using stable isotopes.
Journal of Ornithology 150:459-467.
Prochazka, P., K. A. Hobson, Z. Karcza, and J. Kralj. 2008. Birds of a feather winter together:
migratory connectivity in the Reed Warbler Acrocephalus scirpaceus. Journal of
Ornithology 149:141-150.
Pyle, P. 1997. Identification guide to North American birds, part I: Columbidae to Ploceidae.
Slate Creek Press, Bolinas, CA.
Pyle, P., S. N. G. Howell, and G. M. Yanega. 1997. Molt, retained flight feathers, and age in
North American hummingbirds. Pages 155-166 in R. W. Dickerman, editor. The Era of
Allan R. Phillips: A Fetschrift. Horizon Communications, Albuquerque.
R Development Core Team. 2011. R: A Language and Environment for Statistical Computing.
Version 2.11.1 http://www.R-project.org. R Foundation for Statistical Computing,
Vienna, Austria.
Reudink, M. W., C. E. Studds, P. P. Marra, T. Kurt Kyser, and L. M. Ratcliffe. 2009. Plumage
brightness predicts non-breeding season territory quality in a long-distance migratory
songbird, the American redstart Setophaga ruticilla. Journal of Avian Biology 40:34-41.
Rolshausen, G., G. Segelbacher, K. A. Hobson, and H. M. Schaefer. 2009. Contemporary
Evolution of Reproductive Isolation and Phenotypic Divergence in Sympatry along a
Migratory Divide. Current biology : CB 19:2097-2101.
Rubenstein, D. R., C. P. Chamberlain, R. T. Holmes, M. P. Ayres, J. R. Waldbauer, G. R.
Graves, and N. C. Tuross. 2002. Linking breeding and wintering ranges of a migratory
songbird using stable isotopes. Science 295:1062-1065.
Rubenstein, D. R. and K. A. Hobson. 2004. From birds to butterflies: animal movement patterns
and stable isotopes. Trends in Ecology & Evolution 19:256-263.
Russel, S. M. and R. O. Russel. 2001. The North American banders' manual for banding
hummingbirds. The North American Banding Council, Point Reyes Station, CA.
Sarasola, J. H., J. J. Negro, K. A. Hobson, G. R. Bortolotti, and K. L. Bildstein. 2008. Can a
'wintering area effect' explain population status of Swainson's hawks? A stable isotope
approach. Diversity and Distributions 14:686-691.
Studds, C. E., T. K. Kyser, and P. P. Marra. 2008. Natal dispersal driven by environmental
conditions interacting across the annual cycle of a migratory songbird. Proceedings of the
National Academy of Sciences of the United States of America 105:2929-2933.
45
Stutchbury, B. J. M., S. A. Tarof, T. Done, E. Gow, P. M. Kramer, J. Tautin, J. W. Fox, and V.
Afanasyev. 2009. Tracking Long-Distance Songbird Migration by Using Geolocators.
Science 323:896.
Sulzman, E. W. 2008. Stable Isotope Chemistry and Measurement: A Primer. Pages 1-21 Stable
Isotopes in Ecology and Environmental Science. Blackwell Publishing Ltd.
Wassenaar, L. I. and K. A. Hobson. 1998. Natal origins of migratory monarch butterflies at
wintering colonies in Mexico: New isotopic evidence. Proceedings of the National
Academy of Sciences 95:15436-15439.
Wassenaar, L. I. and K. A. Hobson. 2000. Improved Method for Determining the StableHydrogen Isotopic Composition (δD) of Complex Organic Materials of Environmental
Interest. Environmental Science & Technology 34:2354-2360.
Wassenaar, L. I. and K. A. Hobson. 2001. A stable-isotope approach to delineate geographical
catchment areas of avian migration monitoring stations in North America. Environmental
Science & Technology 35:1845-1850.
Wassenaar, L. I. and K. A. Hobson. 2003. Comparative equilibration and online technique for
determination of non-exchangeable hydrogen of keratins for use in animal migration
studies. Isotopes in Environmental & Health Studies 39:211-217.
Wassenaar, L. I., S. L. Van Wilgenburg, K. Larson, and K. A. Hobson. 2009. A groundwater
isoscape ([delta]D, [delta]18O) for Mexico. Journal of Geochemical Exploration
102:123-136.
Webster, M. S. and P. P. Marra. 2005. The importance of understanding migratory connectivity
and seasonal interactions. Pages 199-209 in R. Greenberg and P. P. Marra, editors. Birds
of two worlds. The ecology and evolution of migration. Johns Hopkins University Press.
Webster, M. S., P. P. Marra, S. M. Haig, S. Bensch, and R. T. Holmes. 2002. Links between
worlds: unraveling migratory connectivity. Trends in Ecology & Evolution 17:76-83.
Welch, B. L. 1951. On the comparison of several mean values: an alternative approach.
Biometrika 38:330-336.
Wennerberg, L. 2001. Breeding origin and migration pattern of dunlin (Calidris alpina) revealed
by mitochondrial DNA analysis. Molecular Ecology 10:1111-1120.
Wikelski, M., R. W. Kays, N. J. Kasdin, K. Thorup, J. A. Smith, and G. W. Swenson. 2007.
Going wild: what a global small-animal tracking system could do for experimental
biologists. Journal of Experimental Biology 210:181-186.
46
Brown_CE-thesis2012.pdf