Intraspecific trait variants determine the nature of interspecific interactions in a habitat-forming species

Authors


Abstract

Although the study of ecological interactions often takes into account functional variation between species, intraspecific variation is commonly ignored. Here, we investigate the importance of an intraspecific polymorphism in shaping interspecific interactions in a habitat-building species. Colonies of the social spider Anelosimus studiosus provide habitat for dozens of arthropod species, and colony members exhibit markedly polymorphic behavioral temperaments (BT): “aggressive” or “docile.” We manipulated the phenotypic compositions of colonies (100% aggressive, 50% aggressive and 50% docile, 100% docile) and measured the nature and magnitude of interactions between A. studiosus and two heterospecific web associates, Larinioides cornutus and Agelenopsis emertoni. We found that BT composition significantly affected the outcome of interspecific interactions, changing the relationship between A. studiosus and its web associates from an ammensalism (where A. studiosus experiences reduced fecundity and survival) to a commensalism or mutualism. Our study successfully illustrates the potential of BTs to impact whole community dynamics, and conversely, for community structure to influence the maintenance of BTs.

Introduction

Although most ecologists accept the notion that individuals vary, the central tendency in ecological studies is to overlook individual variation and suppose that the majority of meaningful functional diversity occurs at the level of the species. However, a growing body of literature challenges this classic view; the result is a greater understanding of how individual variation can influence various ecological phenomena (Bolnick et al. 2003, Crutsinger et al. 2006, Johnson et al. 2006, Bolnick et al. 2011). Individuals are rarely identical. Even with a single population, individuals commonly differ along any number of trait axes and, conceivably, these trait variants could change the way individuals interact with both con- and heterospecifics. Within behavioral ecology, consistent individual differences in behavior are commonly referred to as behavioral types (BTs), behavioral syndromes, personality, and/or temperament (Sih et al. 2004), and these differences are believed to alter the roles that individuals play within populations and, potentially, communities (Sih et al. 2004, Reale et al. 2007).

Consistent individual differences in behavior, or BTs, have been documented in many animal taxa and along countless axes of trait variation (Sih et al. 2004, Reale et al. 2007). To a lesser extent, BTs have been shown to influence the roles that individuals play within communities. For instance, in some species that show consistent individual differences in boldness toward predators, bold BTs tend to reach larger body sizes but experience greater susceptibility to predation (e.g., Biro et al. 2004, 2006; but see Reale and Festa-Bianchet 2003, Harris et al. 2010). Thus, bold BTs contribute more to predators' dietary intake, and predators prevent bold BTs from dominating prey populations. Similarly, variation in aggressiveness has been shown to influence aspects of individuals' foraging ecology, e.g., diet breadth (Riechert 1991), meal size (Pruitt and Krauel 2010), and foraging mode (Johnson et al. 2008). Taken together, a growing body of literature has shown that BTs can change the number or strength of interspecific interactions experienced by individuals. However, whether BTs can change the nature of the interactions experienced (e.g., competition, predation, mutualism) has not been documented. For example, nonaggressive BTs might be more likely to experience positive, facilitative interactions with heterospecifics, whereas aggressive individuals might tend to experience more competitive interactions. We argue that such effects are important because they could change the roles played by individuals within communities and, of course, modify the environments in which various BTs are favored.

To test the hypothesis that BTs may affect the nature of interspecific interactions, we used the spider Anelosimus studiosus as our model, for two reasons. Firstly, A. studiosus spiders build large colonies that serve as habitat to an array of >100 arthropod species (Perkins et al. 2007, Pruitt and Riechert 2011a) and, hence, have the potential to dictate the nature of interspecific interactions occurring within their webs. Secondly, A. studiosus exhibits a markedly bimodal behavioral polymorphism: females exhibit either an aggressive, active (hereafter “aggressive”) or a docile, inactive (hereafter “docile”) BT (Pruitt et al. 2008, Pruitt et al. 2010). Aggressive females exhibit greater aggressiveness toward predators, prey, and mates relative to their docile counterparts (Pruitt et al. 2010). Interestingly, both aggressive and docile BTs are observed in varying frequencies within multi-female colonies (Pruitt and Riechert 2009), and field data suggest that the BT composition of social groups can affect the community of araneofauna that associates with A. studiosus colonies (Pruitt and Riechert 2011a).

Here we investigate the nature of specific interspecific interactions by manipulating the BT composition of A. studiosus colonies and assessing their interactions with two common heterospecific web associates: Larinioides cornutus (Araneae, Araneidae) and Agelenopsis emertoni (Araneae, Agelenidae). We ask the following question: does colony BT composition influence the nature and/or magnitude of interspecific interactions within the web of A. studiosus? In other words, does BT composition significantly (and meaningfully) impact the payoff (mass gain) of associating with A. studiosus colonies for heterospecifics? And, do colonies' BT compositions influence the impact of heterospecifics on colony members' survival and reproductive success (measured by egg case mass)?

