Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Habitat Use and Body Mass Regulation among Warblers in the Sahel Region during the Non-Breeding Season

Habitat Use and Body Mass Regulation among Warblers in the Sahel Region during the Non-Breeding Season

  • James O. Vafidis, 
  • Ian P. Vaughan, 
  • T. Hefin Jones, 
  • Richard J. Facey, 
  • Rob Parry, 
  • Robert J. Thomas


Migratory birds face significant challenges across their annual cycle, including occupying an appropriate non-breeding home range with sufficient foraging resources. This can affect demographic processes such as over-winter survival, migration mortality and subsequent breeding success. In the Sahel region of Africa, where millions of migratory songbirds attempt to survive the winter, some species of insectivorous warblers occupy both wetland and dry-scrubland habitats, whereas other species are wetland or dry-scrubland specialists. In this study we examine evidence for strategic regulation of body reserves and competition-driven habitat selection, by comparing invertebrate prey activity-density, warbler body size and extent of fat and pectoral muscle deposits, in each habitat type during the non-breeding season. Invertebrate activity-density was substantially higher in wetland habitats than in dry-scrubland. Eurasian reed warblers Acrocephalus scirpaceus occupying wetland habitats maintained lower body reserves than conspecifics occupying dry-scrub habitats, consistent with buffering of reserves against starvation in food-poor habitat. A similar, but smaller, difference in body reserves between wet and dry habitat was found among subalpine warblers Sylvia cantillans but not in chiffchaffs Phylloscopus collybita inhabiting dry-scrub and scrub fringing wetlands. Body reserves were relatively low among habitat specialist species; resident African reed warbler A. baeticatus and migratory sedge warbler A. schoenobaenus exclusively occupying wetland habitats, and Western olivaceous warblers Iduna opaca exclusively occupying dry habitats. These results suggest that specialists in preferred habitats and generalists occupying prey-rich habitats can reduce body reserves, whereas generalists occupying prey-poor habitats carry an increased level of body reserves as a strategic buffer against starvation.


Migratory birds face the challenge of finding sufficient food resources on their wintering grounds to avoid starvation during the non-breeding season and to prepare themselves for the return migration to their breeding grounds. To do this, they must select suitable wintering habitat, in competition with other migrant and resident species [1][3]. The causes and consequences of habitat choices made by migratory birds outside the breeding season are widely studied and much debated [4][7]. Although it is well established that competition affects the distribution of individuals on their temperate breeding grounds [8][9], it has been suggested that competition may be less important on the wintering grounds where migrant birds may have lower energetic requirements [10][12]. Migratory birds arriving en masse in the wintering areas must, however, select between habitats that may differ in foraging quality, the density of resident and migrant competitors, and other factors such as predation pressure [13], [14]. The strength of these factors is likely to be affected by variations in climate. Indeed, overwinter survival of some long-distance migrant birds is closely linked to climate-driven measures of broad-scale environmental conditions on their wintering grounds, such as El Niño Southern Oscillation, Sahel Precipitation Index and Normalised Difference Vegetation Index [15][20]. Such correlations imply that survival of some, or many, migrant species may be limited by overwinter foraging conditions.

In a competition-driven system, residents and dominant or early-arriving migrants may be able to secure territories in primary (higher quality) habitat while subordinate or late-arriving individuals may be competitively excluded into secondary habitats: the competitive exclusion hypothesis [18], [21], [22]. Compelling evidence for the role of inter-specific competitive exclusion in the shaping of migrant and resident communities in the wintering grounds is lacking [2], [10], [12]. However, evidence for such competitive exclusion in an intra-specific context was provided by Perez-Tris and Telleria (2002) [23], who showed that resident blackcaps Sylvia atricapilla breeding in southern Spain maintained their breeding territories in primary forest habitats into the non-breeding season, while migrant blackcaps arriving at the start of the wintering season ‘leap-frogged’ into the secondary (lower quality) scrub habitats not occupied by residents.

An alternative to the competitive exclusion hypothesis (though not necessarily mutually exclusive with it) is that migrants are “eurytopic” (i.e. adaptable) and have less restrictive habitat requirements during the non-breeding season, allowing them to use a wider range of winter habitat types than ecologically similar but more specialised resident species [2], [21], [24]. Such adaptability and breadth of foraging strategies has been shown in wintering golden orioles Oriolus oriolus, which were, by virtue of their wider range of foraging tactics, able to survive in lower quality microhabitats than were preferred by resident species [25]. Migrants may therefore be able to exploit lower quality habitats with lower densities of resident competitors [26][29], even though they may be markedly different from the habitat types used on their temperate breeding grounds [12], [24], [30], [31]. Furthermore, migrants may be able to occupy habitats which residents may find unsuitable, either due to the vegetation structure or insufficient food availability to sustain future breeding attempts (i.e. habitats of ‘low breeding currency’; [32], [33]).

It is well-established by both theoretical and empirical studies that the level of body reserves carried by small birds reflects a trade-off between starvation and mass dependent costs of body reserves [34], [35]. These mass dependent costs include increased predation risk, flight costs and increased foraging requirements [36][39]. Balanced against these mass-dependent costs are the benefits of the stored body reserves acting as a strategic buffer against starvation in situations where food is scarce or unpredictable. This hypothesis of ‘adaptive mass-regulation’ or ‘strategic buffering against starvation’ is supported by experimental studies on north-temperate wintering birds such as greenfinches Carduelis chloris [40], European starlings Sturnus vulgaris [41], [42] and great tits Parus major [43]. These species responded to food limitation by accumulating mass (in comparison to controls) and responded to food supplementation by decreasing mass. Strategic buffering has been implicated in the observed fattening of wintering brambling Fringilla montifringilla [44], great tits [45] and Siberian Jays Perisoreus infaustus [46] in response to declining food availability or the risk of prolonged periods of food limitation. While the influence of predation pressure and competition are important determinants of energy reserves carried by birds [39], [47], it is clear that the overall availability of food in a habitat has a major influence on bird foraging and mass regulation strategies. Despite these examples, however, little is known about the mass regulation strategies used by European migrants wintering in Africa.

Wetlands in the western Sahel, mostly within the Inner Niger Delta and Senegal Delta, represent the main overwintering area for many Western Palearctic-African migrant bird species [48]. The area and quality of these wetlands is dependent on the extent of summer rainfall across West Africa. These gradually declined through much of the 20th Century [5], reducing their carrying capacity for wintering songbirds [16], [17]. In addition to climate impacts, habitat loss and degradation from agriculture and irrigation have led to further reduction of available wetland habitat in the Sahel [5]. Many of the areas surrounding the wetlands comprise dry scrub habitats, which are also occupied by migrants.

