Introduction

Since Darwin [1], ecologists have often assumed that competition for food resources is most intense among closely related species [2]. Feeding guilds, however, often include a diverse assemblage of taxa competing for the same resources [3]–[5]. For example, in desert habitats of the American Southwest, ants, diurnal birds, and nocturnal mammals all compete for the same seeds (reviewed in [6]), and the abundance of each species in the guild affects the abundance of the others.

Ants and birds are among the most abundant consumers in many ecosystems [7], [8] and they frequently compete for resources. Competitive relationships between ants and birds can influence the distribution [9], habitat-use [10], foraging behavior [11], [12], and offspring quality [13] of birds. Although ants can impact bird behavior via interference competition, wherein ants actively exclude avian competitors from a given area or resource [11], indirect competition for common food resources appears to underlie most competitive interactions between ants and birds. The change in availability and distribution of nutritional resources brought about by foraging ants can influence the distribution and behavioral decisions of birds such that they avoid settling in areas with high ant densities [9] and avoid specific trees containing high densities of foraging ants [12]. In addition to modifying the behaviors of birds, ants can also reduce the reproductive success of birds [14].

To date, studies exploring the dynamics of ant/bird interactions have focused primarily on species that naturally co-occur (but see [15]). Given the negative impacts that native ants can have on native birds, and the large impacts that invasive species can have on native flora and fauna [16], we were curious about the competitive effects that an introduced ant (red imported fire ants, Solenopsis invicta) might have on a native insectivorous bird (eastern bluebirds, Sialia sialis). Eastern bluebirds are obligate cavity-nesting, double- or triple-brooded passerine birds common throughout eastern North America [17]. Red imported fire ants were introduced into the United States from South America in the 1930's [18] and have subsequently become a serious ecological problem in much of the southern U.S. [19]. Fire ants disrupt natural food webs [19] and negatively impact native flora [20] and fauna [21], [22], frequently devastating local arthropod abundance and diversity [21] (but see [23] for an evaluation of long-term arthropod community recovery).

Because fire ants can dramatically alter arthropod communities [21], their abundance could indirectly influence the foraging and reproductive success of insectivorous animals. Indeed, experimental reduction of fire ant abundance has been shown to increase the abundance of the insectivorous loggerhead shrike (Lanius ludovicianus; [24]). To our knowledge, the study by Allen et al. [24] is the only published experimental study of the competitive effects that fire ant invasions have on insectivorous birds. However, a link between fire ant invasion and reduced reproductive output has previously been suggested for eastern bluebirds (P. Gowaty, unpublished data). Although fire ants occasionally prey on bluebird nestlings, the impact of this invasive ant species on bluebirds due to competition for food is predicted to be substantially greater than losses caused by occasional nest predation [17].

To further examine the impact of invasive fire ants on a native insectivorous bird species, we measured the effects of fire ant abundance on the breeding biology of eastern bluebirds by use of both correlational and experimental methods. First, we examined whether variation in foraging fire ant abundance across bluebird territories was correlated with various parameters of reproductive success (number of fledglings, hatching success) and reproductive strategy (likelihood of producing a second brood in the same territory (re-nesting), offspring sex ratio). Second, we compared the reproductive success and reproductive strategies of bluebirds inhabiting territories with experimentally reduced fire ant abundances to those inhabiting adjacent territories with high fire ant densities. We predicted that resource competition between ants and bluebirds would lead to reduced reproductive success and altered reproductive strategies in bluebird territories with abundant fire ants. Specifically, we predicted that the depletion of arthropod food resources by fire ants would decrease habitat quality in the territories they occupied and cause bluebirds inhabiting such ant-abundant territories to produce female-biased broods [25].

Methods

Results

During the initial wave of nesting attempts by eastern bluebirds at our study site in 2009, bluebirds nested in both control and treated territories in 19 of the 23 dyads of adjacent territories. Of these 19 territory dyads, both members of 15 dyads experienced no predation events. Within this subset of predation-free territory pairs, 11 of 15 nest boxes in control territories were used during the second nesting attempt of the season and 10 of 15 nest boxes in treated territories were re-used (though not necessarily by the same pair).

Fledged Offspring

Bluebird pairs inhabiting treated territories did not fledge significantly more offspring ( = 3.9±0.4) than pairs in control territories ( = 3.0±0.4) during the first breeding attempt of the season (paired t-test, n = 15, t = 1.60, p = 0.13). Additionally, there was no significant difference between the number of young fledged by bluebirds in treated ( = 3.6±0.3) or control ( = 3.4±0.6) territories during second nesting attempts (paired t-test, n = 7, t = 0.31, p = 0.77). However, within dyads, the cumulative fledging success (total number of offspring fledged from a given territory) was significantly different between control and experimental territories. Specifically, when we analyzed the total number of offspring fledged from each territory (regardless of parent identity and excluding nests that experienced predation), we found that treated territories fledged 1.1 more offspring ( = 6.3±0.5) than adjacent control territories ( = 5.2±0.5; paired t-test, n = 15, t = 2.17, p<0.05). This last statistic compares the productivity of territories rather than individuals and takes into account the territories where no second nest was attempted (4/15 control territories and 5/15 treated territories).

