Ecology, 89(6), 2008, pp. 1714–1722 Ó 2008 by the Ecological Society of America

CONSEQUENCES OF OMNIVORY FOR TROPHIC INTERACTIONS ON A SALT MARSH SHRUB CHUAN-KAI HO1

AND

STEVEN C. PENNINGS

Department of Biology and Biochemistry, University of Houston, Houston, Texas 77204 USA

Abstract. Although omnivory is common in nature, its impact on trophic interactions is variable. Predicting the food web consequences of omnivory is complicated because omnivores can simultaneously produce conflicting direct and indirect effects on the same species or trophic level. We conducted field and laboratory experiments testing the top-down impacts of an omnivorous salt marsh crab, Armases cinereum, on the shrub Iva frutescens and its herbivorous and predatory arthropod fauna. Armases is a ‘‘true omnivore,’’ consuming both Iva and arthropods living on Iva. We hypothesized that Armases would benefit Iva through a top-down trophic cascade, and that this benefit would be stronger than the direct negative effect of Armases on Iva. A field experiment on Sapelo Island, Georgia (USA), supported this hypothesis. Although Armases suppressed predators (spiders), it also suppressed herbivores (aphids), and benefited Iva, increasing leaf number, and reducing the proportion of dead shoots. A one-month laboratory experiment, focusing on the most common species in the food web, also supported this hypothesis. Armases strongly suppressed aphids and consumed fewer Iva leaves if aphids were available as an alternate diet. Armases gained more body mass if they could feed on aphids as well as on Iva. Although Armases had a negative effect on Iva when aphids were not present, Armases benefited Iva if aphids were present, because Armases controlled aphid populations, releasing Iva from herbivory. Although Armases is an omnivore, it produced strong top-down forces and a trophic cascade because it fed preferentially on herbivores rather than plants when both were available. At the same time, the ability of Armases to subsist on a plant diet allows it to persist in the food web when animal food is not available. Because omnivores feed on multiple trophic levels, their effects on food webs may differ from those predicted by standard trophic models that assume that each species feeds only on a single trophic level. To better understand the complexity of real food webs, the variable feeding habits and feeding preferences of different omnivorous species must be taken into consideration. Key words: Armases cinereum; food web; herbivory; Iva frutescens; omnivory; Ophraella notulata; Paria aterrima; predation; salt marsh; top-down bottom-up; trophic cascade; Uroleucon ambrosiae.

INTRODUCTION Ecologists have a long-standing interest in how topdown and bottom-up forces (Hairston et al. 1960, Hunter and Price 1992, Denno et al. 2003) and trophic cascades (Paine 1980, Strong 1992, Polis and Strong 1996, Borer et al. 2005) combine with other interactions to structure food webs (Pimm and Lawton 1978, Oksanen et al. 1981, Schmitz et al. 2000). Recent years have seen a substantial advance in this field, with work shifting from championing which force is more important to evaluating what factors affect the relative strength of different trophic interactions (Hunter et al. 1997, Moon and Stiling 2002, Denno et al. 2003). A discrepancy remains, however, between the simple structure of theory and the complexity of many natural food webs: real food webs often contain a large number Manuscript received 29 June 2007; revised 27 September 2007; accepted 4 October 2007. Corresponding Editor: J. T. Cronin. 1 E-mail: [email protected]

of interactive species, frequent omnivory, interactions that change with age structure, trophic loops, and other features that are not included in simple trophic models (Polis 1991, Menge 1995). In this paper, we focus on the issue of omnivory by conducting field and laboratory experiments with a simple coastal food web. Omnivory, feeding on more than one trophic level (Pimm and Lawton 1978), is a common feeding strategy in nature (Polis 1991, Coll and Guershon 2002, Thompson et al. 2007). Omnivory has an advantage over strict carnivory or herbivory in some situations because it allows organisms to feed on both abundant but less nutritious food (e.g., plants) and less abundant but more nutritious food (e.g., animal prey) (Coll and Guershon 2002). Recent studies have identified the importance of omnivory in trophic interactions (Strong 1992, Fagan 1997, Pace et al. 1999, Tanabe and Namba 2005); however, theory does not agree regarding the impact of omnivores on trophic interactions. On the one hand, omnivory can increase food web complexity, which should decrease the strength of trophic cascades

