Brood parasitism increases provisioning rate, and reduces offspring recruitment and adult return rates, in a cowbird host
- First Online:
- Cite this article as:
- Hoover, J.P. & Reetz, M.J. Oecologia (2006) 149: 165. doi:10.1007/s00442-006-0424-1
- 191 Views
Interspecific brood parasitism in birds presents a special problem for the host because the parasitic offspring exploit their foster parents, causing them to invest more energy in their current reproductive effort. Nestling brown-headed cowbirds (Molothrus ater) are a burden to relatively small hosts and may reduce fledgling quality and adult survival. We documented food-provisioning rates of one small host, the prothonotary warbler (Protonotaria citrea), at broods that were similar in age (containing nestlings 8–9 days old), but that varied in composition (number of warbler and cowbird nestlings) and mass, and measured the effect of brood parasitism on offspring recruitment and adult returns in the host. The rate of food provisioning increased with brood mass, and males and females contributed equally to feeding nestlings. Controlling for brood mass, the provisioning rate was higher for nests with cowbirds than those without. Recruitment of warbler fledglings from unparasitized nests was 1.6 and 3.7 times higher than that of fledglings from nests containing one or two cowbirds, respectively. Returns of double-brooded adult male and female warblers decreased with an increase in the number of cowbirds raised, but the decrease was more pronounced in males. Reduced returns of warbler adults and recruitment of warbler fledglings with increased cowbird parasitism was likely a result of reduced survival. Cowbird parasitism increased the warblers’ investment in current reproductive effort, while exerting additional costs to current reproduction and residual reproductive value. Our study provides the strongest evidence to date for negative effects of cowbird parasitism on recruitment of host fledglings and survival of host adults.
KeywordsAdult returnsBrood parasitismFood provisioningNatal philopatryRecruitment
Life-history theory assumes that organisms allocate resources in ways that maximize fitness (Stearns 1989). Maximizing fitness in iteroparous animals involves a balance between investment in current and future reproduction whereby increases in current reproductive effort are balanced by costs that reduce future reproductive potential (Williams 1966; Charnov and Krebs 1974). In avian systems, clutch and brood sizes have been manipulated experimentally in an effort to understand how factors such as incubation efficiency (Moreno and Sanz 1994; Engstrand and Bryant 2002) and the provisioning ability of parents (Hegner and Wingfied 1987; Moreno et al. 1999) limit current reproductive output and the future survival and fecundity of adults (Pugesek 1987; Nur 1988; Siikamaki et al. 1997). Adults can increase food delivery rates to enlarged broods in an attempt to compensate for having more nestlings (e.g., Moreno et al. 1999; Pap and Márkus 2003), but there are potential costs to post-fledging survival of nestlings (Dijkstra et al. 1990; Blondel et al. 1998; Horak 2003) and to survival of adults to the next breeding season (e.g., Nur 1984; Jacobsen et al. 1995; Daan et al. 1996; Saino et al. 1999).
Interspecific brood parasitism in birds presents a special problem for the host because the parasitic offspring exploit their foster parents (Davies 2000), eliciting them to invest dramatically more energy in their current reproductive effort compared to unparasitized individuals (Kattan 1996; Grim and Honza 1997, 2001; Dearborn et al. 1998; Kilner and Davies 1999; Kilpatrick 2002; Grim et al. 2003). Parasitic chicks influence parental behaviors by producing stimuli that are similar in quality but exaggerated in magnitude to those of the hosts’ own brood (Dearborn 1998; Lichtenstein and Sealy 1998; Kilner et al. 1999; Lichtenstein 2001). For example, begging calls of parasitic young serve as a strong signal (Davies et al. 1998; Grim and Honza 2001; Kilner 2003) that promotes a very high rate of provisioning from parents (Dearborn et al. 1998; Grim and Honza 2001; Hauber and Montenegro 2002; Kilner et al. 2004).