Methods

We provide an abridged, general version of our methodology here. However, a more detailed account of our methodology is available in Appendix A.

Collection and laboratory maintenance

Spiders were collected by placing garbage bags over A. studiosus colonies and trimming off the supporting branches. Colonies were then transported to the laboratory and dissected by hand. Juvenile A. studiosus were isolated into 59-mL deli cups, and web associates (Agelenopsis emertoni, Larinioides cornutus) were housed individually in 590-mL enclosures. Spiders were raised under standardized conditions. Upon reaching maturity, female A. studiosus were run through an interindividual distance assay to determine their BT (aggressive vs. docile). Females were then randomly mated with mature males and were assigned to an experimental colony based on their pre-mating body mass.

Behavioral type bioassay

Two females of unknown tendency were individually marked with fluorescent powder and placed in the center of a clear plastic container (13 × 13.5 × 2.5 cm). After 24 h of settling time, we measured the distance between them. All females that exhibited an interindividual distance greater than zero (i.e., they were not in direct contact) were run through a second confirmatory test with a known docile female (i.e., one that previously exhibited an interindividual distance score of zero). This test is necessary to tease apart the two types of females, because aggressive females demand space and chase away docile females. Females that exhibited an interindividual distance <7 cm in the second confirmatory test were categorized as “docile” and females that exhibited an interindividual distance >7 cm were categorized as “aggressive.” Seven centimeters corresponds with a natural break in the distribution of interindividual distance measures between the two phenotypes (Pruitt and Riechert 2009). Interindividual distance scores are repeatable over individuals' lifetimes, heritable, and highly correlated with other aggressiveness and boldness measures (Pruitt et al. 2008).

Artificially reconstituted colonies

Experimental colonies were composed of six size-matched (±3% body mass), randomly mated, and individually marked females of known BT. Experimental colonies were constructed with three BT compositions (100% aggressive, 50% aggressive and 50% docile, and 100% docile). Naturally occurring colonies vary tremendously in their phenotypic composition (range 0–90% aggressive, 10–100% docile; Pruitt and Riechert 2009). Colonies were first established in the laboratory and subsequently were transplanted into the field. Prior to being transplanted, we placed a pre-weighed, individually marked control (glass bead), A. emertoni, or L. cornutus within each colony.

Colony localities were opportunistically selected using the presence of preexisting colonies as an indicator of habitat quality. Colony placement was randomized with regard to treatment. At each location, we removed a preexisting colony and replaced it with an experimental colony. Experimental colonies were checked daily for the next 30 days for egg cases. Egg cases were weighed within 36 h of parturition and returned to their mothers. We used egg case mass as our estimate of individual colony members' fitness because, in spiders, egg case mass is correlated with fecundity (Foelix 1996). At least 10 colonies of each BT composition and treatment (i.e., control, A. emertoni, L. cornutus) were established at our field site (N = 92 colonies total).

After 30 days, colonies were recollected and heterospecifics were reweighed. We used change in body mass over the 30-day period as our estimate of heterospecifics' performance. To confirm that egg case mass was associated with number of offspring produced in A. studiosus, we haphazardly selected one egg case from each colony for dissection, dissected it, and counted the number of eggs therein.

Heterospecific performance in isolation

Heterospecific performance trials were run concurrently with our assessment of colony performance. To test the fitness consequences of associating with A. studiosus colonies, we released individually marked singleton L. cornutus and A. emertoni at locations adjacent to our experimental colonies. Isolated spiders were checked daily for the next 30 days by visually scanning foliage for evidence of webs. After 30 days, spiders were recollected and weighed. We used change in body mass over the 30-day period as our estimate of singleton heterospecifics' performance.

Statistical analyses

We used general linear models to predict change in heterospecifics' body mass, with starting mass as a covariate, treatment (i.e., 100% aggressive, 50% aggressive and 50% docile, 100% docile, isolated) as a fixed effect, and the interaction term starting mass × treatment in the model. For our estimate of colony performance, we used a 3 × 3 design with two response variables: (1) the average egg case mass of all colony constituents, and (2), the number of colony members that survived. We used a general linear model to predict the mass of females' egg cases and ordinal logistic regression (1–6 surviving females) to predict the number of colony members surviving. We included colony members' starting mass (covariate), colony BT composition (100% aggressive, 50% aggressive and 50% docile, 100% docile), treatment (A. emertoni, L. cornutus, control), and all possible interaction terms in our model. We used paired t tests to check for differences in the egg case masses of aggressive and docile females in colonies of mixed phenotypes, and a Pearson's correlation to test for an association between egg case mass and number of eggs therein.

