Evol Biol (2016) 43:48–59

DOI 10.1007/s11692-015-9349-0

RESEARCH ARTICLE

Morphological Adaptations to Migration in Birds

Csongor I. Vágási1,2 • Péter L. Pap1,2 • Orsolya Vincze1,2 • Gergely Osváth2,3 •

Johannes Erritzøe4 • Anders Pape Møller5

Received: 26 February 2015 / Accepted: 14 September 2015 / Published online: 22 September 2015
Ó Springer Science+Business Media New York 2015

Abstract Migratory flight performance has direct or performance might be low, and hence such energetically
carry-over effects on fitness. Therefore, selection is expensive tissues are not associated (pectoral muscle) or
expected to act on minimizing the costs of migratory flight, are inversely associated (heart) with migration distance.
which increases with the distance covered. Aerodynamic
theory predicts how morphological adaptations improve Keywords Aerodynamics
Functional morphology

flight performance. These predictions have rarely been ‘Migratory syndrome’
Organ size
Phylogenetic
tested in comparative analyses that account for scaling and comparative analysis
Wing morphology
phylogenetic effects. We amassed a unique dataset of 149
European bird species and 10 morphological traits. Mass-
adjusted aspect ratio increased, while mass-adjusted heart Introduction
weight and wing loading decreased with increasing
migration distance. These results were robust to whether Flight capacity constitutes complex adaptations in
the analyses were based on the entire species pool or homeothermic vertebrates: animals get airborne to migrate,
limited to passerines or migrants. Our findings indicate that forage, display, hunt or escape from predators. Extraordi-
selection due to migration acts on wing traits that reduce nary locomotion during migratory journeys is the one that
the energetic cost of transportation to increase the flight fascinated scientists for a long time due to impressive
range. Consequently, the demands for high ‘exercise organ’ distances covered (e.g. 64,000 km a year by sooty shear-
waters Puffinus griseus; Shaffer et al. 2006) and high rel-
evance for ecological, evolutionary, conservation and
Electronic supplementary material The online version of this health considerations (e.g. Webster et al. 2002; Alerstam
article (doi:10.1007/s11692-015-9349-0) contains supplementary
material, which is available to authorized users. et al. 2003; Møller et al. 2008; Bowlin et al. 2010; Altizer
et al. 2011; Bauer and Hoye 2014).
& Csongor I. Vágási Birds exhibit a large number of morphological (e.g.
[email protected] body mass, bones, feathered wings, wing-powering mus-
1
MTA-DE ‘‘Lendület’’ Behavioural Ecology Research Group,
cles, streamlined shape) and physiological and visceral
Department of Evolutionary Zoology, University of organic (e.g. respiration, cardiovascular system, metabolic
Debrecen, Debrecen, Hungary rate) adaptations that are thought to enhance flight capacity
2
Evolutionary Ecology Group, Hungarian Department of (Piersma et al. 2005; Videler 2005). However, besides
Biology and Ecology, Babeş-Bolyai University, Cluj Napoca, these ‘general flight adaptations’, Darwinian selection is
Romania expected to add further ‘adaptive layers’ that particularly
3
Museum of Zoology, Babeş-Bolyai University, Cluj Napoca, optimize the net pay-off of migratory flight. This expec-
Romania tation is reasonable considering (1) the energetic and
4
Taps Old Rectory, Christiansfeld, Denmark antioxidant expenses that migratory flight entails, mainly
5
Laboratoire d’Ecologie, Systématique et Evolution, CNRS when covering long distances (Costantini et al. 2007;
UMR 8079, Université Paris-Sud, Orsay Cedex, France Klaassen 1996; Rayner 1990), (2) the substantial mortality

123
Evol Biol (2016) 43:48–59 49

rate during migration (Klaassen et al. 2014; Newton 2004), for catabolism, along with fat, to sustain flight (Battley
and (3) the huge variation in migration distance among et al. 2000; Jenni et al. 2000; Klaassen 1996; Pennycuick
species. Aerodynamic theory predicts how morphological 1998; Piersma et al. 2005). Besides exercise organs,
adaptations might maximize performance through either nutritional organs (gizzard, liver) also contribute to the
energy-efficiency (i.e. transport cost minimization) or time- energy turnover, and hence the physiology of flight, by
efficiency of migratory flight (Hedenström 2008; Norberg regulating food processing. While exercise organs are
1990, 1994). We apply this functional morphology para- functional during flight, nutritional organs are atrophied
digm for studying morphological adaptations to long-dis- during migration to reduce loading and hypertrophied only
tance migratory flight in birds. at stopover refuelling stations. It was shown that migrants
A flying bird should produce lift and thrust that exceed have larger bony surfaces for flight muscle attachment
weight (i.e. pull of gravity) and drag, respectively, to (Calmaestra and Moreno 2000), but similar pectoral muscle
maintain a forward momentum. Both lift and thrust are and heart weights as sedentary species, except that heart
ensured by the feathered wings and their highly complex size was larger in migrant Charadriiformes based on a non-
movement during the stroke cycle. Wing loading (i.e. phylogenetic comparative analysis (Winkler and Leisler
weight/wing area) and aspect ratio (i.e. wingspan2/wing 1992). A phylogenetic analysis based on large sample size
area) are two wing morphology variables considered to and data for the size of multiple organs analysed in a
greatly influence flight capacity and economy. Aerody- unified approach is thus desirable for knowing how organ
namic theory predicts higher body mass and/or smaller weights are related to migration demands.
wing area if time-efficiency (i.e. flight speed maximization) Besides the size of specific organs, total body mass also
is selected for because flight velocity scales to wing load- matters in optimal migration. Body mass is the most
ing to the 0.32 power (Alerstam et al. 2007), while higher important life-history trait, which can adaptively evolve to
aspect ratio and lower wing loading is expected to evolve if meet certain demands or it might constrain the realization
optimal flight economy (longest flight range for smallest of optimal strategies. Migration is not an exception. Body
flight power) is selected (Norberg 1990). Therefore, both mass or drag due to mass are integral parts of functions
temporal and energetic currencies of migratory flight about flight economy (Hedenström 2008; Norberg 1994;
depend on the size and shape of the wing. Long-distance Pennycuick 2008). Therefore, body mass is expected to
migratory birds were reported to have higher aspect ratio, evolve with migration distance, although whether an
longer distal wing, pointed wingtips and lower wing- increase or decrease in body mass is advantageous for
loading than short-distance migrants or residents (Cal- long-distance migrants is still not established (Hedenström
maestra and Moreno 2001; Fiedler 2005; Kaboli et al. 2008). Previous studies found that body mass either
2007; Leisler and Winkler 2003; Lockwood et al. 1998; decreases (Jones and Witt 2014; Kaboli et al. 2007; Sol
Mönkkönen 1995; Voelker 2001; Winkler and Leisler et al. 2005) or increases (Mönkkönen 1992; Winkler and
1992), which hints toward the importance of energy-effi- Leisler 1992) with longer migration. Furthermore, mor-
cient flight. However, these studies do not uniformly detect phological adaptation to migration might be constrained by
such correlations between migration distance and mor- mass due to scaling laws (see e.g. Alerstam et al. 2007),
phological traits and reached contrasting conclusions (re- and hence correction for allometry is recommended.
viewed by Mulvihill and Chandler 1990; Lockwood et al. However, several comparative studies either did not control
1998) about whether migration represents an important for allometry or computed the scaling power of mor-
selective force that shapes avian external morphology phologies without accounting for the similarity due to
(Leisler and Winkler 2003). common descent. If evolution of morphological traits is
Visceral organs also play an important role in migratory governed by selection due to migratory performance, the
flight because they contribute to body mass and metabolic mass-independent scatter in morphology is expected to be
rate (Piersma and Lindström 1997). Given that the per- functional and hence explain variation in migration
formance of organs increases with their size (Piersma and distance.
Lindström 1997), migrants are expected to have well-de- Modern phylogenetic analyses of the correlates of
veloped ‘exercise organs’ such as flight muscles (pectoral migration distance are scarce (Bennett and Owens 2002;
and supracoracoid), heart and lung (Piersma et al. 2005). Leisler and Winkler 2003), in particular those based on
This should be favoured owing to (1) higher mechanical large sample sizes, despite a long history of the topic
power production by larger cross-sectional surface of (Fredericus 1240 cited in Lockwood et al. 1998). Whether
muscles, (2) better respiration and oxygen supply to sustain external morphology is adapted to meet migratory demands
mechanical power generation and to compensate for the is still questioned, while internal morphological adapta-
hypoxic conditions of high altitude flights, and/or (3) tions are poorly understood because only a handful of
steady energy supply by organs providing protein substrate studies addressed this question at the interspecific level.