Brood parasitism by Molothrus cowbirds reduces clutch size and increases hatching failure and nestling mortality in many host species (reviewed in Robinson et al. 1995; Rothstein and Robinson 1998; Hauber 2003; Hoover 2003a). There is extensive evidence that parasitic brown-headed cowbird (Molothrus ater) chicks beg more intensively than their host nest-mates (Woodward 1983; Lowther 1993; Dearborn 1998; Lichtenstein and Sealy 1998), thereby elevating provisioning rates of (Dearborn et al. 1998), and energetic costs to foster parents (Kilpatrick 2002). In addition, if provisioning rates of parents do not keep up with demand, competitively inferior host nestlings may weigh less when they fledge from parasitized broods (Smith 1981; Dearborn et al. 1998; Hoover 2003a). Costs of parasitism paid away from the nest (e.g., reduced juvenile and adult survival to the next breeding season) have been difficult to measure. Numerous studies have documented positive correlations between nestling mass and post-fledging survival in the absence of brood parasitism (e.g., Naef-Daenzer et al. 2001; Monrós et al. 2002), but the cost of cowbird parasitism to post-fledging survival remains poorly known (Smith 1981; Payne and Payne 1998). Similarly, there are scant data on the costs of brood parasitism to adult survival in hosts (but see Payne and Payne 1998). Information on these unmeasured costs is necessary to more accurately assess the effects of brood parasitism on host population dynamics and the evolution of host defences against parasitism.
The costs of cowbird parasitism to the current reproduction of relatively small hosts are disproportionately high, and many small hosts produce few or none of their own offspring when parasitized (Trine et al. 1998; Kilpatrick 2002; Kilner 2003). Some small hosts are, however, capable of raising multiple parasite and host nestlings in broods that are unusually massive (e.g., Hoover 2003a). Such host species may possess a greater capacity to increase their rate of food delivery and thereby reduce the effect of parasitic nestlings on their own offspring. Hosts capable of fledging enlarged broods composed of cowbird and host nestlings provide an excellent opportunity to study some of the poorly known costs of brood parasitism (e.g., reduced offspring recruitment and adult survival).
The prothontary warbler (Protonotaria citrea) is a small (14–16.5 g) Neotropical migratory bird that is commonly parasitized by brown-headed cowbirds, which are three times larger (Petit 1989, 1999; Hoover 2003a). Cowbird parasitism reduces the warblers’ current reproductive effort by lowering hatching success and nestling survival (Petit 1991; Hoover 2003a). Cowbird parasitism does not, however, affect subsequent reproductive effort within the same breeding season as measured by renesting intervals and subsequent clutch sizes (Hoover 2003a). Prothonotary warblers are enigmatic because, for a relatively small host, they can raise large broods (in terms of mass) containing multiple cowbird and warbler nestlings (Hoover 2003a). The mass of individual warbler nestlings that fledge from these parasitized broods is unaffected by the number of warbler nestlings present, but decreases significantly with an increase in the number of cowbird nestlings in the brood (Hoover 2003a). As a result, we made the prediction that the recruitment of warbler fledglings to the breeding population would decrease with an increase in cowbird nest-mates.
Here, we document the provisioning rates of adult prothonotary warblers in relation to brood mass to determine if, over the range of brood sizes (unparasitized and parasitized) occurring naturally, provisioning rates increase with demand (i.e., brood size) and if males and females contribute equally. We also examined the hypothesis that brood parasitism affects offspring recruitment and adult returns in the host by testing the prediction that recruitment of warbler offspring and returns of adult warblers both decrease with increased numbers of cowbirds in warbler broods.
