* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Plant genotypic diversity reduces the rate of consumer resource
Survey
Document related concepts
Gartons Agricultural Plant Breeders wikipedia , lookup
Biological Dynamics of Forest Fragments Project wikipedia , lookup
Biodiversity wikipedia , lookup
Renewable resource wikipedia , lookup
Latitudinal gradients in species diversity wikipedia , lookup
Plant breeding wikipedia , lookup
Transcript
Downloaded from http://rspb.royalsocietypublishing.org/ on June 18, 2017 Plant genotypic diversity reduces the rate of consumer resource utilization Scott H. McArt1,† and Jennifer S. Thaler1,2 1 rspb.royalsocietypublishing.org Research Cite this article: McArt SH, Thaler JS. 2013 Plant genotypic diversity reduces the rate of consumer resource utilization. Proc R Soc B 280: 20130639. http://dx.doi.org/10.1098/rspb.2013.0639 Received: 12 March 2013 Accepted: 18 April 2013 Department of Entomology, and 2Department of Ecology and Evolutionary Biology, Cornell University, Ithaca, NY 14853, USA While plant species diversity can reduce herbivore densities and herbivory, little is known regarding how plant genotypic diversity alters resource utilization by herbivores. Here, we show that an invasive folivore—the Japanese beetle (Popillia japonica)—increases 28 per cent in abundance, but consumes 24 per cent less foliage in genotypic polycultures compared with monocultures of the common evening primrose (Oenothera biennis). We found strong complementarity for reduced herbivore damage among plant genotypes growing in polycultures and a weak dominance effect of particularly resistant genotypes. Sequential feeding by P. japonica on different genotypes from polycultures resulted in reduced consumption compared with feeding on different plants of the same genotype from monocultures. Thus, diet mixing among plant genotypes reduced herbivore consumption efficiency. Despite positive complementarity driving an increase in fruit production in polycultures, we observed a trade-off between complementarity for increased plant productivity and resistance to herbivory, suggesting costs in the complementary use of resources by plant genotypes may manifest across trophic levels. These results elucidate mechanisms for how plant genotypic diversity simultaneously alters resource utilization by both producers and consumers, and show that population genotypic diversity can increase the resistance of a native plant to an invasive herbivore. Subject Areas: ecology Keywords: biodiversity – ecosystem function, community genetics, herbivory, plant – insect interactions, diet mixing, Popillia japonica Author for correspondence: Scott H. McArt e-mail: [email protected] † Present address: Department of Biology, University of Massachusetts – Amherst, Amherst, MA 01003, USA. Electronic supplementary material is available at http://dx.doi.org/10.1098/rspb.2013.0639 or via http://rspb.royalsocietypublishing.org. 1. Introduction Understanding how biodiversity affects the efficiency of resource use in food webs is a central goal of community and ecosystem ecology [1,2]. Following longstanding interest in how plant diversity affects herbivore populations [3,4], recent analyses have begun to reveal how plant species diversity affects resource utilization by herbivores [5–7]. At the same time, a growing body of literature suggests that genotypic diversity within species can have pronounced consequences on populations, communities and ecosystems [8]. However, while several studies have now shown that plant genotypic diversity typically increases the abundance of herbivores [9–12], how changes in herbivore abundance are linked to the consumption of plant resources remains unknown [13,14]—an important gap in our understanding of how plant diversity alters energy transfer across trophic levels. While little theory has addressed how plant genotypic diversity can impact herbivore populations or herbivory, at least four formal hypotheses have been suggested for how species diversity impacts consumers. With regard to herbivore populations, the ‘resource concentration hypothesis’ [4] posits that fewer specialist herbivores will accumulate in diverse plant assemblages owing to reduced plant apparency, herbivore residence time and/or herbivore reproductive output. Alternatively, the ‘enemies hypothesis’ [4] posits that both generalist and specialist natural enemies (i.e. predators of herbivores) will be more abundant in diverse plant assemblages, and therefore suppress herbivore populations in polycultures more than monocultures. Regarding the consumption of plants by herbivores, the ‘variance in edibility hypothesis’ [15,16] posits that a resource base with more species is more likely to contain at least one species that is resistant to consumption, which can then dominate in the presence of consumers and lead to less herbivory. This is analogous to the & 2013 The Author(s) Published by the Royal Society. All rights reserved. Downloaded from http://rspb.royalsocietypublishing.org/ on June 18, 2017 changes in resource utilization at the consumer level related to resource utilization at the producer level? 