The invertebrate-rich wetlands support a mixed community of insectivorous ‘Old World’ warblers (superfamily Sylvoidea), including the resident species African reed warbler Acrocephalus baeticatus and greater swamp warbler A. rufescens, as well as migrant species including sedge warbler A. schoenobaenus and Eurasian reed warbler A. scirpaceus [49][53]. The dry scrub supports dry-habitat breeding migrant warblers, such as subalpine warbler Sylvia cantillans, western olivaceous warbler Iduna opaca and common chiffchaff Phylloscopus collybita (henceforth “chiffchaff” [54]). Most of these species are restricted to either wetland or scrub habitats, but three are relative habitat generalists: Eurasian reed warbler regularly occupies dry scrub, while subalpine warbler and chiffchaff often forage in the ecotone between scrub and wetland habitats. Such use of multiple winter habitat types by warbler species is well known [2], [55], [56]. Furthermore, there appears to be significant intra-specific variation in the body reserves (in the form of fat and pectoral muscle) of birds wintering in West Africa [57], [58], although the reasons for such variation are unknown. However, while several studies have shown intra-specific responses of wintering migrant birds to habitat quality gradients in Neotropical ecosystems [59], [60], [20], no study has attempted to investigate mass regulation strategies among warbler species occupying the two major types of wintering habitats in the Sahel; wetland and dry-scrub.

In the present study, we investigated the distribution, structural size and extent of body reserves (in the form of fat and pectoral muscle) of insectivorous songbirds in a wintering area in Senegal, West Africa, during three successive non-breeding seasons. The aims of the study were to: 1) assess the relative abundance of invertebrate prey resources in wetland and dry scrub habitats, providing a measure of habitat quality; 2) identify habitat preferences of five migratory warbler species and one resident, comprising two wetland specialists, one dry scrub specialist and three habitat generalists; and 3) compare the body mass, fat, pectoral muscle and size of individuals in wet and dry habitats, to test the hypothesis that habitat usage and body condition are driven by strategic responses to the availability of food in the different habitats. We tested predictions about the level of body reserves (in the form of fat and pectoral muscle) expected under conditions of direct food limitation compared to strategic regulation of body reserves, in both wetland and dry-scrub habitats. Specifically, if warblers are strategically regulating their body reserves as a buffer against starvation, we predict that individuals inhabiting dry scrub habitat where food availability is lower (and thus starvation risk is greater), carry higher levels of energy reserves than in habitats where food is more abundant. If, however, warbler body reserves are directly limited by food availability, we predict body reserves to be lower in the habitat with lowest food availability. By including the ‘habitat specialist’ species African reed warbler, sedge warbler and western olivaceous warbler in this study, we test whether this pattern extends across species among habitat specialist and habitat generalist species differing in their migratory tendency and timing of arrival on the wintering grounds.


Ethics Statement

The guidelines promoted by the Association for the Study of Animal Behaviour for the ethical use of animals in research were followed. All species caught as part of this study are common and not registered as an endangered or protected species in any country. All fieldwork was conducted after ethical approval from Parc National des Oiseaux du Djoudj (permit number 021 403) and the British Trust for Ornithology ringing unit (Vafidis license 5475; Facey license 5411).

Study Area

The ‘Parc National des Oiseaux du Djoudj’ (16° 21′ 59″ N, 16° 16′ 26″ W) is located in the semi-arid zone in Senegal, West Africa and covers an area of 16,000 ha of seasonally flooded waterways. Its landscape is Acacia-Commiphora grassland and Tamarix senegalensis scrub savannah interspersed with a range of Phragmites and Typha-dominated wetland habitats. The study area supports wintering populations of a range of warbler species including Eurasian reed warbler, sedge warbler, chiffchaff, subalpine warbler, western olivaceous warbler and African reed warbler (Table 1). Fieldwork was conducted near the village of ‘Diadiem 3’ (16° 21′ 7″ N, 16° 16′ 34″ W), in the south west of the national park in wetland and dry scrub habitats near Marigot du Khar. Permission to undertake fieldwork was obtained from the national park authority.

We used mist nets to sample the community of warblers occupying four locations within an approximately 40 hectare section of wetland (centred at 16° 22′ 25″ N, 16° 16′ 12″ W) over three winter seasons (January 2012, 2013 and 2014) and six locations within an approximate 400 hectare section of dry scrub habitat (centred at 16° 21′ 44″ N, 16° 15′ 49″ W) in the surrounding area in two winter seasons (January 2013 and 2014). Wetland study sites were located 260–1550 m apart. Dry scrub study sites were located 100–2000 m apart. The minimum distance between wetland and dry study sites was 775 m. Although individual birds were occasionally caught in sites adjacent to those in which they were originally ringed, no individuals ringed in wetland sites were recorded in dry scrub sites, or vice-versa. Bird biometric data were collected during 24 morning or evening ringing sessions across the three years, during a period when birds were not migrating (between 17 and 29 January in each year). Each habitat type was subject to the same netting effort per ringing session, over the course of the study (66 m per session, approximately 36 hr/m). Nets were checked at least every 20 minutes.

Bird biometric data

Captured birds were taken in cotton bags to a nearby processing station where species and age (where possible) were recorded following Svensson (1992)[61]. The following biometric measurements taken: wing length (maximum wing chord to 1 mm), tarsus length (tarsal joint to top of flattened foot to 0.1 mm), total head length (back of skull to distal tip of bill, to 0.1 mm) and mass using an electronic balance (to 0.1 g) with time of weighing recorded. The measurement techniques followed the methods described by Svensson [61]. The size of the pectoral muscle was scored following Kaiser (1993)[62] (0 =  emaciated to 3 =  large muscle mass) and subcutaneous fat deposits were estimated following Bairlein (1995 [63]: fat score; 0 =  no visible subcutaneous fat, 8 =  whole belly covered in fat). To eliminate among-observer variability, all measurements were either made or checked by one person (JV). The recording period occurred after the typical winter moulting period for Eurasian reed and sedge warblers, so age was generally not determinable. Each bird was fitted with a British Trust for Ornithology (BTO) issued metal ring.

