Estuaries and Coasts (2012) 35:475–485 DOI 10.1007/s12237-011-9458-7

Effects of an Omnivorous Katydid, Salinity, and Nutrients on a Planthopper-Spartina Food Web Juan M. Jiménez & Kazimierz Więski & Laurie B. Marczak & Chuan-Kai Ho & Steven C. Pennings

Received: 25 May 2011 / Revised: 19 October 2011 / Accepted: 22 October 2011 / Published online: 21 December 2011 # Coastal and Estuarine Research Federation 2011

Abstract Top–down and bottom–up effects interact to structure communities, especially in salt marshes, which contain strong gradients in bottom–up drivers such as salinity and nutrients. How omnivorous consumers respond to variation in prey availability and plant quality is poorly understood. We used a mesocosm experiment to examine how salinity, nutrients, an omnivore (the katydid Orchelimum fidicinium) and an herbivore (the planthopper Prokelisia spp.) interacted to structure a simplified salt marsh food web based on the marsh grass Spartina alterniflora. Bottom–up effects were strong, with both salinity and nutrients decreasing leaf C/N and increasing Prokelisia abundance. Top–down effects on plants were also strong, with both the herbivore and the omnivore affecting S. alterniflora traits and growth, especially when nutrients or salt were added. In

contrast, top–down control by Orchelimum of Prokelisia was independent of bottom–up conditions. Orchelimum grew best on a diet containing both Spartina and Prokelisia, and in contrast to a sympatric omnivorous crab, did not shift to an animal-based diet when prey were present, suggesting that it is constrained to consume a mixed diet. These results suggest that the trophic effects of omnivores depend on omnivore behavior, dietary constraints, and ability to suppress lower trophic levels, and that omnivorous katydids may play a previously unrecognized role in salt marsh food webs. Keywords Top–down and bottom–up interactions . Orchelimum fidicinium . Omnivory . Prokelisia spp. . Salt marsh . Spartina alterniflora

Introduction Electronic supplementary material The online version of this article (doi:10.1007/s12237-011-9458-7) contains supplementary material, which is available to authorized users. J. M. Jiménez : K. Więski : L. B. Marczak : C.-K. Ho : S. C. Pennings (*) Department of Biology and Biochemistry, University of Houston, Houston, TX 77204-5001, USA e-mail: [email protected] Present Address: L. B. Marczak Department of Ecosystem and Conservation Sciences, College of Forestry and Conservation, University of Montana, Missoula, MT 59812, USA Present Address: C.-K. Ho Institute of Ecology and Evolutionary Biology, National Taiwan University 1, Sec. 4, Roosevelt Rd., Taipei 106 Taiwan, Republic of China

Herbivores play a central role in community structure because they both support higher trophic levels and regulate plant community structure and biomass; however, understanding what regulates herbivore abundance is complicated. Herbivore densities can be affected by top–down forcing from predators (Oksanen et al. 1981), side-to-side forces from competitors (Moon and Stiling 2002) and bottom–up forcing by plants (Hunter and Price 1992). Because these potential drivers can interact in many ways, the study of herbivore population dynamics is a challenging task (Power 1992; Strong 1992). In salt marshes, the strong abiotic gradients that occur across the marsh in flooding and salinity, and the variation in nutrient input that different estuaries receive from natural or anthropogenic sources, can all be important bottom–up factors regulating plant biomass and performance (Pennings and Bertness 2001). A number of studies have examined

476

how top–down and bottom–up forces interact to regulate the abundance of herbivores in salt marshes (Bertness et al. 2008; Sala et al. 2008). In particular, top–down control of planthopper herbivores on Borrichia shrubs in Florida saltmarshes is decreased by abiotic stress and increased by eutrophication (Moon and Stiling 2000, 2002, 2004). Similarly, top–down control of planthopper herbivores on Spartina alterniflora grasses in New Jersey saltmarshes is mediated by interactions between nutrients and spider abundance, with Prokelisia spp. able to achieve very high densities when plant nutrition is high and predators rare, conditions typical of tall-form S. alterniflora plants along creek edges (Denno et al. 2002, 2003, 2005; Wimp et al. 2010). These studies and others have laid a foundation for understanding how top–down and bottom–up factors interact to structure communities. These studies, however, examined food webs where the top consumers were obligate predators and parasitoids, feeding solely on herbivores. Both intraguild predators and true omnivores, however, are common in natural communities in general (Polis and Strong 1996) and in salt marshes in particular (Buck et al. 2003; Ho and Pennings 2008; Wason and Pennings 2008). A few studies have examined the trophic effects of intraguild predators, finding that intraguild predators often interfere with each other, weakening trophic cascades (Finke and Denno 2004, Matsumura et al. 2004). Less is known about the trophic effects of true omnivores, which feed on both animals and plants (Eubanks 2005). Because true omnivores feed on both plants and animals, they might respond to bottom–up changes in plant quality by shifting their diet to include more or less plant material. As a result, their trophic impacts might differ considerably from those of strict predators (Bruno and O’Connor 2005; Long et al. 2011), limiting our ability to construct nomothetic theories of ecosystem structure. In this paper, we examine how salinity and nutrient supply interact with the abundance of Orchelimum fidicinium Rehn and Hebard, an omnivorous katydid, to regulate the abundance of Prokelisia spp. (Prokelisia marginata Van Duzee and Prokelisia dolus Wilson) herbivores feeding on the salt marsh grass S. alterniflora Loisel in Georgia salt marshes (Electronic Supplementary Material). We tested three groups of hypotheses. First, we hypothesized that nutrient addition would directly increase Prokelisia spp. herbivore populations by increasing plant nitrogen content. Similarly, because S. alterniflora uses nitrogenous compounds for osmoregulation, we hypothesized that salinity stress would also increase plant nitrogen content and Prokelisia spp. populations. Second, because O. fidicinium is omnivorous, eating both plants and herbivores, we hypothesized that it would have opposing effects on plant size, directly reducing plant size by eating plants, and