123
50 Evol Biol (2016) 43:48–59

Furthermore, although morphological adaptation to Materials and Methods

migration was subject of scientific enquiry (see references
above), most of the previous comparative studies have one Specimens
or more of the following limitations: (1) lack of phyloge-
netic control or comparison of highly contrasting sister Birds were either captured in the field or collected as fresh
taxa, (2) small sample sizes or narrow taxonomic ranges carcasses. Captures took place across Romania by using
(few species or single families studied), (3) migration mist nets (Ecotone, Poland) or rarely traps during the
distance either scored on a two- or four-level scale with breeding seasons 2009–2013. Only adult birds were con-
subjective thresholds or quantified by manual measure- sidered because wing morphologies are predicted to max-
ments of distribution maps, and (4) morphological traits imize fitness in an age-specific manner (Fiedler 2005).
tested in isolation instead of being tested concurrently in Each bird was banded with a unique metal ring and mea-
multiple regression models. sured for standard biometry. We took digital pictures
Everything else being equal, migration distance increa- (Nikon D80) of the left wing and the body; the wing was
ses both energetic and time costs of flight (Hedenström fully unfolded and flattened such that the leading edge was
1993). Aerodynamic theory predicts that long-distance held perpendicular to the body axis. The camera was held
migrants should have high aspect ratio wings to minimize perpendicular to the wing surface and a ruler was used as
energy costs (Rayner 1988). High aspect ratio is predicted metric template. Carcasses from natural deaths (e.g. road
to be accompanied by either low wing loading (i.e. long kill, building collision, electrocution, starvation) were
wingspan) or higher wing loading (i.e. average or short collected in Romania and Denmark for taxidermy. Those
wingspan) depending on whether low cost of transport collected in Romania were photographed for wing archi-
(slow and economic flight) or time-efficient travel (faster, tecture, while those brought to JE in Denmark contributed
but still economic) is the currency under selection, to the organ size data. Only carcasses in good condition
respectively (Norberg 1990). Given that flight power is a were considered. Although there is spatial discrepancy in
function of pectoral muscle weight and flight aerobic the source of wing architecture and organ weight data, we
capacity is enhanced by larger heart size (Bishop 1997; are confident that these do not affect our analyses for the
Rayner 1988), we predicted that exercise organ weights following reasons. First, APM has also measured wing
relative to unit mass will be positively associated with architecture from wing outline drawings of birds captured
migration distance. To test these predictions rooted in in Northern Europe (98 species overlap with the Romanian
aerodynamic theory, we compiled an unprecedented data- dataset). Those measures are strongly positively correlated
set in terms of number of species (149 European bird with those in the present study (phylogenetic generalized
species), taxonomic span (50 families and 20 orders) and least squares regression, aspect ratio: F1,95 = 878.10,
number of morphological traits (size of five organs, four R2 = 0.90, P \ 0.0001; wing loading: F1,94 = 2458.00,
wing size and shape parameters and body mass), which R2 = 0.96, P \ 0.0001). Second, in the dataset presented
allows for generalisation. We analysed migration distance here, within species variance was much smaller than the
as a function of morphology by means of phylogenetic among species variance (see ‘‘Results’’ for intraspecific
models, which constitute a powerful tool for revealing repeatabilities). Third, similarly to wing morphology, it
adaptations, and they are claimed suitable for revealing was previously shown that organ sizes are species-specific
functional adaptations in morphology (Wainwright and and similar between localities within the same species
Reilly 1994; Piersma et al. 2005). Passerine birds might (Garamszegi et al. 2002; Møller et al. 2005). These findings
differ from non-passerines in terms of optimal aerody- together suggest that external and internal morphological
namics, which arise from the difference in basal metabolic traits are species-specific and thus suitable for multispecies
rate between the two groups (Hedenström 1993). This is comparison.
also mirrored in morphological adaptations to flight
expenses (Rayner 1988). For instance, passerines have a Migration Distance
relatively uniform and characteristic flight mode that dif-
fers from other avian taxa (Bruderer et al. 2010), and flight Distribution maps (shape files) were retrieved from http://
type might be related to the morphology of the flight www.birdlife.org/datazone/info/spcdownload (BirdLife
apparatus (Pap et al. 2015; Wang et al. 2011). To assess International and NatureServe 2012). Maps were usually
whether passerines evolved a different set of functional restricted to longitudes between W20° and E60° (Western
morphology than non-passerines, we also tested our models Palearctic), but for certain species map boundaries had to
of migration distance by restricting the analyses to the be modified according to our knowledge of the migratory
order of Passeriformes. behaviour of the subspecies or populations from which