1714

June 2008

OMNIVORY AND TROPHIC INTERACTIONS

via reticulate interactions (Polis and Strong 1996). On the other hand, omnivory may increase the strength of trophic cascades. Given that omnivores can persist during periods of food scarcity by eating alternative foods (Polis and Strong 1996), persistent omnivores might produce strong top-down forces, increasing the likelihood of trophic cascades (Moran et al. 1996, Eubanks and Denno 1999). Predicting the consequences of omnivory in food webs is complicated because omnivores can consume species from multiple trophic levels, producing conflicting direct and indirect effects on the same species or trophic levels (Fig. 1). A number of studies have examined the consequences of omnivory in trophic interactions; however, in many of these studies, ‘‘omnivory’’ refers to intra-guild predators (IGP): top predators that feed on animals from different trophic levels (i.e., predators and herbivores; Diehl 1993, Fagan 1997, Rosenheim and Corbett 2003). Relatively few studies have examined ‘‘true omnivory,’’ or feeding on both plants and animals (Coll and Guershon 2002), though true omnivory is often observed in arthropods (Coll and Guershon 2002), fish (Pringle and Hamazaki 1998), birds (Greenberg 1981), and mammals (Meserve et al. 1988). Only true omnivores have the potential to have direct negative effects on the plant and other trophic levels at the same time. In addition, many previous studies of how omnivory affects trophic cascades have been conducted in aquatic systems (Lodge et al. 1994, Pringle and Hamazaki 1998, Dorn and Wojdak 2004), where interactions may differ from those in terrestrial habitats (Strong 1992, Polis 1999, Shurin et al. 2002). In summary, omnivory is common in natural food webs, but there is a lack of understanding of its impact on food webs, especially in cases involving true omnivory. In order to advance our understanding of how omnivores affect food webs, we examined the role of a true omnivore in a terrestrial food web. We conducted field and laboratory experiments to examine the top-down impacts of an omnivorous salt marsh crab, Armases cinereum, on the shrub Iva frutescens and its associated arthropods. Armases consumes both Iva and many of the predatory and herbivorous arthropods that live on it (Appendix). Because Armases prefers to eat animal food over plants (C.-K. Ho, personal observations) and grows better on an animal than a plant diet (Buck et al. 2003), we hypothesized that Armases would benefit Iva through a top-down trophic cascade, and that this benefit would be stronger than the direct negative effect of Armases on Iva.

1715

FIG. 1. Omnivores can produce conflicting direct and indirect effects on the same species or trophic levels. Here, an omnivore may directly consume an herbivore, leading to an indirect positive effect on a plant. Alternatively, the omnivore may directly consume the plant, leading to a direct negative effect.

Uroleucon ambrosiae is the most abundant herbivore on I. frutescens (Hacker and Bertness 1995). Other herbivores found on I. frutescens include the beetles Ophraella notulata (Futuyma and McCafferty 1990) and Paria aterrima, the grasshopper Hesperotettix floridensis, and a variety of leafhoppers, galling insects, and stem borers (S. C. Pennings and C.-K. Ho, personal observations). The wharf crab, Armases cinereum (¼ Sesarma cinereum), is a common high-marsh omnivore (Teal 1958, Abele 1992, Buck et al. 2003). It is a semiterrestrial crab that ranges from the high intertidal zone to .100 m inland (Seiple 1979, Pennings et al. 1998). It feeds on a wide variety of foods, including live leaves, leaf litter, mammal feces, small fiddler crabs (Uca spp.), terrestrial basidiomycetous fungi, and marsh sediments (Pennings et al. 1998, Buck et al. 2003, Zimmer et al. 2004). During periods of high humidity and moderate temperatures (especially dawn and dusk), it often climbs 1 to ;3 m into the canopy of high-marsh shrubs to feed on leaves, flowers, or canopy arthropods (C.-K. Ho, unpublished data). In the laboratory, Armases readily eats aphids (Uroleucon ambrosiae; Aphididae), leaf beetles (Ophraella notulata and Paria aterrima; Chrysomelidae), spittle bugs (Cercopidae), stink bugs (Pentatomidae), ladybugs (Coccinellidae), and spiders (C.-K. Ho, unpublished data). Ladybugs and spiders both eat aphids (C.-K. Ho, personal observation). All species will be referred to generically hereafter. Field experiment

METHODS Study site and species All research was conducted on Sapelo Island, Georgia, USA (31827 0 N; 81815 0 W). Salt marshes around this island are typical of the southeastern U.S. Atlantic Coast (Pomeroy and Wiegert 1981). Marsh elder, Iva frutescens, is a common high-marsh shrub. The aphid

To examine top-down effects of Armases on the Iva community, we conducted a field experiment with Armases-exclusion and control treatments. We selected 24 Iva plants between 1 to 2 m tall that were not touching adjacent shrubs. Replicate plants were separated by at least 5 m. Armases were excluded from onehalf of the plants by wrapping Iva trunks near the base