Materials and methods
Study sites and study species
We conducted this research within the Cache River Watershed (CRW) in Illinois, USA (37°18′N, 88°58′W). The Cache River has a total length of 176 km and meanders through the southern tip of Illinois to the Ohio River, draining 1,537 km2 of land (Mankowski 1997). Study sites (25) consisted of wet floodplain forests, forested sloughs and backwaters, and bald cypress (Taxodium distichum) and tupelo (Nyssa aquatica) swamps located within a 16×12 km (19,200 ha) portion (referred to as study area hereafter) of the watershed. These wet forested habitats are approximately 3% of the land cover in the watershed and are embedded in a landscape primarily consisting of agriculture (32%), grassland (31%), and upland (26%) and bottomland (6%) forest habitat (Mankowski 1997). Most (76% of 25) study sites were discrete forested wetlands separated from each other by >500 m of dry forest, early successional forest or agriculture (Hoover 2003b) where our study species was not found. The remaining sites were arbitrarily established on portions of large (>500 ha) forested wetlands. Several (>10) pairs of warblers established and defended breeding territories within each study site each year (Hoover 2001, 2003b).
The prothonotary warbler is a migratory bird that winters in the Neotropics and breeds in forested wetlands throughout much of the eastern half of the United States (Petit 1999). This species is territorial and socially monogamous (Petit and Petit 1996), and the presence of unmated “floater” individuals is uncommon in populations that have been studied (Petit 1999; Hoover 2001). Prothonotary warblers build their nests in secondary cavities, and associate closely with standing water in bottomland and swamp forests (Bent 1953; Petit 1999; Askins 2000). These warblers readily use nest boxes when available (Petit 1989; Blem and Blem 1994; Hoover 2003b) and can be studied in great detail during the breeding season (e.g., Petit 1989; Petit and Petit 1996; Hoover 2003a, 2006). The prothonotary warbler is commonly used as a host by the brood parasitic brown-headed cowbird (Petit 1989, 1999; Hoover 2003a, c). All data presented here are taken from pairs of warblers using nest boxes. The addition of nest boxes to study sites does not affect warbler densities (Hoover 2001) or the rate of brood parasitism compared to natural cavities (Hoover 2001). Because the breeding habitat (forested wetlands) of the prothonotary warbler typically occurs in discrete patches in our study area, we knew the identity of >95% of individuals breeding on our study sites every year (Hoover 2003b, 2006).
Monitoring nesting success
We studied the nesting ecology of prothonotary warblers from 15 April to 15 August during 1994–2004. In 1994, we placed nest boxes made from modified 1.9-l cardboard milk and juice cartons (Fleming and Petit 1986) on study sites to cover areas of suitable habitat (inter-box spacing of 25–35 m). Nest boxes had openings that were similar in diameter to natural cavity nests (32–64 mm), and were placed on trees at the average height (1.3 m above the surface of the water) of natural cavity nests in this area (Hoover 2001). There were approximately 1,000 nest boxes distributed among 25 study sites during 1994–2004.
Nest boxes were monitored at 4-day intervals for the duration of the breeding season each year. All active nests were monitored at shorter intervals (1–2 days) during the egg laying, hatching, and late nestling (7–11 days old) periods. On each visit, we recorded the number and status of eggs and nestlings at active nests and determined the fate (i.e., number of offspring fledged or the cause of failure) of every nesting attempt. Starting in 1995, prothonotary warbler nestlings were banded with a uniquely numbered aluminum (United States Fish and Wildlife Service) leg band 2–3 days prior to fledging. All individual adult warblers were captured and color-marked with a unique combination of a numbered aluminum and colored plastic leg bands (Hoover 2003b). We observed adult warblers throughout the breeding season and recorded, for each individual, nest-site location(s), the fate of each nesting attempt, and the number of fledglings (warbler and cowbird) produced.
Food provisioning rates and brood mass
During 1995–1998, we documented rates of food provisioning for male and female warblers at 92 different nests that varied in brood composition (number of warbler and cowbird nestlings) and brood mass. We sampled nests from 10 of the study sites over the 4-year period and measured provisioning rates over the spectrum of brood masses (light to heavy) each year. Provisioning rates of particular male and female warblers were sampled one time only during the study period. To control for the potential effects of nestling age and time of day on food provisioning by adults, we confined each observation to a 30–40 min period between 1000 and 1300 hours (CST) when nestlings were 8–9 days old. We used binoculars to observe nests from a distance of approximately 40–50 m. Males and females have sexually dimorphic plumage (Petit 1999) and, at this distance, we could easily determine the sex of the visiting adult and that a food item had been delivered. We could not, however, determine the number or size of food items adults delivered with each visit to broods. For each nest, our observation period began with delivery of food by one of the adults. This delivery was not included in the calculation of provisioning rates. We then documented the exact times of food delivery during the 30–40 min period for each sex. At the end of the observation period, we used a 100-g spring scale to measure the total mass of the focal brood (to the nearest 0.5 g), and noted the number of warbler and cowbird nestlings present. We determined overall (sexes combined) and sex-specific food provisioning rates (feeds/h) for each nest.