Invertebrate Monitoring

A measure of invertebrate prey availability was determined using sticky traps (yellow, double sided, effective area 100 cm2, Oecos, Hertfordshire, UK). These traps are highly effective for sampling the activity-density of Diptera [64][66], the primary prey taxon of Eurasian reed warbler in Europe (R.A. King, personal communication). To assess stability of wetland invertebrate prey resources over the dry season we monitored the invertebrates on a weekly basis for a period of three months between 16 January and 20 March 2012. Seven traps were set in each of the four sites used for bird monitoring, attached to a mixture of scrub and Phragmites vegetation at heights of between 0.5–1.2 m and at least 10 m away from mist net positions and any regular pathways used by mammals (including humans). These wetland sites and four dry scrub habitats were also monitored using seven traps in each site over the course of one week in 2013 and 2014 (with traps set in the same positions in wetlands as in 2012) and in dry scrub habitats. Invertebrates were not monitored in dry scrub habitats in 2012. Traps were attached around the edges of habitat fragments set at heights between 0.5–1.2 m.

Total captures of Diptera, Arachnidae, Hymenoptera and Hemiptera, as well as other less-frequently encountered taxa (<1%) such as Coleoptera and Lepidoptera recorded in each habitat, providing a cumulative measure of activity-density, were compared between wetland and dry scrub sites. To assess general differences in invertebrate size between habitats, all sampled invertebrates were measured for body length (excluding legs, wings and antennae) and categorised as either small (≤5 mm) or large (>5 mm). This threshold size value is used in other similar studies testing for prey differences for wintering small billed migrant birds for which diet is primarily comprised of small invertebrate prey [33], [67]. The difference in distribution of invertebrate size (small and large) was compared between wetland and dry scrub sites. Mean daily temperatures were recorded in each habitat type using temperature loggers (LASCAR EL USB-1).

Data analysis

All analysis was undertaken using R version 3.0.3. [68].

Assessment of Invertebrate Resources

  1. 2012 Wetland habitat analysis

The invertebrate activity-density across the 2012 study period were analysed with generalized linear mixed-effects models (GLMM; R package “lme4” [69]) using date as a fixed effect and trap location as a random intercept.

  1. 2013 and 2014 between-habitat analysis

Differences in activity-density and the distribution of invertebrate size using the January 2013 and 2014 invertebrate data were analysed using a generalized linear model (GLM) using habitat type (“wetland” or “dry scrub”) and year as factors. Differences in the variances of invertebrate samples between habitat types were tested using an F-test.

Assessment of habitat preferences

We assessed habitat preference among warbler species by modelling probability of occurrence using a binomial GLM with ‘habitat’ as the binary dependant variable (i.e. “dry” or “wet”), and species as an independent variable. Habitat preferences were compared among species using the ‘contrast’ package [70], with positive and negative parameter values representing closer association with wetlands and dry scrub habitats, respectively.

Body mass, muscle, fat and body size comparisons

Intra-specific comparisons of body mass and structural body size (wing length) were made for birds occupying both habitats. Body mass and structural body size was not comparable between species because of differences in overall size and shape of species. Comparisons of fat and muscle reserves of birds using both habitat types were made between species as well as within species.

Body mass was compared between habitats for each generalist species (Eurasian reed warbler, subalpine warbler and chiffchaff) using a generalised additive model (GAM) implemented using the “mgcv” package [70]. Habitat type was modelled as a factor, flattened wing chord, total head length and tarsus length were modelled as linear relationships to control for body size, and time of day was modelled using a cyclic cubic regression spline to allow for the diurnal variation in body mass [71]. We also tested for differences in mass between years, using the same model structure, but with year added as a fixed factor. For birds captured on more than one occasion during the same winter, only the first capture event was included in the analysis, to avoid pseudo-replication. Only six birds were captured in more than one winter, and their recaptures in subsequent years were excluded from the analysis. Residual diagnostic plots from the models were used to verify the assumptions of normality and homogeneity of model residuals, and to test for unduly influential observations [72].

Intra-specific differences in the structural body size of individuals between habitat types were compared using a general linear model (GLM), using flattened wing chord as a dependent variable, and habitat type and year as independent effects. An alternative model using the first principal component (PC1) of a principal components analysis (PCA) of the three measures of bird body size (flattened wing chord, tarsus and total head length) as a dependant variable did not improve the fit of the model.

Inter- and intra-specific comparisons of body condition were conducted using muscle and fat scores, since these scores represent an index of energy reserves and allows for direct comparisons of condition between species of different size and morphology (such comparisons using body mass are problematic because of the difficulty in defining lean body mass in live birds). Comparisons between species of fat and muscle reserves are valid because these are size-independent variables [73][76]. Since the muscle and fat score methods applied in this study used ordered discrete values (0–3) and (0–8) respectively, a proportional odds regression (a special case of ordinal logistic regression) was used to compare muscle and fat scores between habitats, while controlling statistically for time of day [77]. Proportional odds regression models were fitted as ordinal regression models, using the “MASS” package [78]. Before the proportional odds regression was performed, the muscle and fat scores were pooled into one of three levels “0”, “1” and “>1”, so as to distinguish between birds with minimal, low and higher reserves. Pooling together birds with fat score >1 increased the sample size for this category, allowing more powerful contrasts to be made. The ordinal logistic regression analysis generated an odds ratio for each species in a particular habitat. The odds ratio represents the odds of a particular species having a muscle or fat score one unit higher than a reference species, in a particular habitat.


Invertebrate prey resource differences between and within habitats

The longer monitoring period in 2012 (three months) revealed a decline of 10.6% in invertebrate abundance over the winter in wetland habitats (GLMM; slope estimate  = −0.017±0.002, Z = −7.36, P<0.0001; Figure 1a). In both 2013 and 2014, wetland and dry scrub habitats supported a similar invertebrate community at the Order level (Table 2), but mean abundance of invertebrates per sample in the wetlands (135.2±4.15) was more than double that of the dry scrub habitat (57.52±3.07; GLM: F141,140, = 182.141, P<0.0001; Figure 1b). Variances of invertebrate samples in wetlands were significantly higher than dry scrub in 2013 (F27,27 = 5.891, P<0.0001) but not in 2014 (F29,27 = 1.88, P = 0.103). There was significant variation in the occurrence of large invertebrates between wetland and dry scrub habitats (GLM; F1,140 = 37.369, P<0.0001). Large invertebrates comprised 2.3% of all individuals trapped in wetlands in 2013 and 2014, consisting of Arachnidae, Diptera, Hymenoptera, Lepidoptera (and unidentifiable specimens), compared with 1.1% in dry scrub habitats consisting of Arachnidae, Diptera and Hymenoptera (Table 2).

