Can synchronizing feather-based measures

of corticosterone and stable isotopes help

us better understand habitat–physiology
relationships?

Graham D. Fairhurst, Matthias Vögeli,

David Serrano, Antonio Delgado, José
L. Tella & Gary R. Bortolotti

Oecologia

ISSN 0029-8549

Oecologia
DOI 10.1007/s00442-013-2678-8

1 23
Your article is protected by copyright and
all rights are held exclusively by Springer-
Verlag Berlin Heidelberg. This e-offprint is
for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 23
 Author's personal copy
Oecologia
DOI 10.1007/s00442-013-2678-8
PHYSIOLOGICAL ECOLOGY - ORIGINAL RESEARCH
Can synchronizing feather-based measures of corticosterone
and stable isotopes help us better understand habitat–physiology
relationships?
Graham D. Fairhurst • Matthias Vögeli •
David Serrano • Antonio Delgado • José L. Tella •
Gary R. Bortolotti
Received: 1 April 2012 / Accepted: 29 April 2013
Ó Springer-Verlag Berlin Heidelberg 2013
Abstract Physiological mechanisms link the environ-
ment with population dynamics, and glucocorticoid hor-
mones are of particular interest because they respond
adaptively to environmental change and can influence
vertebrate reproduction and fitness. We tested a novel
approach of synchronizing feather-based measures of cor-
ticosterone (the primary avian glucocorticoid; CORTf) and
ratios of stable isotopes (SIs) of C (d13C) and N (d15N) to
provide information about environmental conditions and an
integrated physiological response to those conditions over
the same period of feather synthesis. Using a fragmented
metapopulation of Dupont’s larks Chersophilus duponti, an
endangered steppe songbird, we analyzed interrelationships
among CORTf, d13C, d15N, and the physical environment,
including measures of habitat loss and fragmentation.
Communicated by Peggy Ostrom.
G. D. Fairhurst and M. Vögeli contributed equally.
G. R. Bortolotti: Deceased, 3 July 2011.
Electronic supplementary material The online version of this
article (doi:10.1007/s00442-013-2678-8) contains supplementary
material, which is available to authorized users.
G. D. Fairhurst (&)
M. Vögeli
G. R. Bortolotti
Department of Biology, University of Saskatchewan,
112 Science Place, Saskatoon, SK S7N 5E2, Canada
e-mail: graham.fairhurst@usask.ca
M. Vögeli
D. Serrano
J. L. Tella
Estación Biológica de Doñana (EBD-CSIC), C/ Americo
Vespucio S/N, Sevilla 41092, Spain
A. Delgado
Instituto Andaluz de Ciencias de la Tierra IACT (CSIC-UGR),
Avda. de las Palmeras, 4, Armilla, Granada 18100, Spain
CORTf was not related to any habitat variable measured
directly. However, we detected a significant spatial struc-
ture to CORTf values and food availability, with greater
similarity in both at smaller spatial scales. Using SIs as
proxies for the local environment, we found CORTf was
negatively related to d13C. Values of CORTf, d13C, and the
relationship between the two were likely driven by varia-
tion in agricultural land use surrounding lark habitat pat-
ches. Our feather-based approach revealed that individual
physiology was sensitive to environmental conditions (e.g.,
an interaction of food availability and variation in habitat)
at a local scale, but not patch or landscape scales. Com-
bining CORTf and SIs may be a promising tool because it
can provide individual-based information about habitat,
physiology, and their relationship during the same time
period.
Keywords Biomarker
Physiological conservation