Estuaries and Coasts (2012) 35:475–485

indirectly increasing plant size by eating herbivores, leading to weak net effects. Moreover, we hypothesized that both nutrient addition and salinity stress would reduce top–down control of Prokelisia spp. by O. fidicinium by decreasing the C/N ratio of plants (making plants a more attractive alternate diet for O. fidicinium) and increasing Prokelisia spp. abundance (leading to herbivore escape from control by O. fidicinium). Third, we compared our results with those of a similar study by Ho and Pennings (2008) of an omnivorous crab feeding on aphids and a salt marsh shrub to test the hypothesis that these two omnivores would have similar effects on trophic structure.

Methods S. alterniflora (hereafter, Spartina) is the most common salt marsh plant on the Atlantic and Gulf coasts of the USA, dominating lower and middle marsh elevations (Pennings and Bertness 2001). Spartina stands are exposed to wide gradients of nutrient availability, flooding, and salinity. These gradients are partly responsible for the occurrence of different height forms of S. alterniflora (Pennings and Bertness 2001), with shorter plants (10–40 cm in height) occupying higher marsh elevations than taller plants (1–2 m in height), which have a higher nitrogen content and dominate lower marsh elevations. At a given elevation, plants exposed to high salinities may respond by increasing production of nitrogen-based compounds to reduce osmotic stress (Morris 1984), potentially increasing plant quality for herbivores. The most common herbivores of Spartina are the delphacid planthoppers P. marginata Van Duzee and P. dolus Wilson (Denno et al. 1987). P. marginata and P. dolus (hereafter, Prokelisia) are multivoltine, with a summer generation time of 5–6 weeks (Denno et al. 1985; Stiling et al. 1991). They are phloem sap-feeders that can reach densities of up to 1,000 adults and 100,000 nymphs/m2 (Denno et al. 1987). The most common true omnivore in the Spartina zone of salt marshes on the southeastern coast of the United States is the tettigoniid katydid O. fidicinium (hereafter, Orchelimum) (Smalley 1960; Wason and Pennings 2008). Although historically regarded as an herbivore (Teal 1962), Orchelimum, like most tettigoniids, is omnivorous, and readily eats small arthropods. For example, individual Orchelimum ate 7.8 Prokelisia planthoppers individuals per day (SE, 1.4; n=10) when housed in 1-L jars for 48 h in the laboratory (Wason and Pennings, personal observations) and grew to adulthood when fed a diet consisting only of Prokelisia (see “Results”). Orchelimum is univoltine, with adult densities reaching 9.6 individuals/m2 (Smalley 1960; Stiling et al. 1991).

Estuaries and Coasts (2012) 35:475–485

Experimental work was done in mesocosms on Sapelo Island, GA (81.2699° W, 31.3929° N). Two hundred Spartina plants (~120 cm tall) were collected from the Dean Creek salt marsh at the southern end of the island in late June, 2005, and potted in 3.5-L plastic pots in 50:50 potting soil/sand. After 20 days, 112 healthy plants were haphazardly assigned to four combinations of nutrients (ambient and fertilized) and salinity (low and high) and acclimated for 7 weeks. Nutrient and salinity treatments were initiated with low doses (5 g) of nitrogen-based fertilizer (20:10:5 N/P/K fertilizer pellets, Forestry Suppliers) and low salinity concentrations (10). Nutrient additions and salinity levels were progressively increased every two weeks to final levels of 21 g of fertilizer per month and a salinity of 25. These nutrient and salinity combinations were not designed to mimic particular field situations, but rather to affect plant traits sufficiently to test the hypothesis that bottom–up conditions would alter top–down control. Both treatments were sufficient to affect plant traits and herbivore abundances (see “Results”). Plants were watered daily with fresh or saline water and allowed to drain fully. The omnivore (present/absent) and herbivore (present/ absent) treatments were combined in a factorial design with the nutrient and salinity treatments, for a total of 16 treatment combinations (2 omnivore × 2 herbivore × 2 nutrient×2 salinity), with seven replicates each. Arthropods were enclosed on a single Spartina plant in a 1.4-m high cage made of fine cloth mesh. The cloth reduced PAR (photosynthetically available radiation) levels inside cages to ~70% of ambient. Mesocosms were located outdoors under a plastic roof that sheltered them from direct rain but exposed them to ambient temperature and humidity conditions. Mesocosms representing different treatment combinations were interspersed to prevent any location effects. Prokelisia and Orchelimum were collected from marshes around Sapelo Island. Treatments including Orchelimum contained a single katydid; plants were checked daily and katydids that died were replaced. Treatments including Prokelisia were initiated with 30 nymphs and supplemented weekly with 40 adults to represent immigrants (Denno et al. 1985). Examination of a subset of the Prokelisia individuals suggested that we stocked 80% P. marginata and 20% P. dolus. Once arthropods were stocked, the experiment ran for one month (26 August to 27 September 2005). At the end of the experiment, we counted all the arthropods present on the plants. We measured several plant traits at the beginning and end of the experiment. As indicators of plant size and herbivory, we measured plant height, leaf width, and the number of healthy and damaged leaves, scoring damage from both Prokelisa (visible as yellow spots against the normal dark green leaves) and Orchelimum (visible as missing leaf material). Because Orchelimum damaged every leaf of the Spartina plants when present, we could not