2. Material and methods (a) Study system, plant propagation and field establishment (b) Herbivory surveys and plant productivity We conducted two censuses of Japanese beetles (P. japonica) at the peak of their abundance (once in late July and again in Proc R Soc B 280: 20130639 We manipulated genotypic richness of O. biennis L (common evening primrose, Onagraceae), a native herbaceous plant that is common to old fields and disturbed areas in eastern North America. Oenothera biennis reproduces via a permanent translocation heterozygosity genetic system, which results in seeds that are genetically identical to each other and the parent [29,30]. We collected O. biennis seeds from individual plants in 20 distinct populations around Ithaca (New York). Each genotype used in this experiment was determined to be unique using nine polymorphic microsatellite loci developed for O. biennis [31]. To reduce maternal effects, we first grew the seeds in a common garden in 2007, which was sprayed with insecticide at regular intervals throughout the growing season, and we used seeds collected from these plants (20 genotypes) for our experiment. We cold stratified (48C, 4 days) all seeds for the field experiment in April 2010, sowed them into 96-well trays filled with soil (Pro-mix ‘BX’ with biofungicide, Premier) and thinned germinated seedlings to a single individual per well. Plants were watered ad libitum and fertilized weekly (21-5-20 NPK, 150 ppm) while in the greenhouse (14 L : 10 D cycle, five weeks) and then field-hardened in an outdoor mesh cage (one week) prior to planting in the field. In May 2010, we established the field experiment in an abandoned agricultural field near Ithaca, where the soil was ploughed, but otherwise untreated. As we were interested in invertebrate responses to plant genotypic diversity, we excluded vertebrate herbivores such as deer and rabbits from our plots via a 2.5 m fence (1.5 cm mesh) surrounding our entire experiment. Using our pool of 20 O. biennis genotypes, we constructed two treatments within this enclosure: genotypic monocultures (one O. biennis genotype) and genotypic polycultures (seven O. biennis genotypes). All plots contained seven equally spaced individual plants arrayed in a ring 0.5 m in diameter, and plots were separated by 1.5 m. We clipped encroaching weeds by hand, every two to three weeks, to ensure treatments remained consistent throughout the summer. The original design included 120 plots, but owing to the loss of individuals within plots (which was always only one plant per plot and showed no pattern among genotypes or treatments), we restricted our analyses to the 109 plots that experienced no mortality (monocultures: n ¼ 55; polycultures: n ¼ 54). Every genotype appeared in approximately 19 polycultures and there were three monocultures of each genotype (except for five O. biennis genotypes that had two monocultures each owing to mortality). Oenothera biennis is monocarpic, and every plant in the experiment bolted and produced fruit in the first season, which is a typical response of this plant in disturbed habitats such as our ploughed field [32]. Owing to its large size, we divided our experiment into four spatial blocks to account for potential within-site environmental variation, where each block contained the same proportion of monocultures and polycultures. Although this common garden is located within a homogeneous field, we have typically found significant effects of spatial location within the garden for plant productivity and arthropod responses [9,23] (also this study). 2 rspb.royalsocietypublishing.org selection effect [17] at the resource rather than consumer level. Alternatively, the ‘balanced diet hypothesis’ [18,19] posits that a resource base with higher diversity will result in greater consumption from generalist herbivores via complementary acquisition of deficient nutrients. While some experimental support exists for each of these hypotheses [19 –21], by far the most research has focused on the resource concentration and enemies hypotheses. In a review of 209 studies of 287 herbivorous and 130 predatory arthropod species, Andow [3] found that 51.9 per cent of the herbivorous species examined had lower population densities on plants in polycultures than monocultures (compared with 15.3% of species having lower densities on monocultures). Furthermore, 52.7 per cent of predator species had higher population densities in polycultures (compared with 9.3% having higher densities in monocultures). While the resource concentration and enemies hypotheses are not mutually exclusive, Andow determined that resource concentration was somewhat better at explaining these results. Regardless, it is important to note that these two hypotheses solely consider the population responses of animals to plant diversity and do not explicitly link animal abundance to plant damage [2,3]. This is important because the magnitude of consumption, not simply the abundance of consumers, controls energy transfer across trophic levels [22]. Whether mechanisms altering herbivore populations and herbivory in response to plant species diversity carry over to plant genotypic diversity has received limited attention. The few studies that have focused on herbivore populations in response to plant genotypic diversity