Recruitment of warbler offspring
To determine offspring recruitment, we considered only those warbler nestlings that were banded and known to have fledged from nests in our study sites. In each year, subsequent to the banding of warbler nestlings, we searched systematically on each study site and within a 1,000-m-wide area surrounding each site for any adult warblers bearing an aluminum leg band only (indicating that the bird had been banded as a nestling in a previous year). We captured, determined the identity of, and color-marked all of these individuals that we observed. For each of these returned fledglings, we noted the location of their current breeding territory as well as the exact location and parasitism status of the brood where they were produced. Because we captured nearly all (>95%) males and females breeding on our study sites and knew whether the few we failed to capture were banded or not, we could accurately document the identity of all of the banded fledgling warblers that returned to breed on a study site.
During 2003–2005, we took advantage of the patchy nature of the warbler breeding habitat and expanded our searches for returned fledglings to include >90% of all other (not included in our study sites) suitable breeding habitat within the 19,200 ha study area and >80% of the suitable breeding habitat within a 10-km-wide zone around the study area. We searched for returned fledglings on 20 additional forested wetlands within the study area and 23 within the 10-km-wide zone around the study area in each of the 3 years. We visited each of these additional sites one time during the middle of each breeding season (June) and captured, identified and color-marked any banded adults we observed. We recorded the number of cowbird nestlings (0, 1, 2) present in the brood of origin for each returned warbler fledgling we captured, and also for warbler fledglings that were never seen again. Within each parasitism category, the percent of fledged nestlings that returned to our study area was used as a measure of offspring recruitment.
Between-year returns of adult warblers
Within each study site, we determined the identity of every color-marked individual returning from a previous year and captured and color-marked any individuals that were not banded so that nearly all individuals (>95%) breeding on our study sites were color-marked every year. For the purposes of this paper, we focused on the returns of those males and females that produced two broods in a single breeding season. We limited our analyses to double-brooded individuals for numerous reasons. Several years of correlative and experimental data from this study system demonstrated that the successful production (fledging) of two unparasitized broods in one breeding season was associated with extremely high rates (80%) of between-year returns in male and female warblers (Hoover 2003b). There was little breeding dispersal between years for double-brooded warblers and the vast majority (>80% of returning birds) occupied the same territory as the previous year. Individuals that produced only one or no broods were more likely to move to a different territory or habitat patch between years (Hoover 2001, 2003b), making it more difficult to differentiate between survival and dispersal for these less successful birds. In this study system, there were no other breeding season parameters (e.g., number of offspring produced or nesting attempts made) that explained return rates and between-year territory fidelity as well as the number of broods produced (Hoover 2003b). This earlier research supported the conclusion that double-brooded individuals not returning between years were more likely to have died than to have dispersed from study sites (Hoover 2001, 2003b).
Double-brooded individuals likely expend greater amounts of energy than individuals producing fewer broods and are, therefore, more susceptible to costs associated with producing enlarged broods containing multiple cowbird offspring. The high rate of return (80%) can be viewed as a conservative but relatively close approximation of annual survival for male and female warblers in this study system. Reduced return rates of double-brooded individuals corresponding with increased numbers of cowbirds produced would likely reflect costs of brood parasitism to adult host survival. We identified all individuals that produced two broods in a given breeding season, recorded the number of cowbirds produced by each male and female, and determined whether or not they returned the following breeding season by identifying all adults on study sites the subsequent year and by searching for color-marked birds within a 1,000-m-wide area surrounding each site. Returns of particular male and female warblers were recorded one time only between the first year they were double-brooded in our study system and the very next year. During 2003–2005, we also looked for returned color-marked birds on 20 additional forested wetlands within the study area and 23 within the 10-km-wide zone around the study area.