123
Evol Biol (2016) 43:48–59 51

morphological data were gathered (see Electronic Supple- Statistical Analyses

mentary Material, ESM, Table S1 for modifications). Using
the distribution map shape files, we calculated the centroids All statistical analyses were conducted as implemented in
of the spatial polygons of breeding and wintering ranges R 3.1 (R Core Team 2015). To assess whether morpho-
using the ‘gCentroid’ function of the R package ‘rgeos’ logical traits are species-specific and thus suitable for
(Bivand and Rundel 2013). Migration distance was then multispecies comparison (Møller and Birkhead 1994), we
calculated as the geographic distance between the breeding tested the repeatability of these traits by assessing the
and wintering centroids using a custom function written in importance of among-species compared to within-species
R 3.1 (R Core Team 2015) (see Methodological details in variance (function ‘ICCest’ from R package ‘ICC’; Wolak
ESM). The latitudinal centroid of the breeding range (LCB) et al. 2011). Since variance is dependent on sample size,
was used as a potential ecological confounder in the the estimation of repeatability in comparative studies with
analyses since northerly-breeding birds have a higher ten- unbalanced sample sizes across species is not equivocal. To
dency to be more migratory. The species pool we study cope with this problem, as a first step, species with only
captures a great deal of variation in migration distance one individual sampled were excluded. Subsequently, we
(range 0–8526 km). randomly picked two values out of the total number of
measurements per species for the remaining species and
Wing Architecture partitioned variance to within- and among-group compo-
nents. This was iterated 1000 times, so repeatabilities are
Wing architecture comprises four traits: wingspan (m), the average of the 1000 repeatability estimates. Repeata-
wing area (m2), wing loading (Nm-2) and aspect ratio bility was tested using raw, non-transformed values.
(dimensionless). These were measured based on wing Subsequently, we investigated by means of regression
pictures (see Methodological details in ESM). Wing load- analyses how avian external and internal morphological
ing (weight/wing area) gives the relative size of the wings, traits (1) covary with each other, (2) scale to body mass and
and aspect ratio (wingspan2/wing area) describes both the (3) explain migration distance. However, in regressions
size and shape of the wing (high values indicate a narrow across multiple species the error terms of different species
wing relative to its length). Our dataset covers a huge are correlated such that covariance is higher among closely
variation in body mass (range 5.6–4263.5 g, 760-fold) and related species due to longer shared phylogenetic history.
wing parameters (wingspan: 0.13–1.83 m, 14-fold; wing Therefore, we developed phylogenetically informed mod-
area: 0.0042–0.61 m2, 145-fold; aspect ratio: 3.72–9.14, els that account for the dependence of species by incor-
2.5-fold; wing loading: 10.61–195.84 Nm-2, 18.5-fold; see porating a matrix of covariances among species based on
ESM Table S1). their phylogenetic relationships (Martins and Hansen 1997;
Pagel 1997, 1999), and we estimated the importance of
Organ Size phylogenetic corrections (Freckleton et al. 2002). These
analyses were based on the dated phylogeny published by
Numerous bird specimens were brought frozen to JE, more Thuiller et al. (2011). Because five of our species were
than 95 % of them being found dead and the remaining missing from this tree, our original sample size of 154
were shot by hunters. Danish taxidermists are required by species dropped to 149. We used phylogenetic generalized
law to record the cause of death of all specimens in a log- least squares (PGLS) models as implemented in the ‘pgls’
book. Various visceral organs, labelled as exercise (left function of the R package ‘caper’ (Orme et al. 2011). The
pectoralis major muscle [subsequently pectoral muscle], phylogenetic signal Pagel’s k (Pagel 1997, 1999) was
left supracoracoid muscle, heart) or nutritional organs estimated by maximum likelihood in each model. Strong
(liver and gizzard) were removed during post-mortem and weak signal (i.e. k approaches 1 and 0, respectively)
examination and weighed by JE on a precision balance to indicates that evolution conforms to or deviates from the
the nearest 1 mg. Only birds with undamaged organs were Brownian motion model, respectively. Three transforma-
considered, while storage caused negligible noise in organ tions were adopted prior to the analyses: (1) body mass and
masses (Møller et al. 2005). Data were recorded blindly all morphological traits were log10-transformed (LCB and
with respect to the tested hypothesis. Our dataset covers a migration distance were unaltered because raw values
huge variation in organ mass (gizzard: 0.15–57.22 g, provided better fit to the distribution requirement of model
381-fold; liver: 0.28–68.18 g, 243-fold; pectoral muscle: residuals), (2) all morphological trait values were adjusted
0.35–182.79 g, 522-fold; supracoracoid muscle: to unit body mass to rule out a scaling effect, and (3) all
0.03–15.15 g, 505-fold; heart: 0.08–27.42 g, 343-fold; see fixed effects were standardized to a mean of zero and a SD
ESM Table S1). of one in order to avoid the large difference between the

123
52 Evol Biol (2016) 43:48–59

intervals covered by different traits (cf. wing loading with Results

body mass in ESM Table S1).
First, the pairwise relationships between the nine mor- Repeatability and Scaling
phological characters as well as body mass was assessed by
bivariate PGLS models with one trait as dependent and the Each organ size and wing morphology trait was highly
rest sequentially set as a single fixed term. Second, to esti- repeatable, and repeatability values, expressed as intra-
mate the scaling of morphological characters to body mass, class correlation coefficient (ICC), had a narrow confi-
we used phylogenetically informed reduced major axis dence interval, which together show that these traits are
(PRMA) regressions (‘phyl.RMA’ function from R package highly species-specific (gizzard: ICC = 0.93, 95 % CI
‘phytools’; Revell 2012) because accounting for dependence 0.90–0.96; liver: ICC = 0.78, 95 % CI 0.69–0.84; pec-
due to common descent is highly recommended in scaling toral muscle: ICC = 0.76, 95 % CI 0.64–0.84; supraco-
studies (Harvey 2000). Scaling is expressed as a power racoid muscle: ICC = 0.75, 95 % CI 0.64–0.83; heart:
function of the form Y = a 9 massb, where Y is one of the ICC = 0.92, 95 % CI 0.89–0.95; wingspan: ICC = 0.99,
morphological variables. With log-transformation the 95 % CI 0.99–0.99; wing area: ICC = 0.98, 95 % CI
function becomes log(Y) = log(a) ? b 9 log(mass), where 0.97–0.99; aspect ratio: ICC = 0.91, 95 % CI 0.88–0.94;
a is the PRMA regression coefficient and b is the PRMA note that wing loading has identical repeatability to that
regression slope. Third, we fitted three sets of PGLS models of wing area; for organ weights see also Garamszegi et al.
with migratory flight distance as response variable and the 2002; Møller et al. 2005). Hence, species are confidently
following fixed effect formula: (1) all internal morphologi- characterised in our dataset even if they are represented
cal traits ? log10 body mass ? LCB; (2) all external mor- by only two individuals. Therefore, we included species
phological traits ? log10 body mass ? LCB; and (3) all irrespective of samples size per morphological trait in the
internal and external morphological traits ? log10 body analyses.
mass ? LCB (hereafter referred to as ‘organ size’, ‘wing The scaling exponents of the nine morphological traits are
morphology’ and ‘organ and wing’ models, respectively). shown in Table 1. We derived the mass-adjusted morpho-
These three sets of multiple regression models were reduced logical trait values using the equation mass-adjusted
to minimal adequate models by a backward stepwise elim- trait = original trait/massb, where b is the scaling exponent
ination procedure, discarding the least significant predictor (i.e. PRMA slope). These mass-independent morphological
at each step until only significant predictors remained. Using trait values were used in the subsequent analyses.
the information-theoretic approach and multi-model aver-
aging yielded qualitatively similar results (not shown). We Covariation of Morphological Traits
did not model second-order interactions between morpho-
logical traits since (1) this would enormously increase the External and internal morphological traits were signifi-
number of parameters to be estimated leading to undue cantly associated (all traits log10-transformed but not cor-
overfitting, and (2) the correlation between main effects and rected for body mass, all P B 0.02; not shown). These
interaction terms would imply a serious multicollinearity associations probably stem from the body mass-depen-
issue. dence of morphological traits. However, a large number of
Because morphological traits intercorrelate even after
adjustment for body mass (see ‘‘Results’’), we verified Table 1 Scaling of organ size and wing morphology traits estimated
whether there is a multicollinearity problem in the models by phylogenetic reduced major axis (PRMA) regression
by computing the variance inflation factors (VIFs) within Morphological trait n a b R2 Pagel’s k
each MAM of the three model sets and found that values
were all below the more conservative VIF \5 threshold. Gizzard 107 -1.65 1.02 0.79 0.96
Therefore, multicollinearity does not bias our results. We Liver 112 -1.45 0.94 0.85 0.91
also employed two filtering. First, we repeated the above Pectoral muscle 90 -1.24 1.01 0.89 0.92
three model sets by only using Passeriformes (84 out of Supracoracoid muscle 88 -2.06 0.98 0.85 1.00
149 species, 56.4 %) to assess the robustness of the results Heart 112 -1.97 0.95 0.92 0.94
to difference between passerines and non-passerines. Sec- Wingspan 149 0.87 0.36 0.84 0.97
ond, we repeated the above three model sets by excluding Wing area 149 1.04 0.69 0.85 0.96
resident species with the filtering criterion of migration Aspect ratio 149 0.43 0.14 0.04 0.96
distance [0 km (127 out of 149 species, 85.2 %) to assess Wing loading 149 0.59 0.45 0.65 0.96
whether the morphological predictors of migration distance PRMA coefficient, a, and slope, b, are reported together with the
could arise due to the disparity between resident and coefficient of determination, R2, and phylogenetic signal, Pagel’s k.
migratory species subsets. For details, see ‘‘Materials and Methods’’

