Morphological Adaptations to Migration in Birds
Csongor I. Vágási • Péter L. Pap • OrsolyaVincze • GergelyOsváth • Johannes Erritzøe • Anders PapeMøller
C. I. Vágási (correspondence)
MTA-DE “Lendület” Behavioural Ecology Research Group, Department of Evolutionary Zoology, University of Debrecen, Debrecen, Hungary
e-mail:
Electronic Supplementary Material (ESM)
Methodological details
Code for calculating migration distance
The function was written in R statistical environment by OV.
geo.dist <- function (long1, lat1, long2, lat2)
{
rad <- pi/180
a1 <- lat1 * rad
a2 <- long1 * rad
b1 <- lat2 * rad
b2 <- long2 * rad
dlon <- b2 - a2
dlat <- b1 - a1
a <- (sin(dlat/2))^2 + cos(a1) * cos(b1) * (sin(dlon/2))^2
c <- 2 * atan2(sqrt(a), sqrt(1 - a))
R <- 6378.145
d <- R * c
return(d)
}
Wing architecture
Wing pictures were processed in ImageJ ( and we measured (1) partial wing area by encircling the contour of the wing, (2) wing length (distance between wing root and wingtip), (3) semi-span (distance between backbone and wingtip, i.e. half of the wingspan), and (4) root chord (i.e. wing chord at wing root). Body surface was calculated as 2 × [(wingspan/2 – wing length) × root chord], where (wingspan/2 – wing length) × root chord gives the root box (for calculations and definitions, see Pennycuick 2008). Therefore, wing area = 2 × partial wing area + body surface. These raw measures were used to derive further to traits with relevance to flight. First, wing loading gives the relative size of the wings (i.e. how much weight a unit of wing area should carry in flight) and is calculated as the weight (product of body mass in kg and gravitational acceleration) divided by wing area in m2 (unit: Nm–2). Species-specific body mass was retrieved from ornithological monographs (Cramp 1998; Dunning 2008). Second, aspect ratio describes both the size and shape of the wing. It expresses the narrowness of the wing for a specific length and is computed as wingspan2/wing area (unit: dimensionless). Higher values indicate a narrow wing relative to its length.
The body was not fully visible in some pictures (101 out of 686), and hence we could not measure semi-span and consequently calculate wingspan and wing area, which are necessary for calculating wing loading and aspect ratio. In these instances, we measured the ‘reduced’ wingspan and wing area, which do not include the body width and body surface, respectively. These were found as 2 × wing length and 2 × partial wing area, respectively (see also Andrews et al. 2009). ‘Reduced’ wing loading was calculated as weight/’reduced’ wing area, while ‘reduced’ aspect ratio as (‘reduced’ wingspan)2/‘reduced’ wing area. However, note that there is a very strong correlation between the measures with and without the body both when individual-level data and species averages were regressed (phylogenetically uninformed ordinary least squares regression: wingspan, β (SE) = 1.09 (0.00), t583 = 363.10, R2 = 1.00, P < 0.0001; wing area, β (SE) = 1.11 (0.00), t583 = 783.75, P < 0.0001, R2 = 1.00; phylogenetic generalized least squares regression: wingspan, β (SE) = 1.08 (0.01), t95 = 182.00, R2 = 1.00, P < 0.0001; wing area, β (SE) = 1.10 (0.01), t95 = 155.92, R2 = 1.00, P < 0.0001). Thus, we used the reduced (i.e. without body) measures of wingspan, wing area, wing loading and aspect ratio in subsequent analyses to increase sample sizes. Wing morphology measurements were carried out blindly with respect to the tested hypothesis by CIV and PLP. The measurements by the two authors were pooled because the inter-observer repeatabilities were very high [a subset of wing pictures from 61 individuals for partial wing length, root chord and partial wing area, and 38 individuals for semispan were measured by both authors; all intra-class correlation coefficients (ICC) > 0.99, all lower 95% confidence interval (CI) for ICC ≥ 0.99, all F > 350.8, all P < 0.0001]. Our dataset covers a wide range in body mass (range: 5.6–4263.5 g, 760-fold) and wing parameters (wingspan: 0.13–1.83 m, 14-fold; wing area: 0.0042–0.61 m2, 145-fold; aspect ratio: 3.72–9.14, almost 2.5-fold; wing loading: 10.61–195.84 Nm–2, 18.5-fold; see Table S1).