We compared overall provisioning rates (number of food deliveries per hour) with brood mass using correlation analyses to test the prediction that provisioning rates increased with brood mass. For these analyses we combined male and female food deliveries and analyzed unparasitized and parasitized nests separately. We used ANCOVA to test for differences in provisioning rates between parasitized and unparasitized nests, with brood mass as a covariate. We also used ANCOVA to compare provisioning rates of males and females, with brood mass as a covariate. We used a two-factor ANOVA to test whether the number of cowbird nestlings (1 or 2), number of warbler nestlings (1, 2, or 3), or their interaction had a significant effect on provisioning rates of adults with parasitized broods. Mean values are presented with ±1SE.
We used contingency analyses to test the predictions that offspring recruitment (of individuals known to have fledged from nests) decreased with an increase in the number of cowbird nestmates, and that between-year returns of double-brooded adults would decrease with an increase in the total number of cowbirds produced. We compared recruitment status (yes/no) of warbler nestlings among those fledging from broods containing 0, 1, or 2 cowbird nestlings and compared between-year return status (yes/no) of double-brooded adults (males and females separately) among those producing 0, 1, 2, or ≥ 3 cowbirds.
Food provisioning rates and brood mass
Offspring recruitment and adult returns
We banded 2,534 warbler nestlings that subsequently fledged during 1995–2003, and a total of 198 (7.8%) were observed and captured within our study area in a subsequent year. We captured every warbler bearing a single (aluminum) band that we observed. During 2003–2005, only one of the 1,554 warblers found in the 10-km-wide zone surrounding our study area was banded (a returned fledgling from an unparasitized brood). We did not include this individual in our analysis. Totals of 2,153, 292, and 89 nestlings fledged from broods containing 0, 1, and 2 cowbird nestmates, respectively. The percent of those fledglings we observed in a subsequent year (recruitment to the study area) decreased with an increase in the number of cowbird nest-mates (8.41, 5.14, 2.25, respectively; X22=7.785, P=0.020). Host fledglings from unparasitized nests returned to the study area at a rate 3.7 times greater than fledglings from broods that contained two cowbirds.
In the majority of songbird species, males and females both contribute to care of nestlings and fledglings (Ehrlich et al. 1988). In our study, male and female prothonotary warblers contributed equally to provisioning nestlings and increased their rate of food delivery with increased brood mass. In general, broods with cowbirds elicited a higher delivery rate by adult warblers than unparasitized broods of the same mass (see also Grim and Honza 2001). This is consistent with other studies that have found increased food delivery rates for parasitized broods (Dearborn et al. 1998; Hauber and Montenegro 2002; Kilpatrick 2002; Grim et al. 2003). The presence of cowbirds in warbler broods not only increased the provisioning rates of adult hosts, but also appeared to reduce the recruitment of warbler fledglings and returns of warbler adults.
One cowbird nestling represents the mass equivalent of 2.0–2.5 prothonotary warbler nestlings (Hoover 2003a), and parasitized broods elicit a high rate of provisioning from parents compared to unparasitized broods of the same mass (this study). In an analysis of food delivery rates to parasitized and unparasitized nests of 20 different host species, Kilpatrick (2002) found that cowbird nestlings imposed a substantial energetic demand on smaller host parents. Furthermore, the quantity of energy delivered to parasitized nests demonstrated that parents are often willing to provision nests at a much higher rate than for an average brood of their own (Kilpatrick 2002). The ability of prothonotary warblers to dramatically increase their rate of food delivery to more massive broods demonstrated increased parental investment in current reproductive effort. Had the adults not been able to increase provisioning rates, it is unlikely that many warbler nestlings would fledge from parasitized nests. Prothonotary warblers often were able to raise some of their own nestlings in addition to cowbird nestlings, but these warbler nestlings usually achieved a mass at fledging that was lower than warbler nestlings from unparasitized broods (Hoover 2003a).