123
Evol Biol (2016) 43:48–59 53

relationships remained significant after traits were adjusted neither organ was retained when only the passerines were
to unit body mass (ESM Table S2). considered (Table 2). The MAM of the entire species pool
fitted significantly better the data than the null model,
Migration Distance which modelled only the intercept and did not include any
organ size trait (AIC = 683.48 and 670.97 for null model
The MAMs of each of the three multiple regression PGLS and MAM, respectively; likelihood ratio test, LRT, of null
model sets contain LCB as a positive predictor of distance model vs. MAM: LR = 16.51, df = 3, 5, P = 0.0003).
between breeding and wintering grounds with northerly- The MAM based on only migratory species is the null
breeding species migrating longer distances for all species model, since all morphological traits and the confounding
or passerines only (Tables 2, 3, 4). Body mass was a sig- LCB were rejected during model reduction.
nificant negative predictor of migration distance in the In the ‘wing morphology’ model set (Table 3), wing
‘organ size’ model set and when the analysis was based on loading was the only significant predictor of migration
the entire species pool, while in other models it was distance in the full model involving all studied species,
dropped during model reduction to MAMs (Tables 2, 3, 4). where species making longer journeys having lower load-
In the ‘organ size’ model set (Table 2), mass-adjusted ings. After model simplification to MAMs, however, aspect
heart weight was significantly inversely related to migra- ratio was also retained besides wing loading when the
tion distance in the full models based on both the entire analysis was based on the entire species pool. Species that
species pool and on passerines only. However, it was migrate farther have significantly higher aspect ratio. When
dropped during model simplification to MAMs with other the analysis was restricted to passerines, both wing loading
organ size traits. Therefore, none of the MAMs included and aspect ratio were rejected, and wingspan and wing area
the size of any organ to explain variation in migration became significant explanatory variables. Passerines that
distance among species (Table 2). Similarly, the size of migrate farther have longer wingspan and smaller wing

Table 2 ‘Organ size’ model set: multiple regression PGLS model of migration distance in relation to log10 body mass, latitudinal centroid of the
breeding range (LCB) and all the organ size traits adjusted for body mass
Full model MAM
Predictors b (SE) t P Predictors b (SE) t P

(a) All species

n = 83, Pagel’s k = 0.98 n = 149, Pagel’s k = 0.80
Intercept 3.39 (1.41) 2.41 0.019 Intercept 3.75 (1.10) 3.39 \0.001
Log10 body mass -0.84 (0.49) 1.73 0.088 Log10 body mass -1.26 (0.37) 3.39 \0.001
LCB 0.43 (0.22) 1.94 0.056 LCB 0.45 (0.17) 2.67 0.009
Gizzard -0.32 (0.31) 1.02 0.313
Liver 0.50 (0.36) 1.38 0.173
Pectoral muscle 0.55 (0.35) 1.59 0.115
Supracoracoid muscle -0.61 (0.45) 1.36 0.179
Heart -0.77 (0.35) 2.22 0.03
(b) Only passerine species
n = 55, Pagel’s k = 0.99 n = 84, Pagel’s k = 0.86
Intercept 1.80 (0.93) 1.92 0.06 Intercept 2.11 (0.92) 2.31 0.024
Log10 body mass -0.47 (0.38) 1.22 0.229 LCB 0.61 (0.19) 3.17 0.002
LCB 0.34 (0.25) 1.36 0.182
Gizzard -0.59 (0.39) 1.52 0.135
Liver 0.82 (0.43) 1.92 0.062
Pectoral muscle 0.75 (0.41) 1.83 0.074
Supracoracoid muscle -0.33 (0.46) 0.73 0.468
Heart -0.59 (0.29) 2.02 0.049
Both the full model and the derived minimal adequate model (MAM) are shown for analyses based on either (a) the entire species pool or
(b) restricted to passerines. Significant relationships are highlighted in bold

123
54 Evol Biol (2016) 43:48–59

Table 3 ‘Wing morphology’

Full model MAM
model set: multiple regression
PGLS model of migration Predictors b (SE) t P Predictors b (SE) t P
distance in relation to log10
body mass, latitudinal centroid (a) All species
of the breeding range (LCB) and n = 149, Pagel’s k = 0.76 n = 149, Pagel’s k = 0.75
all wing morphology traits
Intercept 3.47 (0.92) 3.76 \0.001 Intercept 3.20 (0.87) 3.67 \0.001
adjusted for body mass
Log10 body mass -0.92 (0.66) 1.39 0.167 LCB 0.42 (0.15) 2.71 0.008
LCB 0.44 (0.15) 2.90 0.004 Aspect ratio 1.67 (0.22) 7.67 \0.001
Wingspan 0.91 (1.75) 0.52 0.605 Wing loading -0.69 (0.24) 2.88 0.005
Wing area -1.46 (1.28) 1.14 0.257
Aspect ratio 1.15 (0.90) 1.27 0.206
Wing loading -1.34 (0.60) 2.23 0.028
(b) Only passerine species
n = 84, Pagel’s k = 0.83 n = 84, Pagel’s k = 0.86
Intercept 2.57 (0.76) 3.37 0.001 Intercept 2.49 (0.78) 3.20 0.002
Log10 body mass -0.98 (0.91) 1.08 0.286 LCB 0.59 (0.17) 3.57 \0.001
LCB 0.58 (0.17) 3.32 0.001 Wingspan 2.74 (0.51) 5.33 \0.001
Wingspan 2.84 (3.35) 0.85 0.398 Wing area -1.67 (0.48) 3.51 \0.001
Wing area -2.87 (2.01) 1.43 0.157
Aspect ratio -0.16 (2.23) 0.07 0.943
Wing loading -1.24 (1.20) 1.04 0.303
Both the full model and the derived minimal adequate model (MAM) are shown for analyses based on
either (a) the entire species pool or (b) restricted to passerines. Significant relationships are highlighted in
bold