1716

CHUAN-KAI HO AND STEVEN C. PENNINGS

with slippery plastic and surrounding the trunks with inverted funnels. Our observations suggested that these barriers did not exclude small arthropods, like insects and spiders. The experiment began in July 2003 and ended in July 2005. To monitor the effectiveness of the Armases-exclusion treatment, we counted Armases on all 24 Iva plants twice in August 2003, and twice in August 2004. Armases numbers per plant were sharply reduced in the exclusion treatment compared to the control, from 2.25 6 0.37 to 0.04 6 0.04 Armases/plant (mean 6 SE) in August 2003 and 1.08 6 0.27 to 0.12 6 0.08 Armases/plant in August 2004 (P , 0.01 in both cases). To determine impacts of Armases exclusion, shrubs were surveyed monthly during July–October 2003, May–October 2004, and in July 2005. We recorded the number of herbivores (the aphid Uroleucon; all grasshoppers; adults and larvae of beetles Ophraella and Paria), and predators (adults and larvae of Coccinellid ladybugs; all spiders) on each Iva plant. We measured the dimensions (height, width, and depth) of each shrub, counted the number of leaves, and measured the area of three randomly selected leaves as indicators of plant performance. We measured damage to the same three leaves from chewing herbivores (which might include some chewing damage from Armases) as an indicator of herbivore pressure from leaf chewers. On two dates (August 2003 and October 2004) we scored the percentage of terminal shoots that were dead. Canopy volume of Iva plants was calculated as an ellipsoid, 4/3 3 p 3 (height/2) 3 (width/2) 3 (depth/2), and leaf area was calculated as an ellipse, p 3 (leaf length/2) 3 (leaf width/2). Data on herbivore and predator densities are presented as accumulated densities (the sum of all individuals observed on all sampling dates), and were analyzed using ANOVA to compare Armases-exclusion and control treatments. To examine the accumulated treatment effects on plants, we analyzed data from the final sampling period (July 2005), with data from the first sampling period (July 2003) included in the model as a covariate to account for initial differences among shrubs. Plant data were also analyzed with repeatedmeasures ANOVA, yielding very similar conclusions (not shown). Because dead shoots were scored only twice during the experiment, we averaged data from these two dates and compared among treatments using ANOVA. Laboratory experiment To examine and decouple the interactions between the most common taxa from the Iva food web (Armases, Uroleucon, and Iva), we conducted a laboratory experiment in the greenhouse at the University of Georgia Marine Institute on Sapelo Island in July 2004. We established four treatments: (1) Armases and Iva, (2) Iva alone, (3) Uroleucon and Iva, and (4) Armases, Uroleucon, and Iva. Iva plants were placed in individual cages made of polyester mesh fabric (n ¼ 11

Ecology, Vol. 89, No. 6

plants/treatment). All taxa were healthy and behaved normally in cages. Iva plants were germinated from seed collected in October 2002 from marshes in South Carolina, Georgia, and Florida (n ¼ 7 collection sites), potted in a mixture of 60% potting soil and 40% sand, and watered with fresh water. Plants from each collection site were included in each treatment. Plants were 16 months old and 94 6 3 cm (mean 6 SE) tall as the experiment began. Five clones of aphids (;100 individuals total) were collected on Sapelo Island before the experiment, propagated for two weeks on Iva, and pooled. Five juvenile aphids were added to each cage, allowed to increase to .40 individuals, and then thinned to 30 individuals at the start of the experiment. Armases were collected from Sapelo Island, weighed, and immediately added to cages in pairs (one male and one female; average initial mass ¼ 2.0698 6 0.2214 g and 1.4382 6 0.1290 g, mean 6 SE, respectively). Plants were surveyed at the start of the experiment and weekly for four weeks. We measured aphid numbers, the number of Iva leaves newly damaged by Armases during that week, and several plant traits (height, numbers of leaves, average leaf size from three random leaves). The mass of each individual Armases was measured at the start and end of the experiment. At the end of the experiment, we also recorded the percentage of leaves (n ¼ 3 leaves/plant) covered by an epiphyllic fungus, Cladosporium sp. We used two-way ANCOVA to compare results between treatments, with the presence or absence of Armases and aphids treated as the main effects, with data from week 1 included as the covariate. To understand the dynamics of the results over time, we used repeated-measures ANOVA to analyze changes in aphid populations and Armases damage to Iva leaves. Two-way ANOVA was used to analyze the effects of sex and treatment on changes in Armases body mass. Although any laboratory experiment is necessarily an abstraction of the field, we attempted to make the laboratory mesocosms as relevant to field conditions as possible. The plants that we used in the laboratory were about one-third smaller than those studied in the field (94 6 3 and 140 6 4 cm tall, mean 6 SE, respectively), but both were within the range of field plants. Starting densities of aphids in the laboratory experiment (30 per plant with an average of 76 leaves) were within the range of field densities. We assigned two Armases individuals per plant in the laboratory experiment, similar to field densities (2.25 per plant in August 2003). We were concerned that Armases in the laboratory cages might forage in the canopy more often than they did in the field. However, we observed that Armases in cages stayed on the ground most of the time and climbed into the canopy mostly at dawn and dusk, as in the field. Therefore, changes in Armases foraging behavior in the laboratory appeared to be minimal.

