Linking Winter and Summer Events in a Migratory Bird by Using Stable-Carbon Isotopes

See allHide authors and affiliations

Science  04 Dec 1998:
Vol. 282, Issue 5395, pp. 1884-1886
DOI: 10.1126/science.282.5395.1884


For migratory birds, early arrival and physical condition on the breeding grounds are important determinants of reproductive success and fitness. Differences in arrival times often exceed a month, and later arriving individuals are often in poorer condition. Habitat-specific isotopic signatures indicate that the quality of winter habitats occupied by American redstarts (Setophaga ruticilla) determines their physical condition and spring departure dates, which in turn result in variable arrival schedules and condition on temperate breeding grounds. These findings link events in tropical winter grounds with those in temperate breeding areas for a migratory songbird and provide evidence that winter habitats may be limiting.

Natural selection acts on individuals throughout the annual cycle. For migratory animals, understanding these selection processes has been limited by our inability to follow individuals year-round, yet events during each phase of the annual cycle are likely to influence those in subsequent phases. Many long-distance migratory birds, such as the American redstart, spend 3 to 5 months on their temperate breeding grounds, 1 to 2 months on autumn migration, 6 to 7 months on tropical wintering areas, and another month on spring migration (1).

For many migratory species, males arrive at breeding habitats before females (2), and breeding success and physical condition decline with arrival date (3, 4). Early arrival appears to be advantageous because it gives access to the best breeding sites and mates, as well as additional time to replace lost clutches (5). Declining reproductive success for late arriving birds is also attributed to poor physical condition of these individuals (4). Factors that determine arrival time and physical condition of birds in breeding areas are poorly understood.

To test the hypothesis that winter events influence arrival dynamics on the breeding grounds, we studied American redstarts in two habitats in southwestern Jamaica: a black mangrove (Avicennia germinans) forest in which males predominated (65% male and 35% female) and a drier, second-growth scrub habitat in which females were more abundant (30% male and 70% female). Sexual habitat segregation is common in redstarts during the winter period (6) and is produced by the dominance behavior of older males forcing most females and young males into habitats of poorer quality (7–9). In autumn 1995 and 1996, redstarts were captured with mist nets, measured, bled for hormone and stable-isotope assays, color-banded, and released. In late March and early April, those individuals that remained on territory over the winter were recaptured for remeasurement. We found that individuals wintering in the forest habitat, regardless of sex, maintained or gained body mass, whereas individuals in scrub habitat lost up to 11% of their body mass [0.06 ± 0.05 g (mean ± SE) compared with −0.24 ± 0.07 g; two-way analysis of variance: sex F = 0.09, P = 0.77; habitat F = 15.1,P = 0.0004; sex by habitat F = 2.56,P = 0.12]. Individuals in scrub habitats showed other signs of deteriorated physical condition, including elevated plasma corticosterone concentration (9).

The poor physical condition of redstarts in scrub habitat did not lead to lower over-winter survival (8), but it did result in a delay in departure schedules (10). Both males and females departed significantly later from scrub habitat in both years (Fig. 1). Furthermore, departure time was inversely correlated with change in body mass (Fig. 2), implying that redstarts in better physical condition were able to leave sooner.

Figure 1

(top). Spring departure schedules of color-banded American redstarts from their wintering territories in Jamaica, West Indies. In 1995 and 1996 redstarts departed significantly earlier from black mangrove forest compared with second-growth scrub [Kaplan-Meier survivorship analysis (23); 1995, P = 0.0006; 1996, P = 0.0002]. In 1995, there were no significant differences in departure schedules between sexes within a habitat (forest, P = 0.923; scrub, P = 0.34), but in 1996, males from forest departed before females (P = 0.004) and before both males and females from scrub (P = 0.001). Females from forest (1996) departed significantly earlier than both females (P = 0.04) and males (P = 0.008) from scrub.

Figure 2

(bottom). The relation between winter mass change (in grams) of color-banded American redstarts captured in October and then recaptured in April and the number of days until they departed on spring migration (Pearson's r = 0.60,P < 0.004).