area (Table 3). The MAM of the entire species pool fitted df = 3, 6, P \ 0.0001). The MAM based on only migra-
significantly better the data than the null model, which tory species contained the same significant predictors as the
modelled only the intercept and did not include any wing MAM without restriction to migrants (n species = 92,
morphology trait (AIC = 683.48 and 634.86 for null model Pagel’s k = 0.88; LCB: b ± SE = 0.44 ± 0.19, t = 2.35,
and MAM, respectively; LRT of null model vs. MAM: P = 0.02; mass-adjusted heart weight: b ± SE =
LR = 54.62, df = 3, 6, P \ 0.0001). The MAM based on -0.62 ± 0.27, t = 2.30, P = 0.02; mass-adjusted aspect
only migratory species contained the same significant ratio: b ± SE = 1.88 ± 0.31, t = 5.99, P \ 0.0001).
predictors as the MAM without restriction to migrants To verify the robustness of our results, we conducted
(n species = 127, Pagel’s k = 0.74; LCB: b ± SE = two sets of sensitivity analyses. First, because flapping
0.45 ± 0.17, t = 2.63, P = 0.01; mass-adjusted aspect or soaring flight style might favour different morpho-
ratio: b ± SE = 1.58 ± 0.26, t = 6.14, P \ 0.0001; mass- logical traits in long-distance migrants, we tested whe-
adjusted wing loading: b ± SE = -0.76 ± 0.26, t = 2.86, ther exclusion of soaring birds (9 species; flight style
P = 0.005). label ‘2a’ in Bruderer et al. 2010) altered the results, but
In the ‘organ and wing’ model set (Table 4), mass-ad- found qualitatively similar results (see ESM Tables S3–
justed heart mass was negatively related to migration dis- S5). Second, because organ sizes can considerably
tance in the full models of all species and passerines only, change as a function of migratory stage, we excluded all
while mass-adjusted aspect ratio was positively related to individuals with unknown collection date or collected
migration distance in the full model involving all species. during spring or fall migratory periods and once again
After minimization to MAMs, these two traits remained we found qualitatively unchanged results (see ESM
significantly associated with migration distance; species Tables S6 and S7).
that cover longer distances have smaller relative heart
weight and higher aspect ratio (Table 4; Fig. 1). The MAM
of the entire species pool fitted significantly better the data Discussion
than the null model, which modelled only the intercept and
did not include any organ size or wing morphology trait The prevailing view on the evolution of the flight apparatus
(AIC = 502.65 and 464.13 for null model and MAM, suggests that flight-related morphology is evolutionarily
respectively; LRT of null model vs. MAM: LR = 44.52, labile (Piersma et al. 2005) and mirrors better the habitat-

123
Evol Biol (2016) 43:48–59 55

Table 4 ‘Organ and wing’

Full model MAM
model set: multiple regression
PGLS model of migration Predictors b (SE) t P Predictors b (SE) t P
distance in relation to log10
body mass, latitudinal centroid (a) All species
of the breeding range (LCB) and n = 83, Pagel’s k = 0.96 n = 112, Pagel’s k = 0.87
all organ size and wing
Intercept 2.51 (1.16) 2.16 0.034 Intercept 2.84 (0.96) 2.96 0.004
morphology traits adjusted for
body mass Log10 body mass 0.90 (0.95) 0.95 0.346 LCB 0.39 (0.17) 2.33 0.021
LCB 0.30 (0.20) 1.52 0.133 Heart -0.54 (0.23) 2.34 0.021
Gizzard -0.17 (0.27) 0.64 0.521 Aspect ratio 1.83 (0.26) 7.14 \0.001
Liver 0.44 (0.32) 1.36 0.179
Pectoral muscle 0.10 (0.35) 0.30 0.768
Supracoracoid muscle 0.06 (0.41) 0.15 0.883
Heart -0.85 (0.33) 2.58 0.012
Wingspan -2.50 (2.35) 1.06 0.292
Wing area 1.55 (1.90) 0.82 0.417
Aspect ratio 3.20 (1.28) 2.50 0.015
Wing loading -1.07 (0.85) 1.26 0.213
(b) Only passerine species
n = 55, Pagel’s k = 0.91 n = 67, Pagel’s k = 0.83
Intercept 2.38 (0.72) 3.32 0.002 Intercept 2.12 (0.74) 2.84 0.006
Log10 body mass 0.70 (0.83) 0.84 0.403 LCB 0.56 (0.21) 2.70 0.009
LCB 0.25 (0.24) 1.04 0.306 Heart -0.63 (0.27) 2.36 0.021
Gizzard -0.29 (0.34) 0.85 0.40 Aspect ratio 1.74 (0.31) 5.57 \0.001
Liver 0.65 (0.38) 1.69 0.098
Pectoral muscle -0.14 (0.42) 0.33 0.74
Supracoracoid muscle 0.45 (0.42) 1.07 0.292
Heart -0.71 (0.30) 2.32 0.025
Wingspan -1.19 (3.37) 0.35 0.725
Wing area 0.82 (1.87) 0.44 0.662
Aspect ratio 2.63 (2.39) 1.10 0.277
Wing loading -0.62 (1.29) 0.48 0.635
Both the full model and the derived minimal adequate model (MAM) are shown for analyses based on
either (a) the entire species pool or (b) restricted to passerines. Significant relationships are highlighted in
bold

specific feeding behaviour and related flight style than Leisler 1992), which is evidenced by the wide scatter of
migratory behaviour (e.g. Rayner 1988; Norberg 1990; body masses even among long-distance migrants (from
Leisler and Winkler 2003). By contrast, we found that such swallows to large raptors). Because the weight of visceral
traits are conservative as evidenced by a strong phyloge- organs has a considerable contribution to body mass, and
netic signal. Despite this phylogenetic constraint, a few because we measured them mostly outside the migratory
traits were significant predictors of migration distance period, future studies analysing organ sizes measured on
across an ecologically diverse array of species studied. migratory birds could clarify this question.
Body mass was not a consistent predictor of migration Aspect ratio was proposed by many as a key adaptation
distance being retained only in the organ size model set to migration (see ‘‘Introduction’’; but see Rayner 1990).
based on the entire species pool. Our findings thus suggest The single stringent test of this hypothesis was conducted
that the aerodynamic advantages of neither small nor large by Lockwood et al. (1998), who used 27 species pairs to
body mass drive the evolution of body mass to meet the control for phylogenetic effects and showed that birds
challenges imposed by long-distance migratory flights. The categorised as migrants had higher aspect ratio than resi-
weak association between body mass and migration dis- dents. We extended this analysis by using 149 species,
tance is probably the mere consequence of several selective modelling migration distance as a continuous variable and
forces governing the evolution of body size (Winkler and using multiple regression by considering many