Figure 1.

a) Mean numbers of invertebrates per trap (±SE) within wet habitat between 23 January-19 March 2012; and 1b) Wet habitats and dry habitats between 23rd–29th January 2012, 22nd–30th January 2013, and 21st–27th January 2014.

Table 2. Abundances of small and large invertebrates by taxonomic group trapped in wetland and dry scrub sites in Parc National des Oiseaux du Djoudj during January 2013 and 2014 combined.

Testing for species differences in habitat preferences of warblers

Over the course of the study, 116 Eurasian reed warblers, 89 subalpine warblers and 194 chiffchaffs were caught across wetland and dry scrub habitats (Table S1). A total of 151 sedge warblers and 67 African reed warblers were caught in wetlands, and 25 olivaceous warblers were caught in dry scrub habitats. There was no significant variation in habitat preference between generalist species (GLM; F5, 381 = 0.177, P = 0.971) with birds caught in the wetlands (in 2013 and 2014) representing 46.4% of all captures of Eurasian reed warblers, 43.3% of chiffchaffs and 17.9% of subalpine warblers.

Warbler Body Mass

Time of day was an important determinant of body mass for all species, with a significant increase in mass from the early morning to reach a peak in the late afternoon (F7.874,8.667 = 4.872, P<0.0001). All else being equal, the estimated mean difference between birds caught at the beginning of the day and those caught in the late afternoon was greatest for Eurasian reed warblers (1.58 g, representing 15.6% of mean winter body mass) and least for chiffchaff (0.49 g, representing 7.0% of mean winter body mass). After controlling statistically for time of day, year and species effects, mass was positively associated with wing length (GAM; 0.089±0.009 g/mm, t = 10.495, P<0.0001), muscle score (GAM; 0.146±0.049 g/integer, t = 2.971, P = 0.0031) and fat score (GAM; 0.355±0.021 g/integer, t = 16.582, P<0.0001). The effect of habitat type differed between species (Table 3). Eurasian reed warblers and subalpine warblers were significantly heavier in dry scrub than wetlands (by 0.71 g (7.0% of mean winter body mass) ±0.14 g and 0.34 g (4.0% of lean mass) ±0.15 g, respectively). Chiffchaffs also weighed more in scrub on average (by 0.17 g (2.4% of lean mass) ±0.11 g) but the difference was not significant. There was no significant inter-annual variation in body mass for Eurasian reed warbler and subalpine warbler but significant variation between years in chiffchaff with 2014 and 2013 masses significantly greater than 2012 (Table 3; Figure 2). Controlling statistically for time of day, body size and the extent of fat and muscle, chiffchaff mass was strongly positively associated with the Sahel precipitation index for the preceding wet season (GAM: +0.20 g ± 0.04 g, t = 5.764, P<0.0001), indicating greater mass following higher rainfall. Of the habitat specialist species, only African reed warbler showed significant variation in mass, exhibiting an increase in mass between 2012 and 2014 (GAM; +0.399 g±0.177 g, t = 2.261, P = 0.0274; Figure 3).

Figure 2. Body mass (± SE, corrected for structural body size and time of day) and wing length (± SE) of wintering populations of generalist warbler species occupying both wetland and dry scrub habitats.

Figure 3. Body mass (± SE, corrected for structural body size and time of day) and wing length (± SE) of wintering populations of habitat specialist warblers occupying either wetland or dry scrub habitats.

Table 3. Generalised Additive Models of body mass (dependent variable) for each generalist species. Parameter estimates for levels of the factors “Habitat” and “Year”, are relative to the reference levels of “Wetland” and “2012” respectively.

Warbler Structural Body Size

There were no significant differences in structural body size between wetland and dry scrub samples of any of the three generalist species, nor were there any significant differences in size between years (Table 4, Figure 3).

Table 4. General Linear Model of structural body size (dependant variable) for each generalist species.

An equivalent model using 2013 and 2014 data only (excluding 2012 on the grounds that there was no habitat contrast available) showed there was no significant year x habitat interaction influencing body size in Eurasian reed warblers (t83 = −1.683, P = 0.0963), chiffchaff (t105 = −1.264, P = 0.209) or subalpine warbler (t66 = 1.824, P = 0.0729), despite some apparent size differences between 2013 and 2014 in Figure 2. Of the habitat specialist species, only African reed warbler showed significant differences in the population wing length, with birds in 2014 being 1.38±0.561 mm smaller than birds in 2012 (GLM; F68,66 = −3.142, P = 0.0497; Figure 3).

Body reserves

The proportional odds regression revealed that muscle score was not significantly affected by habitat type (t = 0.492, P = 0.6224) or time of day (t = 1.432, P = 0.1521). The ordinal logistic regression analysis generates an odds ratio for each species in a particular habitat, which represent the odds of having a muscle score one unit higher than a reference species, in a particular habitat. The only interspecific difference was the wetland specialist species African reed warbler having 1.50 times greater odds (P = 0.035) of having a muscle score one unit higher than the dry- scrub specialist western olivaceous warbler (Table 5). Fat score was significantly affected by habitat type (t = −3.944, P<0.0001; Table 5, Figure 4) and time of day (t = 3.874, P = 0.0001). Eurasian reed warblers had 2.32 times greater odds of having a fat score one unit higher in the dry-scrub than in the wetlands (Table 6; Figure 4). The wetland specialist species sedge warbler had 1.26 times greater odds of having a fat score one unit higher relative to Eurasian reed warbler occupying the wetland habitat. Conversely, in the dry habitats, Eurasian reed warbler had 1.70 times greater odds of having a fat score one unit higher relative to the dry habitat specialist olivaceous warbler.

Figure 4. Model predictions for the probability of observing birds in wetlands or dry scrub habitat with a fat score of greater than 1.

Table 5. Pairwise comparison of proportional odds of species in a particular habitat having a muscle score of 1 unit higher than another species in a particular habitat, or the same species in a different habitat.

Table 6. Pairwise comparison of proportional odds of species in a particular habitat having a fat score of 1 unit higher than another species in a particular habitat, or the same species in a different habitat.


As expected, wetland habitats supported consistently greater invertebrate resources than dry scrub habitats [79][82]. On the basis of food availability, wintering generalist warblers should therefore favour wetlands over the dry scrub habitats. Eurasian reed warblers occupied both habitats and carried substantially larger body reserves in the dry scrub habitat compared to those occupying the wetland habitat, consistent with greater strategic buffering of energy reserves against starvation in the dry scrub habitat where starvation risk is greatest.