June 2008

OMNIVORY AND TROPHIC INTERACTIONS

1717

FIG. 2. Effects on the Iva food web of excluding Armases in the field. Bars represent means þ SE. P values and sample sizes are indicated above bars. Shown are effects of Armases treatments on densities of (A) spiders, (B) ladybugs, (C) the aphid Uroleucon, (D) beetles (Ophraella and Paria) and grasshoppers (Acridids and Tettigoniids), and effects on (E) area of Iva leaves, (F) number of leaves and plant height, and (G) leaf damage by chewing herbivores and proportion of dead shoots.

RESULTS Field experiment In the field experiment, Armases reduced spider and aphid densities by 75% and 70%, respectively (Fig. 2A, C). Similar trends for ladybugs were not significant due to high variability (Fig. 2B). Armases did not affect beetle or grasshopper densities (Fig. 2D). Armases increased Iva leaf number by 124% (Fig. 2F), but had no effect on leaf area, plant height, or leaf damage from chewing herbivores (one of which is Armases itself) (Fig.

2E–G). Armases reduced the proportion of dead shoots, many of which were probably caused by unidentified stem borers, by 83% (Fig. 2G). Laboratory experiment In the greenhouse, Armases strongly suppressed aphid populations (Fig. 3A). When aphids were present, Armases damaged fewer Iva leaves during the first week of the experiment (Fig. 3B). This difference then disappeared because most of the aphids were quickly consumed (Fig. 3A). Armases with access to both aphids

1718

CHUAN-KAI HO AND STEVEN C. PENNINGS

Ecology, Vol. 89, No. 6

FIG. 3. Effects of manipulating Armases and Uroleucon in the laboratory. (A) Aphid population in Armases-absent vs. Armases-present treatments. (B) Number of Iva leaves newly damaged by Armases each week in aphid-absent and aphid-present treatments. Data in panels (A) and (B) indicate means 6 SE. (C) Body mass changes of male and female Armases over four weeks in aphid-absent and aphid-present treatments. (D) Area of Iva leaves, (E) Iva plant height, and (F) number of Iva leaves at the end of the experiment. Histogram bars in panels (C)–(F) show means þ SE; within panels (D)–(F), bars containing the same letter indicate that the means do not differ significantly (P . 0.05).

and Iva grew faster than those with only Iva to eat (Fig. 3C). Female Armases gained mass faster than males, but there was no interaction between experimental treatments and sex. Aphid feeding induced plastic reductions

in leaf area, but this effect was not as strong when Armases was also present, presumably because aphids were present at very low densities due to predation from Armases (Fig. 3D). Any combination of Armases or

June 2008

OMNIVORY AND TROPHIC INTERACTIONS

aphids reduced plant height compared to controls (Fig. 3E). These effects were not additive (Fig. 3E, significant interaction term), because when Armases was present it consumed most of the aphids. As in the case of leaf area, aphids alone reduced leaf number, but this effect was eliminated when Armases was present (Fig. 3F, significant interaction term) and aphid populations were suppressed. In general, when alone, Armases had weakly negative effects on Iva compared with the strong negative effects of aphids, and Armases had positive effects on Iva if aphids were also present. The epiphyllic fungus Cladosporium sp. was common on the last sampling date in the aphid and Iva treatment where aphids were abundant (fungi on 67% 6 8% [mean 6 SE] of Iva leaves), but absent in the other three treatments where aphids were rare (0 6 0; ANOVA, F3,40 ¼ 73.33, P , 0.0001). This difference in fungal abundance was likely due to fungal growth being stimulated by aphid honeydew deposited on leaves (S. Y. Newell, personal communication). DISCUSSION Because omnivores feed at multiple trophic levels, they can simultaneously produce opposite direct and indirect effects on the same species or trophic level. As a result, omnivory might be expected to increase the complexity of food webs and weaken tropic cascades (Polis and Strong 1996, Pace et al. 1999). If omnivores have stronger effects on one trophic level than another (i.e., a strong diet preference), however, the net result of these opposite effects might still be a strong top-down effect that could lead to a trophic cascade. In our experiments, Armases consumed Iva leaves and had a direct negative effect on Iva; however, Armases also consumed many aphids, which had a stronger negative effect on Iva than Armases did. By eating aphids, Armases created an indirect positive effect on Iva that was stronger than its direct negative effect. This net positive effect of Armases on Iva accords with our observation that Armases prefers to eat aphids over Iva leaves when both are available (C.-K. Ho, personal observation). The beneficial effect of Armases on Iva was present not only in a simplified laboratory food web, but also in the field in a much more complex food web that included a variety of other herbivores and arthropod predators, many of which might be eaten by Armases or eat each other. Thus, omnivores may commonly produce strong top-down effects if they preferentially feed or have a large trophic effect on particular trophic levels or species. Diehl (1993) argued that omnivores (intra-guild predators) are most likely to produce trophic cascades if intermediate consumers are large relative to resources and thus more likely to be consumed by omnivores. In this study, aphids were smaller than Iva plants, but Armases nevertheless preferred to eat the aphids. While agreeing with Diehl (1993) that the trophic consequences of omnivory will depend in large part on which trophic