477

analyze variation in Orchelimum damage, and did not consider final leaf number a useful indicator of plant size. We calculated plant growth as the difference between initial and final plant height. As indicators of leaf quality, we measured chlorophyll content using an optical meter (Opti-Sciences CCM-200 chlorophyll meter) and leaf toughness using a penetrometer (model 516, Chatillon N.Y.) at the end of the experiment. Leaf chlorophyll content was negatively correlated with leaf C/N ratio (n=80, r=0.61, P<0.0001), based on a subset of leaves collected at the end of the experiment (Perkin-Elmer 2400 CHN analyzer). Because we fixed Orchelimum density at one individual per cage, replacing individuals that died, we did not examine bottom–up effects on Orchelimum. Data were analyzed with four-way ANOVA using JMP. Nutrients, salinity, herbivore addition, and omnivore addition were treated as fixed effects. C/N, chlorophyll content and Prokelisia abundance were log-transformed to improve normality and homoscedasticity. For clarity, we show only significant effects in the figures (i.e., data are presented pooled across non-significant terms). To examine multivariate relationships among the response variables, and to examine the generality of our results, we used structural equation modeling (SEM) to compare a subset of our treatments (omitting the nutrient and salt addition treatment combinations) with parallel treatments in a similar study of top–down control by the omnivorous crab Armases cinereum of aphids and a salt marsh shrub (Ho and Pennings 2008). SEM is a multivariate technique that allows one to examine the strength of direct and indirect effects between causally related variables (Grace 2006). SEM is superior to ANOVA for some analyses because it can consider variables (e.g., herbivore damage to leaves) as both response variables and simultaneously as drivers of other response variables. SEM starts with a construction of a path diagram based on a priori known relationships in the system studied. In the next step, the parameters of the SEM model are estimated by scanning the matrix of covariances from the source data over the hypothesized model. Finally, based on the estimated model parameters, a predicted matrix of covariances is calculated and its fit is compared with the observed matrix. As with multiple regression, the “best” SEM model can include some marginally-significant terms. Parameters were estimated with AMOS 7.0 using the maximum likelihood method, and model fit was tested by the likelihood chi-square value. To determine the importance of an omnivorous diet to growth of Orchelimum, we conducted a laboratory growth experiment. We collected 32 Orchelimum nymphs in early July of 2010 from the field at Sapelo Island. Individuals were measured (live mass) and haphazardly assigned to four diet treatments (n=8 reps/treatment): live Spartina (S), live

478

Estuaries and Coasts (2012) 35:475–485

Spartina and Prokelisia (SP), dead Spartina (D), and dead Spartina plus Prokelisia (P). Orchelimum offered only dead Spartina did not feed and died within 10 days; therefore, this treatment was not considered further, and we considered the dead Spartina plus Prokelisia treatment to essentially be a Prokelisia only treatment (with the dead Spartina providing a perch). Individual Orchelimum were housed in 1 L glass jars, stocked with a single ~50-cm Spartina leaf that was either live or dead, as appropriate. A small amount of fresh water was added daily to each replicate to maintain humidity. The experiment was run indoors at ~25°C and a photoperiod of 14:10. For treatments containing Prokelisia, additional Prokelisia adults and nymphs were added every 2 days. Live and dead Spartina leaves were also replaced every 2 days. All diets were ad libitum, with food available in excess of what Orchelimum consumed. Orchelimum were weighed weekly for five weeks.

and Orchelimum on plants, especially when nutrients or salt were added and (3) strong top–down effects of Orchelimum on Prokelisia. Comparison of these results with a parallel study of a crab-aphid-shrub food web indicated (4) that katydids and crabs have different effects as top omnivores. Finally, growth studies indicated (5) that Orchelimum grows best on a mixed diet containing both plant and animal prey. Below, we present these results in this order.

In the mesocosm experiment, we found (1) strong bottom– up effects of nutrients and salt that affected both plants and Prokelisia, (2) strong top–down effects from both Prokelisia

Bottom–up effects Both nutrient addition and salinity had strong bottom–up effects on Spartina and Prokelisia (Table 1). Nutrient addition increased Spartina leaf chlorophyll (Fig. 1a) and reduced leaf C/N ratios (Fig. 1b). Nutrient addition decreased leaf toughness, but only when Orchelimum was present (Fig. 1c). Either salt or nutrient addition alone increased Prokelisia abundance by ~75 % (Fig. 1d) and decreased leaf C/N by 17–36% (Fig. 1e); however, when both salt and nutrients were added in combination, leaf C/N did not decrease much further, and Prokelisia abundance did not increase further (significant salinity x nutrient interactions, Table 1). Finally, nutrient addition moderately decreased the number of leaves damaged by Prokelisia (Fig. 1f). Across all treatments with Prokelisia, the number of leaves damaged

Table 1 Summary of ANOVA results for the effects of salinity stress (low or high), nutrient availability (ambient or addition), herbivore treatment (present or absent), omnivore treatment (present or absent)

and their interactions on C/N ratio, chlorophyll content, leaf toughness, plant growth, leaf width, Prokelisia abundance, and Prokelisia damage