have found no support for either of the two major hypotheses related to plant species diversity—the resource concentration or enemies hypotheses [9–12]. Instead, most of these studies find that herbivore populations tend to increase in response to genotypic diversity, which differs from the decreases that are typical in response to plant species diversity [3,4]. Perhaps this is not surprising, because, due to the inherent differences in trait variation within versus among plant species, fundamentally different mechanisms are responsible for structuring broad-scale patterns of arthropod diversity in response to plant genotypic versus species diversity [9,10,23]. Because trait variation in plants is critically important to herbivores during both the selection of host plants [4,24] and utilization of those hosts [25,26], mechanisms altering herbivore populations and herbivory in response to each type of plant diversity probably differ as well. Here, we address this gap in the literature by linking patterns of herbivore abundance, movement and consumption. In doing so, we provide a mechanism for how plant genotypic diversity increases the abundance of herbivores yet reduces herbivory. Using a manipulative field experiment, feeding bioassays and behavioural observations, we test how genotypic diversity of the common evening primrose (Oenothera biennis) influences the population size and consumption dynamics of its dominant folivore, the Japanese beetle (Popillia japonica). We focus on Japanese beetles—the most widespread and destructive pest of turf grass, landscapes and nursery crops in the eastern United States [27]—because they are responsible for more than 95 per cent of the leaf area consumed on O. biennis in Tompkins County (New York) [28]. We address three main questions in this study. (i) Is resource utilization by P. japonica altered in response to plant genotypic diversity? (ii) What mechanisms are responsible for altering resource utilization by P. japonica? (iii) Are diversity-mediated Downloaded from http://rspb.royalsocietypublishing.org/ on June 18, 2017 where NDRYMi measures the complementarity effect, Ncov ðDRY; MÞ measures the selection effect, N is the number of genotypes in polyculture, Mi is the performance of genotype i in monoculture and DRY ¼ RYOi 2 RYEi is the deviation from expected relative performance of genotype i in the polyculture [17]. Positive complementarity for fruit production implies that genotypes had generally higher fruit production in polyculture than monoculture, which is typically attributed to resource partitioning or facilitation among genotypes. Negative complementarity implies that fruit production is generally lower across genotypes in polyculture than monoculture. Positive selection implies that highly productive genotypes produce proportionally more fruits in polyculture than monoculture, whereas negative selection implies that less productive genotypes produce proportionally more fruits. This method for partitioning net biodiversity effects can also be extended to herbivore damage. In this case, positive complementarity implies that genotypes received more damage on average in polyculture than monoculture, whereas negative complementarity means that they received less damage. We interpret negative complementarity for herbivore damage as evidence for associational resistance [24,33] because proximity to genotypically diverse neighbours buffered genotypes from receiving damage. Positive selection indicates that highly attacked genotypes received disproportionally more damage in polyculture than monoculture, whereas negative selection indicates that less attacked genotypes received disproportionately less damage. For both fruit production and herbivore damage, we tested whether complementarity and selection analyses were positive or negative by observing whether 95% CIs overlapped zero. We tested how changes in herbivore damage were related to changes in fruit production across treatments at both the plant and patch levels. At the plant level, we tested for a genotypic (c) Herbivore movement Because movement among plants has been hypothesized to be a key mechanism altering herbivore populations in response to plant genotypic diversity [12], we conducted an experiment to explicitly evaluate P. japonica movement. We quantified movement of P. japonica within and among O. biennis patches by replicating the 2010 field establishment protocols (described above) for a follow-up experiment in 2011 containing 17 of the original 20 genotypes. Equal numbers of monocultures and polycultures were planted such that each genotype occurred in seven polycultures and one monoculture (n ¼ 34 patches total). Similar to the previous field experiment, monocultures were spatially alternated with polycultures such that two monocultures and two polycultures were present in nine blocks (the final block contained only one monoculture and polyculture each). In late July, we observed beetle movement in each block for periods of 15 min on three successive days. Observations were conducted at the height of beetle activity and feeding (between 09.00 and 12.00 or between 15.00 and 18.00). We randomized which blocks were visited at which time each day. Over