Partial migration in tropical birds: the frontier of movement ecology


Correspondence author. E-mail:


  • image

A. E. Jahn, D. J. Levey, J. A. Hostetler & A. M. Mamani (2010) Determinants of partial bird migration in the Amazon Basin. Journal of Animal Ecology, 79, 983–992.

Partial migration, in which only some individuals of a species migrate, might be central to the evolution of migratory behaviour and is likely to represent an evolutionary transition between sedentariness and complete migration. In one of the few detailed, individual-based migration studies of tropical birds, Jahn et al. study the partial migration system of a South American bird species for the first time. Food limitation forces the large adult males and small, young females to migrate, contrary to the expectations of the body size and dominance hypotheses. This study confirms the importance of food variability as the primary driver of migratory behaviour. There is urgent need for similar studies on the movement ecology of understudied tropical bird species, whose diversity of migratory behaviour can shed light on the evolution of bird migration.

Migration is one of the most spectacular, albeit increasingly threatened, phenomena of the living world (Wilcove 2008). Birds provide impressive examples, with some covering more than 80 000 km annually (Gill et al. 2009; Egevang et al. 2010). Eighteen per cent of the world’s ∼10 000 bird species undertake regular long-distance migrations (Fig. 1; Sekercioglu 2007) and every year 40 billion birds are estimated to migrate (Wikelski et al. 2007). However, research has mostly focused on discovering the migratory cues and mapping migrations, and the ecological reasons for bird migration are still vague (Boyle & Conway 2007; Boyle 2008a). Understanding the drivers of animal migration is particularly challenging due to the difficulty and expense of keeping track of birds covering long distances. Less spectacular but equally important types of migratory behaviour, such as short-distance, altitudinal and partial migration are particularly understudied. This is especially true for birds <100 g that cannot be tracked by satellite even though they make up 65% of the world’s bird species (Dunning 2007; Wikelski et al. 2007). Detailed studies of tropical songbird movement can mean thousands of hours of radiotracking on foot (Sekercioglu et al. 2007), often in rugged, wet and challenging terrain, so even less is known about these birds.

Figure 1.

 Extinction risk as a function of long-distance movement. Bird species with regular long-distance movements are less extinction-prone (threatened, near threatened or extinct) than sedentary birds (Sekercioglu 2007). The number of species known to undertake that type of movement is in parentheses. Reprinted with permission from Sekercioglu (2007) Current Biology, 17, R283–R286.

Although most bird species do not migrate (Fig. 1), even seemingly ‘sedentary’ tropical birds can cover hundreds of kilometres in search of food (Holbrook, Smith & Hardesty 2002), and hundreds of species are involved in short-distance, altitudinal and partial migrations (Berthold 1999; Sekercioglu 2007). However, there are few detailed, long-term, individual-based movement studies of tropical birds. Given the range of migratory behaviour seen in the tropics (Boyle & Conway 2007), such studies are essential for revealing the ecological and evolutionary mechanisms behind bird migration (Levey & Stiles 1992; Chesser & Levey 1998; Boyle & Conway 2007). Understanding avian mobility is also important for guiding the conservation of tropical bird communities and the restoration of their habitats (Sekercioglu et al. 2007), especially in the face of climate change threatening hundreds of tropical bird species with narrow ranges and limited mobility (Sekercioglu et al. 2008).

Partial migration, in which some individuals migrate and others remain sedentary (Jahn et al. 2010), might be central to the evolution of migratory behaviour and likely represents an evolutionary transition between sedentariness and complete migration (Berthold 1999; Boyle 2008b). Although ‘extremely widespread at higher latitudes’ (Berthold 1999), much less is known about partial migration in the tropics where only a few studies have been done (Boyle 2008b; Boyle, Norris & Guglielmo 2010; Jahn et al. 2010). However, most bird species live in the tropics where the diversity of migratory behaviour exhibited makes migratory research especially pertinent. The South American austral migration system, despite being the third major migration system between temperate and tropical regions and the only one in the Southern Hemisphere (Chesser 1994), has been particularly neglected.