Results

Source

P C/N ratio

Chlorophyll content

Leaf toughness

Plant growth

Leaf width

Prokelisia abundance

Prokelisia damage

Salinity (S) Nutrients (N) S*N Herbivore (H) S*H N*H

0.04 <0.0001 0.02 0.14 0.09 0.04

0.20 <0.0001 0.10 0.005 0.18 0.01

0.99 0.045 0.51 0.001 0.78 0.14

0.13 0.57 0.65 0.59 0.46 0.52

0.008 0.81 0.67 0.37 0.14 0.81

0.21 0.04 0.01 NT NT NT

0.22 0.02 0.47 NT NT NT

S*N*H Omnivore (O) S*O N*O S*N*O H*O S*H*O N*H*O S*N*H*O Initial number of leaves

0.76 0.03 0.88 0.15 0.03 0.78 0.14 0.61 0.40 NS

0.69 0.89 0.80 0.46 0.09 0.16 0.87 0.64 0.99 NS

0.48 0.37 0.33 0.047 0.43 0.45 0.89 0.58 0.07 NS

0.02 <0.0001 0.18 0.04 0.61 0.52 0.55 0.07 0.69 NT

0.017 0.96 0.08 0.60 0.006 0.96 0.74 0.017 0.668 NT

NT 0.005 0.71 0.83 0.87 NT NT NT NT NS

NT 0.002 0.92 0.64 0.61 NT NT NT NT 0.008

Full ANOVA results are presented in digital Electronic Supplementary Material. Initial number of leaves was initially included as a covariate in tests where appropriate but dropped if not significant Significant p-values (<0.05) are indicated with italics NT not tested, NS not significant (and dropped from model)

Estuaries and Coasts (2012) 35:475–485

479

Fig. 1 Bottom–up effects of nutrients and salinity, and interactions with presence of herbivorous Prokelisia planthoppers and omnivorous Orchelimum katydids, on plant a chlorophyll content, b C/N ratio, and c leaf toughness (in Newtons); interactive effects of nutrient and salinity on d Prokelisia abundance and e plant C/N ratio; bottom–up effects of nutrient addition on f the number of leaves damaged by Prokelisia; and g relationship between the number of leaves damaged by Prokelisia and leaf chlorophyll content (y=−0.5x+0.19, n=56, R2 =0.16, P=0.002). H+ herbivore present, H− herbivore absent, O+ omnivore present, O− omnivore absent, S+ salt added, So no salt added; black bars, no nutrient addition (No); white bar: nutrient addition (N+). ANOVA statistics are reported in Table 1 and Electronic Supplementary Material. Data are means+1 SE

by Prokelisia decreased with increasing chlorophyll content (Fig. 1g), suggesting that Prokelisia needed to feed less extensively on nutrient-rich plants. There was no interaction between salinity and nutrients for chlorophyll content, leaf toughness or Prokelisia damage (Table 1). Top–down Effects on Plants Prokelisia and Orchelimum both had strong top–down effects on plants, and Orchelimum also had a strong top–down effect on Prokelisia (next paragraph); however, only the top–down effects on plants were mediated by bottom–up conditions (Table 1). Prokelisia reduced leaf toughness (Fig. 2a), and reduced chlorophyll

content (Fig. 1a) and increased C/N ratio when nutrients were added (Fig. 1b, significant nutrient x herbivore interactions, Table 1). Prokelisia reduced plant growth, especially when nutrients and salt were both added (Fig. 2b; salinity×nutrients×herbivore interaction, Table 1). Changes in leaf width depended on three way interactions between salt, nutrients, herbivores, and omnivores: leaf width decreased in all treatments when salt was added; however, in the no salt treatment, leaf width increased when either omnivores or herbivores were present and nutrients were added (significant salinity ×nutrients× omnivore and salinity×nutrients×herbivore interactions,

480

Leaf toughness (N)

No 15

N+

C

A

40 10 20

5 0

0 H-

H+

O-

O+

OS+

D

B

O+

15 10

10

5

5 0

Plant growth (cm)

Plant growth (cm)

So

15

C:N Ratio

Fig. 2 Top–down effects of herbivorous Prokelisia planthoppers and omnivorous Orchelimum katydids and interactions with nutrients and salinity, on a leaf toughness (in Newtons); b plant growth; c plant C/N ratio and d plant growth. H+ herbivore present, H− herbivore absent, O+ omnivore present, O− omnivore absent, S+ salt added, So no salt added; black bars, no nutrient addition (No); white bar, nutrient addition (N+). ANOVA statistics are reported in Table 1 and Electronic Supplementary Material. Data are means+1 SE

Estuaries and Coasts (2012) 35:475–485

0 So

Table 1). Orchelimum increased leaf toughness in unfertilized treatments (Fig. 1c) and interacted with nutrients and salt to affect leaf C/N ratios: fertilization decreased all C/N ratios, but the largest fertilization effect occurred when Orchelimum was absent and salt was not added (Fig. 2c; significant salinity×nutrient×omnivore interaction, Table 1). Additionally, Orchelimum reduced plant growth (change in height), but did so most effectively when nutrients were added (Fig. 2d; significant nutrients × omnivore interaction, Table 1). Nutrient addition had different effects on leaf width depending on the presence of Orchelimum and Prokelisia: leaf width increased with nutrients in the presence of Orchelimum, but decreased when both Orchelimum and Prokelisia were present (significant nutrient×omnivore× herbivore interaction, Table 1). Top–down Effects on Herbivores Orchelimum reduced Prokelisia abundance by ~50% (Fig. 3a) and thereby indirectly reduced the number of leaves damaged by Prokelisia by 25% (Fig. 3b). Top–down control of Prokelisia populations and the trophic cascade on the number of leaves damaged by Prokelisia were strong regardless of bottom–up conditions (no significant bottom–up×top–down interactions for Prokelisia abundance or damage, Table 1). Relationships Among Response Variables and Comparison with Omnivorous Crabs SEM analysis of parallel food web experiments with an omnivorous salt marsh crab (Ho and

