Morphological Shifts of the External Flight Apparatus across the Range of a Passerine (Northern Wheatear) with Diverging Migratory Behaviour

We studied morphological differentiation in the flight apparatus of the four currently recognised sub-species of Northern Wheatears, Oenanthe oenanthe. Considering all measured birds without assigning them a priori to any sub-species we found a clinal morphological shift. Relative wing length, wing pointedness, and the degree of tail forking were positively correlated with migratory distance, whereas tail length (relative to wing length) was negatively correlated. The large-sized, long-distance migrant “Greenland” Wheatear, O. o. leucorhoa, is characterized by relatively longer, broader and more pointed wings and more forked tails, similar to the smaller-sized nominate Northern Wheatear, O. o. oenanthe, from North Europe, Siberia and Russia. In contrast, the short distance migrant “Seebohm's” Wheatear, O. o. seebohmi, from northwest Africa, possesses much rounder wings, and the tail is relatively longer and less forked. Sub-species with intermediate migratory habits (different populations of nominate Northern Wheatear, O. o. oenanthe, and “Mediterranean” Northern Wheatear, O. o. libanotica) show, as expected, intermediate features according to their intermediate migratory behaviour. Our results are congruent with other inter- and intraspecific studies finding similar adaptations for energy-effective flight in relation to migration distance (morphological migratory syndrome).


Introduction
The morphology of the avian wing constitutes a trade-off between various selection pressures that act on its aerodynamic and mechanical properties [1][2][3]. The evolution of wing and tail size and shape is affected by the diverging demands of migratory behaviour, take-off ability in response to predator attacks and by the density of obstacles that constrain flight manoeuvrability in the occupied habitats [4][5][6][7].
Slender and more pointed wings and shorter tails in relation to the wing reduce the induced drag at the wings considerably and are known to produce a larger forward component in flight during migration [1][2][3]8]. Furthermore, more forked tails are known to provide higher uplift and lower drag [3,9]. Consequently, we may assume that the extent of migratory behaviour results in changes in the external morphology of the flight apparatus [10][11][12] which select for energy-efficient flight [13].
The Northern Wheatear, Oenanthe oenanthe (Linnaeus, 1785), is one of the most diverse migratory song birds of the Palaearctic and therefore well suited for an intra-specific study [27]. This species is distributed from North Africa northwards to Iceland and Greenland and continuously from Europe towards eastern Russia [28]. Small populations have even settled the Nearctic region (Canada and Alaska). All populations still overwinter in sub-Saharan Africa and need to migrate large distances in order to reach their winter quarters. However, the distinct populations differ considerably in the distances they have to travel (Figure 1).
Four sub-species of Northern Wheatears are currently recognised [28]. The sub-species O. o. seebohmi (Dixon, 1882; ''Seebohm's'' Wheatear) is restricted to the Atlas mountains of northwest Africa. The male nuptial plumage of this form is quite distinct from all other Northern Wheatears, and it is therefore sometimes treated as a separate species [28]. It shows the shortest migration distances, wintering mainly in southwestern Mauritania and Senegal [29]. Somalia and Ethiopia), has been treated formerly as another sub-species of the Northern Wheatear, but recent genetic studies show that this form is a distinct species [30][31].
The wide distribution of the populations of Northern Wheatears suggests specific adaptations to migration, depending on the distance the birds have to travel [27]. We therefore studied museum specimens to examine how different migratory behaviours correlate with the morphologies of the different subspecies. In particular, we studied which morphological changes of the  [5]. The species has one of the largest breeding ranges for a passerine. The whole population winters in sub-Saharan Africa (in grey; [33]

Results
The four currently recognised sub-species of the Northern Wheatear show clear morphological differentiation in the flight apparatus (Table 1). Our ANOVAs identified various significant differences both in uncorrected and body size corrected analysis ( Table 2, Table 3). Using a PCA on the 9 morphometric variables of the flight apparatus (size corrected; log-transformed; varimax rotation) we obtained two relevant principal components (PCs) with an Eigen-Value .1 explaining 62.6% of total variance (Table 4, Figure 2). PC1 explained 38.9% of the total variance and comprises wing length (maximal wing chord), the distance of first secondary-wing tip, distal primary-wing tip and alula-wing tip. PC2 explained 23. Since obvious clinal variation exists within the single sub-species of the Northern Wheatear and the separation of the sub-species is not always accurate due to distribution overlap, we conducted linear regressions independent of taxonomic status. In these analyses we included only specimens for which we had details on the collection localities (n = 234). We found a significant correlation between both principal components (PC1, PC2) and the migratory distance (Table 5, Figures 3 and 4). Birds with longer migratory pathways had (1) relative longer (WL) and more pointed wings (S1Wt); (2) relatively more narrow wings (WW); (3) a shorter alula and P1 in relation to wing length (AtWt, P1Wt); (4) relatively shorter emarginations on the wing-tip (NoP2, NoP3); and (5) relatively shorter and more forked tails in relation to wing length (TL, TF). Regressions of migratory distance with tail-wing ratio and wing shape index revealed congruent results (Table 5). Birds with longer migration distances showed relatively shorter tails in relation to wing length ( Figure 5) and more pointed wings ( Figure 6).