It is therefore welcome that Jahn et al. (2010) in this issue break the mould and test two important hypotheses of partial migration on a tropical bird species in the Southern Hemisphere, where there is ‘a unique opportunity to better understand the evolutionary ecology of bird migration’ (Dingle 2008). Jahn et al. (2010) conducted a detailed individual-based study of tropical kingbirds (Tyrannus melancholicus) to test the dominance and body size hypotheses. According to the dominance hypothesis, subordinate individuals who are unable to compete effectively with dominant individuals for food are more likely to choose migration over competing for food during times of food scarcity (Gauthreaux 1978; Ketterson & Nolan 1979). The body size hypothesis predicts that larger individuals (usually males) are better able to withstand cold weather and food shortage, which means smaller individuals are more likely to migrate (Belthoff & Gauthreaux 1991). Jahn et al. (2010) were unable to test the arrival time hypothesis, which states that individuals establishing breeding territories (again, usually males) will migrate less than non-territorial individuals because a shorter migration distance will enable a faster return to breeding areas and first access to best territories (Ketterson & Nolan 1976). As all three hypotheses were developed in the north temperate latitudes and mostly tested there, it is especially important to know if they can explain partial migration in other parts of the world (Boyle 2008b).

Tropical kingbirds are common, widespread and mostly insectivorous Neotropical songbirds found in a range of open habitats. They belong to the family Tyrannidae, which is the main bird family in the austral migration system (Chesser 1994). Tropical kingbirds are completely migratory south of 18°S, but at the grassland study site in the Amazon basin of Bolivia (14°49′S), they are partially migratory. Jahn and colleagues combined mark–recapture, radiotracking, and maximum-likelihood modelling in a rigorous framework to tease apart the probabilities of detection, migration, survival and permanent dispersal. Individually colour-banded kingbirds were surveyed every 2 weeks. By radiotracking individuals, they showed that birds were not moving locally, but migrating instead. There was strong support for partial migration, the probability of which was influenced by age, body mass and the interaction of these with sex in an unexpected pattern. Large adult males and small young females had the highest probability of migrating. Because older kingbirds dominate younger kingbirds, the results did not support the dominance hypothesis for males, underlining the importance of sexual differences. Similarly, the body size hypothesis was not supported.

These findings complement those of Boyle & Conway (2007) who conducted a comparative analysis of 379 species of tyrannid flycatchers to test the evolutionary precursor hypothesis of migration (Levey & Stiles 1992). After incorporating phylogenetically independent contrasts, Boyle & Conway (2007) showed the importance of habitat and diet interaction and foraging behaviour in shaping migratory behaviour. Their resource variability hypothesis states that species relying on more seasonal resources are more likely to evolve migratory behaviour (Boyle & Conway 2007). Jahn et al. (2010) explain migratory variation among individuals with their food limitation hypothesis. This modifies the resource variability hypothesis by proposing different threshold values for individuals based on their sex and age (Table 1). The variability of the insect resource forces the largest males and the smallest females to migrate in the dry season, when insect prey is reduced and the basal energy requirements of the former and the territorial defence costs of the latter likely push them above the available energy threshold (Table 1).

Table 1.   A conceptual, simplified framework for the food limitation hypothesis (Jahn et al. 2010) based on hypothetical units of energy required for basal metabolic rate (BMR) and territorial defence. This table is not based on real data. For the same body mass, BMR and other energy needs are assumed to be equal for each sex. Males are, on average, larger than females. Males are assumed to be dominant over females, with the energy cost of territorial defence higher for females with equal body mass. Cost of territorial defence is negatively correlated with body mass, increasing nonlinearly for females and linearly for males. The maximum energy available in the dry season is 70 units per day per bird, with birds requiring more than this threshold value (grey) being forced to migrate Thumbnail image of

Food limitation is thought to be the main driver of migration in most birds (Alerstam, Hedenstrom & Akesson 2003; Boyle 2008a). However, additional environmental conditions such as reduced nest predation, inclement weather or parasites can play critical roles (Alerstam, Hedenstrom & Akesson 2003; Boyle 2010; Boyle, Norris & Guglielmo 2010). The complex interaction of these factors with habitat, sex, body mass and other traits (Boyle & Conway 2007; Jahn et al. 2010) make simple explanations difficult. A variety of research approaches (e.g. bird banding, nest monitoring, hormonal measurements, radiotracking, population models, dietary analysis, parasitology) is necessary, especially outside the well-studied north temperate region. In the Southern Hemisphere, the migration strategies employed are more varied (Jahn et al. 2010) and opportunities abound to study the evolutionary ecology of bird migration (Chesser et al. 1998; Dingle 2008). Therefore, it is remarkable that Jahn et al. (2010) are the first to study the partial migration system of any South American bird species, even though this is ornithologically the richest continent and two-thirds of Neotropical austral migrants are potentially undertaking partial migration (Jahn et al. 2006).