the course of the experiment, each patch contained six P. japonica adults on average, and we observed an average of nine movement events (one beetle moving from one plant to another plant) in each observation period. Thus, we were able to easily record whether each beetle moved to a neighbouring plant within a patch or outside the patch, keeping track of the specific plants the beetle moved between if movement occurred within each four-patch block. We tested whether overall beetle movement within versus between patches differed from a 50 – 50 expectation, and whether within versus between patch movement differed between diversity treatments, via Pearson x2 analyses. (d) Fixed and sequential feeding bioassays We conducted a two-part bioassay to test the resistance of O. biennis genotypes grown in monocultures versus polycultures to P. japonica. Because plant phenotypic traits such as biomass [34,35], C : N ratio [36,37] and chemical defences against herbivores [38] can change when plants are grown in diverse compared with homogeneous communities, we first hypothesized that individual plant resistance (and therefore consumption) could change when genotypes grow in monocultures versus polycultures. Alternatively, owing to the high mobility of P. japonica within plant patches (see §3), we hypothesized that sequential feeding on different genotypes in polyculture compared with sequential feeding on different plants of the same genotype in monoculture could alter rates of consumption. During the peak of P. japonica abundance in late July 2010, we collected individual leaves from six replicate plants of each genotype in each treatment (6 leaves 20 genotypes 2 treatments ¼ 240 bioassays). The first fully expanded leaf from each plant was cut at its petiole, placed in a Petri dish (9 cm diameter) containing a moist sheet of filter paper and immediately transported back to the laboratory. One P. japonica adult (collected from the wild) was placed in each dish and allowed to feed for 24 hours at 208C. At the end of 24 hours, each P. japonica adult was removed, and leaf area consumed (mm2) was assessed on leaves. In this way, we assessed constitutive resistance of individual plants grown in monocultures versus 3 Proc R Soc B 280: 20130639 DY ¼ NDRYMi þ NcovðDRY; MÞ; correlation between reduced herbivore damage (mean damage in monoculture minus mean damage in polyculture) and increased production of fruits (mean number of fruits produced in polyculture minus mean number of fruits produced in monoculture) via linear regression. At the patch level, we tested the relationship between the net biodiversity effect (including complementarity and selection) for herbivore damage and fruit production via linear regression. rspb.royalsocietypublishing.org early August) by visually surveying every plant in the experiment (n ¼ 840 plants). In early September, when all beetles were gone but leaves had not yet dropped from the plants, we surveyed the quantity of beetle leaf damage. We did not observe any leaves dropped due to herbivory over the course of the growing season. We placed an acetate sheet printed with a 1 cm2 grid over each leaf of every plant in the experiment (mean ¼ 109 leaves per plant), quantifying leaf area consumed on each plant to the nearest square centimetre. In early October, when plants stopped producing fruits, we counted the number of fruits on each plant in the experiment (mean ¼ 148 fruits per plant). Because O. biennis is monocarpic, the number of fruits produced is an estimate of lifetime fitness. We note that there is a strong positive relationship between above-ground plant biomass and number of fruits produced by O. biennis [23], so the number of fruits produced is a good estimate of productivity. We tested for differences in plot-level beetle abundance, leaf area consumed and number of fruits produced between diversity treatments by using restricted maximum likelihood (REML) to estimate variance parameters in linear mixed models (LMMs) with spatial block as a random effect (JMP PRO v. 9.0.2). We log-transformed beetle abundance and damage data to improve normality of the residuals. To compare the amount of leaf area consumed between O. biennis genotypes in each treatment, we tested for the effect of genotype, treatment and the genotype treatment interaction via LMM (REML) with spatial block and plot nested within block as random effects. The electronic supplementary material (table S1) details all mixed models used to analyse the data. To understand the mechanisms by which O. biennis genotypic diversity affected the production of fruits, we used the methods of Loreau & Hector [17] to calculate the net biodiversity effect. The net biodiversity effect is calculated using the equation We found 28 per cent more Japanese beetles (P. japonica) in genotypic polycultures versus monocultures of O. biennis (F1,103 ¼ 4.9, p ¼ 0.029; figure 1a; electronic supplementary material, table S1). In contrast to these results, we found a 24 per cent decrease in the amount of leaf area consumed in polycultures compared with monocultures (F1,103 ¼ 4.0, p ¼ 0.044; figure 1b). Because plants tended to be larger in polycultures (see productivity results below), the percentage of