H-

H+

O-

O+

S+

Pennings 2008) and Orchelimum indicated important differences in the way that these two omnivores structure their communities. SEM analysis of the mesocosm exper-

Prokelisia abundance (#)

H+

500 400 300 200 100 0 12

Prokelisia damage (#of leaves)

H-

10 8 6 4 2 0 O-

O+

Fig. 3 Top–down effects from omnivorous Orchelimum katydids on a herbivorous Prokelisia planthoppers and b the number of leaves damaged by Prokelisia. O+ omnivore present, O− omnivore absent. ANOVA statistics are reported in Table 1 and Electronic Supplementary Material. Data are means+1 SE

Estuaries and Coasts (2012) 35:475–485

iment in Ho and Pennings (2008) indicated that the omnivorous crab Armases negatively affected aphids, negatively affected growth of the Iva shrub, and indirectly increased Iva leaf area through its negative effect on aphids (Fig. 4a). Aphids had a significant top–down effect only on Iva leaf area; however, the best SEM model retained nonsignificant top–down effects of aphids on leaf number and plant growth, suggesting that these effects were likely also important. SEM analysis of the no nutrient, no salt treatments in the Spartina mesocosm experiment (this paper) indicated that Orchelimum had stronger direct effects on plants and weaker effects on herbivores than did Armases. Orchelimum reduced Spartina leaf width and growth (change in height), negatively (but weakly) affected Prokelisia numbers and indirectly increased leaf width by reducing damage from Prokelisia (Fig. 4b). Prokelisia did

Fig. 4 SEM path diagrams for a interactions in the experimental Iva food web from Ho and Pennings (2008) (p=0.13, Chi-square/df=1.58) and b a subset of the experimental Spartina food web (this paper) (p=0.49, Chi-square/df=0.96). Path coefficients represent standardized values showing relative effects of variables upon each other; solid lines indicate statistically significant paths; dashed lines indicate non-significant paths retained to improve overall model fit. Arrow width is proportional to path coefficient strength; one-headed arrows depict causal relationships; two-headed arrows depict correlations

481

not affect plant growth, although Prokelisia damage reduced leaf width; this difference between the ANOVA and SEM results occurred because the top–down effects of Prokelisia on plant growth were strongest when salt and nutrients were added, but these treatments were not included in the SEM analysis because we used only treatment combinations that were parallel to those used by Ho and Pennings (2008). Benefits of Mixed Diets In the laboratory, Orchelimum fed a mixed diet of both Spartina and Prokelisia attained a body mass 25% greater than individuals fed either diet alone (Fig. 5). Only 2 of 24 individuals died during the experiment (both from the Spartina-only diet; these were not included in calculations of final mass), and all but one of the survivors had molted into the final instar after 5 weeks.

482

Estuaries and Coasts (2012) 35:475–485

Orchelimum body mass (g)

0.25 0.20

a b

b

0.15 0.10 0.05 0.00

S

PS

P

Fig. 5 Body mass of O. fidicinium after 5 weeks of feeding on Spartina (S), Prokelisia and Spartina (PS), and Prokelisia (P). Data are means+1 SE. ANCOVA with initial mass as the covariate, F(2, 17)=7.18, p<0.01

Discussion We correctly predicted (hypothesis 1) that nutrients and salt would increase the nitrogen content of Spartina and benefit herbivorous planthoppers. In contrast, we largely failed (hypothesis 2) to predict the effects of the omnivorous katydid on the planthopper-Spartina food web, illustrating the difficulty in predicting the effects of omnivores on food web structure and processes. The comparison of the omnivorous katydid and omnivorous crab (hypothesis 3) suggested that variation in diet flexibility may mediate food web effects of omnivores. Hypothesis 1: Effects of Nutrients and Salt on Plants and Herbivores We hypothesized that nutrient and salt additions would directly increase Prokelisia herbivore populations by increasing plant quality. This hypothesis was supported. Nitrogen addition increased leaf chlorophyll content, decreased leaf C/N and toughness, and increased the abundance of Prokelisia. This result was consistent with previous studies that have found that fertilizing Spartina plants in the field decreases C/N ratio and increases Prokelisia abundance (Gratton and Denno 2003; McFarlin et al. 2008). Similarly, salinity addition decreased leaf C/N and increased Prokelisia abundance. This result was also consistent with previous studies that have found that higher salinities decrease Spartina C/N ratio (Moon et al. 2000; Moon and Stiling 2000, 2004). Salt marsh plants are known to increase production of nitrogen-based compounds to reduce osmotic stress when growing in high-salinity soils (Morris 1984). Given that the high-salinity treatment in our experiment (25) was well within the levels that Spartina can tolerate (Richards et al. 2005), it is likely that our “high”salinity treatment was only a moderate stress that did not highly stress the plants. Moon and Stiling (2005), working