123
56 Evol Biol (2016) 43:48–59

Fig. 1 Migration distance as a function of a mass-adjusted wing aspect ratio and b mass-adjusted heart mass. The graphs are drawn based on the
MAM in Table 4a. On panel (b), residual migration distance reflects migration distance corrected for LCB and aspect ratio

morphological traits in addition to aspect ratio. Our find- higher wing loading and decreases with higher aspect ratio,
ings corroborate that of Lockwood et al. (1998). Previous and migrants have high aspect ratio and low to medium
comparative studies did not find a correlation between wing loading. We reconsidered these associations by con-
wing loading and migration distance (Marchetti et al. 1995; trolling for scaling and phylogenetic effects and found that
Winkler and Leisler 1992), but we did when considering long-distance migrants might reduce the maximum range
only external morphological traits and the entire species power via higher aspect ratio, while wing loading has
pool. This study is the first (1) to show a strong relationship limited effects. In the model restricted to passerines, aspect
between migration distance and heart weight and (2) to ratio was discarded from the model, while wingspan was
simultaneously analyse many morphological traits by tak- longer and wing area smaller with increasing migration
ing phylogenetic and scaling effects into account. distance. Note however, that long wingspan in combination
The amount of work done to maintain level flight is a with small wing area is characteristic for high aspect ratio
function of the magnitude of the net lift (i.e. lift-to-drag wings. Therefore, high aspect ratio seems to be a general
ratio), and the size and shape of the wing is decisive in this adaptation to long-distance migration in birds.
respect (Norberg 1990; Pennycuick 2008). Selection might High aspect ratio is acknowledged for reducing the cost
not only minimize workload (i.e. energy or power per unit of transportation (i.e. maximization of flight range per unit
distance, also coined energy-efficiency or cost of trans- energy) because it decreases the induced and profile drag
portation), but time per unit distance flown as well (i.e. (at low-to-medium and high airspeed, respectively), better
time-efficiency). The two currencies are related to each shedding of wingtip vortices, and better glide ratio (sinking
other via flight velocity and because long travels assimilate per horizontal speed) (Norberg 1990, 1994; Pennycuick
more energy (Hedenström 1993). It is still unresolved 2008; Rayner 1988; Shyy et al. 2008; Withers 1981). Note
which currency is more important for migrants (for energy, that high aspect ratio might increase wing loading owing to
see Rayner 1988; Norberg 1994; for time, see Hedenström the narrow wing mean chord (Norberg 1990; Alerstam
1993, 2008). Lindhe Norberg (2002) argued that high et al. 2007; this study) and ultimately flight energetics
aspect ratio combined with short wings (i.e. high wing (Rayner 1988), inasmuch as wingspan does not change.
loading) is favourable for long commuting flight because it Depending on model set, wing loading either decreased or
ensures fast but still economic flight. Rayner (1988) used was not related to migration distance indicating that
principal component analysis to obtain axes that corre- wingspan increased in parallel with aspect ratio, which
spond to size, wing loading and aspect ratio, and found that prevented an elevated wing loading. The positive correla-
maximum range power (i.e. power needed to fly maximum tion of aspect ratio with wingspan, but lack of correlation
range with unit energy) increases with larger size and with wing area, and the only subtle decrease of wing area

123
Evol Biol (2016) 43:48–59 57

with migration distance lend support to this assertion (ESM Given the high demands of long-distance migration and
Table S2). The combination of high aspect ratio with low- a vast array of external and internal morphological traits
to-moderate wing loading is thought to be highly energy- expected to be related to these demands, it was proposed
efficient (Norberg 1990). Taken together, our results sug- that a ‘migratory syndrome’ should exist (Dingle 2006;
gest that flight economy has selective priority over time- Piersma et al. 2005). The existence of syndromes implies
efficiency. that the traits that integrate into a syndrome are correlated
Contrary to our predictions, migration distance was not by being either synergistic or antagonistic (Dingle 2006).
associated with flight muscle sizes and has a strong inverse However, the covariance of internal and external mor-
relationship with heart weight. The supracoracoid muscle phological traits thought to be part of the flight apparatus
may function more in wing strokes related to diverse flight was not scrutinized by accounting for scaling and phylo-
modes and in turn not being shaped by selection due to genetic effects. Two recent reviews concluded that an
migratory flight. In contrast, pectoral muscle and heart integrated migratory syndrome is unlikely to exist because
weight directly determine the power available for flight there are multiple solutions to the same migratory problem,
(Bishop 1997; Rayner 1988). Further tests are required to and, therefore, only a subset of the allegedly migratory trait
clarify this questions, for instance, by collecting flight requirements are found per species (Dingle 2006; Piersma
muscle size data from individuals belonging to several bird et al. 2005). Our results agree with this viewpoint, since
species that are prior to take on wings (exhibit zugunruhe only a few traits were strong predictors of migration dis-
and associated hypertrophied muscles) or are en route tance despite the fact that several morphological traits
(exhibit either hypertrophied or atrophied muscles correlate with each other.
depending on how long they migrated before or on refu- To summarize, we showed that some morphological
elling state). traits are likely adaptations to long-distance migration and
Alternatively, organ weight of long-distance migrants related flight costs. Moreover, small-sized passerines
might mirror the energetic exigencies that these birds face. seemingly are not exempt from this selection pressure.
Evidence from avian comparative studies (e.g. Wiersma Therefore, we disagree with others (see also Voelker 2001),
et al. 2012) and mammalian artificial selection lines who suggested that migratory flight performance is negli-
(Konarzewski and Diamond 1995) show that heart tissue is gible particularly in small birds and at best leads to subtle
energetically expensive. Furthermore, both pectoral muscle differences in morphology (Leisler and Winkler 2003;
and heart weight are considered indicators of aerobic costs Rayner 1988; Winkler and Leisler 1992). Combining the
during extensive locomotion (Bishop 1997; Klaassen results about the predictors of migration distance and
1996), which is supported by species with energetically covariance of morphological traits suggests that selection
expensive flight mode possessing larger relative flight primarily optimizes the preservation of energy assets dur-
muscle and heart weight in order to produce more power ing long travels in order to increase flight range per unit
and to meet the aerobic scope (Bishop 1997; Norberg energy. By virtue of their high aspect ratio that conserves
1990). The average-sized flight muscles of long-distance energy and moderate wing loading that generates sufficient
migrants do not demand a large heart because heart size is lift, migrants might not be subject to strong selection for
optimized to satisfy the oxygen demand of exercise tissues large flight muscle and aerobic capacity. We encourage the
(Bishop and Butler 1995). On the other hand, accepting the collection of a considerable amount of new morphological
higher aspect ratio coupled with moderate wing loading as data from several other bird species because questions like
energy-saving adaptations in long-distance migrants, this how flight style affects functional morphology of migrants
might permit the evolutionary shrinkage of energetically and the scaling rules of these traits, and how organ sizes are
costly exercise organs. A similar case has been reported for adjusted during the migratory season deserve future com-
swordtail fish; species with longer sexually selected sword parative studies.
had higher swimming velocity and smaller heart mass
(Oufiero et al. 2014). It should be noted, however, that the Acknowledgments We appreciate the help during fieldwork by
L}orinc Bărbos, Attila Marton, Krisztina Sándor and Judit Veres-
organ sizes presented here reflect the baseline levels and Szászka, the numerous bird carcasses provided by the members of the
are not related to the size adjustments in the pre- or post- ‘Milvus Group’ Bird and Nature Protection Association, the Museum
migratory periods. Great knots Calidris tenuirostris sub- of Zoology of Babeş-Bolyai University and Costică Adam. László
stantially reduce their organ sizes, except brain and lung, Zsolt Garamszegi kindly aided with statistical analyses and Jácint
Tökölyi with the calculation of migration distances. We thank the
during a 5400 km flight (Battley et al. 2000). Therefore, the administration of the ‘Alexandru Borza’ Botanical Garden of Cluj
energetic hypothesis proposed above should be tested with Napoca for the permission to capture birds. Two anonymous
comparative data for the relationship between migration reviewers provided constructive criticism. This work was licensed by
distance and the magnitude of change in organ sizes the Romanian Academy of Sciences and adhered to recommended
practices for the ringing, measuring, and sampling of wild birds for
between pre- and post-migratory periods.