leaf area consumed was probably even further reduced in polycultures compared with this absolute measure of herbivory. Genotypes differed in leaf area consumed by P. japonica (F19,146 ¼ 21.1, p , 0.001), and there was also a significant genotype treatment interaction for herbivory (F19,146 ¼ 2.2, p ¼ 0.006). However, patterns of herbivory were largely 6 5 4 4 3 2 1 Proc R Soc B 280: 20130639 3. Results (a) rspb.royalsocietypublishing.org 0 (b) 120 leaf area removed (cm2) polycultures by testing for differences in leaf area consumed between treatments via LMM (REML) with genotype as a random effect. Forty-one beetles did not initiate feeding or died during this assay (n ¼ 22 monoculture, n ¼ 19 polyculture); however whether we included these zero values in our analysis did not alter the direction or significance of results. Therefore, we present the data excluding these beetles for consistency with our sequential feeding assay (see below). In order to test whether sequential feeding on the same versus different genotypes altered beetle consumption, we continued this initial bioassay. All beetles from the first assay were immediately transferred to a new leaf in a new Petri dish and allowed to feed for an additional 24 hours. Leaves for the second feeding assay (day 2) were obtained from the field experiment in an identical manner to the first collection. In order to mimic the way sequential feeding might occur within plant patches in the field, beetles assigned to a monoculture sequential feeding treatment were transferred to a leaf from a different plant of the same genotype from the patch, whereas beetles assigned to a polyculture sequential feeding treatment were transferred to a leaf from a different genotype from the patch. Replication of genotypes across bioassay treatments was identical for each assay (six replicate leaves per genotype per treatment) in order for treatments to be balanced at each 24-hour time interval. Assignment of beetles to genotypes within the polyculture sequential feeding assays was random with the exception that no two sequences were identical, while source plants used for the monoculture sequential feeding assays were randomly chosen from each patch. Leaf area consumed during the second assay was assessed at the conclusion of the second 24-hour period (day 2), and this design was repeated for one additional 24-hour period (day 3) such that three Petri dish assays were conducted in sequence for each beetle over 72 hours. Thus, beetles in the monoculture sequential feeding treatment consumed three leaves of the same genotype over 72 hours, whereas beetles in the polyculture sequential feeding treatment consumed three leaves from three different genotypes over 72 hours. We tested for differences in leaf area consumed over all three assays by analysing the effect of treatment, day and treatment day interaction via LMM (REML), with genotype and beetle as random effects. Twelve beetles died over the course of this experiment and 83 beetles stopped feeding during the second or third assays (n ¼ 45 monoculture sequence, n ¼ 50 polyculture sequence). Whether or not we include these zero values did not alter the direction or significance of any results (treatment and day always p . 0.05, treatment day interaction always p , 0.05). Therefore, to most completely assess the effects of sequential feeding on rates of herbivory, here we present only the data where full sequential feeding for all three assays occurred. Popillia japonica abundance Downloaded from http://rspb.royalsocietypublishing.org/ on June 18, 2017 90 60 30 0 monocultures polycultures Figure 1. (a) The abundance of Japanese beetles (Popillia japonica) was 28% greater in Oenothera biennis genotypic polycultures versus monocultures, while (b) the amount of leaf damage was 24% less in polycultures versus monocultures (mean + s.e.m.). Data were log-transformed to improve normality and back-transformed to produce figure. consistent by treatment: the magnitude of damage was reduced on 16 out of 20 genotypes when grown in polycultures (figure 2) and only one genotype received significantly more damage in polycultures versus monocultures (see the electronic supplementary material, table S2). When we partitioned the mechanisms for reduced herbivory in polycultures, we found strong complementarity among genotypes for reduced damage (95% CI ¼ 221.4 + 8.3; figure 2 inset) while there was also weak but significant positive selection (95% CI ¼ 4.1 + 3.2). Negative complementarity implies that decreases in herbivore damage in polycultures are due to overall associational resistance [24,33] to herbivore damage among the majority of plant genotypes, whereas positive selection implies that genotypes that received the greatest amount of damage in monoculture receive proportionally more damage in polyculture compared with low-damage genotypes. In our movement experiment, we observed that beetles were more than twice as likely to move between plants within a patch (i.e. within a monoculture or polyculture) compared with leaving a patch to feed elsewhere. This pattern differed significantly from a 50–50 expectation (within-patch n ¼ 144, out-of-patch n ¼ 67; Pearson x 2 ¼ 28.1, p , 0.001), and there