Starvation risk is a function of both the overall (mean) level of food availability, and its variance. Eurasian reed warblers carried higher body reserves in dry scrub than wetland in both years when this comparison was made, including 2013 when variance in prey availability was higher in wetland habitats than dry scrub but the mean abundance remained approximately twice as high in the wetland. This indicates that even the lower end of the range of prey availability in wetlands in 2013 was sufficient to allow birds to maintain low fat reserves without an excessive risk of starvation.

Although muscle was a significant predictor of body mass in this study, the effect-size was low muscle score differed less than fat between habitats. This pattern supports the concept that adaptive changes to mass are likely to involve dynamic fat reserves while loss of functional muscle is more likely to be a consequence of physiological stress [83]. It is also possible that the method of muscle measurement (a three-point scale) provides insufficient resolution to detect a difference between habitats in muscle score. Similarly, the extent to which visible fat scores reflect the relative body-lipid reserves may limit conclusions about adaptive fat deposition [76]. There were no significant differences in structural body size of reed warblers between habitats or years and, since we were unable consistently to determine the age and sex of the warblers in our study, we were not able to test for intra-specific differences (e.g. age and sex variation) in habitat effects as found in several other studies of migrants in their wintering habitats [4], [31], [84], [85]. It is thus possible that the distribution of individuals is a result of age or sex-biased habitat segregation, despite the absence of any significant differences in wing length between habitats in our dataset.

These results suggest that mass is regulated strategically among warblers overwintering in the Sahel, rather than being limited directly by food availability. Experimental manipulations of food availability and predictability have shown that birds can strategically increase mass through fat accumulation to buffer against energetic shortfalls [35], [38], [40], [42]. A similar effect is evident in the context of migratory fuelling, where larger energetic reserves are accumulated in anticipation of low food availability ahead of large ecological barriers [86].

Although higher energy reserves are better for avoiding starvation and for fuelling migration, the extent of muscle and fat accumulation must be balanced against the costs of maintaining higher body mass, such as increased wing-loading costs on flight performance and longer foraging periods, both of which increase susceptibility to predation [87]. The differences in predator density across wetland and dry scrub habitats in this study are unknown but it is likely that birds inhabiting the dry scrub habitats would have higher exposure to predators as they must more frequently cross open spaces between habitat patches than in the wetlands. Subalpine warblers and chiffchaffs exhibited similar but less pronounced habitat differences in body reserves than observed in the Eurasian reed warbler (though this habitat difference was non-significant in chiffchaffs). Although subalpine warblers and chiffchaffs were regularly captured in wetlands, it is notable that these captures occurred mainly in the scrub-Phragmites interface, rather than in the pure stands of Phragmites where Eurasian and African reed warblers and sedge warblers were all regularly captured. This is consistent with the less pronounced habitat difference in body reserves in subalpine warblers and chiffchaffs being due to their more limited exploitation of the invertebrate-rich Phragmites stands compared to Eurasian reed warblers and the two wetland-specialist species.

The significantly lower mass for chiffchaffs in 2012 compared with 2013 and 2014 may be due to direct food limitation following the low rainfall (as measured by the Sahel precipitation index) recorded in the preceding wet season, of 2011. This response is contrary to the strategic buffering strategy expected for birds in food-reduced habitats, suggesting direct food limitation in drought conditions. This association with Sahel precipitation index was not observed in the other species in this study, although studies of survival of trans-Saharan migrants indicate that drought can directly limit survival. For example, Sahel rainfall has been found to be significantly associated with survival in the sedge warbler [16], [88], reed warbler [89], sand martin [90], [17] and common nightingale Luscinia megarhynchos [91], indicating that drought can impair survival, presumably via the mechanism of food limitation.

We found that habitat specialist species maintained lower energetic reserves than generalist species in their preferred habitats. This was the case for African reed warbler and western olivaceous warbler, but not for sedge warbler. The analysis suggested that sedge warblers were more likely to carry higher fat levels than any other species in the wetland habitat. A possible testable explanation for this is that sedge warblers rely on a narrower range of prey than Eurasian and African reed warblers, with a foraging strategy adapted to exploiting slow moving or sedentary prey [92]. Although such adaptation is beneficial during seasonal super-abundances of such prey in the pre-migratory grounds [93][95], it may restrict their wintering habitat to wetlands and limit their foraging time to the cooler parts of the day when active aerial invertebrates are easier to catch. Such limitations may require the accumulation of fat reserves despite the occupation of invertebrate-rich wetland habitat. Another important consideration in the interpretation of these results is that the data were collected during the middle of the dry season, when wetlands have undergone contraction since the time of migrant arrival at the end of the wet season [5], [96], [97]. It is therefore possible that the availability of prey suitable for sedge warblers is particularly low at this time of year. Scarcity of certain prey taxa may explain why sedge warbler populations are particularly vulnerable in the winters following low wet-season rainfall [16], [88]. In contrast to sedge warblers, Eurasian reed warblers do not suffer such dramatic population crashes in response to Sahelian drought events [49], [98], [99]. Our results suggest that this may be due to their ability to exploit invertebrate resources in dry habitats in addition to the high quality but intensely competitive and drought-prone wetland habitats.

Although our results suggest an explanation of the presence of different species in a particular habitat or multiple habitats by their ability to exploit resources and survive in these habitats, we are not yet able to explain the process by which community assembly occurs. It is conceivable that habitat choice may be determined by order of arrival, with earliest arriving migrants (e.g. sedge warblers) occupying the highest quality habitats, while later arriving migrants (e.g. Eurasian reed warbler) must select the secondary habitats. However, this may be unlikely considering the findings of studies of interspecific territorialism and competition for space between reed and sedge warblers on the breeding grounds [100], [101] in which early arriving sedge warblers are often displaced by the later arriving reed warblers.

In conclusion, the observed patterns of habitat occupation and body condition described in our study are consistent with the hypothesis that when individuals are in food-poor habitats, they utilise strategic buffering to avoid starvation. This mechanism may only be effective above a critical food availability threshold. Below this threshold, bird mass is likely to be directly limited by food availability [30], [84], [85]. Although there was little evidence for such direct limitation of body reserves during our study period (other than among chiffchaffs), such patterns of nutritional stress are observed in wintering sand martins Riparia riparia [17] and barn swallows Hirundo rustica [102] in the Sahel during drought years and are implicated as a major cause of population-level declines through increased mortality and carry over effects on productivity [103][105]. Similarly, impacts of physiological stress have been shown in other contexts, such as in American redstarts Setophaga ruticilla [31], [85] northern waterthrush Seiurus noveboracensus [60] and ovenbirds S. aurocapillus [30], [106] occupying low quality non-breeding habitats in the Neotropics. Such studies suggest that if foraging conditions in the Sahel become worse as a result of climate-driven and anthropogenic habitat degradation, migrant warblers may become increasingly constrained in their ability to use strategic buffering to reduce the threat of starvation and survive the winter.