123
58 Evol Biol (2016) 43:48–59

research purposes. Logistics and data collection between 2010 and over site. Comparative Biochemistry and Physiology C, 144,
2013 was financed by a CNCSIS Grant (PN II. RU TE 291/2010) of 363–371.
the Romanian Ministry of Education and Research. CIV and OV were Dingle, H. (2006). Animal migration: Is there a common migratory
supported by the European Union and the State of Hungary, co-fi- syndrome? Journal of Ornithology, 147, 212–220.
nanced by the European Social Fund in the framework of TÁMOP Fiedler, W. (2005). Ecomorphology of the external flight apparatus of
4.2.4.A/2-11-1-2012-0001 ‘National Excellence Program’. During blackcaps (Sylvia atricapilla) with different migration behavior.
writing, CIV was financed by the Postdoctoral Fellowship Programme Annals of the New York Academy of Sciences, 1046, 253–263.
and PLP by the János Bolyai Research Scholarship of the Hungarian Freckleton, R. F., Harvey, P. H., & Pagel, M. (2002). Phylogenetic
Academy of Sciences. analysis and comparative data: A test and review of evidence.
American Naturalist, 160, 712–726.
Compliance with Ethical Standards Garamszegi, L. Z., Møller, A. P., & Erritzøe, J. (2002). Coevolving
avian eye size and brain size in relation to prey capture and
Conflict of interest The authors declare that they have no conflict nocturnality. Proceedings of the Royal Society of London B, 269,
of interest. 961–967.
Harvey, P. H. (2000). Why and how phylogenetic relationships should
be incorporated into studies of scaling. In J. H. Brown & G.
B. West (Eds.), Scaling in biology. New York, NY: Oxford
University Press.
References Hedenström, A. (1993). Migration by soaring or flapping flight in
birds: The relative importance of energy cost and speed.
Alerstam, T., Hedenström, A., & Åkesson, S. (2003). Long-distance Philosophical Transactions of the Royal Society of London B,
migration: Evolution and determinants. Oikos, 103, 247–260. 342, 353–361.
Alerstam, T., Rosén, M., Bäckman, J., Ericson, P. G. P., & Hellgren, Hedenström, A. (2008). Adaptations to migration in birds:
O. (2007). Flight speeds among bird species: Allometric and Behavioural strategies, morphology and scaling effects. Philo-
phylogenetic effects. PLoS Biology, 5, e197. sophical Transactions of the Royal Society of London B, 363,
Altizer, S., Bartel, R., & Han, B. A. (2011). Animal migration and 287–299.
infectious disease risk. Science, 331, 296–302. Jenni, L., Jenni-Eiermann, S., Spina, F., & Schwabl, H. (2000).
Battley, P. F., Piersma, T., Dietz, M. W., Tang, S., Dekinga, A., & Regulation of protein breakdown and adrenocortical response to
Hulsman, K. (2000). Empirical evidence for differential organ stress in birds during migratory flight. American Journal of
reductions during trans-oceanic bird flight. Proceedings of the Physiology, Regulatory, Integrative and Comparative Physiol-
Royal Society of London B, 267, 191–195. ogy, 278, R1182–R1189.
Bauer, S., & Hoye, B. J. (2014). Migratory animals couple Jones, M. R., & Witt, C. C. (2014). Migrate small, sound big:
biodiversity and ecosystem functioning worldwide. Science, Functional constraints on body size promote tracheal elongation
344, 1242552. in cranes. Journal of Evolutionary Biology, 27, 1256–1264.
Bennett, P. M., & Owens, I. P. F. (2002). Evolutionary ecology of Kaboli, M., Aliabadian, M., Guillaumet, A., Roselaar, C. S., &
birds. Life histories, mating systems and extinction. New York, Prodon, R. (2007). Ecomorphology of the wheatears (genus
NY: Oxford University Press. Oenanthe). Ibis, 149, 792–805.
BirdLife International, & NatureServe. (2012). Bird species distribu- Klaassen, M. (1996). Metabolic constraints on long-distance migra-
tion maps of the world, version 2.0. Cambridge, UK and tion in birds. Journal of Experimental Biology, 199, 57–64.
Arlington, TX: BirdLife International and NatureServe. Klaassen, R. H., Hake, M., Strandberg, R., Koks, B. J., Trierweiler,
Bishop, C. M. (1997). Heart mass and the maximum cardiac output of C., Exo, K.-M., et al. (2014). When and where does mortality
birds and mammals: Implications for estimating the maximum occur in migratory birds? Direct evidence from long-term
aerobic power input of flying animals. Philosophical Transac- satellite tracking of raptors. Journal of Animal Ecology, 83,
tions of the Royal Society of London B, 352, 447–456. 176–184.
Bishop, C. M., & Butler, P. J. (1995). Physiological modelling of Konarzewski, M., & Diamond, J. (1995). Evolution of basal
oxygen consumption in birds during flight. Journal of Experi- metabolic rate and organ masses in laboratory mice. Evolution,
mental Biology, 198, 2153–2163. 49, 1239–1248.
Bivand, R., & Rundel, C. (2013). rgeos: Interface to geometry Leisler, B., & Winkler, H. (2003). Morphological consequences of
engine—open source (GEOS). R package version 0.3-2. http:// migration in passerines. In P. Berthold, E. Gwinner, & E.
CRAN.R-project.org/package=rgeos. Sonnenschein (Eds.), Avian migration (pp. 175–186). Heidel-
Bowlin, M. S., Bisson, I.-A., Shamoun-Baranes, J., Reichard, J. D., berg, Germany: Springer.
Sapir, N., Marra, P. P., et al. (2010). Grand challenges in Lindhe Norberg, U. M. (2002). Structure, form, and function of flight
migration biology. Integrative and Comparative Biology, 50, in engineering and the living world. Journal of Morphology, 252,
261–279. 52–81.
Bruderer, B., Peter, D., Boldt, A., & Liechti, F. (2010). Wing-beat Lockwood, R., Swaddle, J. P., & Rayner, J. M. V. (1998). Avian
characteristics of birds recorded with tracking radar and cine wingtip shape reconsidered: Wingtip shape indices and morpho-
camera. Ibis, 152, 272–291. logical adaptations to migration. Journal of Avian Biology, 29,
Calmaestra, R. G., & Moreno, E. (2000). Ecomorphological patterns 273–292.
related to migration: A comparative osteological study with Marchetti, K., Price, T., & Richman, A. (1995). Correlates of wing
passerines. Journal of Zoology, 252, 495–501. morphology with foraging behaviour and migration distance in
Calmaestra, R. G., & Moreno, E. (2001). A phylogenetically-based the genus Phylloscopus. Journal of Avian Biology, 26, 177–181.
analysis on the relationship between wing morphology and Martins, E. P., & Hansen, T. F. (1997). Phylogenies and the
migratory behaviour in passeriformes. Ardea, 89, 407–416. comparative method: A general approach to incorporating
Costantini, D., Cardinale, M., & Carere, C. (2007). Oxidative damage phylogenetic information into the analysis of interspecific data.
and anti-oxidant capacity in two migratory bird species at a stop- American Naturalist, 149, 646–667.