Although habitat, diet and their interaction can be important in determining which birds migrate and which do not (Boyle & Conway 2007), the variability of access to food resources has emerged as the primary driver of migratory behaviour (Levey & Stiles 1992; Chesser 1998; Boyle & Conway 2007; Boyle, Norris & Guglielmo 2010). In fact, in South America, open country insectivores show the highest migratory tendency (Boyle & Conway 2007), whereas insectivorous species of ‘buffered’ tropical forest interior environments are not known to migrate (Chesser et al. 1998). With this study, Jahn et al. (2010) provide important empirical support for the importance of resource variability in austral bird migration by explicitly linking migratory behaviour in a Neotropical insectivorous bird to seasonal changes in insect numbers. Tropical forests, commonly thought to be ‘stable and constant’, can also experience substantial resource variability that has consequences for avian reproductive seasonality (Wikelski, Hau & Wingfield 2000). Many tropical forest insectivorous bird species faced with resource fluctuations and increased habitat fragmentation (Sekercioglu et al. 2002) might be undertaking partial migrations yet to be discovered.

In the only other Neotropical (Cost Rica) partial migration system studied to date, Boyle, Norris & Guglielmo (2010) also showed the importance of the interaction of environmental variation with intrinsic variables like age and sex in determining migration likelihood. However, unlike Jahn and colleagues’ system, Boyle et al. found that inclement weather (storms and heavy rains) reducing foraging opportunities, not food shortage per se, was the main driver of migration in white-ruffed manakins (Corapipo altera). In this system, adult males were also more likely to migrate, but this was because they were smaller and did not have the fasting advantages of the larger females. In both systems, migration likelihood is influenced by the effect of weather on food obtained, and different ages and sexes vary in their responses to food shortage based on their body mass. While the Brazilian dry season reduces insect food abundance and forces some kingbirds to migrate, in the wet mountains of Costa Rica storms and heavy rains prevent manakins from foraging, forcing them to choose between fasting or migrating.

Because increased migratory behaviour in birds reduces extinction likelihood (Sekercioglu 2007), advances in avian movement ecology will also advance bird conservation biology, especially in the face of climate change. While climate change is expected to affect sedentary birds more negatively than migratory species (Sekercioglu et al. 2008), migratory birds will face their own challenges, such as longer migrations and simultaneous impacts on wintering, breeding and stopover sites (Huntley et al. 2006).

Unfortunately, detailed movement studies involving long-term bird banding, radiotracking, and satellite telemetry are lacking for most bird species, especially in the tropics. With accurate satellite transmitters dropping in price, there is great promise of testing various migration theories (Alerstam, Hedenstrom & Akesson 2003), but the high cost (£1000s/animal) of satellite transmitters and data collection fees make it prohibitively expensive for most scientists, especially in tropical, developing countries. Satellite transmitters have been getting smaller and are now down to 5 g, but this is estimated to be their minimum size limit (Wikelski et al. 2007). Because the transmitter should ideally not exceed 5% of the body mass of the study species (Wikelski et al. 2007), the migrations of birds <100 g cannot be followed, excluding at least 65% of bird species (Dunning 2007). Wikelski et al. (2007) proposed the establishment of a global small-animal tracking system based on the launch of a dedicated satellite ( The cost of £33–66 million is more than justified by the revolutionary implications for research on animal ecology, evolution, conservation, agriculture and public health.

Nevertheless, Jahn and colleagues’ elegant and painstaking study shows that we need not wait for a dedicated satellite to conduct ground-breaking research on bird migration, especially in the tropics. Thousands of tropical bird species have never been studied, basic truths about bird migration are waiting to be discovered, and movement ecology is a scientific frontier that is wide open.