Fecundity-Survival Trade-Offs and Parental Risk-Taking in Birds

See allHide authors and affiliations

Science  20 Apr 2001:
Vol. 292, Issue 5516, pp. 494-497
DOI: 10.1126/science.1059379


Life history theory predicts that parents should value their own survival over that of their offspring in species with a higher probability of adult survival and fewer offspring. We report that Southern Hemisphere birds have higher adult survival and smaller clutch sizes than Northern Hemisphere birds. We subsequently manipulated predation risk to adults versus offspring in 10 species that were paired between North and South America on the basis of phylogeny and ecology. As predicted, southern parents responded more strongly to reduce mortality risk to themselves even at a cost to their offspring, whereas northern parents responded more strongly to reduce risk to their offspring even at greater risk to themselves.

Should parents place themselves, or their offspring, at greater risk of mortality when threatened with predation? Theoretical models of life history evolution predict that the resolution to this dilemma will vary among species depending on offspring number and the probability of survival for the parents (1–5). Parents should tolerate greater risk to themselves, but not their young, in species with many offspring and reduced adult survival because the fitness value of the current brood is high and the probability of surviving to breed in the future is low (i.e., residual reproductive value is low). In contrast, parents of species with fewer offspring and higher probability of adult survival should tolerate less risk to themselves, even at a cost to their young, because the fitness value of current offspring is lower and prospects for producing young in the future are greater. This choice between responding to the risk of mortality directed toward offspring versus parents in the face of predation is a fundamental challenge faced by all organisms that provide parental care and is highlighted by birds with altricial young that require frequent feeding visits to the nest.

Parental feeding visits to the nest can increase the risk of predation to both parents and young by attracting the attention of predators (5–11). When faced with an immediate risk of predation, parents temporarily decrease visitation rates to reduce risk to themselves or their young (5, 6), but such responses should vary among species according to their life histories. In particular, when predation risk is directed at parents, species with fewer young and higher adult survival should reduce risk to themselves by reducing feeding visits relatively more than species with many young and lower survival, even if offspring suffer from reduced food delivery (1–5) (Fig. 1A). Conversely, when offspring are at risk of predation, species with more young and lower adult survival should reduce feeding visits relatively more than less fecund species to reduce risk toward their greater investment in current reproduction (1–5) (Fig. 1A). These predicted differences in parental care tactics have not been tested, but contrasting life histories of birds from southern and northern latitudes provide a broad-scale opportunity to test them.

Figure 1

Annual adult survival probability and parental risk-taking relations relative to clutch size. (A) Life history theory (1–5) predicts that parents of species with fewer young and higher adult survival probability, as typical of southern regions, should minimize risk of mortality to themselves even at a cost to their young. In contrast, parents of species with more young and lower adult survival probability should minimize risk of mortality to their young, even at a cost to themselves. (B) Clutch size is negatively related to estimated annual adult survival probability from studies (17) that used resighting in addition to capture/recapture techniques (r = −0.63, P < 0.0001, n = 182), and Southern Hemisphere species (open symbols) show higher survival and smaller clutches than north temperate species (closed symbols). This relation remains significant after controlling for phylogeny with independent contrasts (r = −0.55, P < 0.0001,n = 172). (C) Clutch size is also negatively related to estimates of annual adult survival on our Arizona and Argentina study sites (r = −0.92, P < 0.0001) (18), even when controlled for phylogeny (r = −0.94, n = 9 independent contrasts). Species from Argentina (open symbols) have consistently smaller clutch sizes and higher annual adult survival compared with species from Arizona (closed symbols).