was no difference in this within-patch bias for movement in monocultures (within-patch n ¼ 62, out-of-patch n ¼ 29) versus polycultures (within-patch n ¼ 82, out-of-patch n ¼ 38; Pearson x 2 ¼ 0.001, p ¼ 0.98). Given the strong complementarity for reduced damage in O. biennis polycultures, we tested two mechanisms for how complementarity could occur. We first tested whether Downloaded from http://rspb.royalsocietypublishing.org/ on June 18, 2017 60 sel comp –5 40 –20 30 –35 20 10 0 S18 S1 S6 S36 S21 S7 S15 S11S20 S43 S9 A23S17A29 S38 A4 S32 S33 S5 S44 genotype Figure 2. Least-squares mean values of damage on 20 O. biennis genotypes grown in genotypic monocultures (dark grey bars) versus polycultures (light grey bars; error bars not shown for clarity). Reduced herbivory in polycultures resulted from negative complementarity (comp) and positive selection (sel) among O. biennis genotypes for the amount of damage received (inset). Negative complementarity implies that decreases in herbivore damage in polycultures were due to beneficial associational effects among plant genotypes [24,33], whereas positive selection implies that the most resistant genotypes in monocultures were even more resistant in polycultures. Complementarity and selection analyses performed following the methods outlined by Loreau & Hector [17]; mean + 95% CIs shown. 3.2 * leaf area consumed hr–1 (mm2) individual plant resistance was different for genotypes when grown in monocultures versus polycultures. When we performed a bioassay allowing beetles to consume leaf tissue from the same genotypes grown in monocultures versus polycultures, we found no difference in consumption between treatments (F1,178 ¼ 1.9, p ¼ 0.17; figure 3, day 1), implying that constitutive plant resistance to herbivore damage was not affected by diversity treatment. Next, owing to the high degree of movement by P. japonica among plants within patches, we tested whether sequential feeding on different genotypes in polycultures compared with sequential feeding on different plants of the same genotype in monoculture resulted in a reduced rate of herbivory. Consistent with this hypothesis, we found that beetles increased their rate of consumption when feeding on leaves from the same genotype in sequence (monoculture sequence), whereas beetles tended to have a reduced rate of consumption when feeding on leaves from three different genotypes in sequence ( polyculture sequence) over a period of 72 hours (treatment day interaction: F2,279 ¼ 6.9, p ¼ 0.001; post hoc contrast day 3: p ¼ 0.010; figure 3). Consistent with previous studies in this system and others, we found an 11 per cent increase in the productivity (number of fruits produced) in O. biennis genotypic polycultures (F1,103 ¼ 5.9, p ¼ 0.017), and there was a positive relationship between plot-level fruit production and P. japonica abundance (F1,103 ¼ 9.1, p ¼ 0.003, R 2 ¼ 0.13). Increased fruit production was a result of strong positive complementarity (95% CI ¼ 158.0 + 63.0), while there was also weak but significant negative selection (95% CI ¼ 29.2 + 7.5). Because fruit production increased in polycultures compared with monocultures, whereas herbivore damage decreased in polycultures compared with monocultures, we tested whether increases in plant productivity depended on decreases in herbivore damage. At the plant level, we found a negative relationship between the decrease in herbivory by genotypes in polycultures versus the increase in fruit 3.1 3.0 2.9 2.8 2.7 2.6 day 1 day 2 day 3 Figure 3. Leaf area consumed by P. japonica when fed leaves from different plants of the same genotype (monoculture: black squares, solid line) or leaves from different genotypes ( polyculture: open squares, dashed line) in sequence (least-squares means + s.e.). Day 1 tested the effect of diversity treatment on individual plant resistance to P. japonica, while days 1 – 3 tested the effect of genotype mixing in their diet. Data were log-transformed to improve normality and back-transformed for the figure. *p , 0.05 post hoc contrast between treatments for individual assays. production in polycultures (F1,19 ¼ 22.3, p , 0.001, R 2 ¼ 0.55; electronic supplementary material, figure S1). At the patch level, we also found a negative relationship between the net biodiversity effects for decreased herbivory versus increased fruit production (F1,53 ¼ 11.3, p ¼ 0.002, R 2 ¼ 0.18; figure 4). When we partitioned the net biodiversity effect in complementarity and selection, we found this latter result was driven by the negative relationship between complementarity for reduced damage versus increased fruit production (F1,53 ¼ 3.1, p ¼ 0.003, R 2 ¼ 0.16); there was no relationship between selection for herbivore damage versus fruit production (F1,53 ¼ 1.4, p ¼ 0.23). Proc R Soc B 280: 20130639 leaf area removed (cm2) 10 rspb.royalsocietypublishing.org 50 5 Downloaded from http://rspb.royalsocietypublishing.org/ on June 18, 2017 500 300 100 –100 –300 60 20 –20 –60 –100 net biodiversity effect (herbivore damage) Figure 4. Relationship between net biodiversity effects [17] for leaf area removed by herbivores (herbivore damage) and plant productivity (number of fruits produced). Negative values along the x-axis correspond to less damage in genotypic polycultures than predicted from monocultures, whereas positive values along the y-axis correspond to greater productivity than predicted. 