In addition to comparisons of territory dyads, we compared total fledging success across all territories (including unpaired territories) relative to fire ant abundance. Among nest boxes that were never depredated (in either first or second nesting attempts), we found that increased fire ant abundance was associated with decreased reproductive output (R² = 0.17, F_{1, 31} = 5.98, p = 0.02; Fig. 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Seasonal reproductive success of bluebirds as a function of fire ant abundance.

Relationship between fire ant abundance (mean number of ants captured per sample vial) and the seasonal reproductive success (total number of fledged offspring) of eastern bluebirds (R² = 0.17, F_{1, 31} = 5.98, p = 0.02). Each point represents a single territory (n = 31).

https://doi.org/10.1371/journal.pone.0022578.g001

Hatching Success

We found that hatching success was not related to fire ant abundance during first broods (n = 32, p = 0.40) but that the hatching success of mothers who re-nested in their original territories (and for whom we had fire ant data, n = 21) was negatively related to the abundance of foraging fire ants (p = 0.051, Fig. 2).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Hatching success of bluebirds as a function of fire ant abundance.

Relationship between fire ant abundance and hatching success of second nests of eastern bluebirds (n = 21). Logistic regression equation (Y = e^{2.4788 -0.0085(fire ant abundance)}/1+ e^{2.4788 - 0.0085(fire ant abundance)}), where Y is equal to the hatching success, represented by the line.

https://doi.org/10.1371/journal.pone.0022578.g002

Likelihood of Re-Nesting

To determine whether bluebird parents re-used the same nest box during their second breeding attempt of the season we had to capture both parents at the nest box during both breeding attempts. After analyzing the 39 nest locations across the study population where either a) we positively identified mothers and fathers from both first and second broods or b) no second nesting attempt occurred, we determined that re-nesting by one or both of the original parents at a nest was equally likely in both treated and untreated territories (overall probability of re-nesting  =  0.64, chi-square between control and treated territories  =  0.03, d.f. = 1, p = 0.86). However, when we used a logistic regression to analyze the likelihood of re-nesting across territories for which we had data on abundance of foraging fire ant and re-nesting information (n = 29 nests), we found a significant negative relationship between fire ant abundance and likelihood of re-nesting (n = 29, p = 0.045). This effect may be driven by territories with very high fire ant numbers and the fact that parents at these nests did not undertake a second breeding attempt at the same site (Fig. 3).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Probability of bluebirds re-nesting in their original territory as a function of fire ant abundance.

Relationship between fire ant abundance (mean number of ants captured per sample vial) and the calculated probability of re-nesting by the same eastern bluebird parents (n = 29). Logistic regression equation (Y = e^{1.9204 -0.00958(fire ant abundance)}/1+ e^{1.9204 - 0.00958(fire ant abundance)}), where Y is equal to the probability of re-nesting, represented by the line.

https://doi.org/10.1371/journal.pone.0022578.g003

Discussion

The abundance of foraging fire ants had a marked effect on the breeding success and reproductive strategies of eastern bluebirds. Bluebirds that nested in territories with abundant fire ants produced fewer fledglings, had more hatching failures during second broods, produced fewer male offspring, and were less likely to re-nest in their original territories than neighboring bluebirds on territories with fewer ants. These results confirm the importance of cross-phyletic competition in shaping ecological community dynamics, and demonstrate the strong impact that invasive species can have on native animals.

The effects of fire ants on the reproductive success of bluebirds were not significant during either the first or second nesting attempts when each was considered alone, although the trends during each reproductive bout were in the predicted directions with treated territories producing more offspring than control territories. There was, however, a significant difference between treated and control territories in total reproductive output (number of fledged offspring) over the course of the entire breeding season. These observations are similar to those reported by Aho et al. [14], who found that differences in reproductive output between Eurasian treecreepers (Certhia familiaris) inhabiting territories with experimentally manipulated levels of red wood ants (Formica rufa) were most pronounced when treecreepers were double rather than single-brooded. For both treecreepers and bluebirds, prolonged competition with ants appears to take a gradual toll, which becomes most evident after repeated bouts of reproduction.

The mechanisms by which greater fire ant abundance reduces reproductive success in bluebirds are not entirely clear. Likely, ants affect bluebirds primarily by competing for arthropod food, but there are at least two mechanisms by which lower food resources could impact reproductive success. First, less food on a territory may subject parents to additional stress, which could shift time budgets towards self-maintenance behaviors and away from offspring care [34]. Second, reduced food availability could increase the distance that bluebird parents must travel to provide sufficient resources to their offspring (sensu [35]), increasing the work that parents must perform to successfully raise offspring. Previously, it has been shown that higher workloads, in the form of experimentally enlarged broods, reduce the likelihood of female bluebirds producing second clutches [36]. If providing food to many offspring (as in [36]) and providing adequate food when resources are scarce put similar strains on bluebird parents, both stressors could result in decreased seasonal reproductive output.