Offspring recruitment and juvenile survival
Several studies have shown a positive correlation between mass at fledging and juvenile survival and recruitment in unparasitized nests (Naef-Daenzer et al. 2001; Monrós et al. 2002; Keedwell 2003) whereas others have not (Anders et al. 1997; Brown and Roth 2004; Kershner et al. 2004). Competitively superior cowbird nestlings may cause significant reductions in mass or growth of host nestlings (Smith 1981; Payne and Payne 1998; Burhans et al. 2000; Hoover 2003a) that are particularly costly to post-fledging survival of host offspring. For example, we documented previously that the average mass of warbler nestlings at the time of fledging was reduced by 9 and 16% in the presence of one or two cowbird nestlings, respectively, compared to individuals from unparasitized broods (Hoover 2003a). Therefore, the warbler nestlings from parasitized nests likely began their post-fledging period at a disadvantage.
Cowbird fledglings are known to usurp much time and effort of their foster parents for weeks after leaving the nest. Woodward (1983) summarized feeding rates (feeds per hour) to cowbirds and host fledglings for eight parasitized host species. For all species, the feeding rate to host fledglings with equivalent total mass of a cowbird fledgling was significantly lower than the feeding rate to a single cowbird, and the cowbirds did not become independent until they were 25–39 days old. Adult prothonotary warblers continue to provision warbler fledglings for up to 35 days after they leave the nest (Petit 1999). On numerous (>10) occasions we have observed male prothonotary warblers feeding juvenile cowbirds 20–30 days after the fledging of parasitized broods (J. Hoover, personal observation). These observations indicate that the negative effects of cowbird parasitism on the host offspring may continue well beyond the nest if cowbird fledglings continue to divert food from warbler fledglings. Radio-telemetry studies of newly fledged birds have documented that survival rates are relatively low during the first few weeks but higher after that period (Sullivan 1989; Anders et al. 1997; Naef-Daenzer et al. 2001; Yackel et al. 2001), suggesting that the first 2–3 weeks out of the nest is a critical period for juvenile birds. If, as we suspect, caring for cowbird fledglings reduces the warblers’ provisioning of their own fledglings during the first 3 weeks out of the nest, the survival of warblers fledging from parasitized broods may be reduced further.
We are aware of only two other studies that present data on the effects of cowbird parasitism on natal philopatry (i.e., natal returns) or juvenile survival. Each study considered differences between two categories of parasitism (parasitized and not parasitized). Payne and Payne (1998) documented a reduction in natal philopatry in indigo buntings (Passerina cyanea) relative to parasitism status and found that 151 of 2,442 (6.18%) nestlings from unparasitized nests returned to breed whereas only 1 of 95 (1.05%) returned from parasitized nests. In a study of an island population of song sparrows (Melospiza melodia) in British Colombia, Smith (1981) found that survival of juvenile sparrows was not related to parasitism status, but suggested that predation of recently fledged cowbirds by crows (Corvus caurinus) may have reduced the potential for cowbird fledglings to negatively affect the survival of sparrow fledglings.
Prothonotary warbler fledglings recruit into the breeding population only after enduring the post-fledging period, their first fall migration to the tropics, an over-wintering period, and a spring migration from the tropics back to the breeding grounds. Natal philopatry is typically low (1–10%) for many species of migratory passerines (Weatherhead and Forbes 1994) and, consequently, juvenile survival rates and natal dispersal distances for these species are poorly known. Unparasitized prothonotary warbler fledglings were 1.6 and 3.7 times more likely to return to our study area than fledglings from broods with 1 or 2 cowbirds, respectively. We argue that these differences likely reflect differences in survival rather than differences in natal dispersal.