4. Discussion In this study, we show that despite an increase in herbivore abundance, the consumption of resources by herbivores is reduced in plant genotypic polycultures. These results are consistent with recent analyses showing that herbivory can be reduced in response to plant species [5–7] and genotypic [13,14] diversity. Our results also support a growing recognition that plant genotypic diversity increases the abundance of herbivores [9–12], which contrasts with the suppression of herbivores that is typical in response to plant species diversity [3,21]. By linking patterns of herbivore abundance, movement and consumption in a single study, we provide a mechanism to reconcile how plant genotypic diversity can increase the abundance of herbivores, yet reduce consumption: by reducing the rate of resource utilization. Our results do not provide strong support for any of the four main published hypotheses posed for altering herbivore populations or herbivory in response to plant species diversity. Because we found greater numbers of P. japonica in polycultures and they consumed less plant tissue, our data do not support the resource concentration hypothesis [4] or balanced diet hypothesis [18,19]. In this system and others, the increased abundance of herbivores in response to genotypic diversity has been found to track plant biomass [9,10], which suggests herbivores are primarily responding to the increase in available plant resources. We also find no support for the enemies hypothesis [4]; P. japonica adults are relatively resistant to predation [27], and we did not find an increase in predator abundance or richness in response to O. biennis genotypic diversity in a previous study at the same site [9]. Finally, we found there was positive selection for plant damage in polycultures (figure 2), meaning that the most resistant genotypes in monoculture were even more resistant in polyculture. While this result is consistent with the variance in edibility hypothesis [15,16], the increased resistance of particular O. biennis genotypes in polycultures (i.e. positive selection for plant damage) was weak compared with the strong overall associational resistance (i.e. negative complementarity for plant damage) in genotypic polycultures (figure 2 inset). This suggests that mechanisms other than 6 Proc R Soc B 280: 20130639 –500 variance in edibility were the primary drivers of reduced herbivore damage in response to O. biennis genotypic diversity. We tested two alternative hypotheses that could explain how strong complementarity for reduced herbivory could occur in O. biennis genotypic polycultures. First, in our bioassay with P. japonica (figure 3), we found that individual plant resistance did not differ between genotypes grown in monocultures versus polycultures (figure 3, day 1). This suggests that individual plant quality traits that can change when plants are grown in diverse species mixtures [36 –38] were either not affected by genotypic diversity or unimportant for resistance to P. japonica. In support of this result, we previously found no effect of O. biennis genotypic diversity on the density of trichomes, leaf toughness or specific leaf area of leaves from plants grown in monocultures versus polycultures ( p . 0.2 in each case; S.H.M. 2009, unpublished data). Instead, in this study, we found that sequential feeding on different O. biennis genotypes resulted in a reduced rate of herbivory compared with sequential feeding on the same genotype (figure 3). Because our results show that P. japonica is a mobile herbivore that preferentially moves and feeds on plants within patches compared with moving out of patches, the reduced rate of herbivory that beetles experience on a mixed-genotype diet provides a probably mechanistic link between the opposing patterns of increased P. japonica abundance and reduced damage observed in O. biennis polycultures. While sequential feeding on different plants is not typically considered a mechanism of associational resistance to herbivores [24,33], this may be a common response of mobile herbivores that feed on multiple neighbouring plants. For example, in the most detailed study to date regarding the population response of herbivores to plant genotypic diversity, movement among plants was hypothesized to be a major driver of aphid population growth [12]. Compensatory feeding can occur when animals are restricted to diets suboptimal for their target intake of different nutrients, such as suboptimal protein : carbohydrate ratios [39,40]. Because the genotypes used in this experiment are known to differ significantly in nutritional characteristics such as C : N ratio [41], individual P. japonica beetles may have compensated for suboptimal nutrition in single-genotype monocultures by consuming more leaf tissue compared with mixed-genotype polycultures (where diet mixing among genotypes occurred). Alternatively, the genotypes we used in this experiment are also known to differ significantly in the abundance of secondary compounds, including ellagitannins and