Supporting Information

Table S1.

Number of birds captured in each habitat in January 2012 (wet only), 2013 and 2014.



Many thanks to Jez Smith, Khady Gueye, Katharina Kühnert, Adam Seward, Paul Robinson, Laura Robertson and Rich Howells, Station Biologique du Djoudj and Hotel du Djoudj for help with trip planning and fieldwork.

Author Contributions

Conceived and designed the experiments: JV RT IV RF THJ RP. Performed the experiments: JV RT RF. Analyzed the data: JV RT IV. Contributed reagents/materials/analysis tools: JV RT RF. Wrote the paper: JV RT IV THJ RF RP.


  1. 1. Hutto R (1985) Habitat selection by nonbreeding, migratory land birds. Pages 455–476 In M. L. Cody, ed. Habitat selection in birds. Academic Press, Orlando, Fl
  2. 2. Leisler B (1992) Habitat selection and coexistence of migrants and Afrotropical residents. Ibis 134:77–82.
  3. 3. Jones P, Vickery J, Holt S (1996) A preliminary assessment of some factors influencing the density and distribution of Palearctic passerine migrants wintering in the Sahel zone of West Africa. Bird Study 43:37–41.
  4. 4. Fernández G, Lank DB (2006) Sex, age, and body size distributions of Western Sandpipers during the nonbreeding season with respect to local habitat. Condor 108:547–557.
  5. 5. Zwarts BL, Bijlsma RG, Kamp J Van Der, Wymenga E (2009) Living on the edge: wetlands and birds in a changing Sahel. KNNV Publishing. Utrecht.
  6. 6. Hinde RA (1956) The biological significance of the territories of birds. Ibis 98:340–369.
  7. 7. Rogers CM (2005) Food limitation among wintering birds: a view from adaptive body mass models. In: Greenberg, R. and Marra, P. P. (eds). Birds of two worlds: the ecology and evolution of migration. The Johns Hopkins University Press, Baltimore, pp.106–113.
  8. 8. Newton I (1998) Population limitation in birds. London: Academic press.
  9. 9. Dhondt AA (2011) Interspecific Competition in Birds. Oxford Avian Biology. Oxford University Press, Oxford, UK.
  10. 10. Greenberg R (1986) Competition in migrant birds in the nonbreeding season. Current Ornithology 3:281–307.
  11. 11. Salewski V, Bairlein F, Leisler B, Radolfzell V, Obstberg A (2003) Niche partitioning of two Palearctic passerine migrants with Afrotropical residents in their West African winter quarters. Behavioural Ecology 14:493–502.
  12. 12. Salewski V, Jones P (2006) Palearctic passerines in Afrotropical environments: a review. Journal of Ornithology 147:192–201.
  13. 13. Lima SL, Dill LM (1990) Behavioral decisions made under the risk of predation: a review and prospectus. Canadian Journal of Zoology 68:619–640.
  14. 14. Sutherland WJ (1996) From individual behaviour to population ecology. Oxford University Press.
  15. 15. Den Held JJ (1981) Population changes in the Purple Heron in relation to drought in the wintering area. Ardea 69:185–191.
  16. 16. Peach W, Baillie S, Underhill L (1991) Survival of british sedge warblers Acrocephalus schoenobaenus in relation to west African rainfall. Ibis 133:300–305.
  17. 17. Norman D, Peach WJ (2013) Density-dependent survival and recruitment in a long-distance Palaearctic migrant, the Sand Martin Riparia riparia. Ibis 15:284–296.
  18. 18. Hutto R (1980) Winter habitat distribution of migratory land birds in western Mexico, with special reference to small, foliage-gleaning insectivores. Pages 181–203 in A. Keast and E.S. Morton, editors. Migrant birds in the Neotropics: ecology, behaviour, distribution and conservation.
  19. 19. Sillett TS, Holmes RT, Sherry TW (2000) Impacts of a global climate cycle on population dynamics of a migratory songbird. Science 288:2040–2042.
  20. 20. Johnson MD, Sherry TW, Holmes RT, Marra PP (2006) Measuring habitat quality for a migratory songbird wintering in natural and agricultural areas. Conservation Biology 20:1433–1444.
  21. 21. Levey D, Stiles F (1992) Evolutionary precursors of long-distance migration: Resource availability and movement patterns in Neotropical landbirds. American Naturalist 140:447–476.
  22. 22. Herremans M (1997) European Golden Oriole Oriolus oriolus. Atlas of South African birds 2:96–97.
  23. 23. Pérez-tris J, Tellería JL (2002) Migratory and sedentary blackcaps in sympatric non-breeding grounds: implications for the evolution of avian migration. Journal of Animal Ecology 71:211–224.
  24. 24. Salewski V, Jones P, Vickery J (2002) partitioning between immigrant Palearctic willow warblers Phylloscopus trochilus and resident Afrotropical warblers in three woodland habitats in Zimbabwe. Avian Science 2:1–9.
  25. 25. Baumann S (2001) Observations on the coexistence of Palearctic and African Orioles Oriolus spec. in Zimbabwe. Vogelwelt 122:67–79.
  26. 26. Cox GW (1968) The role of competition in the evolution of migration. Evolution 22:180–192.
  27. 27. MacArthur RH (1972) Coexistence of species. In J. H. Bethnke (ed.), Biological problems, directions towards their solution. Oxford University Press, New York. Pp. 253–259
  28. 28. Karr JR (1976) Seasonality, resource availability, and community diversity in tropical bird communities. American Naturalist 110:973–994.
  29. 29. Cox GW (1985) The evolution of avian migration systems between temperate and tropical regions of the New World. American Naturalist 126:451–474.
  30. 30. Strong AM, Sherry TW (2000) Habitat-specific effects of food abundance on the condition of ovenbirds wintering in Jamaica. Journal of Animal Ecology 69:883–895.
  31. 31. Marra PP, Holmes RT (2001) Consequences of dominance-mediated habitat segregation in American Redstarts during the nonbreeding season. Auk 118:92–104.
  32. 32. Greenberg R (1995) Insectivorous migratory birds in tropical ecosystems: the breeding currency hypothesis. Journal of Avian Biology 26:260–264.