with a salt marsh shrub, found that intermediate levels of chronic salinity stress stimulated a planthopper herbivore (Pissonotus quadripustulatus, Homoptera: Delphacidae) but that high levels of chronic salinity had negative effects. Given that we did not examine extremely high salinity levels (Spartina occurs in soils with porewater salinities of >60), it is possible that negative effects of salinity on herbivores would have prevailed at higher salinities. Because both nutrients and salinity decreased plant C/N ratio and increased Prokelisia abundance, one might have expected that their combination would affect plant C/N and planthopper abundance even further, but this did not occur. Moon and Stiling (2000) and Stiling and Moon (2005) obtained a similar result for P. quadripustulatus feeding on the shrub Borrichia. Although the mechanism underlying this result was not explored in any of these cases, the fact that nitrogen and salinity have less than additive effects on herbivores is important, because it means that the effects of nitrogen addition on herbivores cannot be predicted without also considering the salinity regime experienced by the plants. Because both nutrients and salinity decreased plant C/N ratio and increased Prokelisia densities, one might expect to find increased suppression of plant growth in these treatments. This did happen, but only when both nutrients and salt were added. Previous studies with the salt marsh shrub Borrichia frutescens (Moon and Stiling 2002, 2004; Stiling and Moon 2005) and with Spartina (Denno et al. 2002; Sala et al. 2008) also found positive effects of nutrients on the top–down effect of herbivores on plants. Similarly, some studies have found increased effects of herbivores on salt-stressed S. alterniflora (Silliman et al. 2005). Why we only found increased top–down effects when both nutrients and salt were added is not clear, but the general pattern of the results is consistent with these previous studies. Hypothesis 2: Effects of Orchelimum on Plants and Herbivores Our hypotheses about top–down effects from Orchelimum were not as often correct. Because Orchelimum is omnivorous, eating both plants and herbivores, we hypothesized that it would have both direct negative and indirect positive effects on plants, leading to weak net effects. In contrast to our expectations, Orchelimum strongly suppressed plant growth (determined as change in height). The SEM analysis confirmed that Orchelimum did have an indirect positive effect on Spartina by reducing Prokelisia abundance, which in turn reduced the number of leaves damaged by Prokelisia. However, although Prokelisia affected leaf chlorophyll content, width, and toughness, it did not affect plant growth, and thus the beneficial effect of Orchelimum on Spartina (reduced leaf damage from Prokelisia) did not

Estuaries and Coasts (2012) 35:475–485

translate into a growth benefit, and its direct negative effects were far stronger than its indirect positive ones (Fig. 4). As a result, Orchelimum had strong net negative effects on plants, not weak net effects. In general, it has proven easier to demonstrate trophic cascades that affect herbivore damage to plants than to demonstrate cascades that affect plant growth (Schmitz et al. 2000). Other studies have demonstrated negative effects of Prokelisia on Spartina growth (Denno et al. 2002; Finke and Denno 2004); however, and it is likely that we would have found such an effect had the experiment continued longer than one month. We further hypothesized that both nutrient addition and salinity stress would reduce top–down control of Prokelisia by Orchelimum by decreasing the C/N ratio of plants (making plants a more attractive alternate diet for Orchelimum) and increasing Prokelisia abundance (leading to herbivore escape from control by Orchelimum). Only some aspects of this hypothesis were supported. As discussed above, nutrient and salt addition did decrease plant C/N ratios and did increase Prokelisia abundances; however, the top–down effect of Orchelimum on Prokelisia was not affected by nutrient or salinity additions (Table 1). The top– down effect of Orchelimum on plants, however, strongly increased in the fertilized treatment, and in this way the effects of Orchelimum paralleled those of herbivores. A number of studies have shown that predators (mostly spiders) can exert strong top–down control on Prokelisia populations (Denno et al. 2003, 2005; Finke and Denno 2004). The role of Orchelimum and related katydids as top– down consumers in salt marshes, however, has been overlooked since the foundational work of Teal (1962) and Smalley (1960) mis-identified these species as strict herbivores. Based on the premise that Orchelimum was a strict herbivore, subsequent experimental work that detected decreases in Prokelisia when Orchelimum was present attributed the results to competition (Stiling et al 1991). We observed, however, that Orchelimum vigorously hunts and consumes Prokelisia, that Prokelisia actively hides from Orchelimum behind Spartina leaves, and we documented that Orchelimum benefits from including both animal and plant material in its diet (Fig. 5). Several species of Orchelimum and related katydids in the genus Conocephalus occur in North American salt marshes (Wason and Pennings 2008). All are likely omnivorous (Goeriz Pearson et al. 2011). We suggest that these katydids may play a previously overlooked role in mediating the densities of small arthropods in salt marsh food webs, and that these effects merit additional study. Previous studies looking for interactions between nutrient or salt addition and predator pressure on salt marsh herbivores found that nutrient additions can strengthen top–down forces on herbivores, while salinity stress can weaken them (Denno