123
Evol Biol (2016) 43:48–59 59

Møller, A. P., & Birkhead, T. R. (1994). The evolution of plumage R Core Team. (2015). R: A language and environment for statistical
brightness in birds is related to extrapair paternity. Evolution, 48, computing. Vienna, Austria. http://www.R-project.org/: R
1089–1100. Foundation for Statistical Computing.
Møller, A. P., Erritzøe, J., & Garamszegi, L. Z. (2005). Covariation Rayner, J. M. V. (1988). Form and function in avian flight. In R.
between brain size and immunity in birds: Implications for F. Johnston (Ed.), Current Ornithology (Vol. 5, pp. 1–66). New
brain size evolution. Journal of Evolutionary Biology, 18, York, NY: Plenum Press.
223–237. Rayner, J. M. V. (1990). The mechanics of flight and bird migration
Møller, A. P., Rubolini, D., & Lehikoinen, E. (2008). Populations of performance. In E. Gwinner (Ed.), Bird migration. Psyiology
migratory bird species that did not show a phenological response and ecophysiology (pp. 283–299). Heidelberg, Germany:
to climate change are declining. Proceedings of the National Springer.
Academy of Sciences of the USA, 105, 16195–16200. Revell, L. J. (2012). phytools: An R package for phylogenetic
Mönkkönen, M. (1992). Life history traits of palaearctic and nearctic comparative biology (and other things). Methods in Ecology and
migrant passerines. Ornis Fennica, 69, 161–172. Evolution, 3, 217–223.
Mönkkönen, M. (1995). Do migrant birds have more pointed wings?: Shaffer, S. A., Tremblay, Y., Weimerskirch, H., Scott, D., Thompson,
A comparative study. Evolutionary Ecology, 9, 520–528. D. R., Sagar, P. M., et al. (2006). Migratory shearwaters
Mulvihill, R. S., & Chandler, C. R. (1990). The relationship between integrate oceanic resources across the Pacific Ocean in an
wing shape and differential migration in the Dark-eyed Junco. endless summer. Proceedings of the National Academy of
Auk, 107, 490–499. Sciences of the USA, 103, 12799–12802.
Newton, I. (2004). Population limitation in migrants. Ibis, 146, Shyy, W., Lian, Y., Tang, J., Viieru, D., & Liu, H. (2008).
197–226. Aerodynamics of low Reynolds number flyers. New York, NY:
Norberg, U. M. (1990). Vertebrate flight: Mechanisms, physiology, Cambridge University Press.
morphology, ecology and evolution. Berlin, Germany: Springer. Sol, D., Lefebvre, L., & Rodrı́guez-Teijeiro, J. D. (2005). Brain size,
Norberg, U. M. (1994). Wing design, flight performance, and habitat innovative propensity and migratory behaviour in temperate
use in bats. In P. C. Wainwright & S. M. Reilly (Eds.), Palaearctic birds. Proceedings of the Royal Society of London B,
Ecological morphology (pp. 205–239). Chicago, IL: University 272, 1433–1441.
of Chicago Press. Thuiller, W., Lavergne, S., Roquet, C., Boulangeat, I., Lafourcade, B.,
Orme, C. D. L., Freckleton, R. P., Thomas, G. H., Petzoldt, T., & & Araujo, M. B. (2011). Consequences of climate change on the
Fritz, S. A. (2011). caper: Comparative analyses of phylogenet- tree of life in Europe. Nature, 470, 531–534.
ics and evolution in R. R package version 0.5.2. http://CRAN.R- Videler, J. J. (2005). Avian flight. New York, NY: Oxford University
project.org/package=caper. Press.
Oufiero, C. E., Meredith, R. W., Jugo, K. N., Tran, P., Chappell, M. Voelker, G. (2001). Morphological correlates of migratory distance
A., Springer, M. S., et al. (2014). The evolution of the sexually and flight display in the avian genus Anthus. Biological Journal
selected sword in Xiphophorus does not compromise aerobic of the Linnean Society, 73, 425–435.
locomotor performance. Evolution, 68, 1806–1823. Wainwright, P. C., & Reilly, S. M. (Eds.). (1994). Ecological
Pagel, M. (1997). Inferring evolutionary processes from phylogenies. morphology: Integrative organismal biology. Chicago, IL:
Zoologica Scripta, 26, 331–348. University of Chicago Press.
Pagel, M. (1999). The maximum likelihood approach to reconstruct- Wang, X., McGowan, A. J., & Dyke, G. J. (2011). Avian wing
ing ancestral character states of discrete characters on phyloge- proportions and flight styles: First step towards predicting the
nies. Systematic Biology, 48, 612–622. flight modes of Mesozoic birds. PLoS One, 6, e28672.
Pap, P. L., Osváth, G., Sándor, K., Vincze, O., Bărbos, L., Marton, A., Webster, M. S., Peter, P., Haig, S. M., Bensch, S., & Holmes, R. T.
et al. (2015). Interspecific variation in the structural properties of (2002). Links between worlds: Unraveling migratory connectiv-
flight feathers in birds indicates adaptation to flight requirements ity. Trends in Ecology & Evolution, 17, 76–83.
and habitat. Functional Ecology, 29, 746–757. Retrieved from Wiersma, P., Nowak, B., & Williams, J. B. (2012). Small organ size
http://doi.wiley.com/10.1111/1365-2435.12419. contributes to the slow pace of life in tropical birds. Journal of
Pennycuick, C. J. (1998). Computer simulation of fat and muscle burn Experimental Biology, 215, 1662–1669.
in long-distance bird migration. Journal of Theoretical Biology, Winkler, H., & Leisler, B. (1992). On the ecomorphology of migrants.
191, 47–61. Ibis, 134(S1), 21–28.
Pennycuick, C. J. (2008). Modelling the flying bird. London, UK: Withers, P. C. (1981). An aerodynamic analysis of bird wings as fixed
Academic Press. aerofoils. Journal of Experimental Biology, 90, 143–162.
Piersma, T., & Lindström, Å. (1997). Rapid reversible changes in Wolak, M. E., Fairbairn, D. J., & Paulsen, Y. R. (2011). Guidelines
organ size as a component of adaptive behaviour. Trends in for estimating repeatability. Methods in Ecology and Evolution,
Ecology & Evolution, 12, 134–138. 3, 129–137.
Piersma, T., Pérez-Tris, J., Mouritsen, H., Bauchinger, U., & Bairlein,
F. (2005). Is there a ‘‘migratory syndrome’’ common to all
migrant birds? Annals of the New York Academy of Sciences,
1046, 282–293.

123