  33. 33. Johnson MD, Sherry TW, Strong AM, Medori A (2005) Migrants in Neotropical bird communities: an assessment of the breeding currency hypothesis. Journal of Animal Ecology 74:333–341.
  34. 34. Verboven N, Visser ME (1998) Seasonal variation in local recruitment of great tits: the importance of being early. Oikos 81:511–524.
  35. 35. McNamara JM, Houston AI (1990) The value of fat reserves and the tradeoff between starvation and predation. Acta Biotheoretica 38:37–61.
  36. 36. Thomas RJ (2000) Strategic diel regulation of body mass in European robins. Animal Behaviour 59:787–791.
  37. 37. Blem CR (1975) Geographic variation in wing-loading of the house sparrow. Wilson Bulletin 87:543–549.
  38. 38. Witter MS, Cuthill IC (1993) The ecological costs of avian fat storage. Philosophical Transactions of the Royal Society of London Bioliogical Sciences 340:73–92.
  39. 39. Gosler AG, Greenwood JJD, Perrins C (2002) Predation risk and the cost of being fat. Nature 377:621–623.
  40. 40. Ekman JB, Hake MK (1990) Monitoring starvation risk: adjustments of body reserves in greenfinches (Carduelis chloris L.) during periods of unpredictable foraging success. Behavioral Ecology 1:62–67.
  41. 41. Witter MS, Swaddle JP, Cuthill IC (1995) Periodic food availability and strategic regulation of body mass in the european starling, Sturnus vulgaris. Functional Ecology 9:568–574.
  42. 42. Witter MS, Swaddle JP (1997) Mass regulation in juvenile starlings: response to change in food availability depends on initial body mass. Functional Ecology 11:11–15.
  43. 43. Bednekoff PA, Krebs JR (1995) Great tit fat reserves: effects of changing and unpredictable feeding day length. Functional Ecology 9:457–462.
  44. 44. Jenni L, Jenni-Eiermann S (1987) Body-weight and energy reserves of bramblings in winter. Ardea 75:271–284.
  45. 45. Gosler AG (1996) Environmental and social determinants of winter fat storage in the great tit Parus major. Journal of Animal Ecology 65:1–17.
  46. 46. Ratikainen II, Wright J (2013) Adaptive management of body mass by Siberian jays. Animal Behaviour 85:427–434.
  47. 47. Witter MS, Swaddle JP (1995) Dominance, competition and energetic reserves in the European starling (Sturnus vulgaris). Behavioural Ecology 6:343–348.
  48. 48. Grimmett R (1987) A Review of the Problems Affecting Palearctic Migratory Birds in Africa: Cambridge: Int. Council Bird Preservation Migratory Birds Prog. 25 pp.
  49. 49. Leisler B, Schulze-Hagen K, Quinn D, Hillcoat B, Martin MA (2011) The Reed Warblers: Diversity in a Uniform Bird Family. KNNV Publishing.
  50. 50. Flade M, Diop I, Haase M, Nevé A, Oppel S, et al. (2011) Distribution, ecology and threat status of the Aquatic Warblers Acrocephalus paludicola wintering in West Africa. Journal of Ornithology 152:129–140.
  51. 51. Moreau RE (1972) The Palaearctic-African bird migration systems. New York, London: Academic.
  52. 52. Aidley DJ, Wilkinson R (1987) The annual cycle of six Acrocephalus warblers in a Nigerian reed-bed. Bird Study 34:226–234.
  53. 53. Sauvage A, Rumsey S, Rodwell S (1998) Recurrence of Palaearctic birds in the lower Senegal river valley. Malimbus 20:33–53.
  54. 54. Borrow N, Demey R (2001) Birds of western Africa. Christopher Helm London.
  55. 55. Rabøl J (1987) Coexistence and competition between overwintering Willow Warblers Phylloscopus trochilus and local warblers at Lake Naivasha, Kenya. Ornis Scandinavica 18:101–121.
  56. 56. Salewski V, Bairlein F, Leisler B (2002) Different wintering strategies of two Palearctic migrants in West Africa - a consequence of foraging strategies? Ibis 144:85–93.
  57. 57. Aidley DJ, Wilkinson R (1987) Moult of some Palaearctic warblers in northern Nigeria. Bird Study 34:219–225.
  58. 58. Loske K (1990) Spring weights and fat deposition of palaearctic passerine migrants in Senegal. Ringing & Migration 11:23–30.
  59. 59. Sherry TW, Holmes RT (1996) Winter Habitat Quality, Population Limitation, and Conservation of Neotropical-Nearctic Migrant Birds. Ecology 77:36–48.
  60. 60. Smith JAM, Reitsma LR, Marra PP (2010) Moisture as a determinant of habitat quality for a nonbreeding neotropical migratory songbird. Ecology 91:2874–2882.
  61. 61. Svensson L (1992) Identification Guide to European Passerines. Naturhistoriska Riksmuseet, Stockholm.
  62. 62. Kaiser J (1993) A New Multi-Category Classification of Subcutaneous fat deposits of Songbirds. Journal of Field Ornithology 64:246–255.
  63. 63. Bairlein F (1995) Manual of field methods. European-African songbird migration network. Institut für Vogelforschung, Wilhelmshaven.
  64. 64. Black W, Krafsur E (1985) Use of sticky traps to investigate seasonal trends in the spatial distribution of house flies and stable flies (Diptera: Muscidae). Journal of Medical Entomology 22(5):550–557.
  65. 65. Hogsette J, Jacobs R, Miller R (1993) The sticky card: device for studying the distribution of adult house fly (Diptera: Muscidae) populations in closed poultry houses. Journal of Economic Entomology 86:450–454.
  66. 66. Goulson D, Derwent LC, Hanley ME, Dunn DW, Abolins SR (2005) Predicting calyptrate fly populations from the weather, and probable consequences of climate change. Journal of Applied Ecology 42:795–804.
  67. 67. Jedlicka JA, Greenberg R, Perfecto I, Philpott SM, Dietsch TV (2006) Seasonal shift in the foraging niche of a tropical avian resident: resource competition at work? Journal of Tropical Ecology 22:385–395.
  68. 68. R Core Team (2014) R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing. Vienna, Austria.
  69. 69. Bates D, Maechler M, Bolker B, Walker S, (2013) lme4: Linear mixed-effects models using Eigen and S4. R package version 1.1–6.
  70. 70. Kuhn M, Weston S, Wing J, Forester J (2013) contrast: A collection of contrast methods. R package version 0.19.