Dupont’s lark
Fragmentation
Stress hormone
Introduction
Physiology mediates the relationship between animals and
their environment (Ricklefs and Wikelski 2002). Physio-
logical mechanisms that underpin responses to environ-
mental change are particularly important to conservation
because they can inform us about how animals respond to
and function within changing habitat. Habitat loss, frag-
mentation, and degradation (hereafter ‘‘habitat change’’)
are main drivers of species declines and extinctions
(Andrén 1994; Saunders et al. 1991), yet we know very
little about how individuals perceive, cope with, and
respond physiologically to habitat change and how this
may contribute to population trends. Complex interactions
123
 Author's personal copy
among habitat structure, food availability, predation, and
competition make it difficult to interpret the influences of
habitat change on individual animals.
The need for an effective biomarker of habitat change
has led to the use of measures of glucocorticoid hormones
(GCs) in habitat studies (Busch and Hayward 2009; Marra
and Holberton 1998; Suorsa et al. 2003; Wikelski and
Cooke 2006). Although exposure to stressors (e.g., severe
weather, predation, parasite infestation) generally results in
an increase in GCs aimed at helping the animal cope with
adverse conditions (Raouf et al. 2006; Romero 2004;
Romero et al. 2000; Sapolsky et al. 2000; Scheuerlein et al.
2001), there is flexibility in how individuals respond. For
example, some birds can suppress GC responses to chal-
lenges such as dietary restrictions (e.g., Fairhurst et al.
2012; Kitaysky et al. 2005). Importantly, the sensitivity of
GCs to environmental change (Romero 2004; Sapolsky
et al. 2000), and significant relationships between GCs and
components of vertebrate health, reproduction, and fitness
(Blas et al. 2007; Ouyang et al. 2011; Romero and Wi-
kelski 2001), make GCs an important physiological link
among habitat change, population dynamics, and ulti-
mately conservation (Wikelski and Cooke 2006).
Limited research with mixed results indicates that the
relationship between habitat change and GCs is far from
clear. Our understanding of how vertebrates respond
physiologically to habitat change is therefore limited
(Busch and Hayward 2009; Homyack 2010). This is due, in
part, to how previous GC studies have measured habitat.
Virtually all previous habitat–GC studies have used either:
(1) a dichotomous habitat classification (disturbed/not
disturbed, fragmented/continuous; but see Suorsa et al.
2003), which sheds little light on the actual factors that
influence individual levels of GCs; or (2) habitat measures
at scales which may not be the most appropriate scale for
physiological studies (for a review see Busch and Hayward
2009). Furthermore, all previous studies of habitat–GC
relationships have used blood or fecal measures of GCs,
which only provide short-term measures over hours or days
and can be logistically challenging in the field. Thus,
conservation would benefit from a long-term integrative
GC biomarker of habitat, and particularly one that is more
easily used in the field.
Here, we suggest that the novel approach of synchro-
nizing feather-based measures of GCs and stable isotopes
(SIs) may clarify the interplay of habitat and GCs. Feathers
contain a record of both corticosterone (CORT; the primary
avian GC) and SIs over the period of feather growth
(Bortolotti 2010; Bortolotti et al. 2008; Hobson and Was-
senaar 2008; Inger and Bearhop 2008). CORT in feathers
(CORTf) is positively correlated with blood levels of
CORT (Bortolotti et al. 2008; Lattin et al. 2011) and
integrates the magnitude and duration of GC responses,
123
Oecologia
including both baseline and responses to stressors. Thus,
CORTf provides a long-term measure of avian physiolog-
ical responses to environmental variation. CORTf values
are positively correlated with capture and restraint and
short-term environmental enrichment (Bortolotti et al.
2008; Fairhurst et al. 2011), and can negatively vary with
nutritional condition (Fairhurst et al. 2012). However, no
previous study has addressed how CORTf relates to habitat
change.
Ratios of SIs of C (d13C) and N (d15N) from feathers are
commonly used in avian spatial and dietary studies (Inger
and Bearhop 2008) and complement CORTf by providing
an individual measure of local environment. When tissues
such as feathers are synthesized, they incorporate C and N
that is ultimately derived from food sources. In terrestrial
systems, C3 plants have lower (i.e., more negative) d13C
values than C4 plants, so d13C values can be used to dif-
ferentiate photosynthetic pathways (Kelly 2000). Thus,
d13C values in feathers can reliably reflect the food web
from which individuals were feeding (Inger and Bearhop
2008; Kelly 2000; Marra et al. 1998). Feathers can have
higher d15N values with each step in trophic level (Hobson
1999; Inger and Bearhop 2008; Kelly 2000).
Agricultural intensification influences SI values.
Increased fertilizer inputs, animal waste, and soil tillage can
result in significantly higher d15N values (Alguacil et al.
2011; Hebert and Wassenaar 2001; Hobson 1999). Irriga-
tion, which can increase the abundance and diversity of
arthropods in native and farmed landscapes (Frampton et al.
2000; Wenninger and Inouye 2008), can lower the d13C
values of plants through changes in water use efficiency
(Choi et al. 2005; Wright et al. 1994). Moreover, the d13C
values of arthropods increase with the percentage of C4
crops (e.g., corn; Briones et al. 2001). The SI composition
of feathers can reflect SIs in the local environment, making
feather-based SIs useful biomarkers for detecting shifts in
food webs due to agriculturally driven habitat change.
Our understanding of how animals respond physiologi-
cally to anthropogenic change would benefit from a more
holistic perspective, which may be provided by a combi-
nation of individual-based biomarkers, particularly those
from feathers (Lens and Eggermont 2008). Collecting
information about environmental conditions, and physio-
logical responses to those conditions, easily and non-
invasively is an important step toward an early-warning
system for conservation (Clarke 1995; Lens and Eggermont
2008). Synchronizing feather-based measures of CORT
with SIs is a strong approach because it can supply such
information.
Here, we test for the first time the utility of combining
CORTf with d13C and d15N as a measure of individual
responses to environmental conditions and habitat change.
Our study system is a fragmented metapopulation of
 Author's personal copy
Oecologia
Dupont’s larks Chersophilus duponti (Vieillot) (henceforth
‘‘lark’’), an endangered songbird. The lark’s semi-arid
steppe habitat has suffered large loss and fragmentation
due to agricultural conversion, resulting in population
declines, decreased song repertoire, and increased
inbreeding in local populations (Laiolo and Tella 2006;
Méndez et al. 2011; Vögeli et al. 2010).
Habitat change and individual physiology may vary at
different spatial scales, so we begin by using a spatial
autocorrelation analysis to better understand the influence
of spatial scale on the distribution of CORTf in the meta-
population. We then test four predictions of the hypothesis
that CORTf should be related to direct and indirect (i.e.,
d13C and d15N from feathers) measures of the local envi-
ronment. Extensive research has shown that adverse envi-
ronmental conditions generally result in an increase in GCs
(for reviews see Romero 2004; Sapolsky et al. 2000), and
we reason that agricultural encroachment should be chal-
lenging to habitat specialists, such as larks, that are highly
dependent on natural vegetation. Thus, we predict that:
1. Larks surrounded by a relatively greater proportion of
agriculture should have relatively higher CORTf
values.
2. Related to the latter, we expect values of both d13C and
d15N to positively correlate with the extent of agricul-
ture because corn (a C4 grass) is grown throughout the
study area and can strongly influence d13C values there
(Carrete et al. 2009).
3. Therefore, CORTf should be positively related to both
d13C and d15N.
4. Larks are strict habitat specialists that tend to crowd in
remnant habitat patches (Vögeli et al. 2010), and
crowding can adversely affect the physiology of free-
living birds (Kilgas et al. 2006; Mazerolle and Hobson
2002). Thus, we predict that larks from smaller, more
isolated, and more densely populated habitat patches
should have relatively higher CORTf.
To provide context to our CORTf and SI results, we
consider the influence of weather and how arthropod bio-
mass, species richness, and diversity varied spatially
throughout the study area. We conclude by integrating all
of our datasets to contrast how GCs relate to habitat metrics
at two spatial scales (local and landscape) and with varying
environmental conditions.
Materials and methods
Study system and field methods
Fieldwork was carried out in the Ebro Valley (northeast
Spain; Fig. 1) between 2004 and 2006. The climate is
semiarid with little seasonal precipitation and maximum
temperatures regularly exceeding 40 °C (GDN 2011; IAE
2011). Lark habitat is restricted to flat steppes with natural
vegetation, and habitat change over the past few decades
has led to a scattered distribution of predominantly small
local populations in discrete habitat patches within a
landscape matrix dominated by agriculture (Fig. 1) (Vögeli
et al. 2010). Larks are almost exclusively insectivorous but
feed on a broad spectrum of arthropod families (Cramp
1988; Herranz et al. 1993). Habitat change is negatively
affecting lark population dynamics, and dispersal appears
to be very limited (Laiolo 2008; Laiolo et al. 2008; Vögeli
et al. 2008, 2010).
Larks were captured, ringed, and measured as described
in Vögeli et al. (2007). Because we used playbacks of
recorded vocalizations to capture larks, over 90 % of
captured individuals were male. Thus, we omitted juvenile
and female larks from our analyses. During ringing, we
collected a fully grown (dead) outer tail feather (6th rec-
trix) from each bird and additionally extracted a drop of
blood for molecular sexing (Vögeli et al. 2007). For both
the 2004 and 2005 breeding seasons, we trapped some
individuals in the fall after the post-breeding molt and
others the following spring. Thus, the feathers collected
from both seasons within one ‘‘feather-year’’ (i.e., fall
2004/spring 2005 and fall 2005/spring 2006) were grown
during the same time period because larks molt only once
per year in the fall (Svensson 1992). We sampled 102 and
42 larks in feather-years 2004 and 2005, respectively.
Larks in the Ebro Valley have extremely low natal and
breeding dispersal (Laiolo and Tella 2008; Vögeli et al.
Fig. 1 Map of the Ebro Valley study area in northeast Spain, where
the second largest European population of Dupont’s lark Chersoph-
ilus duponti remains (inset photo of Dupont’s lark). Remnant habitat
patches occupied by larks are shown in black. Solid lines encircling
patches identify the local populations sampled
123
 Author's personal copy
2008), so we are confident that feathers were grown in the
same area where they were ultimately collected. The sta-
bility of CORTf is not affected by feather age (Bortolotti
et al. 2009). All field procedures complied with the laws of
the government of Spain.
To characterize habitat and spatial structure of the Ebro
Valley metapopulation, we used variables previously cal-
culated by Vögeli et al. (2010) and described and listed in
Table 1. The vegetation of lark habitat patches (veg6) was
classified according to phytosociological criteria (Braun-
Blanquet and de Bolós 1957). Lark habitat patch sizes (area)
were identified by the intersection of all lark locations
obtained during field work in the Ebro Valley from 2004 to
2007 (n = 2,035) (Vögeli et al. 2010) with highly detailed
land-use maps (SIGPAC 2011). Based on high-resolution
aerial orthophotographs (1:5,000, year 2006), we adjusted
these areas by discarding non-suitable steppe habitat for
larks because of exceeding slopes (i.e., steeper than 5 %;
Seoane et al. 2006). Steppe patches are easily recognizable
at a landscape scale since they are island-like remnants in a
landscape dominated by agriculture with evident boundaries
to other land uses. Finally, all resulting habitat patches were
digitized to calculate their patch size in ArcGIS 9 (ESRI
2004). Using CORINE land cover digital maps from the
European Environmental Agency (year 2000, map resolu-
tion 100 m), we calculated the percentage of land-use cover
in the landscape matrix inside a buffer ring with 20-km
radius [i.e., the maximum dispersal distance of larks
detected in our study area (Vögeli et al. 2010)] from the
patch edge. The main land uses of the landscape matrix were
categorized as: (1) intensive agriculture (INTA), (2) non-
irrigated arable land (NIAL), and (3) natural steppe vege-
tation (NVEG). The isolation (or inversely, connectivity) of
each habitat patch was characterized as: (1) the distance to
the nearest occupied habitat patch (isolation1), and (2) the
isolation index Ii = -Rexp(-dij)Nj, where dij is Euclidean
distance between patches i and j, and Nj is the number of
occupied territories of patch j (isolation2) (Hanski et al.
1994). Lark population sizes (Npop) were calculated as the
number of occupied territories through territory mapping of
males aided by its acoustic identification and observations of
individually color-ringed birds (Vögeli et al. 2008). Lark
breeding densities in each habitat patch (density) were
characterized as the number of occupied territories divided
by patch size (Laiolo 2008; Vögeli et al. 2010).
Data on the total measured precipitation during the lark
molting period (PREC) of July–September, and average
maximum daily temperatures (TMAX) during the same
period, were obtained from regional administration web-
sites (GDN 2011; IAE 2011). Weather data were collected
from nine meteorological stations across the study area,
and we used weather data from the station nearest to each
local population (mean distance 10.5 km, range 0–20 km).
123
Oecologia
Arthropod abundance was calculated from previously
published data; for complete methodology, see Vögeli et al.
(2010). Briefly, in May and June 2005 we installed five
pitfall traps (filled with 70 % ethyl alcohol) in each of 52
linear transects throughout 17 local populations. Trap
locations were determined randomly but separated by at
least 25 m. Traps were emptied after 3 days and the con-
tents filtered, washed, and stored in 70 % ethyl alcohol. All
specimens were identified by the same investigator (M. V.)
to species level when possible or to morphospecies and
categorized as recognizable taxonomic units (RTUs) (Oli-
ver and Beattie 1993). For each transect we counted the
number of RTUs trapped as a proxy for species richness
and, by weighing the RTU abundance against its respective
mean body length, we calculated a biomass index as the
average value obtained for all sampling points. We then
computed a species diversity index of each transect
(Shannon and Weaver 1949) based on the number of
RTUs.
Corticosterone analysis
CORTf assays followed methods in Bortolotti et al. (2008)
and have been replicated with other species (Fairhurst et al.
2011, 2012; Harms et al. 2010). Each feather was measured
from end to end (i.e., proximal calamus to distal tip of the
vane) by flattening and straightening the entire feather
length against a metal ruler. The calamus was then
removed and discarded and the remaining sample feather
length was re-measured as above. Each feather was then
placed in a separate glass vial and cut into pieces \5 mm2
with scissors. We extracted CORT from feathers by adding
10 mL of methanol (high performance liquid chromatog-
raphy grade; Fisher Scientific, Fair Lawn, NJ) to each vial
and placing the samples in a sonicating water bath at room
temperature for 30 min, followed by incubation at 50 °C
overnight. Methanol was separated from feather material
by vacuum filtration, using synthetic polyester fibre in the
filtration funnel. Extracts were placed in a 50 °C water bath
and subsequently evaporated. Extract residues were
reconstituted in a small volume of phosphate buffer system
(0.05 M, pH 7.6) and frozen at -20 °C until analyzed by
radioimmunoassay (RIA). We assessed the recovery effi-
ciency using feather samples spiked with a small amount
(approximately 5,000 counts min-1) of 3H-corticosterone
in the extraction. Approximately 96 % of the radioactivity
was recoverable in the reconstituted samples. For more
information about validation, see Supplementary Appendix
S1 in Bortolotti et al. (2008).
CORTf levels were determined by RIA, and all CORTf
assays were performed at the University of Saskatchewan,
Canada. Measurements were performed on reconstituted
methanol extracts and were duplicated. Samples were
 Author's personal copy
Oecologia