Variable ANOVA Bonferroni
Tail-wing ratio   Birds with longer migratory pathways possess relatively longer, more pointed, and more slender wings, shorter emarginations on the wing tip, and show relatively shorter tails in relation to wing length and a more forked tail.
The large sub-species O. o. leucorhoa shows the strongest adaptations to long-distance migration, because it is the only form which needs to cross a large water body (north Atlantic) during migration. These adaptations include relative longer, broader and more pointed wings and stronger forked tails, which may help to stabilise the bird during migration in harsh climatic conditions over the sea. Similar results were obtained in a recent study by Delingat and colleagues [27], who showed by means of isotopic analyses that presumed Greenlandic Northern Wheatears of the sub-species O. o. leucorhoa have more pointed wings than their congeners from other European breeding areas. However, in our study we found that,   Table 4). doi:10.1371/journal.pone.0018732.g002 Table 5. Regression analyses between migratory distance and PC1 (WL, ATWT, P1Wt, S1Wt), PC2 (TF, TL, WW, NoP2, NoP3.), tail-wing ratio, and wing shape index in Northern Wheatear Oenanthe oenanthe (n = 235). The adaptations in the flight apparatus observed in our study follow the general predictions of the so-called migratory syndrome [32]. Similar to the study of Fiedler [13], we found birds with a more ''migratory type'' flight apparatus to have developed a more efficient morphology of the external flight apparatus than their less migratory conspecifics. Studies on aerodynamics of bird flight [1,2,8] have demonstrated that the observed morphological shift with increasing migratory distances is well suited to produce a larger forward component in flight due to a more prominent distal part of the wing. The more slender and pointed wings and the shorter tail in relation to the wing reduce the induced drag at the wings and produce greater uplift . Relationship between PC1 (WL, AtWT, P1Wt, S1Wt) and migratory distance. Populations of Northern Wheatears with longer migration pathways have relatively longer wings (for statistics, see Table 5). doi:10.1371/journal.pone.0018732.g003  and thrust [2,3]. Additionally, the short and stronger forked tails provide high lift and low drag [3,9].
As a possible trade-off, the adaptations for migration constrain the manoeuvrability of the birds. A decrease of Reynolds number due to a higher aspect ratio of the wing and a reduced ability of the tips to bend and generate lift due to relatively short notches at the wing tip result in a reduced capacity for very slow flights under high angles of attack [2,13]. Additionally, relatively short tails generate less lift in slow flights and reduce the ability of the tail to start or stop roll manoeuvres [9].
Besides, it is likely that other factors might have influenced the morphological differentiation of the flight apparatus as well, such  Table 5). doi:10.1371/journal.pone.0018732.g005 Figure 6. Relationship between wing pointedness (wing shape index [13]) and migratory distance. Populations of Northern Wheatears with longer migration pathways have more pointed wings (for statistics, see Table 5). doi:10.1371/journal.pone.0018732.g006 as differences in foraging, breeding habitat or sexual selection of different sub-populations. However, because all sub-species live in very similar habitat types (open, rocky areas) and show equivalent breeding and foraging behaviour, we believe that the demands for migration are the main driving forces for the morphological shift of the flight apparatus.
To summarize, the intraspecific patterns in flight apparatus that we found in the Northern Wheatear nicely follow the expectations drawn from other work [11,13,32], indicating that in birds travelling longer distances the traits for energy-effective flight (in terms of distance travelled per energy expended) are obviously more strongly developed then the traits for manoeuvrability. Future work needs to reveal how these changes in external flight morphology are linked to other physiological, behavioural and internal morphological adaptations to migration and how fast these morphological shifts may appear in the evolutionary history of a species.

Materials and Methods
We measured external morphological traits of the flight apparatus to compare between Northern Wheatears of the four currently recognised sub-species with different migratory behaviour ( Figure 1). Specimens from the following European museum collections were used (Appendix S1): Zoologisches Forschungsmu Nine external characters of the flight apparatus were measured to the nearest 0.1 mm [11] (Table 6). Furthermore, we calculated tail-wing ratio and wing shape index. The latter was derived by the following formula: Wing shape index = (differences between longest primary and innermost primary -difference between longest primary and outermost primary)/wing length following Fiedler [13]. A higher value indicates a more pointed wing. In order to guarantee comparability between specimens we used only skins of adult male specimens in spring or summer plumage collected from breeding areas. We calculated the distance between collection place and main wintering area [33] following the method of Imboden & Imboden [34].
In total, we obtained data from 242 male Northern Wheatears, Oenanthe oenanthe. Finally we conducted regression analyses between migratory distance and the two principal components and the two indices (the latter two showed normal distribution). We did not correct for phylogeny, because all forms are closely related and currently no comprehensive tree exists for the genetic relationship between the different sub-populations and sub-species. All analyses were done in SPSS 12.0.

Supporting Information
Appendix S1 Specimens of Northern Wheatear Oenanthe oenanthe measured in different museums. Given are the collection numbers and the assigned sub-species group. (DOC) Tail-wing ratio ratio of tail length to maximum wing chord Wing shape index wing pointedness [13] Primaries are numbered from the outermost to the innermost. doi:10.1371/journal.pone.0018732.t006