Tropical and Southern Hemisphere birds have small clutch sizes that historically were thought to be associated with increased adult survival compared with northern species (7–10), but latitudinal differences in adult survival have been challenged in recent years (12, 13). These challenges, however, came from studies that relied on passive capture and recapture of marked birds in nets, which may underestimate survival probabilities, particularly in southern regions (14, 15). Supplementing recapture with resighting data of breeding territorial adults provides improved estimates of survival (14,15). We obtained data on clutch size and adult survival from published studies that used such resighting methods for 182 passerine species of Northern (Europe and North America) and Southern (Australia, New Zealand, and South Africa) Hemispheres (16). We found a strong negative relation between clutch size and adult survival within and between regions, where southern species have smaller clutch sizes and higher adult survival than northern species (Fig. 1B). To verify that this general relation applied to our study systems, we analyzed data for species that were intensively color-banded and resighted at our sites in Argentina and Arizona (17). Again, we found a strong negative relation, with species from Argentina having higher adult survival associated with smaller clutch sizes compared with birds from Arizona (Fig. 1C). Three pairs of species (Atlapetes-Pipilo,Arremon-Junco, and Basileuterus-Vermivora) are closely matched phylogenetically and ecologically between the two sites (10), and these six species show survival to be higher in Argentina (Fig. 1C) independent of phylogeny (pairedt = 12.4, P < 0.003). At the same time, clutch sizes of Southern Hemisphere species are smaller when closely related species that have the same numbers of broods are compared between regions (Table 1) [also (10, 18)]. Given these clutch size and adult survival differences between northern and southern regions (Table 1, and Fig. 1, B and C) [also (10,15, 18)], we tested the predictions that southern parents should more strongly reduce risk to themselves whereas northern parents should reduce risk to their offspring (Fig. 1A).

Table 1

Species pairs used for experimental tests in North and South America.

View this table:

We manipulated predation risk to parents versus offspring at 61 nests of 10 species that were paired between North and South America (19) on the basis of ecology and phylogeny and that had smaller clutch sizes and higher adult survival in South America (Table 1 and Fig. 1C). We presented a common predator of adults (hawk) and nestlings (jay) and a control that represented no predation threat (tanager) to examine the extent that parents risk themselves versus their offspring when making feeding trips to the nest (20). We found that parents did not change the rate that they visited the nest to feed offspring during control presentations (Fig. 2A). In contrast, parents strongly reduced visitation rate when presented with either a nestling (Fig. 2B) or adult (Fig. 2C) predator. Thus, parents differentiated among presentations and reacted appropriately to reduce potential threats to themselves and their offspring by reducing feeding visits to the nest. Responses to predation threats differed between regions and types of predators. Parents decreased feeding rates more in North than South America when the predation threat (i.e., jay) was directed at the offspring (Fig. 2B). However, this result was reversed when the predation threat (i.e., hawk) was directed at the parent; South American species decreased feeding rate more than North American species (Fig. 2C). In short, North American species reacted more strongly to reduce risk to their offspring, whereas South American species reacted more strongly to reduce risk to themselves. These differences in parental responses between regions are consistent with theoretical predictions that life history differences should predict variation in parental behavior (Fig. 1 and Table 1). However, because such responses could be confounded by environmental effects unique to North and South America, it is important to examine species-specific responses as a function of their life histories.

Figure 2

Log-transformed rates (/90 min ± 1 SE) that parents visit the nest to feed young before presentation (prepresentation) and during presentation (presentation) of control and predator models for 10 species of birds paired between North and South America (Table 1). (A) Feeding rates did not change when a control model (tanager) that presents no predation threat was presented [F(1, 59) = 3.13, P > 0.05]. (B) Feeding rates decreased in North and South America [F(1, 57) = 140.19, P < 0.001] in response to increased nestling predation risk from jay presentations. The interaction between treatment and location was not significant, but feeding rates during presentations were significantly lower in North than South American species (one-tailed phylogenetically pairedt = 2.42, df = 4, P = 0.036). (C) Feeding rates also decreased in response to increased adult predation risk from hawk presentations in North and South America [F(1, 56) = 105.71, P < 0.001]. As above, the interaction term was not significant, but feeding rates during presentations were lower in South than North American species (one-tailed phylogenetically pairedt = −3.41, df = 4, P = 0.013).