Table 1 Variables characterizing the habitat and spatial structure of the Ebro Valley metapopulation of Dupont’s lark Chersophilus duponti
Description Definition

Vegetation community of habitat
occupied by larks (veg6)
Matrix land-use cover
Intensive agriculture (INTA)
Non-irrigated arable land (NIAL)
Natural steppe vegetation (NVEG)
Patch size (area)
Local population size (Npop)
Local population density (density)
Nearest neighbor distance (isolation1)
Isolation index (isolation2)
All variables were previously described by Vögeli et al. (2010) or Laiolo and Tella (2008)
Six phytosociological classes based on Braun-Blanquet and de Bolós (1957)
Percentages of land-use cover in the landscape matrix within 20 km (maximum dispersal distance of
Dupont’s larks) from each population edge
These three main land-use categories described 83–97 % of the total matrix cover
Size (ha) of habitat patches
Number of occupied territories within each local population
Number of occupied territories per ha
Mean straight-line distance between the nearest singing neighbor (Laiolo and Tella 2008)
Ii = -Rexp(-dij)Nj, where dij is the Euclidean distance between populations i and j, and Nj is the
number of occupied territories of population j (Hanski et al. 1994)

randomized throughout ten assays that had an average
(±SD) intra-assay coefficient of variation of 7.1 (±3.03)
%. Average (±SD) minimum detectability (ED80) was
11.19 (±1.11) pg/assay tube, but we had no undetectable
samples and all assay values were above this limit. Data
values are expressed as pg CORT mm-1 of feather, which
gives a valid estimate of CORT per unit time of feather
growth (for validation see Bortolotti 2010; Bortolotti et al.
2008, 2009).
SI analysis
A small (*1 mm2) piece of feather was cut from the distal
vane of each feather prior to CORT analysis. SI values can
vary considerably within feathers, such that a small piece
of vane might not be representative of an entire feather
(Wiley et al. 2010). Therefore, we calculated repeatability
estimates (Nakagawa and Schielzeth 2010) for d13C and
d15N values from proximal and distal portions of a large
subsample of lark feathers. The average (±SD) isotopic
difference between proximal and distal portions was
0.433 ± 1.234 % for d13C (n = 91) and 0.073 ± 1.089
for d15N (n = 95), and the values from proximal and distal
portions were statistically highly repeatable (d13C,
R = 0.563, v2 = 32.7, df = 1, P \ 0.0001; d15N,
R = 0.843, v2 = 113.5, df = 1, P \ 0.0001). We thus
concluded that SI values varied little during the time that it
took the 6th rectrix to grow and represent lark diet gener-
ally during the time frame of our study. Each feather
sample was cleaned by ultrasonic shaking for 20 min in 2:1
methanol:chloroform (three times) and in deionised water
in glass vials, then dried at 70 °C. Approximately
0.5–1 mg of each feather piece was loaded into a tin cup
and combusted at 1020 °C in an elemental analyzer (Carlo
Erba EA1500 NC; Carlo Erba, Milan) connected to a mass
spectrometer (Finnigan Delta Plus XL; Finnigan, Bremen).
SI abundance was expressed in standard notation relative to
Vienna Pee Dee belemnite and atmospheric N for C and N,
respectively: d13C or d15N = [(Rsample/Rstandard) -
1] 9 1,000, where R = 13C/12C or 15N/14N of the sample
and international reference materials, respectively. All
samples were analyzed three times on different days, and
two secondary isotopic references materials were measured
for every *ten samples. Commercial CO2 and N2 were
used as the internal standard for the C and N isotopic
analyses. For C, phthalic acid (-30.63 %) and sucrose
(-11.65 %) were used as internal reference materials. For
N, urea (-1.02 %) and shark cartilage (?16.01 %) were
used as internal reference materials. Precision was calcu-
lated, after correcting for the mass spectrometer daily drift,
from reference materials systematically interspersed in
analytical batches. Precision was better than ±0.1 % for d
15
N and d 13C.
Statistical analyses
We first analyzed patterns of spatial associations of CORTf
across the entire metapopulation using Moran’s I index
(Sokal and Oden 1978). We assumed a null hypothesis of
random spatial distribution of CORTf where a statistically
significant positive deviation from the expected value of
Moran’s I would represent a higher similarity of CORTf
among individuals than expected by chance alone. First, we
performed a global analysis with the entire dataset for each
feather-year. We calculated all pair-wise distances between
individuals, and grouped them in distance classes (bins)
with a lag of 5 km from 0 to 40 km. Due to low sample
sizes, the rest of the individuals were merged into two bins
(40–100 and [100 km) for feather-year 2004, and into one
bin ([40 km) for feather-year 2005. The number of pair-

123
 Author's personal copy
Oecologia

wise comparisons ranged from 244 to 3,938 (2004), and 38
to 850, respectively (2005). We conducted a second anal-
ysis at small spatial scale with only individuals from
feather-year 2004 and used bins with a lag of 500 m from 0
to 5 km. We then calculated Moran’s I for arthropod bio-
mass, species richness, and diversity using the same bins as
for CORTf analyses. Spatial analyses were conducted using
SAM v. 4 (Rangel et al. 2010), and test significance
(a = 0.05) was calculated with 999 permutations.
We analyzed relationships between CORTf, d13C, d15N,
and the physical environment by generalized linear mixed
models (GLMMs) using PROC GLIMMIX in SAS v. 9.2
(SAS Institute, Cary, NC). d13C and d15N were each used
= 1.46, P = 0.23). PREC was greater in 2004 than in 2005
(mean ± SE: 2004 = 45.97 ± 6.10 mm, 2005 = 23.88 ±
3.41 mm; F1,10 = 9.98, P = 0.01; Fig. 2c), whereas TMAX
did not differ between years (mean ± SE: 2004 =
29.68 ± 0.61 °C, 2005 = 30.06 ± 0.35 °C; F1,19 = 0.21,
P = 0.65). PREC and TMAX were not correlated in either
year (Pearson correlations: 2004, r = -0.105, P = 0.67,
n = 19 populations; 2005, r = 0.367, P = 0.22, n = 13
9.0
a
8.5
8.0