1719

levels are preferentially eaten, we would argue that these preferences are likely to be based on the relative food quality of the prey types, which is a function of their nutritional quality and capture success, rather than their size per se. For true omnivores, an animal diet is likely to always be more nutritious than a plant diet (Coll and Guershon 2002, Buck et al. 2003). In this case, because aphids were both more nutritious than plants and had a limited repertoire of escape behaviors, there would have been little benefit for Armases to preferentially eat plants if aphids were present. These results might have differed if the primary herbivore had been more difficult for Armases to catch, and in fact Armases did not have significant negative effects on the more mobile beetles or grasshoppers in the field. These results highlight the importance of including omnivory in food web models. Otherwise, the predictions of trophic models might often be misleading. For example, the HSS-Oksanen-Fretwell model predicts a negative cascading effect of top predators on plants in a system with four trophic levels (Power 1990). For our field experiment, this model predicts that consumption of predators by Armases should release herbivores and in turn negatively affect Iva plants (Fig. 4). This prediction was incorrect: Armases positively affected Iva plants by suppressing both predators and herbivores (Fig. 4). Some studies with intra-guild predators, which do not consume plants, also obtained similar results, in which the fourth-level predators benefited plants by suppressing herbivores (Spiller and Schoener 1990, Moran et al. 1996). In other cases, however, intra-guild predators may reduce top-down control by eating or interfering with each other rather than by consuming herbivores (Finke and Denno 2003). Because omnivores can feed at multiple trophic levels, understanding omnivore feeding preferences and the strength of their different trophic effects, rather than an artificial attempt to force omnivores into discrete trophic ‘‘levels’’ (Thompson et al. 2007), is likely to be the key to understanding their effects on food webs. In the case of true omnivores, it is likely that omnivores will often prefer to eat animals over plants. As was found here, the omnivores may act as a ‘‘super-predator,’’ causing an apparently complicated food web to collapse to three effective trophic levels, thereby producing a positive effect on plants. Whether intra-guild predators suppress or release herbivores is likely to depend on the details of IGP ‘‘omnivore’’ feeding behavior and vulnerability of prey. In these cases, IGP omnivores might either increase or decrease top-down control of herbivores (Finke and Denno 2002, 2003). In the case of the Iva system, the laboratory experiment helped identify mechanisms by which Armases benefited Iva plants in the field. In the laboratory, Armases had a weakly negative effect on Iva when aphids were absent. When aphids were present, however, Armases strongly reduced aphid numbers, thereby releasing Iva from aphid herbivory and having

1720

CHUAN-KAI HO AND STEVEN C. PENNINGS

Ecology, Vol. 89, No. 6

FIG. 4. A food web model predicts that Armases should have a negative effect on Iva plants in the field, by releasing herbivores from predators. In reality, however, the omnivorous Armases produced a positive effect on Iva plants by suppressing herbivores. Solid arrows indicate direct top-down forces; dashed arrows indicate trophic cascades.

a net positive effect on Iva (i.e., a trophic cascade). In the field, aphids were typically present, and Armases reduced aphid populations by 70%, benefiting Iva. After three growing seasons, Iva plants visited by Armases had twice as many leaves as Iva plants with Armases excluded. Thus, it is clear that top-down forces from Armases critically mediate the performance of Iva in the field. In addition, although we did not detect significant negative effects of Armases on beetle and grasshopper densities, it is possible that the threat of Armases predation altered the behavior of these herbivores, and that this contributed to the positive effect of Armases on Iva (i.e., a behaviorally mediated trophic cascade [Schmitz et al. 1997]). Theoretically, it is possible that the impact of Armases on herbivores was due to competition for their shared host plant, rather than to predation. However, we consider this possibility unlikely because we observed Armases aggressively consuming aphids in the laboratory cages. Trophic cascades promoted by Armases did not end with plants, but in the laboratory extended to epiphyllic fungi, Cladosporium sp. This fungus was probably common in the field at low densities but prospered in the laboratory where aphids became an order of magnitude more abundant than in the field, and stimulated fungal growth by producing abundant honeydew. Because the fungus only appeared near the end of the laboratory experiment, it probably had little effect on plant performance (as an epiphyte, it would affect Iva mainly by shading leaves). A number of selective advantages may promote an omnivorous diet over a specialized one (Eubanks and Denno 1999, Coll and Guershon 2002, Diehl 2003). Armases grew faster if they had access to both aphids and Iva rather than just Iva (Fig. 3C), paralleling previous results that Armases grew better on a diet including animal protein than on a plant diet alone (Buck et al. 2003). However, as an omnivore, Armases