  71. 71. Wood S (2011) Fast stable restricted maximum likelihood and marginal likelihood estimation of semiparametric generalized linear models. Journal of the Royal Statistical Society (B) 73(1):3–36.
  72. 72. Zuur AF, Ieno EN, Elphick CS (2010) A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution 1:3–14.
  73. 73. Johnson DH, Krapu GL, Reinecke KJ, Jorde DG (1985) An Evaluation of Condition Indices for Birds. Journal of Wildlife Management 49:569.
  74. 74. Brown ME (1996) Assessing body condition in birds. Current Ornithology 13:67–135.
  75. 75. Green AJ, May N (2001) Mass/Length Residuals: Measures of Body Condition or Generators of Spurious Results? Ecological Society of America 82:1473–1483.
  76. 76. Bergstrom B.J. Sherry TW (2008) Estimating lipid and lean body mass in small passerine birds using TOBEC, external morphology and subcutaneous fat-scoring. Journal of Avian Biology 39:507–513.
  77. 77. Lemeshow S, Hosmer D (2000) Applied Logistic Regression. Wiley Series in Probability and Statistics. Wiley-Interscience; 2 Sub edition.
  78. 78. Venables WN, Ripley BD (2002) Modern applied statistics with S. Springer. 3rd edition.
  79. 79. Hawke C, José P (1996) Reedbed management for commercial and wildlife interests. Royal Society for the Protection of Birds, Sandy, Bedfordshire, UK.
  80. 80. Tscharntke T (1999) Insects on common reed (Phragmites australis): community structure and the impact of herbivory on shoot growth. Aquatic Botany 64:399–410.
  81. 81. Bedford AP, Powell I (2005) Long-Term Changes in the Invertebrates Associated with the Litter of Phragmites australis in a Managed Reedbed. Hydrobiologia 549:267–285.
  82. 82. Schmidt MH, Lefebvre G, Poulin B, Tscharntke T (2005) Reed cutting affects arthropod communities, potentially reducing food for passerine birds. Biological Conservation 121:157–166.
  83. 83. Neto JM, Gosler AG (2010) Variation in body condition of breeding Savi'sWarblers Locustella luscinioides: the reproductive stress and flight adaptation hypothesis revisited. Journal of Ornithology 151:201–210.
  84. 84. Marra P (2000) The role of behavioral dominance in structuring patterns of habitat occupancy in a migrant bird during the nonbreeding season. Behavioural Ecology 11:299–308.
  85. 85. 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.
  86. 86. Jenni-Eiermann S, Almasi B, Maggini I, Salewski V, Bruderer B, et al. (2011) Numbers, foraging and refuelling of passerine migrants at a stopover site in the western Sahara: diverse strategies to cross a desert. Journal of Ornithology 152:113–128.
  87. 87. Blem CR (1990) Avian energy storage. Current Ornithology 7:59–113.
  88. 88. Foppen R, Ter Braak CJF, Verboom J, Reijnen R (1999) Dutch sedge warblers Acrocephalus schoenobaenus and West-African rainfall: empirical data and simulation modelling show low population resilience in fragmented marshlands. Ardea 87:113–127.
  89. 89. Salewski V, Hochachka WM, Fiedler W (2013) Multiple Weather Factors Affect Apparent Survival of European Passerine Birds. PLoS ONE 8:e59110.
  90. 90. Szep T (1995) Survival rates of Hungarian sand martins and their relationship with Sahel rainfall. Journal of Applied Statistics 22:891–904.
  91. 91. Boano G, Bonardi A, Silvano F (2004) Nightingale Luscinia megarhynchos survival rates in relation to Sahel rainfall. Avocetta 28:77–85.
  92. 92. Lima SL (1986) Predation risk and unpredictable feeding conditions: determinants of body mass in birds. Ecology 67:377–385.
  93. 93. Bibby C, Green R (1981) Autumn migration strategies of Reed and Sedge Warblers. Ornis Scandinavica 12:1–12.
  94. 94. Green RE, Davies NB (1972) Feeding ecology of reed and sedge warblers. Wicken Fen Group Report 4:8–14.
  95. 95. Bayly NJ (2007) Extreme fattening by sedge warblers, Acrocephalus schoenobaenus, is not triggered by food availability alone. Animal Behaviour 74:471–479.
  96. 96. Janzen DH (1973) Sweep samples of tropical foliage insects: effects of seasons, vegetation types, elevation, time of day, and insularity. Ecology 54:687–708.
  97. 97. Wolda H (1978) Seasonal fluctuations in rainfall, food and abundance of tropical insects [Panama Canal Zone]. Journal of Animal Ecology 47:369–381.
  98. 98. Redfern C, Alker P (2002) Eurasian Reed Warbler. In The Migration Atlas: Movements of the Birds of Britain and Ireland (eds Wernham CV, Toms MP, Marchant JH, Clark JA, Siriwardena GM & Baillie SR) pp 548–551. T. & A.D. Poyser, London.
  99. 99. Thaxter CB, Redfern CPF, Bevan RM (2006) Survival rates of adult Reed Warblers Acrocephalus scirpaceus at a northern and southern site in England. Ringing & Migration 23:65–79.
  100. 100. Catchpole C (1973) Conditions of co-existence in sympatric breeding populations of Acrocephalus warblers. Journal of Animal Ecology 42:623–635.
  101. 101. Svensson S (1978) Territorial exclusion of Acrocephalus schoenobaenus by A. scirpaceus in reedbeds. Oikos 30:467–474.
  102. 102. Van den Brink B, Bijlsma RG, van der Have TM (2000) European Swallows Hirundo rustica in Botswana during three non-breeding seasons: the effects of rainfall on moult. Ostrich 71:198–204.
  103. 103. Møller AP (1989) Population dynamics of a declining swallow Hirundo rustica population. Journal of Animal Ecology 58:1051–1063.
  104. 104. Baillie SR, Peach WJ (1992) Population limitation in palaearctic-African migrant passerines. Ibis 134:120–132.
  105. 105. Bryant DM, Jones G (1995) Morphological changes in a population of sand martins Riparia riparia associated with fluctuations in population size. Bird Study 42:57–65.
  106. 106. Brown DR, Sherry TW (2006) Food supply controls the body condition of a migrant bird wintering in the tropics. Oecologia 149:22–32.