CORTf (pg mm )
as main effects in separate models with CORTf as the

-1
dependent variable. All GLMMs included feather-year as a 7.5

covariate, a main effect 9 feather-year interaction term, 7.0

and population identity as a random term. CORTf values 6.5
were log transformed to improve normality. All GLMMs
6.0
were fitted with a normal error distribution and an identity
link function. A Laplace correction for computing the 5.5

marginal log likelihood was employed to allow the use of 5.0

likelihood-based inference (Bolker et al. 2009). Model 2004 2005

selection procedures were implemented to assess the 5.8

b
strength of evidence for the relative influence of the dif- 5.6
ferent independent variables included in the models for
each tested prediction (Bolker et al. 2009; Burnham and 5.4

Anderson 2002). We included the entire set of independent 5.2

δ N

variables and covariates separately for the full models of

15

5.0
each tested prediction. All candidate models (i.e., all pos-
sible combinations of variables) were then derived from the 4.8
full models. Then, because we were interested in popula-
4.6
tion-level effects, we used conventional information-theo-
retic inference (Vaida and Blanchard 2005) based on the 4.4
Akaike’s information criterion corrected for small sample -22.4 -22.2 -22.0 -21.8 -21.6 -21.4 -21.2
δ C
13
sizes (AICc) and ranked the candidate models according to
their differences in AICc. Candidate models within two 31.0
AICc points of the most parsimonious model (i.e., smallest c
AICc) were considered as having similar statistical support.
30.5
TMAXmoult (°C)

30.0
Results
29.5
Overall, CORTf values were significantly higher in feather-
year 2004 than they were in 2005 (mean ± SE: 2004 =
29.0
8.06 ± 0.35 pg mm-1, 2005 = 6.27 ± 0.28; F1,135 = 8.54,
P = 0.004; Fig. 2a). Values of d13C (range: 2004 = -24.57
28.5
to -19.62 %, 2005 = -22.71 to -20.12 %) were, on 15 20 25 30 35 40 45 50 55
average, significantly lower in 2004 than in 2005 PREC (mm)
(mean ± SE: 2004 = -22.10 ± 0.08 %, 2005 = -21.38
± 0.10 %; F1,126 = 28.60, P \ 0.001; Fig. 2b), whereas Fig. 2 Mean (±SE) plots of a corticosterone (CORTf) and b ratios of
stable isotopes (SIs) of C (d13C) and N (d15N) from feathers of
d15N values (range: 2004 = 1.08–12.45 %, 2005 = 1.21–
Dupont’s larks, and c total precipitation (PREC) and average
11.21 %) did not differ between feather-years (mean ± SE: maximum daily temperature (TMAX) during molting in 2004 (filled
2004 = 4.73 ± 0.17 %, 2005 = 5.16 ± 0.40 %; F1,128 circles) and 2005 (open circles)

123
 Author's personal copy
Oecologia
populations), suggesting that inter-annual differences in
PREC operated independently of temperature.
Spatial patterns of CORTf and arthropods
Using spatial autocorrelation analysis, we detected a
positive effect of proximity on CORTf (Fig. 3). The global
model for feather-year 2004 revealed statistically signifi-
cant positive spatial autocorrelation in the first bin
(0–5 km) only (expected Moran’s I = -0.009, observed
Moran’s I = 0.195, P = 0.001). Analyses failed to detect
any significant spatial effects (all P [ 0.32) for feather-
year 2005, likely owing to smaller sample size. At a
small spatial scale we detected positive spatial autocorre-
lation in three bins (0–500 m, I = 0.506, P = 0.001;
2,000–2,500 m, I = 0.353, P = 0.002; 3,000–3,500 m,
I = 0.799, P = 0.001). When we removed the largest
populations ([40 occupied territories) from this analysis,
so that between-individual distances of [3.5 km implied
that individuals were in separate local populations, only the
effects in the first bin were maintained (0–500 m,
I = 0.500, P = 0.002; all other bins, P [ 0.17). In sum-
mary, we found higher similarity of CORTf in pairs of
individuals at very limited spatial scale and, importantly,
only within discrete local populations.
Applying the same spatial scales used for analysis of
CORTf to the arthropod data, we found a statistically sig-
nificant positive spatial autocorrelation in the first bin
(0–5 km) for biomass (expected Moran’s I = -0.02,
observed Moran’s I = 0.413, P = 0.001), but not for
species richness or diversity. This pattern is similar to that
0.2 ***
0.1
Moran's I
0.0
-0.1
-0.2
0 20 40 60 80 100 120 140
Distance between Dupont's larks (km)
Fig. 3 Correlogram demonstrating the geographic structure of
CORTf similarity. Only pairs of individuals separated by straight-
line distances \5 km in 2004 (marked with asterisks) showed
significantly similar values (P = 0.001). Filled circles Values for
2004, open circles values for 2005. Lines (solid 2004, dotted 2005)
show expected values of Moran’s I
of CORTf, suggesting a significantly high similarity of
arthropod biomass at a local spatial scale.
Relationships among CORTf, d13C and d15N,
and the physical environment
Of the candidate models explaining variation in CORTf in
terms of traditional habitat measures, the most parsimonious
model included only feather-year (Supplemental Appendix
1). Although this suggests that traditional habitat measures
did not improve the model fit compared to feather-year at
explaining the variation in CORTf, a model including
both NIAL and NVEG received statistically similar sup-
port, indicating weak relationships with these variables.
Regarding the influence of weather during the molting
period on CORTf, the most parsimonious model included
PREC only (Supplemental Appendix 1). However, the
model with overwhelming support explaining the variation
in CORTf included d13C, feather-year, and a significant
interaction term (Table 2). When we analyzed the rela-
tionship between d13C and CORTf separately by feather-
year, we found a significant negative relationship for 2004
(F1,81 = 17.24, P \ 0.0001; Fig. 4), but not for 2005
(F1,27 = 0.60, P = 0.44). Our sample size was considerably
smaller in 2005, which likely reduced statistical power.
Including d15N as a predictor variable did not improve the fit
of any model explaining significant variation in CORTf.
Similarly, combining habitat, weather, and SI ratio predictor
variables in a single model did not produce a model fit better
than analyzing variables as separate groups (Table 2).
The best-supported model explaining variation in d13C in
terms of habitat included negative effects of INTA and
NVEG (Supplemental Appendix 2). The next most parsi-
monious model received nearly identical statistical support
and also included NIAL (Supplemental Appendix 2), sug-
gesting that d13C values reflected the landscape matrix in
general (Fig. 5). The most parsimonious model explaining
variation in d13C in terms of weather during molting inclu-
ded effects of TMAX and feather-year, but the estimate on
the former parameter was poor (Supplemental Appendix 2).
A model receiving similar statistical support included only
feather-year (Supplemental Appendix 2). Thus, it appears
that the influence of weather on d13C was likely weak.
Variation in d15N was explained most parsimoniously
by NVEG (Supplemental Appendix 2), but the second and
third best-supported models additionally included NIAL
160
and INTA, respectively. Adding any predictor variable to
the null model did not improve the model fit when ana-
lyzing the variation in d15N in terms of weather during
molting (Supplemental Appendix 2). Thus, similar to d13C,
variation in d15N was best explained by land use sur-
rounding lark habitat patches, and was apparently not
related to weather.
123
 Author's personal copy
Oecologia