can survive in the absence of its preferred food(s). This diet flexibility of omnivores can have important consequences for trophic cascades (Polis and Strong 1996, Eubanks and Denno 1999). When aphid populations decrease, Armases populations likely will not. As a result, when aphids increase, Armases can suppress their numbers without a time lag. In contrast, populations of specialized predators (e.g., ladybugs) will decrease when their prey are rare, and do not increase rapidly when prey become abundant. As a result, the population increase of specialized predators often lags behind that of their prey (Showler and Greenberg 2003), and the top-down effect of specialized predators will occur only following a significant time lag. Therefore, omnivores might provide immediate and more consistent top-down pressure on prey populations than specialized predators, likely resulting in stronger trophic cascades. For this reason, Murdoch et al. (1985) argued that generalist consumers were often more effective biocontrol agents than specialist ones because a higher consumer population could be maintained despite low densities of the targeted prey species. Our field and laboratory experiments produced consistent results, suggesting that the laboratory experiment captured the most important mechanisms operating in the field. We, however, would like to emphasize two caveats that come with any experiment involving a simplified, caged, food web. First, the laboratory experiment was a simplified system compared to the field one; therefore, Armases had no alternative arthropod prey other than aphids. However, because aphids were the most vulnerable prey in laboratory feeding trials, and were an order of magnitude more abundant than other arthropods, it is likely that Armases primarily fed on aphids in the field even when other arthropods were present. In fact, aphids were the only herbivores reduced by Armases in the field experiment. Second, the field experiment was not conducted in cages, and all

June 2008

OMNIVORY AND TROPHIC INTERACTIONS

arthropods (except for Armases in the exclusion treatment) were free to come and go from marked plants, possibly enhancing or minimizing treatment effects. Nevertheless, the treatment effects were similar in the field and laboratory experiments. Salt marsh food webs on the Atlantic Coast of the USA have long been assumed to be dominated by bottom-up forces, such as flooding and the salinity, sulfide, and nitrogen content of soils (Sullivan and Daiber 1974, Valiela and Teal 1974, King et al. 1982). More recently, however, ecologists have documented strong top-down effects from consumers on the food webs of smooth cordgrass, Spartina alterniflora, which dominates the low marsh (Silliman and Bertness 2002, Denno et al. 2005), and the food webs of the shrubs Borrichia frutescens and Iva frutescens, which dominate the terrestrial border of the marsh (Hacker and Bertness 1995, Stiling and Rossi 1997). Top-down effects have also been found to impact species succession in Dutch marshes (Kuijper and Bakker 2005). Therefore, the traditional view of salt marsh food webs clearly needs to be replaced by a more nuanced view that considers interactions between top-down and bottom-up factors and how these interactions are affected by environmental heterogeneity (Hunter and Price 1992) across marsh landscapes (Denno et al. 2002, Goranson et al. 2004, Denno et al. 2005). ACKNOWLEDGMENTS We thank NSF (DEB-0296160 and OCE99-82133) and the NOAA NERR GRF program (NA04NOS4200137) for funding, and the ACE Basin Reserve, South Carolina (USA), for serving as the host reserve for the NOAA GRF. We thank the University of Georgia Marine Institute (UGAMI) for facilitating our field work. This paper is UGAMI contribution number 958. We thank R. Denno, E. Preisser, E. Siemann, B. Cole, M. Travisano, and three anonymous reviewers for advice and comments on the manuscript, and S. Newell for identifying the epiphyllic fungus. This work would not be done without support from C.-Y. Ho, F.-J. Sha, H.-C.-M. Ho, Y. Chung, S. Wu, S. Liaw, J. Lin, I. Jian, H. Yen, D. Gill, D. Inouye, J. Dietz, C. Salgado, N. Dave, C. Gormally, and I. Sellers. This work is a product of the Georgia Coastal Ecosystems LTER program. LITERATURE CITED Abele, L. G. 1992. A review of the grapsid crab genus Sesarma (Crustacea: Decapoda: Grapsidae) in America, with the description of a new genus. Smithsonian Contributions to Zoology 527. Borer, E. T., E. W. Seabloom, J. B. Shurin, K. E. Anderson, C. A. Blanchette, B. Broitman, S. D. Copper, and B. S. Halpern. 2005. What determines the strength of a trophic cascade? Ecology 86:528–537. Buck, T. L., G. A. Breed, S. C. Pennings, M. E. Chase, M. Zimmer, and T. H. Carefoot. 2003. Diet choice in an omnivorous salt-marsh crab: different food types, body size, and habitat complexity. Journal of Experimental Marine Biology and Ecology 292:103–116. Coll, M., and M. Guershon. 2002. Omnivory in terrestrial arthropods: mixing plant and prey diets. Annual Review of Entomology 47:267–297. Denno, R. F., C. Gratton, H. Dobel, and D. Finke. 2003. Predation risk affects relative strength of top-down and