References
Andrews, C. B., Mackenzie, S. A., & Gregory, T. R. (2009). Genome size and wing parameters in passerine birds. Proceedings of the Royal Society of London B, 276, 55–61.
Cramp, S. (Ed.). (1998). The complete birds of the Western Palearctic on CD-ROM. Oxford, UK: Oxford University Press.
Dunning, J. B. J. (2008). CRC handbook of avian body masses (2nd ed.). Boca Raton, FL: CRC Press.
Pennycuick, C. J. (2008). Modelling the flying bird. London, UK: Academic Press.
Table S1 Data on wing morphology, organ sizes, body mass and migration distance of bird species used in data analyses. Values in parentheses indicate sample sizes per species.Note that wing area and wing morphology variables derived from wing area does not contain the body area intercepted by the wings (see Methods). For the sake of comparison with other studies, values including the body between the wings are on average greater by 0.056 m for wingspan, 0.008 m2 for wing area and 0.56 for aspect ratio, and smaller by 4.42 Nm–2 for wing loading compared with values without the body between the wings
See the dataset file (.xlsx) in a separate ESM file.
1
Table S2Pairwise regressionsof organ size and wing morphology traits and migration distance across all species using PGLS models with Pagel’sλ estimated by maximum likelihood. The slope, β and F-valuen are shown, with significant relationships marked with boldface (* P ≤ 0.05, ** P ≤ 0.01, *** P ≤ 0.001).In the lower matrix, row names denote dependent and column names denote explanatory variables, while in the upper matrix vice versa
log mass / gizzard / liver / pectoral muscle / s.coracoid muscle / heart / wingspan / wing area / aspect ratio / wing loading / log migrationdistancelog mass / – / β = –0.10
F107 = 4.34
* / β = –0.07
F112 = 2.04 / β = –0.07
F90 = 1.51 / β = –0.11
F88 = 1.67 / β = –0.05
F112 = 0.98 / β = –0.12
F149 = 4.02
* / β = –0.11
F149 = 3.53 / β = –0.40
F149 = 75.92
*** / β = –0.21
F149 = 11.90
*** / β = –0.05
F149 = 7.69
**
gizzard / β = –0.40
F107 = 6.55
** / – / β = 0.36
F104 = 18.31
*** / β = 0.23
F85 = 5.34
* / β = 0.12
F83 = 0.70 / β = 0.27
F106 = 7.06
** / β = –0.11
F107 = 0.82 / β = –0.15
F107 = 1.62 / β = 0.22
F107 = 3.97
* / β = 0.26
F107 = 5.87
* / β = 0.01
F107 = 0.02
liver / β = –0.31
F112 = 3.87
* / β = 0.42
F104 = 18.22
*** / – / β = 0.08
F90 = 0.76 / β = 0.10
F88 = 0.61 / β = 0.59
F110 = 46.75
*** / β = –0.26
F112 = 5.42
* / β = –0.41
F112 = 13.42
*** / β = 0.32
F112 = 7.94
** / β = 0.48
F112 = 19.99
*** / β = 0.07
F112 = 5.83
*
pectoral muscle / β = –0.30
F90 = 3.28 / β = 0.20
F85 = 3.36 / β = 0.06
F90 = 0.24 / – / β = 0.74
F88 = 65.60
*** / β = 0.51
F90 = 19.26
*** / β = 0.06
F90 = 0.26 / β = –0.05
F90 = 0.14 / β = 0.08
F90 = 0.35 / β = –0.01
F90 = 0.01 / β = 0.02
F90 = 0.38
s.coracoid muscle / β = –0.17
F88 = 1.67 / β = 0.08
F83 = 0.99 / β = 0.06
F88 = 0.65 / β = 0.36
F88 = 44.18
*** / – / β = 0.19
F88 = 6.02