Table 2 Best-supported models, explaining variation in lark feather corticosterone (CORTf), from each of the groups of predictor variables:
traditional measures of habitat, weather during molting, stable isotope ratios, and metapopulation characteristics
Model Parameters Estimate SE df t P AICc DAICc
13
1 d C 0.0268 0.0353 1,119 2.14 0.147 -134.14 0
Feather-year -2.4031 0.8328 1,119 8.33 0.005
d13C 9 feather-year -0.1134 0.0386 1,119 8.61 0.004
2 d13C 0.0295 0.0361 1,117 1.59 0.21 -133.02 1.12
d15N -0.0033 0.0101 1,117 1.88 0.173
Feather-year -2.3006 0.8649 1,117 7.08 0.009
d13C 9 feather-year -0.1115 0.0394 1,117 8 0.006
d15N 9 feather-year -0.0125 0.0126 1,117 0.98 0.324
3 PREC 0.0028 0.0009 1,121 9.9 0.002 -118.87 15.27
4 TMAX -0.0217 0.0191 1,120 1.3 0.257 -117.98 16.16
PREC 0.003 0.0009 1,120 10.97 0.001
5 Feather-year 0.0798 0.0284 1,121 7.89 0.006 -116.98 17.16
6 NIAL -0.0053 0.0026 1,119 4.16 0.04 -116.66 17.48
NVEG -0.0036 0.0036 1,119 1 0.32
Feather-year 0.0798 0.0283 1,119 7.96 0.006
7 Isolation2 -0.0533 0.045 1,120 1.4 0.239 -116.18 17.96
Feather-year 0.0758 0.0286 1,120 7.05 0.009
8 Isolation1 0.0027 0.0043 1,120 0.39 0.531 -115.23 18.91
Feather-year 0.0788 0.284 1,120 7.7 0.006
9 Area -0.00004 0.0001 1,120 0.12 0.728 -114.96 19.18
Feather-year 0.0801 0.0284 1,120 7.95 0.006
10 Density -0.0479 0.2142 1,120 0.05 0.823 -114.88 19.26
11 Null -111.42 22.72
Models are ranked by Akaike’s information criterion corrected for small sample sizes (AICc), and the null model is included for comparison
PREC Total measured precipitation during the lark molting period (July–September), TMAX average maximum daily temperatures during the
lark molting period (July–September); for other abbreviations, see Table 1

25.0 feather-year (Supplemental Appendix 3). Population size,

isolation, and density variables failed to improve model fit.
20.0 Thus, we lack statistical evidence that any of these vari-
ables were significantly related to CORTf.
CORTf (pg/mm)

15.0

Discussion
10.0

We tested the novel approach of synchronizing measures of

5.0
0.0
-24.5 -24.0 -23.5 -23.0 -22.5 -22.0 -21.5 -21.0 -20.5 -20.0
δ13C
Fig. 4 Relationship between feather-based measures of d13C and
CORTf from Dupont’s larks in 2004 (filled circles) and 2005 (open
circles). Line shown is for 2004 data only (P \ 0.0001)
Relationships between CORTf and the spatial structure
of the metapopulation
The best-supported models explaining CORTf in terms of
population size, isolation, and density included only
CORTf with SIs to study habitat–physiology relationships
in a fragmented metapopulation of birds. We detected a
significant spatial structure to CORTf values, driven by
similarity among individuals within discrete habitat pat-
ches, suggesting that CORTf was sensitive to local envi-
ronmental conditions. However, contrary to our prediction,
CORTf was not related to any of the traditional habitat
variables we measured. Despite the importance of the local
environment, traditional measures of habitat may not
quantify individual habitat requirements or constraints. The
strength of SIs is that they provide information about the
resources actually used by individuals. Our study presents
an innovative way to combine that information with a
measure of individual physiology during the same period

