Males migrate farther than females in a differential migrant: an examination of the fasting endurance hypothesis

Patterns of migration including connectivity between breeding and non-breeding populations and intraspecific variation in the distance travelled are important to study because they can affect individual fitness and population dynamics. Using data from 182 band recoveries across North America and 17 light-level geolocators, we examined the migration patterns of the northern flicker (Colaptes auratus), a migratory woodpecker. This species is unusual among birds because males invest more in parental care than females. Breeding latitude was positively correlated to migration distance because populations in the north appeared to travel farther distances than southern populations to find wintering locations with little snow cover. Connectivity was strong for populations west and east of the Continental Divide. Contrary to the three main hypotheses for intraspecific variation in migration distance, females wintered, on average, farther north than males, although there was overlap throughout their non-breeding range. This pattern contradicts those of other species found to date and is most consistent with the fasting endurance hypothesis if investment in parental care depletes the energy reserves of male flickers more than females. We thus propose a new factor, parental effort, which may influence optimal wintering areas and migration strategies within birds, and encourage future experimental studies to test the relationship between parental care roles and migration strategies of the sexes.


Summary
Patterns of migration including connectivity between breeding and non-breeding populations and intraspecific variation in the distance travelled are important to study because they can affect individual fitness and population dynamics. Using data from 182 band recoveries across North America and 17 light-level geolocators, we examined the migration patterns of the northern flicker (Colaptes auratus), a migratory woodpecker. This species is unusual among birds because males invest more in parental care than females. Breeding latitude was positively correlated to migration distance because populations in the north appeared to travel farther distances than southern populations to find wintering locations with little snow cover. Connectivity was strong for populations west and east of the Continental Divide. Contrary to the three main hypotheses for intraspecific variation in migration distance, females wintered, on average, farther north than males, although there was overlap throughout their nonbreeding range. This pattern contradicts those of other species found to date and is most consistent with the fasting endurance hypothesis if investment in parental care depletes the energy reserves of male flickers more than females. We thus propose a new factor, parental effort, which may influence optimal wintering areas and migration strategies within birds, and encourage future experimental studies to test the relationship between parental care roles and migration strategies of the sexes.

Introduction
Birds exhibit diverse strategies for migrating between their breeding and wintering sites, such as variation in routes, timing and distances travelled. Patterns of migration can be described at different taxonomic and spatial scales and likely energy) into parental care they would be more likely to migrate farther than females to seek out the more benign southerly habitats. By contrast, the body size hypothesis predicts that females will migrate farther because males are slightly larger than females [16] and can withstand harsher conditions. The arrival time hypothesis also predicts males should winter farther north; although flickers do not defend feeding territories, males defend cavities [22] and are more philopatric to nest sites [23] than females. However, there is no evidence of different arrival timing between males and females on breeding sites [16,24]. Finally, although one sex does not appear to be dominant in flickers (E. A. Gow and K. L. Wiebe 2014, unpublished data), in several species of woodpeckers males are dominant over females [25,26], predicting that males may force females to move farther south (dominance hypothesis).
Currently, little is known about migration patterns of northern flickers. Northern populations are certainly migratory, whereas southern-breeding flickers appear to be sedentary [16] but this has not been thoroughly examined. Flickers are not currently a species of conservation concern, but the Breeding Bird Survey (BBS) shows that populations are declining at an average rate of −1.36% per year since 1966, across North America, with higher rates of decline in eastern compared with western North America [27]. Hence, in addition to examining sex-related differences in migration distance, we were interested in describing continent-wide movements of flickers, an ecologically valuable keystone species for cavity nesting and roosting animals [28], in order to better understand environmental factors which may be affecting the population dynamics of flickers across their geographical range.