We acknowledge that it is possible that reduced offspring recruitment with increased brood parasitism was not the result of differences in juvenile survival but instead a product of an effect of brood parasitism on natal dispersal distances. If parasitized fledglings returned and bred farther away, on average, from their natal location than unparasitized fledglings, then our ability to locate returned warbler fledglings from parasitized broods might have been reduced. However, mean natal dispersal distances did not increase with parasitism (mean distances were 2.7±0.26, 2.8±0.65, and 1.6±0.69 km for warbler fledglings raised with 0, 1, or 2 cowbirds, respectively; F2, 192=0.115, P=0.89). In addition, during the last 3 years of the study period we expanded our search for returned offspring to a 10-km-wide zone around the study area and found only one additional returned fledgling, which came from an unparasitized brood. We found no additional returned fledglings from broods that contained one or two cowbirds, indicating that our ability to locate recruited offspring from parasitized nests was not reduced by differences in dispersal at the scale of our 32×36 km study system. Differences in survival, rather than dispersal, appear to provide a better explanation for reduced recruitment with increased cowbird parasitism. When exactly (e.g., pre-migration, migration, on wintering grounds) warbler fledglings that had cowbird nest-mates pay this survival cost is not yet known.
Observational and experimental work in this study system has previously shown an apparent lack of adaptive responses by the warblers (e.g., using nests with openings too small for cowbirds, defending nests during egg laying, deserting parasitized nests) to costly cowbird parasitism (Hoover 2003a, c). Adult prothonotary warblers may, however, adaptively disperse away from territories where they were heavily parasitized. If the differences in between-year returns that we documented were the result of adaptive dispersal, territory fidelity of individuals returning to the study system should decrease (i.e. local dispersal increase) with increased numbers of cowbirds raised. The territory fidelity of returning males and females did not, however, decrease with increasing parasitism for males (89, 83, 86, and 100%; X23=1.122, P=0.772) or females (72, 81, 81, and 86%; X23=2.142, P=0.543) after producing 0, 1, 2, or ≥3 cowbirds, respectively. In addition, for the relatively few returning adults that moved off of their previous territory, mean breeding dispersal distances did not increase with increased parasitism for females producing 0, 1, 2, or ≥3 cowbirds (197±22, 178±60, 207±56, and 190±90 m; F3, 28=0.293, P=0.83), or for males producing 0, 1, or 2 cowbirds (147±31, 132±20, and 60 m; F2, 9=0.528, P=0.61), respectively. Dispersal away from territories by warblers that produced multiple cowbirds does not, therefore, explain the differences in adult returns we documented.
If the energetic costs of raising cowbirds to the host foster parents are significant, parasitism may influence between-year returns of adults by increasing mortality. Differences in between-year returns of double-brooded individuals relative to the number of cowbirds produced should, therefore, reflect differences in adult survival. Results from our study support this hypothesis and demonstrate how producing multiple cowbirds is particularly detrimental (Fig. 3). Additional support for this hypothesis is that we have never documented a double-brooded individual switching to a different study site between years, disappearing for a year or more and subsequently reappearing, or showing up on other patches of suitable breeding habitat within the entire study area and the 10-km-wide zone around it. We were unable to determine if parasitized individuals were more likely to disperse outside our 115,000-ha study system, but it is an unlikely possibility given the lack of any correlation between dispersal distance and parasitism in our study. The warblers in this study system accept parasitism and employ no apparent anti-parasitism defenses that are likely less costly than dispersal to an unfamiliar area (Hoover 2003c). Thus, adaptive dispersal out of our study system by parasitized individuals is less likely.
Few other studies have attempted to measure the cost of raising cowbirds to the survival of host adults. Brood parasitism had no apparent effect on the survival of adult female song sparrows (Smith 1981), or on the between-year fidelity of male and female indigo buntings (Payne and Payne 1998). We were unable to ascertain whether or not the samples in these two studies included birds that were double-brooded. Collectively, these studies (along with our own) suggest that raising a single cowbird in a breeding season may not be very costly to the survival of adult hosts. Our study does indicate, however, that raising multiple cowbirds was much more costly.