123
 Author's personal copy
Oecologia
-19 -19
-20 -20
-21 -21
δ13C
-22 -22
-23 -23
-24 -24
-25 -25
10 20 30 40 50 60 70
Non-irrigated arable land (%)
Fig. 5 Relationships between amount of matrix land-use cover and the ratio of d13C from feathers of Dupont’s larks in 2004 (filled circles) and
2005 (open circles)
of feather synthesis. We used SIs as proxies for local
resource use and detected a strong significant negative
relationship between d13C and CORTf. The direction of
this relationship suggests that larks feeding on diets with
relatively lower d13C values may have been coping with
relatively greater energetic challenges. The amount of food
locally available to larks could have been related to both
CORTf and d13C, and the significant spatial structure of
arthropod biomass provides some support for this expla-
nation. We also found significantly lower CORTf values in
the second year of our study, highlighting the temporal
sensitivity of this physiological measure.
CORTf and the spatial structure of the metapopulation
Why were none of the habitat and spatial structure vari-
ables related to CORTf? We believe this is an important
negative result. For the same inter-individual distance,
larks within suitable habitat patches showed significant
similarity in CORTf, whereas individuals separated by
unsuitable habitat did not. In other words, the closer indi-
viduals were spatially, the more similar their CORTf values
were, but only within local populations. Habitat loss and
fragmentation in our study area have led to highly isolated
and island-like local populations that have independent
population dynamics (Laiolo et al. 2008; Vögeli et al.
2010, 2011). Consequently, CORTf patterns that represent
physiological responses to local conditions may no longer
correlate with habitat variables that operate beyond the
patch environment. Although this may be a reasonable
explanation, when we considered a more appropriate scale
and variables that are more likely to directly influence
physiology (i.e., local vegetation community, size of local
populations), we still failed to find relationships with
CORTf. Similarly, we lack evidence to support an influence
of conspecific density on CORTf, and add to a very small
number of studies addressing this in the wild (e.g., Fokidis
et al. 2011).
-19
-20
-21
-22
-23
-24
-25
5 10 15 20 25 30 35 15 20 25 30 35 40
Intensive agriculture (%) Natural vegetation (%)
Lark food availability likely played an important role in
determining spatial patterns of CORTf. Larks feed on a
wide variety of ground-dwelling arthropods (see below),
and CORT can be released in response to both the quality
and quantity of food, and can facilitate foraging behaviors
(Astheimer et al. 1992; Kitaysky et al. 2001). Because
arthropod biomass, but not species richness or diversity,
varied similarly to CORTf (i.e., at the local, but not land-
scape, scale), it is likely that larks were responding phys-
iologically to local variation in the amount of food (i.e.,
hunger), rather than particular prey types. Indeed, CORTf
values may reflect food availability and the energetic
exertion of foraging (Carrete et al. 2013). Patterns of
CORTf are also likely to reflect, at least partly, genetic
effects because there is a genetic basis to avian responses to
stressors (Evans et al. 2006). A combination of heteroge-
neous habitat change and local environmental pressures in
the study area may have resulted in spatial variation in lark
responses to stressors. Increased population isolation has
led to greater inbreeding and genetic relatedness among
larks within local populations of our study area (Méndez
et al. 2011), adding some credence to a genetics-based
explanation. Regardless of the mechanism, although patch-
and landscape-level variables such as patch size, isolation,
and composition of the landscape matrix influence popu-
lation dynamics in highly fragmented populations such as
this one (Vögeli et al. 2010), we have shown that individual
physiology is likely determined at much smaller spatial
scales and may operate separately from these processes.
Interplay among CORTf, SIs, and the physical
environment
Values of d13C were related most strongly to the land uses
surrounding lark habitat patches (i.e., in the landscape
matrix). This indicates that larks were likely influenced by
diet subsidies originating beyond the local environment
(Polis et al. 1997), because larks forage exclusively within
123
 Author's personal copy
habitat patches (Cramp 1988; Herranz et al. 1993). Overall,
variation in d13C values was low, and values were not high
enough in either year to indicate that larks were feeding
mainly from a C4-based food web (Kelly 2000; see also
Carrete et al. 2009). Thus, corn is apparently not grown
close enough to lark habitat patches to have significantly
influenced lark diet.
We argue that patterns of habitat transformation in the
study area may explain CORTf, d13C, and the relationship
between them. The majority of lark habitat patches arose
from historical ploughing that isolated natural vegetation
on small plateaus and hillsides that differ in topography,
orientation, and soil depth, while a few large patches of
continuous untransformed habitat remain in more homo-
geneous and flat protected areas. Variation in topography
and soil depth influences the micro-availability of water in
our study area (Peñuelas et al. 1999). Related to this,
increased water use efficiency (WUE) within and among
plant species has been shown to result in significant
increases in d13C values in our study area and elsewhere
(Peñuelas et al. 1999; Wright et al. 1994), and soil char-
acteristics can additionally affect plant d13C (Weitner et al.
2007). Therefore, lark d13C values may reflect spatial
variation in drought tolerance of plants and soil charac-
teristics within and near lark habitat patches. Thus, patterns
of CORTf and its relationship with d13C could have been
driven by an interaction between local arthropod abun-
dance and variation in WUE of plants or soils where they
live. For example, if drought-resistant plants (i.e., high
WUE) and arthropods feeding on them were both partic-
ularly abundant within or near some lark habitat patches,
they would have supplemented the diet of larks, leading to
higher d13C values and lower CORTf. Indeed, arthropod
biomass and CORTf both varied significantly at the same
spatial scale, providing some evidence that the abundance
of local prey was influential.
Values of d15N most strongly reflected the amount of
natural vegetation surrounding lark habitat patches, but
were not related to CORTf. Despite a lack of difference
between years, there was substantial variation in d15N
values among individuals (1–12 %), suggesting that larks
may have been feeding at three different trophic levels
(Post 2002). This is not surprising, as lark diet comprises a
wide range of invertebrate prey, including Gastropoda,
Araneae, Hymenoptera, Diptera, Lepidoptera, Orthoptera,
and Coleoptera (Cramp 1988; Herranz et al. 1993; M. V.,
D. S., J. L. T., unpublished data), as well as occasional
seeds (Aragüés 1992). Importantly, larks consume carniv-
orous and omnivorous members of some taxa (e.g., Ara-
neae, Carabidae, Staphylinidae; Herranz et al. 1993) and
d15N values can be higher in arthropod species feeding at
higher trophic levels (Bennett and Hobson 2009; Ostrom
et al. 1997). Thus, larks relying on an abundance of high
123
Oecologia
trophic-level prey would have feathers with significantly
higher d15N values (Inger and Bearhop 2008; Kelly 2000).
Because larks are insectivorous generalists spatially
restricted to remnant habitat patches (Vögeli et al. 2010),
they are expected to eat whatever prey items are local and
abundant; to the extent that we can infer diet diversity from
variation in d15N values among individuals, this was
apparently the case. However, if d15N values were a proxy
for diet diversity, a lack of relationship between CORTf
and d15N suggests that diet diversity per se was not
strongly related to CORT physiology. This parallels a
previous result showing that prey diversity is not related to
lark occurrence (Vögeli et al. 2010).
Weather was related to CORTf values in our study, as
has been shown previously, particularly during molt
(Romero et al. 2000; Wingfield et al. 1983), but not as we
expected. Wetter, and to a lesser extent cooler, weather
during molting was associated with increased CORTf,
indicating that this may have been challenging for larks.
Increased precipitation tends to improve conditions for
birds in arid regions, but can interfere with activities such
as foraging (Newton 1998), which would result in
increased CORT. Precipitation during molting in our study
area during 2004 was nearly twice that experienced during
2005, so it is possible that reduced precipitation in 2005
ameliorated conditions for larks. It is also possible that
larks suppressed CORT secretion in response to increased
temperature (e.g., Wingfield et al. 1992), but maximum
temperatures were not high enough to indicate that larks
were coping with excessively hot weather. Instead, cooler
temperatures likely exacerbated the effects of precipitation.
In summary, we reason that historical agriculturally
driven changes to the local food web influenced CORTf
and feather SI values. Spatial and annual variation in
agricultural practices can alter local plant species assem-
blages, water availability, soil characteristics, isotope ratio,
and availability of arthropod prey (Benton et al. 2002; Choi
et al. 2005; Frampton et al. 2000; Vickery et al. 2001;
Wright et al. 1994). This, in turn, can influence SI values
(Diggs et al. 2011; Hobson and Bairlein 2003; Podlesak
and McWilliams 2006) and blood levels of CORT
(Kitaysky et al. 2001, 2006; Sahin et al. 2002), providing
an explanation of how land uses were significantly related