Finally, we looked at the R2 and AICc scores of the above-mentioned models with and without (classic ecological approach) including BTs in the models. These comparisons allowed us to quantitatively measure the benefits (if any) of including intraspecific traits in our models. For heterospecifics, we tested performance of individuals associating with A. studiosus colonies (irrespective of composition) vs. isolated individuals; for A. studiosus, we tested the effects of associating with heterospecifics, but again, irrespective of BT composition. All of our statistics were run using JMP 8.0 (SAS Institute 2009).

Results

Web associates' body mass

Our results indicate a significant effect of treatment (F3,48 = 13.5, P < 0.0001) on mass gain by A. emertoni, where mass gain was positively associated with the presence of docile A. studiosus (Fig. 1). We failed to detect a significant effect of individuals' starting mass (F1,48 = 0.2, P = 0.63) or an interaction between starting mass and treatment (F3,48 = 0.4, P = 0.74). Mass gain for L. cornutus was also affected by treatment (F3,47 = 21.1, P < 0.0001), and as with A. emertoni, L. cornutus mass gain was positively associated with docile A. studiosus (Fig. 1). Again, we did not detect a significant effect of individuals' starting mass (F1,47 = 0.02, P = 0.92) or a significant interaction between starting mass and treatment (F3,47 = 0.6, P = 0.62). For both A. emertoni and L. cornutus, irrespective of colony composition, individuals associated with A. studiosus colonies gained at least as much mass as isolated individuals in similar habitats (Fig. 1).

Figure 1.

Boxplots of the relative mass gain (change in mass/starting mass) of spiders Agelenopsis emertoni and Larinioides cornutus in isolation or in experimental colonies of the social spider Anelosimus studiosus of three behavioral type (BT) compositions: six aggressive females, three aggressive and three docile females, or six docile females. The center line designates the median, vertical gray bars give the interquartiles range, and vertical lines are the 90th and 10th percentiles; dots represent outliers. Boxplots not sharing the same lowercase letters are significantly different at α = 0.05. Photo credits: Troy Bartlett (Agelenopsis) and Rich Hoyer (Larinioides).

A. studiosus fitness

We detected a significant effect of treatment (control, A. emertoni present, L. cornutus present) (F2,74 = 17.2, P < 0.0001) and BT composition (F2,74 = 10.4, P = 0.0001) on A. studiosus egg case masses. However, as evidenced by the significant interaction treatment × BT composition (F4,74 = 16.9, P < 0.0001) and treatment × starting body mass (F2,74 = 5.0, P = 0.01), the effects of heterospecific web associates differed among colonies of varying BT composition and starting body mass. We obtained qualitatively similar results from our data on colony member survival (Table 1, Fig. 2). Individuals in colonies containing a mixture of aggressive and docile females produced the largest egg cases in the absence of heterospecific invaders (Fig. 2); however, the addition of A. emertoni eliminated this advantage by reducing the egg case masses of females in groups containing the aggressive BT, but had no significant effect on groups of all docile females. Similarly, L. cornutus had a negative impact on colony members' egg masses in colonies containing aggressive females, but L. cornutus significantly increased the fitness of females in colonies of all-docile females. There was a significant positive association between egg case mass and number of eggs therein (r = 0.64, P < 0.0001, df = 91); see Appendix B. Within colonies of mixed BT, the fitness of aggressive and docile females did not differ significantly in the control (t20 = 1.22, P = 0.23) or L. cornutus treatment (t20 = −0.79, P = 0.44); however, aggressive females experienced higher fitness in the presence of A. emertoni (t20 = 2.67, P = 0.01); see Appendix B.

Figure 2.

Boxplots of (A) the relative reproductive success (egg case mass/female pre-mating body mass) of female Anelosimus studiosus, and (B) the proportion of A. studiosus colony members surviving in colonies of various behavioral type compositions and treatments (Agelenopsis emertoni, L. cornutus, control bead). Boxplot components are as in Fig. 1. Within BT compositions, boxplots not sharing the same lowercase letters are significantly different at α = 0.05. Photo credit: Terry Babin (Anelosimus).

Table 1. Summary of effects tests for the combined model predicting average egg case mass and survival of colony members of the social spider Anelosimus studiosus.Thumbnail image of

Summative statistics

Inclusion of BT always resulted in an unambiguously superior, more informative model, even after penalizing for additional terms: AICc values were always lower (ΔAICc ≥ 14.9), Akaike weights of BT-included models closely resembled 1.0, and adjusted R2 values of the BT-included models (Table 2) were much greater (Akaike 1987, Burnham and Anderson 2002).