* / β = –0.21
F88 = 5.84
* / β = –0.19
F88 = 4.49
* / β = –0.05
F88 = 0.26 / β = 0.18
F88 = 4.10
* / β = 0.01
F88 = 0.01
heart / β = –0.26
F112 = 3.60 / β = 0.23
F106 = 7.34
** / β = 0.50
F110 = 46.26
*** / β = 0.35
F90 = 18.83
*** / β = 0.25
F88 = 5.35
* / – / β = –0.17
F112 = 2.90 / β = –0.35
F112 = 12.12
*** / β = 0.32
F112 = 10.67
** / β = 0.36
F112 = 14.07
*** / β = 0.04
F112 = 2.52
wingspan / β = –0.22
F149 = 4.02
* / β = –0.04
F107 = 0.44 / β = –0.09
F112 = 1.98 / β = 0.07
F90 = 0.71 / β = –0.30
F88 = 5.84
* / β = –0.07
F112 = 1.02 / – / β = 0.85
F149 = 574.44
*** / β = 0.39
F149 = 30.18
*** / β = –0.74
F149 = 146.97
*** / β = 0.09
F149 = 15.80
***
wing area / β = –0.22
F149 = 3.53 / β = –0.05
F107 = 0.45 / β = –0.19
F112 = 9.20
** / β = 0.03
F90 = 0.20 / β = –0.26
F88 = 4.62
* / β = –0.21
F112 = 9.89
** / β = 0.93
F149 = 574.44
*** / – / β = 0.07
F149 = 0.68 / β = –0.82
F149 = 196.78
*** / β = 0.03
F149 = 1.67
aspect ratio / β = –0.86
F149 = 75.92
*** / β = 0.14
F107 = 3.24 / β = 0.20
F112 = 8.71
** / β = 0.08
F90 = 1.18 / β = –0.07
F88 = 0.40 / β = 0.24
F112 = 10.88
*** / β = 0.44
F149 = 30.18
*** / β = 0.07
F149 = 0.68 / – / β = 0.23
F149 = 6.86
** / β = 0.15
F149 = 45.79
***
wing loading / β = –0.36
F149 = 11.90
*** / β = 0.12
F107 = 3.80
* / β = 0.20
F112 = 11.16
*** / β = –0.07
F90 = 0.79 / β = 0.25
F88 = 4.10
* / β = 0.18
F112 = 7.81
** / β = –0.68
F149 = 146.97
*** / β = –0.69
F149 = 196.78
*** / β = 0.19
F149 = 6.86
** / – / β = 0.01
F149 = 0.01
log migration distance / β = –1.04
F149 = 7.69
** / β = –0.01
F107 = 0.01 / β = 0.63
F112 = 5.21
* / β = 0.27
F90 = 0.60 / β = –0.16
F88 = 0.14 / β = 0.43
F112 = 1.97 / β = 1.06
F149 = 15.80
*** / β = 0.34
F149 = 1.67 / β = 1.55
F149 = 45.79
*** / β = 0.01
F149 = 0.01 / –
1
Table S3 The same ‘Organ size’ model set as in Table 2 in the MS except that the 9 species that have soaring flight style were excluded(7 Accipitriformes: Accipiter gentilis, Accipiter nisus, Aquila chrysaetos, Buteobuteo, Buteolagopus, Circaetusgallicus and Circus cyaneus; 2 Ciconiiformes: Ciconiaciconia and Ciconianigra). Because all these 9 species were non-passerines, the analysis with restriction to passerines only (i.e. Table 2b) was not repeated. Significant relationships are highlighted in bold
Full model / MAMpredictors / β (SE) / t / P / predictors / β (SE) / t / P
all species except soarers
n = 79, Pagel’sλ = 0.97 / n = 140, Pagel’sλ = 0.79
intercept / 3.48 (1.42) / 2.45 / 0.02 / intercept / 3.87 (1.10) / 3.53 / 0.0006
log10 body mass / –1.08 (0.51) / 2.10 / 0.04 / log10 body mass / –1.41 (0.39) / 3.56 / 0.0005
LCB / 0.47 (0.24) / 1.98 / 0.05 / LCB / 0.53 (0.18) / 2.92 / 0.004
gizzard / –0.34 (0.32) / 1.05 / 0.30
liver / 0.53 (0.38) / 1.40 / 0.17
pectoral muscle / 0.63 (0.36) / 1.73 / 0.09
supracoracoid muscle / –0.63 (0.46) / 1.37 / 0.18
heart / –0.81 (0.36) / 2.26 / 0.03