to d13C, d15N, and CORTf values. Agricultural intensifi-
cation is an important driver of habitat change and popu-
lation dynamics for our study species (Vögeli et al. 2010),
and our results add to this by suggesting a physiological
effect of habitat change.
Conclusions and future research
Our study provides the first evidence that combining
feather-based measures of CORT and SIs can provide
 Author's personal copy
Oecologia
insight into habitat–GC relationships. Clearly, additional
research is needed to better understand what factor(s) drive
variation in CORTf and d13C in our study area. Future
studies will benefit from using multiple feathers molted
over different time periods for SI analyses to characterize
greater temporal variability in environmental conditions.
Moreover, measuring SI values in vegetation and arthro-
pods, and how these vary with intrinsic WUE, soil features,
and specific agricultural practices (e.g., crop diversity and
irrigation), will provide a basis for understanding isotopic
change through the food web. Relating CORT with addi-
tional SIs (e.g., 2H) may help better characterize the local
environment. Future work is also needed to clarify the
relative influences, and possible interaction of, environ-
ment and genetics on patterns of CORTf.
Despite these unknowns, our approach helped us better
understand habitat–physiology relationships by suggesting
that factors operating at the individual level, rather than
patch or landscape levels, were most important in deter-
mining physiological responses. We also demonstrated that
measures of habitat and population structure likely to
influence population dynamics in larks were not related to
an integrated measure of CORT physiology. By synchro-
nizing CORTf with d13C, we found evidence of a physio-
logical response to agriculturally driven variation in food
supply. This approach is especially well suited to species
that have low dispersal, such as larks in our study area, or are
territorial year-round because such species can be captured
in the same habitat where feathers were grown. This will aid
in spatially synchronizing measures of the environment with
feather-based measures of SIs and CORTf. Nonetheless, SIs
can help to identify where feathers were grown in the case of
long-distance migrants (Carrete et al. 2013).
In conclusion, our minimally invasive feather-based
approach shows promise for providing an individual-based
perspective on the relationships between habitat and
physiology in birds. The wealth of individual-based mea-
sures that feathers can provide (e.g., Bond and Lavers
2011; Bortolotti et al. 2002; Burger and Gochfeld 2000;
Fairhurst et al. 2011; Grubb 2006; Inger and Bearhop 2008;
Jovani and Diaz-Real 2012; Kelly 2000; Lens and Egger-
mont 2008) make them an ideal source of synchronized
biomarkers. Our work, and that of others (Jovani and Diaz-
Real 2012; Lens and Eggermont 2008), suggests that
combining such measures will provide a richer and more
integrated understanding of avian ecophysiology.
Acknowledgments We thank all collaborators during fieldwork,
especially M. Méndez, P. Laiolo, M. A. Carrero and I. Afán. We
thank T. Marchant for generously allowing us to use her lab and I.
Luque and V. Fachal for their assistance with the corticosterone
analyses. We are particularly grateful to P. Ostrom, K. Hobson, A.
Bond, and two anonymous reviewers for helping improve the man-
uscript. G. D. F. was supported by a Dr. Ruby Larson Student
Research Fund in Biology, a Malcolm A. Ramsay Memorial Award, a
Nature Saskatchewan Graduate Student Award, and a Candace
Savage and Keith Bell Fellowship in Grassland Ecology Studies.
M. V. was supported by a pre- and postdoctoral fellowship (I3P-CSIC
and MICINN). Funds were provided by Junta de Andalucı́a Excel-
lence Project P07RNM 02918 (M. V., D. S., J. L. T.), the Natural
Science and Engineering Research Council of Canada (G. R. B.), and
the University of Saskatchewan (G. D. F., G. R. B.).
References
Alguacil MD, Roldan A, Salinas-Garcia JR, Querejeta JI (2011) No
tillage affects the phosphorus status, isotopic composition and
crop yield of Phaseolus vulgaris in a rain-fed farming system.
J Sci Food Agric 91:268–272
Andrén H (1994) Effects of habitat fragmentation on birds and
mammals in landscapes with different proportions of suitable
habitat: a review. Oikos 71:355–366
Aragüés A (1992) Estudio de la alondra de Dupont en la región
Aragonesa (Study of the Dupont’s lark in the region of Aragón).
Ph.D. University of Zaragoza, Zaragoza
Astheimer LB, Buttemer WA, Wingfield JC (1992) Interactions of
corticosterone with feeding, activity and metabolism in passerine
birds. Ornis Scand 23:355–365
Bennett PM, Hobson KA (2009) Trophic structure of a boreal forest
arthropod community revealed by stable isotope (d13C, d15N)
analyses. Entomol Sci 12:17–24
Benton TG, Bryant DM, Cole L, Crick HQP (2002) Linking
agricultural practice to insect and bird populations: a historical
study over three decades. J Appl Ecol 39:673–687
Blas J, Bortolotti GR, Tella JL, Baos R, Marchant TA (2007) Stress
response during development predicts fitness in a wild, long
lived vertebrate. Proc Natl Acad Sci USA 104:8880–8884
Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens
MHH, White JSS (2009) Generalized linear mixed models: a
practical guide for ecology and evolution. Trends Ecol Evol
24:127–135
Bond AL, Lavers JL (2011) Trace element concentrations in feathers
of flesh-footed shearwaters (Puffinus carneipes) from across
their breeding range. Arch Environ Contam Toxicol 61:318–326
Bortolotti GR (2010) Flaws and pitfalls in the chemical analysis of
feathers: bad news-good news for avian chemoecology and
toxicology. Ecol Appl 20:1766–1774
Bortolotti GR, Dawson RD, Murza GL (2002) Stress during feather
development predicts fitness potential. J Anim Ecol 71:333–342
Bortolotti GR, Marchant TA, Blas J, German T (2008) Corticosterone
in feathers is a long-term, integrated measure of avian stress
physiology. Funct Ecol 22:494–500
Bortolotti GR, Marchant T, Blas J, Cabezas S (2009) Tracking stress:
localisation, deposition and stability of corticosterone in feath-
ers. J Exp Biol 212:1477–1482
Braun-Blanquet J, de Bolós O (1957) Les groupement végétaux du
bassin moyen de l’Ebre et leur dynamisme. Anal Estac Exp Aula
Dei 5:1–266
Briones MJI, Bol R, Sleep D, Allen D, Sampedro L (2001) Spatio-
temporal variation of stable isotope ratios in earthworms under
grassland and maize cropping systems. Soil Biol Biochem
33:1673–1682
Burger J, Gochfeld M (2000) Metal levels in feathers of 12 species of
seabirds from Midway Atoll in the northern Pacific Ocean. Sci
Total Environ 257:37–52
Burnham KP, Anderson DR (2002) Model selection and multiple
inference: a practical information-theoretic approach, 2nd edn.
Springer, Heidelberg
123
 Author's personal copy
Busch DS, Hayward LS (2009) Stress in a conservation context: a
discussion of glucocorticoid actions and how levels change with
conservation-relevant variables. Biol Conserv 142:2844–2853
Carrete M, Serrano D, Illera JC, López G, Vögeli M, Delgado A,
Tella JA (2009) Goats, birds, and emergent diseases: apparent
and hidden effects of exotic species in an island environment.
Ecol Appl 19:840–853
Carrete M, Bortolotti GR, Sánchez-Zapata JA, Delgado A, Cortés-
Avizanda A, Grande JM, Donázar JA (2013) Stressful conditions
experienced by endangered by endangered Egyptian vultures on
African wintering areas. Anim Conserv. doi:10.1111/acv.12001
Choi WJ, Chang SX, Allen HL, Kelting DL, Ro HM (2005) Irrigation
and fertilization effects on foliar and soil carbon and nitrogen
isotope ratios in a loblolly pine stand. For Ecol Manage
213:90–101
Clarke GM (1995) Relationships between developmental stability and
fitness: application for conservation biology. Conserv Biol
9:18–24
Cramp S (ed) (1988) Handbook of the birds of Europe, the Middle
East and North Africa. Oxford University Press, Oxford
Diggs NE, Marra PP, Cooper RJ (2011) Resource limitation drives
patterns of habitat occupancy during the nonbreeding season for
an omnivorous songbird. Condor 113:646–654
ESRI (2004) ArcGIS 9. ESRI, Redlands
Evans MR, Roberts ML, Buchanan KL, Goldsmith AR (2006)
Heritability of corticosterone response and changes in life
history traits during selection in the zebra finch. J Evol Biol
19:343–352
Fairhurst GD, Frey MD, Reichert JF, Szelest I, Kelly DM, Bortolotti
GR (2011) Does environmental enrichment reduce stress? An
integrated measure of corticosterone from feathers provides a
novel perspective. PLoS ONE 6:e17663
Fairhurst GD, Navarro J, González-Solı́s J, Marchant TA, Bortolotti
GR (2012) Feather corticosterone of a nestling seabird reveals
consequences of sex-specific parental investment. Proc R Soc B
Biol Sci 279:177–184
Fokidis HB, Orchinik M, Deviche P (2011) Context-specific territo-
rial behavior in urban birds: no evidence for involvement of
testosterone or corticosterone. Horm Behav 59:133–143
Frampton GK, van den Brink PJ, Gould PJL (2000) Effects of spring
drought and irrigation on farmland arthropods in southern
Britain. J Appl Ecol 37:865–883
Gobierno de Navarra (GDN) (2011) Meteorologı́a y climatologı́a de
Navarra. http://meteo.navarra.es/. Accessed April 2011
Grubb TC (2006) Ptilochronology: feather time and the biology of
birds. Oxford University Press, Oxford
Hanski I, Kuussaari M, Nieminen M (1994) Metapopulation structure
and migration in the butterfly Melitaea cinxia. Ecology
75:747–762
Harms NJ, Fairhurst GD, Bortolotti GR, Smits JEG (2010) Variation
in immune function, body condition, and feather corticosterone
in nestling tree swallows (Tachycineta bicolor) on reclaimed
wetlands in the Athabasca oil sands, Alberta, Canada. Environ
Pollut 158:841–848
Hebert CE, Wassenaar LI (2001) Stable nitrogen isotopes in
waterfowl feathers reflect agricultural land use in western
