Flight Morphology and Foraging Behavior of Male and Female Cory's Shearwaters

Abstract.

In some seabirds sexually dimorphic in size, males and females segregate at sea or diverge in other aspects of foraging behavior. We examined factors influencing foraging strategies of Cory's Shearwater and compared the sexes' flight morphology and activity patterns. Trip duration, incubation-shift length, total mass gain, and rate of mass gain at sea of birds from our two study colonies differed. The colonies are situated in regions of contrasting oceanographic conditions: Selvagem Grande, a remote subtropical oceanic island, and Berlengas, an island on the Portuguese continental shelf. Although the wing loading and wing span of males and females breeding at Selvagem Grande differed significantly, sex did not consistently influence activity patterns of Cory's Shearwaters foraging at sea during the incubation period. Moreover, both sexes breeding at Selvagem Grande foraged in areas with similar sea-surface temperatures. Our study suggests that sexual differences in size and shape may be poor predictors of differentiation in the ways male and female pelagic seabirds use the marine environment.

Resumen.

En algunas aves marinas que presentan dimorfismo sexual con relación al tamaño, los machos y las hembras se segregan en el mar o divergen en otros aspectos de su comportamiento de forrajeo. Examinamos factores que influencian las estrategias de forrajeo de Calonectris diomedea y comparamos la morfología del vuelo y los patrones de actividad entre los sexos. La duración de los viajes, el largo del periodo de intercambio durante la incubación y el aumento de la masa corporal en el mar difirieron entre las aves de nuestras dos colonias de estudio. Las dos colonias se encuentran en regiones con condiciones oceanográficas contrastantes: Selvagem Grande, una isla oceánica remota, y Berlengas, una isla en la plataforma continental de Portugal. A pesar de que la carga y la envergadura alar de machos y hembras reproductivas de Selvagem Grande se diferenciaron significativamente, el sexo no influenció de forma consistente en los patrones de actividad de los individuos de C. diomedea que forrajean en el mar durante el periodo de incubación. Además, individuos de ambos sexos que se reproducen en Selvagem Grande forrajearon en áreas con temperaturas de superficie marina similares. Nuestro estudio sugiere que las diferencias sexuales de tamaño y forma no podrían predecir la diferenciación de las formas de uso del ambiente marino por parte de los machos y las hembras de aves marinas pelágicas.

Introduction

Sexual differences in foraging behavior, resource use, and habitat use have been described in many birds, including passerines (Gosler 1987, Aho et al. 1997), birds of prey (Marquiss and Newton 1982), and seabirds (Phillips et al. 2004, Weimerskirch 1995). Sexual segregation is usually considered to result either from social dominance and competitive exclusion (generally of females by larger males) or from niche specialization resulting from differences in morphology or reproductive role (Petit et al. 1990, Marra 2000, Phillips et al. 2004). Segregation can occur at various spatial scales, ranging from broad geographical differences in distribution to local differences in habitat and microhabitat use (Catry et al. 2005). In seabirds, the larger sex may dominate intraspecific interactions close to the colony, forcing the smaller and less competitive sex to travel farther to obtain food (González-Solís et al. 2000).

Sexual differences in the foraging behavior of seabirds appear in many cases to be related to sexual size dimorphism, found in species such as the Giant Petrels (Macronectes spp.; Gonzáles-Solís et al. 2000), Wandering Albatross (Diomedea exulans; Weimerskirch et al. 1993, Weimerskirch 1995), Black-browed Albatross (Thalassarche melanophrys), and Gray-headed Albatross (T. chrysostoma; Phillips et al. 2004). This divergence in behavior may be related to differences in wing loading; female albatrosses are smaller, their foraging and flight efficiency potentially higher in the lighter winds typical of lower latitudes (Shaffer et al. 2001, Phillips et al. 2004). Sexual differences in foraging behavior, however, have also been described for sexually monomorphic species, including the Manx Shearwater (Puffinus puffinus; Gray and Hamer 2001) and Northern Gannet (Morus bassanus; Lewis et al. 2002), although Garthe et al. (2007) found no foraging differences between the sexes of the latter species,. Therefore, differences between the sexes in foraging behavior may arise independently of the degree of dimorphism. Other studies suggest that sexual size dimorphism may have evolved from the specialization of each sex in reproduction. In birds of prey, for instance, larger females play a more active role in nest defense (Andersson and Norberg 1981) and/or have higher capacity to produce and incubate larger eggs (Reynolds 1972, Snyder and Wiley 1976).