Table S4 The same ‘Wing morphology’ model set as in Table 3 in the MS except that the 9 species that have soaring flight style were excluded(7 Accipitriformes: Accipiter gentilis, Accipiter nisus, Aquila chrysaetos, Buteobuteo, Buteolagopus, Circaetusgallicus and Circus cyaneus; 2 Ciconiiformes: Ciconiaciconia and Ciconianigra). Because all these 9 species were non-passerines, the analysis with restriction to passerines only (i.e. Table 3b) was not repeated. Two MAMs are presented, the upper was found by redoing the stepwise elimination procedure, while the lower is identical to the MAM in Table 3a of the MS. Significant relationships are highlighted in bold
Full model / MAMpredictors / β (SE) / t / P / predictors / β (SE) / t / P
all species except soarers
n = 140, Pagel’sλ = 0.74 / n = 140, Pagel’sλ = 0.73
intercept / 3.54 (0.90) / 3.95 / 0.0001 / intercept / 3.55 (0.89) / 3.99 / 0.0001
log10 body mass / –0.90 (0.67) / 1.36 / 0.18 / log10 body mass / –0.96 (0.47) / 2.05 / 0.04
LCB / 0.51 (0.16) / 3.14 / 0.002 / LCB / 0.51 (0.16) / 3.16 / 0.002
wingspan / –0.21 (1.81) / 0.12 / 0.91 / wing area / –1.07 (0.47) / 2.26 / 0.03
wing area / –0.93 (1.31) / 0.71 / 0.48 / aspect ratio / 1.58 (0.25) / 6.20 / < 0.0001
aspect ratio / 1.68 (0.93) / 1.81 / 0.07 / wing loading / –1.68 (0.50) / 3.37 / 0.001
wing loading / –1.73 (0.64) / 2.69 / 0.008
same MAM structure as in Table 3a of the MS
n = 140, Pagel’sλ = 0.72
intercept / 3.22 (0.85) / 3.77 / 0.0002
LCB / 0.49 (0.16) / 3.01 / 0.003
aspect ratio / 1.70 (0.22) / 7.67 / < 0.0001
wing loading / –0.65 (0.25) / 2.63 / 0.01
Table S5 The same ‘Organ & wing’ model set as in Table 4 in the MS except that the 9 species that have soaring flight style were excluded(7 Accipitriformes: Accipiter gentilis, Accipiter nisus, Aquila chrysaetos, Buteobuteo, Buteolagopus, Circaetusgallicus and Circus cyaneus; 2 Ciconiiformes: Ciconiaciconia and Ciconianigra). Because all these 9 species were non-passerines, the analysis with restriction to passerines only (i.e. Table 4b) was not repeated. Significant relationships are highlighted in bold
Full model / MAMpredictors / β (SE) / t / P / predictors / β (SE) / t / P
(a) all species
n = 79, Pagel’sλ = 0.94 / n = 106, Pagel’sλ = 0.85
intercept / 2.46 (1.11) / 2.21 / 0.03 / intercept / 2.79 (0.93) / 2.99 / 0.004
log10 body mass / 0.90 (0.93) / 0.97 / 0.34 / LCB / 0.47 (0.18) / 2.57 / 0.01
LCB / 0.30 (0.21) / 1.43 / 0.16 / heart / –0.55 (0.24) / 2.28 / 0.02
gizzard / –0.14 (0.27) / 0.52 / 0.61 / aspect ratio / 1.87 (0.26) / 7.25 / < 0.0001
liver / 0.44 (0.33) / 1.33 / 0.19
pectoral muscle / 0.14 (0.37) / 0.38 / 0.70
supracoracoid muscle / 0.09 (0.41) / 0.22 / 0.83
heart / –0.84 (0.35) / 2.39 / 0.02
wingspan / –3.58 (2.32) / 1.54 / 0.13
wing area / 2.19 (1.89) / 1.16 / 0.25
aspect ratio / 3.81 (1.27) / 3.01 / 0.004
wing loading / –1.30 (0.86) / 1.51 / 0.14