1721

bottom-up impacts on insect herbivores. Ecology 84:1032– 1044. Denno, R. F., C. Gratton, M. A. Peterson, G. A. Langellotto, D. L. Finke, and A. F. Huberty. 2002. Bottom-up forces mediate natural-enemy impact in a phytophagous insect community. Ecology 83:1443–1458. Denno, R. F., D. Lewis, and C. Gratton. 2005. Spatial variation in the relative strength of top-down and bottomup forces: causes and consequences for phytophagous insect populations. Annales Zoologici Fennici 42:295–311. Diehl, S. 1993. Relative consumer sizes and the strengths of direct and indirect interactions in omnivorous feeding relationships. Oikos 68:151–157. Diehl, S. 2003. The evolution and maintenance of omnivory: dynamic constraints and the role of food quality. Ecology 84: 2557–2567. Dorn, N. J., and J. M. Wojdak. 2004. The role of omnivorous crayfish in littoral communities. Oecologia 140:150–159. Eubanks, M. D., and R. F. Denno. 1999. The ecological consequences of variation in plants and prey for an omnivorous insect. Ecology 80:1253–1266. Fagan, W. F. 1997. Omnivory as a stabilizing feature of natural communities. American Naturalist 150:554–567. Finke, D. L., and R. F. Denno. 2002. Intraguild predation diminished in complex habitats: implications for top-down suppression of prey populations. Ecology 28:643–652. Finke, D. L., and R. F. Denno. 2003. Intra-guild predation relaxes natural enemy impacts on herbivore populations. Ecological Entomology 28:67–73. Futuyma, D., and S. S. McCafferty. 1990. Phylogeny and the evolution of host plant associations in the leaf beetle genus Ophraella (Coleoptera, Chrysomelidae). Evolution 44:1885– 1913. Goranson, C. E., C.-K. Ho, and S. C. Pennings. 2004. Environmental gradients and herbivore feeding preferences in coastal salt marshes. Oecologia 140:591–600. Greenberg, R. 1981. The abundance and seasonality of forest canopy birds on Barro Colorado Island, Panama. Biotropica 13:241–251. Hacker, S. D., and M. D. Bertness. 1995. A herbivore paradox: why salt marsh aphids live on poor-quality plants. American Naturalist 145:192–210. Hairston, N. G., F. E. Smith, and L. B. Slobodkin. 1960. Community structure, population control, and competition. American Naturalist 94:421–425. Hunter, M. D., and P. W. Price. 1992. Playing chutes and ladders: heterogeneity and the relative roles of bottom-up and top-down forces in natural communities. Ecology 73: 724–732. Hunter, M. D., G. C. Varley, and G. R. Gradwell. 1997. Estimating the relative roles of top-down and bottom-up forces on insect herbivore populations: a classic study revisited. Proceedings of the National Academy of Sciences (USA) 94:9176–9181. King, G. M., M. J. Klug, R. G. Wiegert, and A. G. Chalmers. 1982. Relation of soil water movement and sulfide concentration to Spartina alterniflora production in a Georgia salt marsh. Science 218:61–63. Kuijper, D. P. J., and J. P. Bakker. 2005. Top-down control of small herbivores on salt-marsh vegetation along a productivity gradient. Ecology 86:914–923. Lodge, D. M., M. W. Kershner, and J. E. Aloi. 1994. Effects of an omnivorous crayfish (Orconectes rusticus) on a freshwater littoral food web. Ecology 75:1265–1281. Menge, B. A. 1995. Indirect effects in marine rocky intertidal interaction webs: patterns and importance. Ecological Monographs 65:21–74. Meserve, P. L., B. K. Lang, and B. D. Patterson. 1988. Trophic relationships of small mammals in a Chilean temperate rainforest. Journal of Mammalogy 69:721–730.