To determine if habitat segregation during winter influences the arrival schedules of birds onto breeding sites, we used δ13C signatures in redstart tissue as an indicator of habitat occupancy. Stable-carbon isotopes are suitable for such applications because the rate of diffusion of 13C from the atmosphere into plant tissues differs between plants with C3, C4, and Crassulacean acid metabolism (CAM) photosynthetic pathways (C3 and C4plants produce a three-carbon or four-carbon acid, respectively, as the first product of photosynthesis). C3 plants are typically associated with cooler, moister habitats and have more depleted δ13C values, whereas C4 and CAM plants are often associated with hotter and drier environments and have more enriched δ13C values (11). Similar13C enrichment patterns in C3 plant tissues may also result from differences in plant water use efficiency (11). Mangrove (12) and tropical lowland forests are both C3 habitats (13, 14). More xeric tropical habitats containing grasses (for example, our scrub habitat in Jamaica) typically have more C4 plants (11, 14). These relations were substantiated with isotopic assessments of insects collected from the two Jamaican habitats (15).

Previous studies have shown that animal tissues reflect the isotopic composition of their supporting food web (14, 16). Thus, we expected American redstarts as obligate insectivores (1) to incorporate a habitat-specific δ13C signature into their tissues from the phytophagous insects they consumed. An analysis of δ13C values in tissue (17) collected from redstarts in Jamaica and in a second geographically distinct locality in Honduras revealed that individuals in wet, forested habitats had significantly depleted δ13C values relative to individuals in drier, scrub habitats, regardless of sex or locality (Fig. 3).

Figure 3

Stable-carbon isotope values (δ13C) (mean ± SE) taken from blood samples of American redstarts in three habitat types in Jamaica, West Indies, and in two habitat types in Honduras, Central America. These habitats contained different sex ratios of wintering American redstarts (wet lowland forest, 95% male; black mangrove forest, 65% male; second-growth scrub, 30% male). In both localities, isotope values differed significantly across habitat types (F = 77.34,P < 0.0001) but did not differ between geographic localities (F = 1.91, P = 0.20). No effects of sex were found within a habitat when comparing isotope values of redstarts in mangrove forest versus scrub (F = 0.16,P < 0.69).

To determine if habitats occupied in winter influence arrival dates in North America, we collected tissue from American redstarts as they arrived in spring 1997 and settled on breeding areas at the Hubbard Brook Experimental Forest, central New Hampshire, U.S.A. We found that later arriving redstarts had tissue δ13C values more enriched relative to earlier arrivals (Fig. 4) (18). This suggests that males arriving on breeding grounds early were those originating from wetter tropical habitats, whereas those arriving later were from drier tropical habitats (Fig. 4). The period of arrival that we sampled may have been too short to adequately test this for females. Using body mass corrected for skeletal size as an index to physical condition, we also found that the physical condition of redstarts arriving on the breeding grounds declined from early to later arrival (r = −0.52, P = 0.016, n = 21) (19).

Figure 4

Stable-carbon isotope values (δ13C) taken from muscle tissue of American redstarts as they arrived in spring at the Hubbard Brook Experimental Forest, West Thornton, New Hampshire, U.S.A. (sexes combined: Spearman ρ = 0.47, P = 0.01; males only: Spearman ρ = 0.80,P = 0.002; females only: Spearman ρ = 0.11,P = 0.72).

The specific winter ground origin of redstarts breeding at Hubbard Brook is not known (1). However, because habitat segregation is pervasive throughout most of the winter range of redstarts (9), we believe the carbon isotope signatures from Jamaican and Honduran habitats to be representative of the major habitat types occupied by redstarts throughout their winter distribution. The application of stable-carbon isotope methodology has allowed us to link two separate periods in the annual cycle of this migratory species.

These results implicate events during the preceding winter, namely intraspecific competition for optimal winter habitat mediated through behavioral dominance (8, 20), as an important factor determining arrival times and condition upon arrival of redstarts in their north temperate breeding areas. This finding is important because arrival time at the breeding ground is a major determinant of fitness in migratory birds (3, 21). Furthermore, our evidence that later arriving birds wintered in drier habitats and that physical condition declined with arrival date suggests that optimal winter habitats for redstarts may be saturated and therefore limiting. If optimal winter habitats (more mesic sites) were always available, then all redstarts should have occupied them, and we would have found no relation between δ13C values and the physical condition of redstarts over the arrival period. This conclusion, that winter habitats are limiting, has important conservation implications for the long-term stability of migratory bird populations, many of which are declining and of conservation concern (22).

  • * To whom correspondence should be addressed. E-mail: pmarra{at}

  • Present address: Smithsonian Migratory Bird Center, National Zoological Park, Washington, DC 20008, USA.


View Abstract

Navigate This Article