Canada. Environ Sci Technol 35:3482–3487
Herranz J, Yanes M, Suárez F (1993) First record on the diet of
Dupont’s lark Chersophilus duponti nestlings in the Iberian
Peninsula. Ardeola 40:77–79
Hobson KA (1999) Stable-carbon and nitrogen isotope ratios of
songbird feathers grown in two terrestrial biomes: implications
for evaluating trophic relationships and breeding origins. Condor
101:799–805
Hobson KA, Bairlein F (2003) Isotopic fractionation and turnover in
captive garden warblers (Sylvia borin): implications for
123
Oecologia
delineating dietary and migratory associations in wild passerines.
Can J Zool Rev Can Zool 81:1630–1635
Hobson KA, Wassenaar LI (eds) (2008) Tracking animal migration
with stable isotopes. Academic Press, New York
Homyack JA (2010) Evaluating habitat quality of vertebrates using
conservation physiology tools. Wildl Res 37:332–342
Inger R, Bearhop S (2008) Applications of stable isotope analyses to
avian ecology. Ibis 150:447–461
Instituto Aragonés de Estadistica (IAE) (2011) Clima. http://www.
aragon.es/DepartamentosOrganismosPublicos/Organismos/
InstitutoAragonesEstadistica/. Accessed April 2011
Jovani R, Diaz-Real J (2012) Fault bars timing and duration: the
power of studying feather fault bars and growth bands together.
J Avian Biol 43:97–101
Kelly JF (2000) Stable isotopes of carbon and nitrogen in the study of
avian and mammalian trophic ecology. Can J Zool Rev Can Zool
78:1–27
Kilgas P, Mand R, Magi M, Tilgar V (2006) Hematological
parameters in brood-rearing great tits in relation to habitat,
multiple breeding and sex. Comp Biochem Physiol Mol Integr
Physiol 144:224–231
Kitaysky AS, Kitaiskaia EV, Wingfield JC, Piatt JF (2001) Dietary
restriction causes chronic elevation of corticosterone and
enhances stress response in red-legged kittiwake chicks.
J Comp Physiol B Biochem Syst Environ Physiol 171:701–709
Kitaysky AS, Romano MD, Piatt JF, Wingfield JC, Kikuchi M (2005)
The adrenocortical response of tufted puffin chicks to nutritional
deficits. Horm Behav 47:606–619
Kitaysky AS, Kitaiskaia EV, Piatt JF, Wingfield JC (2006) A
mechanistic link between chick diet and decline in seabirds?
Proc R Soc B Biol Sci 273:445–450
Laiolo P (2008) Characterizing the spatial structure of songbird
cultures. Ecol Appl 18:1774–1780
Laiolo P, Tella JL (2006) Landscape bioacoustics allow detection of
the effects of habitat patchiness on population structure. Ecology
87:1203–1214
Laiolo P, Tella JL (2008) Social determinants of songbird vocal
activity and implications for the persistence of small populations.
Anim Conserv 11:433–441
Laiolo P, Vögeli M, Serrano D, Tella JL (2008) Song diversity
predicts the viability of fragmented bird populations. PLoS ONE
3:e1822
Lattin CR, Reed JM, DesRochers DW, Romero LM (2011) Elevated
corticosterone in feathers correlates with corticosterone-induced
decreased feather quality: a validation study. J Avian Biol
42:247–252
Lens L, Eggermont H (2008) Fluctuating asymmetry as a putative
marker of human-induced stress in avian conservation. Bird
Conserv Int 18:S125–S143
Marra PP, Holberton RL (1998) Corticosterone levels as indicators of
habitat quality: effects of habitat segregation in a migratory bird
during the non-breeding season. Oecologia 116:284–292
Marra PP, Hobson KA, Holmes RT (1998) Linking winter and
summer events in a migratory bird by using stable-carbon
isotopes. Science 282:1884–1886
Mazerolle DF, Hobson KA (2002) Physiological ramifications of
habitat selection in territorial male ovenbirds: consequences of
landscape fragmentation. Oecologia 130:356–363
Méndez M, Tella JL, Godoy JA (2011) Restricted gene flow and
genetic drift in recently fragmented populations of an endan-
gered steppe bird. Biol Conserv 144:2615–2622
Nakagawa S, Schielzeth H (2010) Repeatability for Gaussian and
non-Gaussian data: a practical guide for biologists. Biol Rev
85:935–956
Newton I (1998) Population limitation in birds. Academic Press,
London
 Author's personal copy
Oecologia
Oliver I, Beattie AJ (1993) A possible method for the rapid
assessment of biodiversity. Conserv Biol 7:572–578
Ostrom PH, Colunga-Garcia M, Gage SH (1997) Establishing
pathways of energy flow for insect predators using stable isotope
ratios: field and laboratory evidence. Oecologia 109:108–113
Ouyang JQ, Sharp PJ, Dawson A, Quetting M, Hau M (2011)
Hormone levels predict individual differences in reproductive
success in a passerine bird. Proc R Soc B Biol Sci 278:2537–2545
Peñuelas J, Filella I, Terradas J (1999) Variability of plant nitrogen
and water use in a 100-m transect of a subdesertic depression of
the Ebro valley (Spain) characterized by leaf d13C and d15N.
Acta Oecol 20:119–123
Podlesak DW, McWilliams SR (2006) Metabolic routing of dietary
nutrients in birds: effects of diet quality and macronutrient
composition revealed using stable isotopes. Physiol Biochem
Zool 79:534–549
Polis GA, Anderson WB, Holt RD (1997) Toward an integration of
landscape and food web ecology: the dynamics of spatially
subsidized food webs. Annu Rev Ecol Syst 28:289–316
Post DM (2002) Using stable isotopes to estimate trophic position:
models, methods, and assumptions. Ecology 83:703–718
Rangel TF, Diniz JAF, Bini LM (2010) SAM: a comprehensive
application for spatial analysis in macroecology. Ecography
33:46–50
Raouf SA, Smith LC, Brown MB, Wingfield JC, Brown CR (2006)
Glucocorticoid hormone levels increase with group size and
parasite load in cliff swallows. Anim Behav 71:39–48
Ricklefs RE, Wikelski M (2002) The physiology/life-history nexus.
Trends Ecol Evol 17:462–468
Romero LM (2004) Physiological stress in ecology: lessons from
biomedical research. Trends Ecol Evol 19:249–255
Romero LM, Wikelski M (2001) Corticosterone levels predict
survival probabilities of Galapagos marine iguanas during El
Nino events. Proc Natl Acad Sci USA 98:7366–7370
Romero LM, Reed JM, Wingfield JC (2000) Effects of weather on
corticosterone responses in wild free-living passerine birds. Gen
Comp Endocrinol 118:113–122
Sahin K, Sahin N, Onderci M, Gursu F, Cikim G (2002) Optimal
dietary concentration of chromium for alleviating the effect of
heat stress on growth, carcass qualities, and some serum
metabolites of broiler chickens. Biol Trace Elem Res 89:53–64
Sapolsky RM, Romero LM, Munck AU (2000) How do glucocorti-
coids influence stress responses? Integrating permissive, sup-
pressive, stimulatory, and preparative actions. Endocr Rev 21:
55–89
Saunders DA, Hobbs RJ, Margules CR (1991) Biological conse-
quences of ecosystem fragmentation: a review. Conserv Biol
5:18–32
Scheuerlein A, Van’t Hof TJ, Gwinner E (2001) Predators as
stressors? Physiological and reproductive consequences of
predation risk in tropical stonechats (Saxicola torquata axillar-
is). Proc R Soc Lond Ser B Biol Sci 268:1575–1582
Seoane J, Justribo JH, Garcia F, Retamar J, Rabadan C, Atienza JC
(2006) Habitat-suitability modelling to assess the effects of land-
use changes on Dupont’s lark Chersophilus duponti: a case study
in the Layna Important Bird Area. Biol Conserv 128:241–252
Shannon CE, Weaver W (1949) A mathematical model of commu-
nication. University of Illinois Press, Urbana
SIGPAC (2011) Sistema de información geográfica de parcelas
agrı́colas (SIGPAC) http://www.marm.es/es/cartografia-y-sig/
temas/sistema-de-informacion-geografica-de-parcelas-agricolas-
sigpac-/default.aspx
Sokal RR, Oden NL (1978) Spatial autocorrelation in biology. 1.
Methodology. Biol J Linn Soc 10:199–228
Suorsa P et al (2003) Forest management is associated with
physiological stress in an old-growth forest passerine. Proc R
Soc Lond Ser B Biol Sci 270:963–969
Svensson L (1992) Identification guide to European passerines.
British Trust for Ornithology, Stockholm
Vaida F, Blanchard S (2005) Conditional Akaike information for
mixed-effects models. Biometrika 92:351–370
Vickery JA et al (2001) The management of lowland neutral
grasslands in Britain: effects of agricultural practices on birds
and their food resources. J Appl Ecol 38:647–664
Vögeli M, Serrano D, Tella JL, Méndez M, Godoy JA (2007) Sex
determination Duponts lark Chersophilus duponti using molec-
ular sexing and discriminant functions. Ardeola 54:69–79
Vögeli M, Laiolo P, Serrano D, Tella JL (2008) Who are we
sampling? Apparent survival differs between methods in a
secretive species. Oikos 117:1816–1823
Vögeli M, Serrano D, Pacios F, Tella JL (2010) The relative importance
of patch habitat quality and landscape attributes on a declining
steppe-bird metapopulation. Biol Conserv 143:1057–1067
Vögeli M, Lemus JA, Serrano D, Blanco G, Tella JL (2011) An island
paradigm at mainland: population fragmentation impairs the
community of avian pathogens. Proc R Soc B Biol Sci 278:
2668–2676
Weitner A, Dupouey JL, Lefèvre Y, Bréda N, Badeau V, Ferhi A,
Duquesnay A, Thimonier A (2007) Roles of soil chemistry and
water availability in site-related d13C variations in French beech
forests. Tree Physiol 27:1043–1051
Wenninger EJ, Inouye RS (2008) Insect community response to plant
diversity and productivity in a sagebrush-steppe ecosystem.
J Arid Environ 72:24–33
Wikelski M, Cooke SJ (2006) Conservation physiology. Trends Ecol
Evol 21:38–46
Wiley AE, Ostrom PH, Stricker CA, James HF, Gandhi H (2010)
Isotopic characterization of flight feathers in two pelagic
seabirds: sampling strategies for ecological studies. Condor
112:337–346
Wingfield JC, Moore MC, Farner DS (1983) Endocrine responses to
inclement weather in naturally breeding populations of white-
crowned sparrows (Zonotrichia leucophrys pugetensis). Auk
100:56–62
Wingfield JC, Vleck CM, Moore MC (1992) Seasonal changes of the
adrenocortical response to stress in birds of the Sonoran desert.
J Exp Zool 264:419–428
Wright GC, Rao RCN, Farquhar GD (1994) Water-use efficiency and
carbon isotope discrimination in peanut under water deficit
conditions. Crop Sci 34:92–97
123