In Cory's Shearwater (Calonectris diomedea) sexual dimorphism in bill size and shape, wing length, and overall body size is considerable (Granadeiro 1993, Thibault et al. 1997, Navarro et al. 2009). Hence sexual segregation in foraging habitat and activity patterns might be expected. But Navarro et al. (2009) found that foraging movements, feeding areas, and signatures of carbon and nitrogen isotopes were similar in both sexes and did not correspond to variation in bill dimensions. That study, however, took place in a single breeding season and colony. Given the strong annual and seasonal variability in oceanographic conditions that breeding Cory's Shearwaters experience (Granadeiro et al. 1998, Ramos et al. 2003), the possibility that males and females might segregate or differ in foraging strategies under other circumstances deserves further scrutiny. Therefore, we examined annual variation in several aspects of foraging behavior of Cory's Shearwaters from a coastal and an offshore colony.

In this study we investigated whether sexual dimorphism in Cory's Shearwater results in sexual differences in foraging behavior. The variables we quantified were (1) trip duration, an indirect measure of the distance from the colony to the foraging area, (2) number of landings, as a measure of foraging effort, (3) mass gained at sea, as a measure of the profitability of foraging trips, and (4) sea-surface temperature, as an indirect measure of marine productivity. We begin by quantifying differences between the sexes in wing-loading and confirming that there is a marked dimorphism in this relevant trait. We then examine the factors that could influence the duration of foraging trips by male and female Cory's Shearwaters breeding in two contrasting colonies of the North Atlantic: Berlengas, on the continental shelf off Portugal, and Selvagem Grande, a remote oceanic subtropical island between Madeira and the Canary Islands. Oceanographic conditions surrounding these two colonies are very different (Huntsman and Barber 1977, Ambar and Dias 2008): food should be relatively abundant around the first colony and relatively scarce around the second (Granadeiro et al. 1998, Alonso et al. 2009). These colonies also differ considerably in size, with a few hundred pairs at Berlenga (M. Lecoq, pers. comm.) and about 30000 at Selvagem Grande (Granadeiro et al. 2006), so there is potential for much greater intraspecific competition at Selvagem Grande. Overall, this framework is suitable for comparing different selective pressures that may promote sexual segregation in feeding ecology. We examine and compare the factors that influence the duration of foraging trips by male and female Cory's Shearwaters breeding in Selvagem Grande and Berlengas. We also compare the activity patterns and flight morphology of males and females at Selvagem Grande, where constraints on reproduction and foraging should be greater.

Methods

We carried out this study on Selvagem Grande, Savage Islands (30° 09′ N, 15° 52′ W), and Berlengas, North Atlantic Ocean (39° 23′ N, 9° 36′ W), between 20 June and 15 July 2006 and 2007. The areas around these two colonies differ markedly in oceanographic conditions (Fig. 1): during the study period mean sea-surface temperature (SST) within a 200-km radius of Berlengas and Selvagem Grande was 18.9° C and 21.2° C, respectively (extracted from http://poet.jpl.nasa.gov/). There is also a strong thermal gradient between Selvagem Grande and the Canaries upwelling along the African coast (Fig. 1), which is about 400 km from Selvagem Grande.

We selected three study areas on Selvagem Grande and one on Berlengas. At both Selvagem Grande and Berlengas, we measured both members of the pairs studied and painted each with a distinguishing color mark on the chest and neck. We measured wing length, culmen length, maximum bill height (at base of bill), and bill depth at the gonydeal angle to the nearest 0.1 mm with calipers. We determined the sex of 31% of the birds by their vocalizations (Thibault et al. 1997). For birds that did not vocalize during handling, we determined sex as per Granadeiro (1993), with bill measurements by a discriminant function (DS = 0.087 culmen length + 1.206 bill height + 0.598 gonys — 40.063): individuals with a score >0.1 were classified as males, those with a score <-0.1 as females). During the 2 years of the study, we sexed 141 individuals by the two methods, and the discriminant function misclassified only two individuals (1.4%). By using these procedures and cross-checking with partners, we accurately sexed more than 99% of the birds handled.