Raising multiple cowbirds reduced the returns of male warblers more than females. Male and female prothonotary warblers are known to care for juveniles after they fledge from the nest for up to 5 weeks (Petit 1999; Hoover 2001). In many species, adult birds may divide broods so that each sex cares for approximately half of fledged young (Skutch 1976; Smith 1978; Kopachena and Falls 1991; Anthonisen et al. 1997). However, females may stop delivering food to fledglings prior to independence in order to initiate a new nesting attempt, particularly earlier in the nesting season (Ogden and Stutchbury 1997; Vega Rivera et al. 2000). Female prothonotary warblers incubated their next clutch of eggs within 7–9 days following the successful fledging of a first brood (Hoover 2003a), leaving the male to provide nearly all of the care for the fledglings.
The first broods of double-brooded males that raised multiple cowbirds in our study were more likely to contain multiple cowbird nestlings (19 of 29) than second broods (7 of 29) (X21=10.04, P=0.002). This resulted in parasitized first broods having a mean mass 42% higher than unparasitized first broods of the typical size of five warbler nestlings (84.5±2.11 vs 59.9±0.40 g on days 8–9 of the nestling period; Hoover 2003a). Males were therefore responsible for providing the majority of the post-fledging care of the largest broods because the females of these earlier broods were engaged in new nesting attempts. Survival of male songbirds has been shown to decrease with increased number of young cared for post-fledging (Wheelwright et al. 2003). Thus, care for post-fledging broods enlarged by parasitism likely represents a significant burden to males, and may explain why the cost of brood parasitism appeared to be higher in males than females (Fig. 3).
No studies have previously attempted to determine the extent to which rearing a brood parasite lowers a host’s reproductive value (Rothstein and Robinson 1998). Basic life history theory assumes that current reproductive effort reduces future reproductive value (Stearns 1992) either by increasing adult mortality during or after the breeding season, or through a decrease in the ability to invest in future offspring. The increased investment of adult warblers to parasitized nests appears to be a trade-off whereby the extra investment in current reproductive effort was associated with a negative effect on adult survival. That any host young at all fledged from heavily parasitized nests was likely a result of the warblers’ ability to significantly increase provisioning rates to heavier, more demanding broods. However, our data suggest that the host fledglings from these nests were diminished in quality and survived poorly to adulthood. Our study is the first to document that cowbird parasitism reduces fledgling recruitment and adult returns in the host, likely as a result of reduced survival. This underscores the need for and importance of long-term studies of marked populations to elucidate the effects of brood parasitism on reproductive tradeoffs and lifetime reproductive success in host species.
The tremendous efforts of many dedicated field assistants greatly improved this research, especially those of E. Whetsell, A. Corso, B. Holliday, C. Kelly, J. Stahl, L. Rodman, B. Franklin, R. Schmitz, A. Spencer, M. Mckim-Louder, and D. Robertson. We also thank the members of the Cache River Joint Venture (the United States Fish and Wildlife Service, The Nature Conservancy, and the Illinois Department of Natural Resources) for their assistance with setting up the field research in southern Illinois. The Illinois Natural History Survey, in particular the staff of the Center for Wildlife and Plant Ecology, provided critical logistic support. Scott Robinson provided valuable insights and ideas that greatly improved the research. Comments from two anonymous reviewers strengthened the manuscript. The research presented here was described in Animal Research Protocol No. N6C107/7093 approved on 8 April 1997 and No. N8C046 approved on 2 February 1998 by the Institutional Animal Care and Use Committee of the University of Illinois at Urbana-Champaign. Financial support for this work was provided by the United States Fish and Wildlife Service (INT 1448-0003-95-1007), The Nature Conservancy, the Illinois Department of Natural Resources Wildlife Preservation Fund, The National Fish and Wildlife Fund, the University of Illinois (Dissertation Completion Fellowship and Travel Grant), the North American Bluebird Society, the Champaign County and Decatur Audubon Societies, and Sigma Xi. Any opinions, findings and conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the agencies and organizations that supported the research.