Material and methods
We determined the autumn migration movements of northern flickers by using pooled data from both banding records and geolocators. The Canadian Bird Banding Office provided 1086 band recoveries for flickers throughout North America from 1915 to 2013 and we used 175 of these records from birds marked during breeding (15 May-31 July) and re-sighted/recovered during winter  to determine the movement of individual flickers from breeding to wintering sites. We eliminated those birds that were re-trapped within the same season (i.e. breeding or winter) because such movements are not migratory. We probably also excluded juvenile dispersal by excluding recoveries from the postbreeding autumn migratory period (1 August-14 October). There were 41 records from birds banded as part of a long-term study (1997)(1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009)(2010)(2011)(2012)(2013) at Riske Creek, British Columbia, Canada (52.03 • N, 122.515 • W). In addition, seven re-sighting records from Riske Creek that were not reported to the banding office were included. One bird that was banded at Riske Creek and recovered in Newfoundland after flying across the continent for 4835 km was excluded from analysis because it was not a typical migration. In summary, we used a combination of continent-wide band recovery data (n = 142), band recoveries/re-sightings from a single population at Riske Creek (n = 41) and geolocator locations from 17 flickers tagged at Riske Creek to map the wintering areas, migratory distances and general wintering patterns of flickers.
In 2010-2012, we attached geolocators to 51 male and 25 female flickers during breeding between 15 May and 15 July, 2010-2012 at Riske Creek, and thus had additional migration and wintering details from 17 individuals (16 male and one female). Return rates of birds with geolocators were similar to natural return rates for our study site (see details in [29]). Flickers were caught at the nest cavity by flushing them into a net placed over the hole and then banded with a United States Geological Survey aluminium band and a unique colour band combination, sexed and aged as yearling (SY) versus older (ASY) based on moult [30]. The geolocator models (MK12 in 2010, MK20AS in 2011 and MK10 in 2012 weighing 0.9, 1.0 and 1.6 g, respectively; British Antarctic Survey, Cambridge, UK) were less than 1% of the average body mass (157 g ± 0.2 s.e., n = 2161). We attached the devices using a leg loop backpack harness [31] with 45 lb test braided nylon cord tied and glued.
Prior to deploying the geolocators, we placed them outdoors with a similar view of the horizon for at least 7 days. The pre-deployment data were used to calculate a light threshold level and sun elevation angle for sunrise and sunset transitions for the stationary geolocators. The corresponding sun altitude for the calibration period was −4.3 • ± 0.36 s.d., n = 18 and the light threshold level was 16. Because flickers roost in cavities throughout the year [29], adjustments were made to the light-level transitions in the morning and evening to more accurately reflect the true sunrise and sunset times (E. A. Gow 2014, unpublished data). This method produced a total location error of mean 94 ± 167 km (s.d.) or mean 60 ± 107 km (s.d.) in latitude and mean 71 ± 142 km (s.d.) in longitude, as determined from birds at a known location on their breeding territory. To obtain wintering locations from birds with geolocators, we first averaged locations over a period of 5 days and if the bird then appeared to be stationary (i.e. differences  . Flickers tended to winter in regions with little snowfall and females wintered at significantly higher latitudes on average than males. Squares, male; filled circles, female; crosses, unknown. less than 2 s.d. of the error across several 5 day intervals), we determined location by averaging all days in the stationary period. Using a pooled dataset (unless otherwise indicated) of both the banding data from across North America (n = 41 from Riske Creek, n = 142 from elsewhere) and geolocator data from Riske Creek (n = 17), we examined whether wintering latitude was influenced by the fixed effects of sex, age and breeding latitude, and breeding longitude using a generalized linear model (GLM) with a Gaussian distribution and identity link function. We initially included the interaction effects between breeding longitude and latitude with sex and age, but because none was significant we removed the interactions from the model to increase power. Additionally, because there were no significant effects of longitude on wintering latitude, we removed longitude from our model. All statistical analyses were conducted in R 3.03 [32], error is expressed as ± s.d., unless indicated otherwise. All statistical tests are two-tailed and significance is p ≤ 0.05.