483

et al. 2002; Moon and Stiling 2002, 2004). One mechanism that might have reduced the importance of interactions between top–down and bottom–up factors at the herbivore level in this experiment is that the mesocosm cages prevented immigration and emigration of arthropods. In particular, migration of Prokelisia from low to high-quality plants is known to mediate the strength of top–down interactions in the field (Denno et al. 2003, 2005). It is likely that Orchelimum also moves across the marsh to follow high-quality plants or abundant prey, with similar effects, but this has not been documented. Hypothesis 3: Different Omnivores have Similar Trophic Effects Finally, we compared our results with those of a similar study by Ho and Pennings (2008) of an omnivorous crab feeding on aphids and a salt marsh shrub to test the hypothesis that these two omnivores would have similar effects on trophic structure. This hypothesis was rejected, with Orchelimum having strongest effects on plants whereas the crab had similar effects on plants and herbivores (Fig. 4). One benefit of an omnivorous diet is that the consumer can potentially switch between various food types depending on their availability and quality. Thus, the omnivorous crab A. cinereum consumed fewer Iva frutescens leaves when aphids were present as an alternative, preferred food source (Ho and Pennings 2008). We expected Orchelimum to switch in the same way from feeding on Spartina when Prokelisia were absent to feeding mostly on Prokelisia when it was present. Contrary to this expectation, Orchelimum always heavily damaged plant leaves, and suppressed Spartina growth whether or not Prokelisia was present. Whether omnivory should in general strengthen or weaken trophic cascades is unclear (Polis and Strong 1996; Eubanks 2005). The answer likely depends in part on the omnivore’s feeding preferences for plant versus animal prey, and on the strength of its negative effects on different trophic levels. In our study, Orchelimum suppressed both Prokelisia and Spartina, but the direct negative effects on Spartina dominated the food web (Fig. 4). Conversely, the omnivorous crab Armases preferred to eat aphids, and switched foods to essentially act as a predator whenever aphids were present, producing a trophic cascade that benefitted Iva plants (Ho and Pennings 2008). This difference is reflected in the stronger direct link from omnivore to herbivore in the SEM analysis of the Iva food web versus the Spartina food web. The different food web effects of these two omnivores are a function of their dietary flexibility: Armases grows as well on a diet of animal prey as on a mixed diet (Buck et al. 2003) and so can readily switch to

484

an animal-only diet when animal preys are abundant. In contrast, Orchelimum grows best on a mixed diet (Fig. 5) and therefore is less flexible in its feeding behavior. In other words, whereas plants and herbivores may be somewhat substitutable foods (Van-Rijn and Sabelis 2005) for Armases, they are complimentary for Orchelimum, meaning that both are required in the diet. We suggest that one key to understanding the trophic effects of omnivores will be developing the ability to predict the flexibility of the diet of different omnivore species. In addition to our comments about possible cage effects above, we mention two caveats to our results. First, the katydid densities that we used were at the high end of natural densities. Thus, the top–down effects that we documented likely represent maximum levels, although the fact that we supplemented Prokelisia stocks weekly ameliorates this concern. Second, the lower salinity treatment (0 ppt) rarely occurs in the field except temporarily following heavy rain, and both the salinity and nutrient treatments likely did not encompass the full range of variation in these factors that occurs in the field. Had we used a wider range of bottom–up conditions, we might have seen stronger bottom–up effects. Despite these caveats, our results are both consistent with and shed new insight into past field studies.

Conclusions In summary, we found that both top–down and bottom–up forces were strong in this salt marsh system. Salinity and nutrients, important bottom–up factors in salt marshes, interacted in a less than additive way to decrease plant C/N ratios and increase herbivore populations. Herbivores and omnivores both suppressed plants, although the strength of this effect varied with nutrient and salinity conditions. In contrast to our expectations, top–down control of herbivores was strong, independent of bottom–up conditions. The omnivorous katydid Orchelimum has a less flexible diet and plays a different trophic role than the omnivorous crab Armases that occurs in the same salt marshes, suggesting that generalizations about the trophic effects of omnivores cannot be made without a detailed understanding of the omnivore’s dietary constraints and ability to suppress lower trophic levels. Acknowledgments We thank Chris Craft for generously performing C/N analyses and B. DeLong, E. Wilkinson and the GCE-LTER schoolyard teachers for help with experiments. Financial support was provided by NSF (DEB-0296160) and the Biology and Biochemistry department of the University of Houston. We are grateful to A. Armitage, R. Azevedo, A. Frankino, S. Murphy, E. Siemann, G. Wimp, and anonymous reviewers for comments on the manuscript. This is a contribution of the Georgia Coastal Ecosystems LTER program, and contribution number 1010 from the University of Georgia Marine Institute.