Because fire ants are known to disrupt the native arthropod communities [21] upon which bluebirds rely for food, and because food abundance is known to influence hatching success in a number of birds (e.g. carrion crows Corvus corone, [37]; white storks Ciconia ciconia, [38]; black-legged kittiwakes Rissa tridactyla, [39]), we predicted that a reduced availability of food for parents in territories with high fire ant densities would depress hatching success. However, we found reduced hatching success among broods in territories with high densities of fire ants only during second breeding attempts. The reason we failed to detect differences in hatching success in the first nests of the season but did detect differences in second nests is likely a result of cumulative effects of insufficient food on the condition of breeding bluebirds. Consequently, significant negative effects on hatching success may only manifest after prolonged exposure to fire ant-mediated food stress [40]. Alternatively, it is possible that the arthropod communities on which bluebirds depend did not have time to respond to the experimental reduction of fire ants by the time of the first nesting attempt. Because the success of individual bluebird nests was typically an all-or-none response, with respect to hatched nestlings, the reduction in hatching rate is likely largely responsible for the decreased fledgling output of control territories.

Fire ants not only reduced the reproductive output of bluebirds and reduced hatching success; they also triggered a shift in the sex ratio of bluebird offspring. Although the mechanisms of primary sex-allocation are currently unknown for birds [41], numerous studies support the idea that females can alter the sex-ratio of their offspring and do so in response to both social [42], [43] and environmental [25], [44] factors. Because reproductive variance is usually greater for males than females [45], [46], and because high-quality males will produce more offspring than high-quality females, sex ratio theory suggests that parents should increase investment in male offspring when conditions are favorable for producing high-quality offspring [47], [48]. In concordance with these predictions, bluebird clutches from territories with experimentally reduced fire ant abundance had, on average, 22% more males than clutches in adjacent territories. Although this difference was statistically significant only for nests in the first half of the breeding season, our sample size during second nesting attempts severely limited our statistical power to detect real differences. It is interesting to note that the significant differences in sex ratio between adjacent territories in first nests arose not from the reduction of male offspring in habitats with more fire ants (average proportion of males  =  0.49), but from an increase in the proportion of male offspring in territories with fewer ants (average proportion of males  =  0.71). The direction of this response indicates that an experimental reduction of fire ants improves local resources and maternal condition, both of which should favor the increased production of male offspring.

Invasive species can occasionally exert selection pressure strong enough to induce rapid evolutionary changes in native biota (e.g. [49], [50]). With respect to fire ants, any adaptive responses by bluebirds are likely to include changes in habitat and selection of breeding sites. To avoid the possibility that bluebird used post-manipulation fire ant abundance as a criterion in choosing territories, we applied the hydramethylnon fire ant bait only after bluebirds had already chosen territories. Although our treatment regime precluded the possibility that bluebirds chose territories based on experimentally manipulated levels of fire ants, it did not prevent bluebirds from relocating after their first breeding attempt. Re-nesting by one or both parents, however, was equally likely (∼64%) in both treated and control territories. Nevertheless, the probability that bluebirds would stay in their original territories was negatively correlated with the fire ant abundance in that territory. This relationship appears to be driven by the fact that bluebirds in territories with high abundance of foraging fire ants (>250 ants/vial) did not re-nest. Relocation may be a simple behavioral strategy in response to reduced food abundance in territories with large numbers of fire ants, rather than an active assessment of fire ant abundance.

The primary insight from this study is not the negative effect of invasive fire ants on the abundance of eastern bluebirds, which is a common songbird throughout their range and not a species of conservation concern [17]. Rather, the more important implication of our study is that invasive fire ants can be a significant competitor of insectivorous birds, and perhaps other insectivorous vertebrates, and can reduce the reproductive success of such species. Several other insectivorous birds that nest in pasture and farmland in the southeastern United States where fire ants have been introduced, including the eastern meadowlark (Sturnella magna), northern mockingbird (Mimus polyglottus), and brown thrasher (Toxostoma rufum), have shown persistent declines in recent decades [51]–[53]. The declines of these species are often attributed to habitat alteration but, based on the observations from our study, competition with invasive fire ants should be included among the factors that might be contributing to the decline of these insectivorous bird species. Given the predicted global expansion of fire ants [54], understanding the detrimental effects that fire ants may have on insectivorous birds could become even more important in the future.

Acknowledgments

We thank B. Rolek for valuable intellectual contributions concerning the study design, as well as V. Rebal, A. Langston, E. Dunavant, and J. Hudson for their hard work and assistance in the field. We also thank F. Graham for providing his valuable insight and knowledge of fire ants throughout our study. Finally, we thank N.E. Davis, P.A. Gowaty, K.J. McGraw, J.D. Ligon and one anonymous reviewer for helpful comments on earlier editions of this manuscript.