Results
There were band recoveries from flickers that bred as far north as 53.08

Discussion
Banding and geolocator recoveries revealed several patterns in the migratory movements of northern flickers. At the continental scale, populations breeding in the north migrated farther than those breeding in the south. Additionally, populations west of the Rocky Mountains showed little overlap in wintering areas with populations breeding east of the Rockies. Most interesting, female flickers wintered farther north than males, which contradicts the pattern found in all other species of birds studied to date and is most consistent with the fasting endurance hypothesis. At the species level, northern flickers are partial migrants, because some individuals in the south are non-migratory, but northern populations would be considered short-distance migrants as all individuals leave the breeding area. The location of wintering areas on the continent suggests the main attribute of suitable habitat is snow-free ground favourable for foraging (figure 1). Flickers feed almost exclusively on ground-dwelling ants during breeding [33] and such ground-dwelling insect prey are also the mainstay of the diet in the winter although more fruit may be incorporated [34,35]. Other climatic variables, such as rain, and cold temperatures are less likely than snow cover to restrict access to ground-dwelling arthropods. Our data show that migration in flickers is not leapfrog or chain migration and indicate that more northern populations must travel longer distances to reach locations that remain snow-free during winter creating an overlap in the wintering sites of both southern-and northern-breeding individuals. Among species with a large geographical breeding range, it is common for migration propensity to vary with latitude as we found in flickers [6]. Newton [1] suggests that at least 69 species of birds show similar latitude-based migration tendencies with northern populations tending to be more migratory and southern ones sedentary (e.g. red-tailed hawks (Buteo jamaicensis [36]) and European robins (E. rubecula [7]). More intensive studies of focal populations of flickers are needed, especially of those breeding in the south, to determine to what extent partial migration and residency exists within certain populations.
Migratory connectivity varied along the continuum between strong and weak. Flicker populations east and west of the Continental Divide did not overlap, which suggests the divide can act as a physical barrier to dispersal [37]. Flockhart & Wiebe [24] hypothesized that such a migratory divide keeps the red-and-yellow-shafted subspecies of flicker from introgressing. Studies of other North America species such as Swainson's thrushes (Catharus ustulatus [37]) and yellow warblers (Dendroica petecha [38]) have also found that populations do not cross the Continental Divide during migration which leads to segregation of western and eastern populations and thus migratory connectivity on the scale of the species entire geographical range. For flickers east of the Rocky Mountains, the central population and eastern population overlapped to a small extent on the wintering grounds pointing to moderate connectivity between the two populations. However, at a finer scale within a single breeding population of flickers in the west (i.e. Riske Creek), individuals wintered over a considerable latitudinal range suggesting weaker connectivity at the scale of a breeding population. We found variation in migratory distances within the population at Riske Creek and among flickers at the continental scale, which suggests that decisions about the distance to travel are based on different costs and benefits to individuals. The variation was not based on age class but was related to sex, with males travelling farther south than females on average, although with considerable overlap in the wintering ranges of the sexes. Sex differences in migration distances are well documented in other bird species but all reported cases suggesting females migrate farther than males [7,12,15]. Support for the arrival time hypothesis in relation to sex has been found in European blackbirds [15] and in relation to age class in lesser black-backed gulls (Larus fuscus [39]). The dominance hypothesis has been supported for such animals as blue tits (Cyanistes caeruleus [40]), red-spotted newts (Notophthalmus viridescens [41]), red deer (Cervus elaphus [42]) and American dippers (Cinclus mexicanus [43]), and the body size hypothesis is supported in fishes (reviewed in [44]), manakins [45] and dark-eyed juncos (Junco hyemalis [12]). However, the pattern in flickers contradicts expectations from these three main hypotheses for differential migration. Instead, the more southern wintering locations of male flickers are most consistent with the fasting endurance hypothesis, assuming that their higher investment in care is physiologically more costly than that experienced by females. The recoveries of geolocators were male-biased and were associated with longer documented movements, so recovery methodology may have exaggerated the sex bias. However, methodology cannot entirely explain the sex bias in migration distance because when we removed the 17 geolocator birds, females still wintered farther north than males.
Reproductive differences may contribute to migratory tendency in a polygynous lekking species with female-only parental care (white-ruffed manakins, Corapipo altera; an altitudinal migrant) because males tended to be more migratory than females [45]. However, sex roles in reproduction were not examined in this species, instead it was suggested that migration differences could be attributed to the 14% larger body size in females [45]. In comparison, flicker males migrated farther than females, but body size differences are less than 3% [16], suggesting another factor drives the greater flight distances of male flickers. Here, we stress for the first time to our knowledge, that sex-biased investment in parental care may influence the optimal migration strategies of the sexes.
Other explanations for the sex differences in migration distance have been proposed but probably do not apply to flickers. For example, a difference in diet and habitat preference between the sexes such as is the case for hooded warblers (Wilsonia citrina [46]) may lead to a selection of different wintering habitats. However, there was no difference in the diet of male and female flickers at least during the breeding season [33]. Neither do the sexes have different habitat preferences when foraging [47]. If the fasting endurance hypothesis explains migratory distance in flickers, any additional energetic costs of flying further south among males must be outweighed by the increased access to food at the more benign southerly locations. As there is currently no data on energy expenditure during migration for flickers, this hypothesis should be tested by quantifying the physiological costs of flying different distances.
In conclusion, we documented strong migratory connectivity in flickers at the continental scale, with eastern and western populations divided. This division has probably had evolutionary significance by reducing hybridization between the eastern and western colour forms of flickers. It may also provide an explanation for the different rates of decline in the east and west (see BBS trend data [27]). We are the first to examine migration patterns in a species with partially reversed sex roles in parental care and showed that, in contrast to other studies, females wintered farther north than males. Hence, we propose that parental effort may be another factor that influences adaptive migratory strategies of individuals by affecting the physiological stress experienced by males versus females and encourage future studies that quantify energetic costs of parental care to more directly test this hypothesis.