Estuaries and Coasts (2012) 35:475–485

References Bertness, M.D., C. Crain, C. Holdredge, and N. Sala. 2008. Eutrophication and consumer control of New England salt marsh primary productivity. Conservation Biology 22: 131–139. Bruno, J.F., and M.I. O’Connor. 2005. Cascading effects of predator diversity and omnivory in a marine food web. Ecology Letters 8: 1048–1056. 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. Denno, R.F., L.W. Douglass, and D. Jacobs. 1985. Crowding and host plant nutrition—Environmental determinants of wing-form in Prokelisia marginata. Ecology 66: 1588–1596. Denno, R. F., M. E. Schauff, S. W. Wilson, and K. L. Olmstead. 1987. Practical diagnosis and natural-history of 2 sibling salt marsh-inhabiting planthoppers in the genus Prokelisia (homoptera, delphacidae). In: Proceedings of the Entomological Society of Washington, vol 89, pp 687–700 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., C. Gratton, H. Dobel, and D. Finke. 2003. Predation risk affects relative strength of top–down and bottom–up impacts on insect herbivores. Ecology 84: 1032–1044. Denno, R.F., D. Lewis, and C. Gratton. 2005. Spatial variation in the relative strength of top–down and bottom–up forces: Causes and consequences for phytophagous insect populations. Annales Zoologici Fennici 42: 295–311. Eubanks, M. D. 2005. Predaceous herbivores and herbivorous predators: the biology of omnivores and the ecology of omnivore–prey interactions. In Ecology of predator-prey interactions, eds. P. Barbosa and I. Castellanos, 3–16. Oxford: Oxford University Press. Finke, D.L., and R.F. Denno. 2004. Predator diversity dampens trophic cascades. Nature 429: 407–410. Goeriz Pearson, R.E., S.T. Behmer, D.S. Gruner, and R.F. Denno. 2011. Effects of diet quality on performance and nutrient regulation in an omnivorous katydid. Ecological Entomology 36: 471–479. Grace, J.B. 2006. Structural equation modelling and natural systems. Cambridge: Cambridge University Press. Gratton, C., and R.F. Denno. 2003. Seasonal shift from bottom–up to top–down impact in phytophagous insect populations. Oecologia 134: 487–495. Ho, C.-K., and S.C. Pennings. 2008. Consequences of omnivory for trophic interactions on a salt marsh shrub. Ecology 89: 1714– 1722. 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. Long, Z.T., J.F. Bruno, and E.J. Duffy. 2011. Food chain length and omnivory determine the stability of a marine subtidal food web. Journal of Animal Ecology 80: 586–594. Matsumura, M., G.M. Trafelet-Smith, C. Gratton, D.L. Finke, W.F. Fagan, and R.F. Denno. 2004. Does intraguild predation enhance predator performance? A stoichiometric perspective. Ecology 85: 2601–2615. McFarlin, C.R., J.S. Brewer, T.L. Buck, and S.C. Pennings. 2008. Impact of fertilization on a salt marsh food web in Georgia. Estuaries and Coasts 31: 313–325. Moon, D.C., and P. Stiling. 2000. Relative importance of abiotically induced direct and indirect effects on a salt-marsh herbivore. Ecology 81: 470–481.

Estuaries and Coasts (2012) 35:475–485 Moon, D.C., and P. Stiling. 2002. The effects of salinity and nutrients on a tritrophic salt-marsh system. Ecology 83: 2465–2476. Moon, D.C., and P. Stiling. 2004. The influence of a salinity and nutrient gradient on coastal vs. upland tritrophic complexes. Ecology 85: 2709–2716. Moon, D.C., and P. Stiling. 2005. Effects of nutrients and parasitism on the density of a salt marsh planthopper suppressed by within-trophiclevel interactions. Ecological Entomology 30: 642–649. Moon, D.C., A.M. Rossi, and P. Stiling. 2000. The effects of abiotically induced changes in host plant quality (and morphology) on a salt marsh planthopper and its parasitoid. Ecological Entomology 25: 325–331. Morris, J.T. 1984. Effects of oxygen and salinity on amonia uptake by Spartina alterniflora Loisel. and Spartina patens (Aiton) Muhl. Journal of Experimental Marine Biology and Ecology 78: 87–98. Oksanen, L., S.D. Fretwell, J. Arruda, and P. Niemela. 1981. Exploitation ecosystems in gradients of primary productivity. American Naturalist 118: 240–261. Pennings, S.C., and M.D. Bertness. 2001. Salt marsh communities. In Marine community ecology, ed. M.D. Bertness, S.D. Gaines, and M.E. Hay, 289–316. Sunderland: Sinauer Associates. Polis, G.A., and D.R. Strong. 1996. Food web complexity and community dynamics. American Naturalist 147: 813–846. Power, M.E. 1992. Top–down and bottom–up forces in food webs— Do plants have primacy. Ecology 73: 733–746. Richards, C.L., S.C. Pennings, and L.A. Donovan. 2005. Habitat range and phenotypic variation in salt marsh plants. Plant Ecology 176: 263–273. Sala, N.M., M.D. Bertness, and B.R. Silliman. 2008. The dynamics of bottom–up and top–down control in a New England salt marsh. Oikos 117: 1050–1056.

485 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. Silliman, B.R., J. Van de Koppel, M.D. Bertness, L.E. Stanton, and I. A. Mendelssohn. 2005. Drought, snails, and large-scale die-off of southern U.S. salt marshes. Science 310: 1803–1806. Smalley, A.E. 1960. Energy-flow of a salt-marsh grasshopper population. Ecology 41: 672–677. Stiling, P., and D.C. Moon. 2005. Are trophodynamic models worth their salt? Top–down and bottom–up effects along a salinity gradient. Ecology 86: 1730–1736. Stiling, P., B.V. Brodbeck, and D.R. Strong. 1991. Population increases of planthoppers on fertilized salt-marsh cord grass may be prevented by grasshopper feeding. Florida Entomologist 74: 88–97. Strong, D.R. 1992. Are trophic cascades all wet? Differentiation and donor–control in speciose ecosystems. Ecology 73: 747–754. Teal, J.M. 1962. Energy-flow in salt-marsh ecosystem of Georgia. Ecology 43: 614. Van-Rijn, P.C.J., and M.W. Sabelis. 2005. Impact of plant-provided food on herbivore–carnivore dynamics. In Plant-provided food for carnivorous insects: A protective mutualism and its applications, ed. F.L. Wackers, P.C.J. Van Rijn, and J. Bruin, 223–226. Cambridge: Cambridge University Press. Wason, E.L., and S.C. Pennings. 2008. Grasshopper (Orthoptera: Tettigoniidae) species composition and size across latitude in Atlantic Coast salt marshes. Estuaries and Coasts 31: 335–343. Wimp, G.M., S.M. Murphy, D.L. Finke, A.F. Huberty, and R.F. Denno. 2010. Increased primary production shifts the structure and composition of a terrestrial arthropod community. Ecology 91: 3303–3311.