Life history differences exist between regions, but life histories also vary along a continuum within as well as between regions (Fig. 1, B and C). Hence, we predicted that parental investment tactics should vary among species on the basis of their position along this continuum (Fig. 1). To test this prediction, we examined parental responses of each species relative to its clutch size. Clutch size reflects current investment in reproduction and is correlated with adult survival (Fig. 1C), a measure of residual reproductive value. Clutch size therefore reflects both factors that should influence how parents respond to a juvenile versus an adult predator (Fig. 1A, see earlier). We found strong predicted relations between clutch size and parental responses to juvenile and adult predators across all 10 species (Fig. 3). Specifically, the magnitude of responses to a juvenile predator increased with increasing clutch sizes, as typical of northern species (Fig. 3, A and B). In contrast, magnitude of responses to an adult predator intensified with decreasing clutch sizes, reflecting higher annual adult survival (Fig. 1C), as typical of southern species (Fig. 3, C and D). Thus, parental responses to juvenile and adult predator manipulations can be predicted from their life histories independent of whether species are from northern or southern regions but yield differences between regions because of latitudinal differences in clutch size and survival.

Figure 3

Partial regression plots of parental responses, measured as the percentage (arc-sin transformed) that feeding rates decreased, relative to clutch size for 10 bird species paired between North (closed symbols) and South (open symbols) America. Larger responses reflect greater decreases in feeding rates and risk. (A) Species with larger clutch sizes had larger responses to nestling predator (jay) presentations (one-tailedrp = 0.83, df = 7, P = 0.003) controlling for differences in the natural risk of nest predation among species (6, 10). (B) Independent contrasts that control for any phylogenetic effects and nest predation showed the same response to jay presentations (one-tailed rp = 0.87, df = 6,P = 0.003). (C) In contrast, species with smaller clutch sizes had larger responses to adult predator (hawk) presentations (one-tailed rp = −0.60, df = 7, one-tailed P = 0.04) controlling for body mass. (D) Independent contrasts showed the same response to hawk presentations when controlling for body mass (one-tailedrp = −0.77, df = 6, P = 0.01).

These data provide two important sets of results. First, all species reduce risk when faced with a predation threat by reducing the rate that they visit the nest to feed offspring (Fig. 2, B and C). Hence, parents appear to trade off the costs associated with reduced food delivery to their young against a reduction in the risk of mortality to themselves or their offspring. Species differentially express this trade-off depending on whether risk is directed at the parents or their offspring, and expression is predicted by clutch size and associated adult survival among species (Figs. 1 and 3). Interspecific covariation of life history traits such as clutch size and adult survival has been commonly explored (3,4, 21–23), but covariation of these traits with parental behavior, as shown here, has been previously undocumented. Second, higher adult survival rates in southern species have been debated in recent years (12–15), potentially because of methodological differences in how survival is measured (14,15, 18). Our review of published studies that incorporate resighting (Fig. 1B) and our own efforts (Fig. 1C) indicate that southern birds have higher survival than related northern birds. Additionally, our finding that parents of southern species respond more strongly than northern species to risk of adult mortality (Figs. 2C and3, C and D) follows from higher adult survival in southern birds (Fig. 1A). This increased adult survival of southern birds is correlated with smaller clutch sizes (Fig. 1, B and C). Yet, no previous study has shown that clutch size and adult survival are correlated along a continuum that includes hemispheric differences (i.e., Fig. 1, B and C). Our demonstration that life history differences among southern and northern species accurately predict differences in parental risk-taking behavior provides empirical support for generalized models of life history evolution and offers insight into the underlying factors responsible for interspecific variation in parental care tactics within and among latitudes.

  • * Present address: Department of Biology, University of California, Riverside, CA 92521, USA.

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


View Abstract

Navigate This Article