Table S6 The same ‘Organ size’ model set as in Table 2 in the MS except that the organ size data of individuals with unknown date of collection or collected during the spring or autumn migratory period were excluded. Significant relationships are highlighted in bold
Full model / MAMpredictors / β (SE) / t / P / predictors / β (SE) / t / P
(a) all species
n = 83, Pagel’sλ = 0.98 / n = 149, Pagel’sλ = 0.80
intercept / 3.39 (1.41) / 2.41 / 0.02 / intercept / 3.75 (1.10) / 3.39 / 0.0009
log10 body mass / –0.84 (0.49) / 1.73 / 0.09 / log10 body mass / –1.26 (0.37) / 3.39 / 0.0009
LCB / 0.43 (0.22) / 1.94 / 0.06 / LCB / 0.45 (0.17) / 2.67 / 0.009
gizzard / –0.32 (0.31) / 1.02 / 0.31
liver / 0.50 (0.36) / 1.38 / 0.17
pectoral muscle / 0.55 (0.35) / 1.59 / 0.12
supracoracoid muscle / –0.61 (0.45) / 1.36 / 0.18
heart / –0.77 (0.35) / 2.22 / 0.03
(b) only passerine species
n = 49, Pagel’sλ = 0.96 / n = 84, Pagel’sλ = 0.86
intercept / 2.02 (1.00) / 2.02 / 0.05 / intercept / 2.11 (0.92) / 2.31 / 0.02
log10 body mass / –0.82 (0.45) / 1.83 / 0.07 / LCB / 0.61 (0.19) / 3.17 / 0.002
LCB / 0.32 (0.32) / 1.02 / 0.31
gizzard / –0.54 (0.41) / 1.32 / 0.19
liver / 0.35 (0.44) / 0.80 / 0.43
pectoral muscle / 0.39 (0.48) / 0.82 / 0.42
supracoracoid muscle / 0.24 (0.51) / 0.47 / 0.64
heart / –0.62 (0.39) / 1.57 / 0.12
Table S7 The same ‘Organ & wing’ model set as in Table 2 in the MS except that the organ size data of individuals with unknown date of collection or collected during the spring or autumn migratory period were excluded. Significant relationships are highlighted in bold
Full model / MAMpredictors / β (SE) / t / P / predictors / β (SE) / t / P
(a) all species
n = 79, Pagel’sλ = 0.94 / n = 79, Pagel’sλ = 0.94
intercept / 2.46 (1.11) / 2.21 / 0.03 / intercept / 2.50 (1.00) / 2.51 / 0.01
log10 body mass / 0.90 (0.93) / 0.97 / 0.34 / heart / –0.71 (0.27) / 2.59 / 0.01
LCB / 0.30 (0.21) / 1.43 / 0.16 / aspect ratio / 2.00 (0.30) / 6.75 / < 0.0001
gizzard / –0.14 (0.27) / 0.52 / 0.61
liver / 0.44 (0.33) / 1.33 / 0.19
pectoral muscle / 0.14 (0.37) / 0.38 / 0.70
supracoracoid muscle / 0.09 (0.41) / 0.22 / 0.83
heart / –0.84 (0.35) / 2.39 / 0.02
wingspan / –3.58 (2.32) / 1.54 / 0.13
wing area / 2.19 (1.89) / 1.16 / 0.25
aspect ratio / 3.81 (1.27) / 3.01 / 0.004
wing loading / –1.30 (0.86) / 1.51 / 0.14
(b) only passerine species
n = 49, Pagel’sλ = 0.91 / n = 60, Pagel’sλ = 0.88
intercept / 2.47 (0.82) / 3.02 / 0.005 / intercept / 2.23 (0.70) / 3.21 / 0.002
log10 body mass / 0.56 (1.03) / 0.55 / 0.59 / heart / –0.50 (0.25) / 2.06 / 0.04
LCB / 0.11 (0.32) / 0.35 / 0.73 / aspect ratio / 1.66 (0.28) / 5.93 / < 0.0001
gizzard / –0.32 (0.39) / 0.82 / 0.42
liver / 0.34 (0.40) / 0.84 / 0.41
pectoral muscle / –0.34 (0.50) / 0.68 / 0.50
supracoracoid muscle / 0.89 (0.50) / 1.76 / 0.09
heart / –0.92 (0.38) / 2.42 / 0.02
wingspan / –0.57 (3.97) / 0.14 / 0.89
wing area / 1.02 (2.22) / 0.46 / 0.65
aspect ratio / 2.24 (2.82) / 0.79 / 0.43
wing loading / 0.09 (1.63) / 0.05 / 0.96
1