We examined differences in morphology of the sexes at Selvagem Grande only. We used a tape to measure maximum body circumference and shoulder width. For wing area, we extended the right wing onto a sheet of paper attached to a flat wooden board, photographed it with a digital camera, then calculated the area with GIS software. We estimated total wing area by doubling the area of the digitized wing and adding the area between the wings, the latter equal to the product of root chord (= wing width at the junction with the body, measured in the field) and shoulder width (Pennycuick 1989). We calculated wing span by doubling the distance from the tip of the outermost primary to the junction with the body and adding the shoulder width. Wing loading describes the weight (force) per unit area of wing, and we calculated it as body mass × gravitational constant/wing area (N m^-2). We calculated aspect ratio, a dimensionless variable that describes the shape of the wing, as wing span²/wing area. Higher values of aspect ratio indicate a comparatively narrow wing, lower values a comparatively broad wing (Pennycuick 1989).

To measure changes in body mass and determine the duration of foraging trips and incubation shifts, we inspected each nest daily during the morning and recorded the presence of color-marked individuals. We noted birds that were incubating for the first time (i.e., the first day of the incubation shift after arrival from the sea) and weighed these birds on the following day to allow time for meal digestion. Thereafter we weighed the bird every 1 or 2 days until the sexes exchanged incubating. After we weighed the birds, we returned them immediately to their nests to resume incubation; we found no evidence that periodic weighing influenced the behavior of adults (see also Granadeiro et al. 1999). For birds not weighed on the day preceding a changeover, we estimated mass at departure by using the last mass recorded and the mean sex-specific rate of mass loss. To retain statistical independence, we included in further analyses one foraging trip per individual.

On Selvagem Grande, we fitted birds with a 3.5-g logger (model MK7, manufactured by the British Antarctic Survey) that recorded light (for geolocation), salt-water immersion (activity), and temperature. We fitted 28 males and 30 females in 2006 and 30 males and 19 females in 2007. We attached the devices to the metal leg band with a plastic cable-tie and deployed them for a single foraging trip. The loggers represented less than 0.5% of the adult body mass. Studies, including one on Cory's Shearwater, indicate that devices of this relative size have no deleterious effects (Phillips et al. 2003, Igual et al. 2005).

The loggers measured the intensity of visible light every 60 sec, logging the highest value in every 10-min period. They measured water temperature after immersion for a continuous period of 20 min and a new value only after exposure to dry air and immersion for a further 20 min. They measured the time of all changes of state (from immersed to dry and from dry to immersed) that lasted >6 sec, by testing every 3 sec. From these data, we derived detailed activity patterns of Cory's Shearwater and compared them by sex. We calculated the total number of immersions (landings), the number of landings during darkness and daylight each day, the proportion of total time spent in flight overall, as well as the proportion of time spent in flight during darkness and daylight. We considered daylight the period between 05:00 and 22:00 local time. To avoid pseudoreplication and the potential influence of bright sunlight on the temperature reading, we compared the sea-surface temperature in areas used by males and females by using the value recorded nearest midnight for each individual each day.

We used t-tests to compare sexual differences in measurements, including those related to flight. By using a factorial ANOVA, we assessed the influence of (a) sex and location (Selvagem Grande vs. Berlengas), and their interaction, on foraging-trip characteristics and (b) sex and year (2006 vs. 2007), and their interaction, on activity patterns of Cory's Shearwaters from Selvagem Grande. Whenever necessary, we transformed variables before analysis (logarithmic for continuous variables, square root for counts, and arcsine for proportions) to attain normality and homogeneity of variances.

Results

Sexual Size Dimorphism in Body Measurements and Flight Morphology at Selvagem Grande

The sexes of Cory's Shearwater differed strongly in all standard morphological characters; values for males were always higher than those for females (Table 1). Wing loading and wing span, but not wing area or aspect ratio, were also significantly greater in males than in females (Table 1).