1722

CHUAN-KAI HO AND STEVEN C. PENNINGS

Moon, D. C., and P. Stiling. 2002. The effects of salinity and nutrients on a tritrophic salt-marsh system. Ecology 83:2465– 2476. Moran, M. D., T. P. Rooney, and L. E. Hurd. 1996. Top-down cascade from a bitrophic predator in an old-field community. Ecology 77:2219–2227. Murdoch, W. W., J. Chesson, and P. L. Chesson. 1985. Biological control in theory and practice. American Naturalist 125:344–366. Oksanen, L., S. D. Fretwell, J. Arruda, and P. Niemela. 1981. Exploitation ecosystems in gradients of primary productivity. American Naturalist 118:240–261. Pace, M. L., J. J. Cole, S. R. Carpenter, and J. F. Kitchell. 1999. Trophic cascades revealed in diverse ecosystems. Trends in Ecology and Evolution 14:483–488. Paine, R. T. 1980. Food webs: linkage, interaction strength and community infrastructure. Journal of Animal Ecology 49: 667–685. Pennings, S. C., T. H. Carefoot, E. L. Siska, M. E. Chase, and T. A. Page. 1998. Feeding preferences of a generalist salt marsh crab: relative importance of multiple plant traits. Ecology 79:1968–1979. Pimm, S. L., and J. H. Lawton. 1978. On feeding on more than one trophic level. Nature 275:542–544. Polis, G. A. 1991. Complex trophic interactions in deserts: an empirical critique of food-web theory. American Naturalist 138:123–155. Polis, G. A. 1999. Why are parts of the world green? Multiple factors control productivity and the distribution of biomass. Oikos 86:3–15. Polis, G. A., and D. R. Strong. 1996. Food web complexity and community dynamics. American Naturalist 147:813–846. Pomeroy, L. R., and R. G. Wiegert. 1981. Ecology of a salt marsh. Springer-Verlag, New York, New York, USA. Power, M. E. 1990. Effects of fish in river food webs. Science 250:811–814. Pringle, C. M., and T. Hamazaki. 1998. The role of omnivory in a neotropical stream: separating diurnal and nocturnal effects. Ecology 79:269–280. Rosenheim, J. A., and A. Corbett. 2003. Omnivory and the indeterminacy of predator function: can a knowledge of foraging behavior help? Ecology 84:2538–2548. Schmitz, O. J., A. P. Beckerman, and K. M. O’Brien. 1997. Behaviorally mediated trophic cascades: effects of predation risk on food web interactions. Ecology 78:1388–1399.

Ecology, Vol. 89, No. 6

Schmitz, O. J., P. A. Hamback, and A. P. Beckerman. 2000. Trophic cascades in terrestrial systems: a review of the effects of carnivore removals on plants. American Naturalist 155: 141–153. Seiple, W. 1979. Distribution, habitat preferences and breeding periods in the crustaceans Sesarma cinereum and S. reticulatum (Brachyura: Decapoda: Grapsidae). Marine Biology 52:77–86. Showler, A. T., and S. M. Greenberg. 2003. Effects of weeds on selected arthropod herbivore and natural enemy populations, and on cotton growth and yield. Environmental Entomology 32:39–50. Shurin, J. B., E. T. Borer, E. W. Seabloom, K. Anderson, C. A. Blanchette, B. Broitman, S. D. Copper, and B. S. Halpern. 2002. A cross-ecosystem comparison of the strength of trophic cascades. Ecology Letters 5:785–791. Silliman, B. R., and M. D. Bertness. 2002. A trophic cascade regulates salt marsh primary production. Proceedings of the National Academy of Sciences (USA) 99:10500–10505. Spiller, D. A., and T. W. Schoener. 1990. A terrestrial field experiment showing the impact of eliminating top predators on foliage damage. Nature 347:469–472. Stiling, P., and A. M. Rossi. 1997. Experimental manipulations of top-down and bottom-up factors in a tri-trophic system. Ecology 78:1602–1606. Strong, D. R. 1992. Are trophic cascades all wet? Differentiation and donor-control in speciose ecosystems. Ecology 73: 747–754. Sullivan, M. J., and F. C. Daiber. 1974. Response in production of cord grass, Spartina alterniflora, to inorganic nitrogen and phosphorus fertilizer. Chesapeake Science 15(2):121–123. Tanabe, K., and T. Namba. 2005. Omnivory creates chaos in simple food web models. Ecology 86:3411–3414. Teal, J. M. 1958. Distribution of fiddler crabs in Georgia salt marshes. Ecology 39:185–193. Thompson, R. M., M. Hemberg, B. M. Starzomski, and J. B. Shurin. 2007. Trophic levels and trophic tangles: the prevalence of omnivory in real food webs. Ecology 88:612– 617. Valiela, I., and J. M. Teal. 1974. Nutrients limitation in salt marsh vegetation. Pages 547–563 in R. J. Reimold and W. H. Queen, editors. Ecology of halophytes. Academic Press, New York, New York, USA. Zimmer, M., S. C. Pennings, T. L. Buck, and T. H. Carefoot. 2004. Salt marsh litter and detritivores: a closer look at redundancy. Estuaries 27:753–769.

APPENDIX Iva food web based on observations on Sapelo Island, Georgia, USA (Ecological Archives E089-102-A1).