Table 2. Side-by-side comparisons of our combined models with and without consideration of intraspecific trait variants (i.e., behavioral types).Thumbnail image of

Discussion

Although ecological studies commonly emphasize the importance of species' functional roles (e.g., guilds), relatively few studies consider the effects of functional diversity occurring at the level of the individual. In the present study, we documented within-species behavioral variants that influenced both the magnitude and nature of species interactions (i.e., turning ammensalisms into either a commensalism or mutualism) in a habitat-forming species (Figs. 1 and 2, Table 1). In terms of our statistical models, including information about BTs always resulted in better statistical explanations of the performance of web associates and colony members' fitness variables. Although it is intuitive that adding explanatory variables would lead to a higher R2, the magnitude of the improvement of the model fit indicates that intraspecific effects are at least as important, if not more so, than species identity in determining the protagonists' fitness. This interpretation is further supported by the finding (Table 2) that BT-included models were deemed more informative using likelihood-based information metrics, which penalize for superfluous terms (Burnham and Anderson 2002).

The socially polymorphic spider A. studiosus builds large, multi-female colonies along riparian habitats throughout the eastern United States. This species is ecologically important because its colonies serve as hubs for a rich community of arthropods, particularly araneofauna, whose net effects on A. studiosus are negative (Pruitt and Riechert 2011a). Here we demonstrate that the impacts of heterospecifics on colony member fitness differed among colonies with different BT compositions. In colonies with aggressive females, heterospecifics imposed large negative effects on colony member fitness (i.e., reducing egg case mass by 32%), whereas their effects were neutral or positive (i.e., increasing egg case mass by 50%) for colonies of all-docile females (Fig. 2). Thus, colonies containing a higher percentage of docile females appear to play a greater facilitative role within their community: heterospecifics gained more mass when associating with colonies of docile females (Fig. 1), and docile females were not negatively impacted by these interactions (Fig. 2). These results are important because they suggest that intraspecific trait variants can alter the functional roles played by individuals within their community by (1) changing the ecological relationships among species, and (2) affecting the quality of ecological services provided by key species within the community (e.g., habitat-forming species, ecosystem engineers).

From the point of view of heterospecifics, the payoff of associating with A. studiosus colonies differed among colonies with different BT compositions. Associating with colonies containing aggressive females had no significant effect on mass gain for either heterospecific (Fig. 1); however, both A. emertoni and L. cornutus experienced increased mass gain by associating with colonies of all-docile females (Fig. 1).Thus, BT composition appears to influence the payoff of associating with A. studiosus colonies, and from the standpoint of heterospecifics, colonies of all-docile females seem to be more profitable habitat. This trend probably stems from the fact that docile females rarely attack or molest heterospecific web associates (Pruitt and Riechert 2011a) and may instead benefit from their presence by stealthily consuming heterospecifics' uneaten prey (Pruitt and Riechert 2009). This strategy could, in part, explain why colonies of docile females experienced increased fitness in the presence of L. cornutus (Fig. 2).

Finally, our data suggest that the presence of heterospecifics may influence the BT compositions favored in different environments. In previous work, Pruitt and Riechert (2011b) documented a trend in which colonies of mixed BTs experienced increased fitness under standardized laboratory conditions. Concordantly, in the absence of heterospecifics, our data here confirm that females in colonies of mixed BTs tend to experience greater survival and fecundity (Fig. 2; also see Appendix B). Thus, in the absence of heterospecifics, within-colony behavioral variation confers increased fitness. However, the advantage of within-colony behavioral variation is lost when heterospecifics are present (Fig. 2): colonies of all three BT compositions experienced statistically indistinguishable fitness in the presence of A. emertoni, and colonies of all-docile females enjoyed the highest fitness in the presence of L. cornutus (Fig. 2). Thus, the relative performance of various BT compositions appears to depend on the presence of heterospecifics (i.e., an aspect of the biotic environment). One plausible inference from these data is that variation in the presence or abundance of heterospecifics could generate the environmental variation needed to maintain the diversity of BT compositions observed in nature.

Acknowledgments

We thank an anonymous reviewer for helpful comments on a previous version of our manuscript. Funding to J. N. Pruitt was provided by grants from the University of Tennessee and the University of California–Davis Center for Population Biology. Funding to M. C. O. Ferrari was provided by the Natural Sciences and Engineering Research Council of Canada and the University of Saskatchewan.

  1. Corresponding Editor: F. S. Dobson.

APPENDIX A

Detailed methods for establishing the artificially reconstituted colonies of the social spider Anelosimus studiosus (Ecological Archives E092-162-A1).

APPENDIX B

Figures depicting the relationship between egg case mass and number of eggs therein, and the egg case masses of aggressive vs. docile females in colonies of mixed behavioral type (Ecological Archives E092-162-A2).

Ancillary