Comparison by Sex and Breeding Location of Characteristics of Foraging Trips

Foraging-trip duration, mass gained at sea, and duration of incubation shifts differed by colony but not by sex (Table 2). At the subtropical oceanic island of Selvagem Grande, Cory's Shearwaters foraged for longer periods during incubation, gained more mass at sea, and incubated their eggs longer. On average, males and females spent similar periods foraging at sea, but for both sexes birds from Berlengas spent only about half as much time on foraging trips as birds from Selvagem Grande (Table 2).

For all birds except males at Selvagem Grande, the total mass gained during a foraging trip showed a positive and highly significant relationship with trip duration (Fig. 2). The length of the previous incubation shift, however, did not influence the duration of the subsequent trip at sea, again except for males at Selvagem Grande (Fig. 3).

Comparison by Sex of Activity Patterns and Sea-Surface Temperature at Selvagem Grande

The total number of landings and the number of landings in darkness and in daylight in 2006 differed from those in 2007. There was a significant interaction between sex and year on the total number of landings, number of landings at night, and percentage of time spent in flight at night (Table 3). This interaction arose because the number of landings by females in total and during the night was greater than the corresponding number of landings by males in 2006 and the opposite was true in 2007. Both males and females foraged in waters with similar mean SST in both years (Table 3). Trip duration for both sexes was significantly negatively correlated with the minimum SST value (r = -0.53, n = 33, P = 0.001). This correlation was negative because only in longer trips can birds reach the cold waters of the Canaries upwelling near the African coast (Fig. 1), as suggested by the logger measurements of SST. This area is used also by birds from the nearby Canary Islands (Navarro and González-Solís, 2009, Navarro et al. 2009). Flying at a mean ground speed of 36 km hr^-1 (Rosén and Hedenström 2001), a Cory's Shearwater would take about 11 hr to reach the nearest points along the African coast.

Discussion

Our study confirmed considerable sexual size dimorphism in Cory's Shearwater, with males 4% to 14% larger than females in linear dimensions and body mass. The differences were greatest for culmen length and bill depth at gonydeal angle (also see Granadeiro 1993, Navarro et al. 2009). Sexual dimorphism in wing loading and wing span paralleled that of linear measurements, but only for wing loading was the difference between males and females particularly marked (9%). In other seabirds, sexual dimorphism in body mass and wing morphology is thought to play a functional role in flight performance and to explain differences in distribution of males and females at sea at least for part of the year (Shaffer et al. 2001, Phillips et al 2004). Wing loading has been used to predict interspecific variation in foraging radius, as the energetic cost of flight is likely to be positively correlated with wing loading (Hertel and Ballance 1999), and to predict the use of foraging strategies that require low flight speeds, as wing loading is negatively correlated with glide/ flight speed (Pennycuick 1989). As male Cory's Shearwaters have greater wing loading and a longer wing span, we would expect them to range farther to forage, land fewer times (to reduce energetic costs), exploit more windy areas, and fly at greater speeds than females. Our data on foraging trips, activity patterns, and SST, however, did not show differences between males and females in feeding ecology.

Our results also show that regional oceanographic characteristics have a much stronger influence than sex on the duration of incubation shifts and foraging trips and on total mass gained at sea. The long duration of foraging trips from Selvagem Grande and low SSTs the birds encounter during these trips suggest that adults probably exploit the cold waters of the Canaries upwelling near the African coast (Davenport et al. 2002). Given that foraging trips from Berlenga were shorter, we suggest that birds breeding at this island do not need to make long foraging trips to recover body reserves used in the previous incubation shift. This mass gain can presumably be achieved by exploiting productive upwelling over the Portuguese continental shelf close to the colony. We also suggest that SST can be useful as an indicator of foraging areas, especially for Selvagem Grande, where discrete foraging areas are identifiable by distinctive SST signatures. The strong thermal gradient between Selvagem Grande and the Canaries upwelling along the coast of Africa allows an easy distinction between birds feeding over the continental shelf of Africa and birds feeding in pelagic waters. Overall, SST reflects food resources for most pelagic birds (Abraham and Sydeman 2004, Peck et al. 2004).

At each island, the duration of males' and females' incubation shifts and foraging trips and rates of mass gain at sea did not differ significantly. This finding suggests that for both sexes the costs of incubation and foraging are similar, as is the subsequent rate of restoration of body mass, implying a comparable level of foraging efficiency and suggesting that both sexes can regain their body reserves readily. Nevertheless, at Selvagem Grande there appear to be slight differences in the foraging strategies of males and females, because the duration of foraging trips by males was not related to mass gained at sea but was influenced by the duration of the previous incubation period. It may be that females encounter prey more predictably and that males need to regulate their foraging trips in relation to mass lost during incubation. Such subtle differences between the sexes may arise only at colonies where surrounding oceanographic conditions are less favorable, as we detected no differences at Berlengas. However, as we found no significant differences in activity patterns between sexes of birds from Selvagem Grande, nor in the SST of the waters they used, it appears that males and females largely invest the same effort in feeding and probably exploit the same oceanic areas and food resources.

Although seabirds with pronounced sexual size dimorphism often show sex-specific differences in foraging behavior (González-Solís et al. 2000, Phillips et al. 2004, Weimerskirch et al. 2006), this correlation is far from universal (Phalan et al. 2007, Navarro et al. 2009). Moreover, sexual differences in distribution and/or behavior have been noted in several monomorphic species (Lewis et al. 2002, Gray and Hamer 2001). Our study indicates that Cory's Shearwater is a further example of a species in which pronounced sexual size dimorphism has no apparent connection to foraging behavior. This is the case at least during incubation but cannot be excluded for other stages of the life cycle. The winds around Selvagem Grande are much lighter than those typical of polar and temperate oceans, where sexual segregation in foraging areas appears to be more common (González-Solís et al. 2000, Phillips et al. 2004). Cory's Shearwater breeds in subtropical waters of the North Atlantic and migrates to tropical and subtropical waters in the South Atlantic (González-Solís et al. 2007), where wind conditions are different and food resources presumably more dispersed. Under those conditions, lower wing loading may allow more efficient exploitation of local wind conditions (Spear and Ainley 1998), so in those areas males and females may differ in foraging strategy. In general, individual variation in measurements of tropical petrels is greater, suggesting that they are less specialized than polar species, presumably because they experience a wider range of foraging conditions (Spear and Ainley 1998).

Our study adds to previous evidence (Navarro et al. 2009) that sexual differences in size and shape may be poor predictors of differentiation in the way male and female seabirds exploit the marine environment. The disproportionately larger bill of male Cory's Shearwaters (Navarro et al. 2009) implies that there was direct selection pressure on bill size in the past (Shine 1989). The sexual differences in bill morphology (Navarro et al. 2009) and wing loading (this study) may be explained by behavioral differences related to mate acquisition and nest defense. Sexual differences in foraging behavior in species dimorphic in size, however, may be related to the partitioning of food resources on the basis of prey size and do not necessarily involve different search methods or foraging strategies (Shine 1989). In several diving species of seabirds, males forage at a trophic level higher than do females, which may be related to differences in physiological performance, as males dive deeper than females (Bearhop et al. 2006). Cory's Shearwater is a plunge diver adapted to surface feeding but may dive if necessary (Martin 1986). Therefore, it is possible that subtler differences in the diving behavior of males and females may exist but are difficult to detect because they occur infrequently. Therefore, although our results suggest that ecological factors play a relatively minor role in the maintenance of sexual size dimorphism in Cory's Shearwater, the ultimate cause of this dimorphism remains obscure.

Acknowledgments

Parque Natural da Madeira, and particularly Dília Menezes, provided permission for us to work on Selvagem Grande and, together with the wardens at the Nature Reserve, gave important logistical support. The Instituto de Conservação da Natureza e Biodiversidade provided permission for our work on Berlengas. Hany Alonso, Rafael Matias, Maria Dias, Miguel Lecoq, Rui Rebelo, Filipe Moniz, Ana Leal, Ricardo Martins, and many others helped with field work. This study is an output from a project on the ecology and senescence of Cory's Shearwaters (PDCT/MAR/58778/2004) supported by Fundação para a Ciência e a Tecnologia (FCT—Portugal). P. Catry benefited from postdoctoral fellowships from FCT (BPD/11631/02 and SFRH/BPD/30031/2006), and further support was received through Programa Plurianual (UI&D 331/94). A previous draft benefited from the comments of Vitor Paiva and two anonymous referees.

Literature Cited