flavonoids [41]. Although diet mixing among plant species that contain different toxins is predicted to allow increased consumption by generalist herbivores [42], there is remarkably little experimental evidence supporting this detoxification limitation hypothesis [43]. Furthermore, we are aware of no studies that have tested the hypothesis for different plant genotypes, which may differ more in the abundance of different toxins as opposed to their presence or absence. Understanding how primary and secondary metabolites interact to affect consumption is an important but understudied area of chemical and nutritional ecology [44,45], and probably underlies the specific physiological mechanism responsible for our results. A few recent studies suggest that plant diversity can alter the strength of top-down control over plant productivity [5,6,14]. This is not surprising since plant diversity can modify patterns of herbivory [7,13,14], and herbivory often affects the outcome of plant competitive interactions such rspb.royalsocietypublishing.org net biodiversity effect (fruit production) 700 Downloaded from http://rspb.royalsocietypublishing.org/ on June 18, 2017 We thank A. Agrawal, S. Cook-Patton, A. Kessler and S. Meldau for helpful comments on the manuscript, and F. Vermeylan for advice on statistical models. The NSF (IGERT small grant in Biogeochemistry and Environmental Biocomplexity to S.H.M.), Grace Griswold Endowment (to S.H.M.) and USDA-NRI (2006-35302-1731 to J.S.T.) supported this work. S.H.M. and J.S.T. designed the experiments, S.H.M. conducted the experiments and analysed the data, and S.H.M. and J.S.T. wrote the manuscript. The authors declare no competing interests. References 1. 2. 3. Loreau M. 2010 From populations to ecosystems: theoretical foundations for a new ecological synthesis. Princeton, NJ: Princeton University Press. Duffy JE, Cardinale BJ, France KE, McIntyre PB, Thébault E, Loreau M. 2007 The functional role of biodiversity in ecosystems: incorporating trophic complexity. Ecol. Lett. 10, 522 –538. (doi:10.1111/j. 1461-0248.2007.01037.x) Andow DA. 1991 Vegetational diversity and arthropod population response. Annu. Rev. Entomol. 36, 561–586. (doi:10.1146/annurev.en.36.010191. 003021) 4. 5. 6. Root RB. 1973 Organization of a plant –arthropod association in simple and diverse habitats: fauna of collards (Brassica oleracea). Ecol. Monogr. 43, 95 –120. (doi:10.2307/1942161) Edwards KF, Aquilino KM, Best RJ, Sellheim KL, Stachowicz JJ. 2010 Prey diversity is associated with weaker consumer effects in a meta-analysis of benthic marine experiments. Ecol. Lett. 13, 194 –201. (doi:10.1111/j.1461-0248.2009.01417.x) Hillebrand H, Cardinale BJ. 2004 Consumer effects decline with prey diversity. Ecol. Lett. 7, 192– 201. (doi:10.1111/j.1461-0248.2004.00570.x) 7. 8. 9. Jactel H, Brockerhoff EG. 2007 Tree diversity reduces herbivory by forest insects. Ecol. Lett. 10, 835 –848. (doi:10.1111/j.1461-0248.2007.01073.x) Hughes AR, Inouye BD, Johnson MTJ, Underwood N, Vellend M. 2008 Ecological consequences of genetic diversity. Ecol. Lett. 11, 609–623. (doi:10.1111/j. 1461-0248.2008.01179.x) Cook-Patton SC, McArt SH, Parachnowitsch AL, Thaler JS, Agrawal AA. 2010 A direct comparison of the consequences of plant genotypic and species diversity on communities and ecosystem function. Ecology 92, 915–923. (doi:10.1890/10-0999.1) 7 Proc R Soc B 280: 20130639 relationship between complementarity ( p ¼ 0.003); there was no relationship between selection for herbivore damage versus fruit production ( p ¼ 0.23). Diversity-mediated increases in productivity are typically driven by complementarity and increases in the number of bolting stalks per plant in this system [9,23]. Thus, as beetles move among plants in the most productive polycultures, they are more likely to encounter a stalk from the same (self) genotype than a stalk from a different genotype compared with less productive polycultures. Therefore, diet mixing among genotypes by P. japonica may actually be reduced in the most productive polycultures compared with the least productive polycultures, suggesting a cost of high amounts of complementarity for productivity. This could explain why we observed increases in herbivory in the most productive polycultures ( p ¼ 0.007, R 2 ¼ 0.13). Our data cannot support this complementarity cost hypothesis with certainty; however, it is an extension of the same sequential feeding mechanism we outlined earlier. Understanding the mechanisms that drive the functional consequences of plant biodiversity is currently an area of much interest [1,49], and we suggest that a mechanistic understanding at multiple trophic levels will greatly improve predictability in this field of research. In summary, by linking behavioural observations of herbivore movement with abundance surveys, plant damage patterns and manipulative feeding bioassays, we were able to gain insight into how plant genotypic diversity increases the abundance of a dominant herbivore, yet decreases herbivory. The reduced consumption of resources by a herbivore in response to plant genotypic diversity is consistent with several recent analyses showing that herbivory is reduced in response to plant species diversity [5–7]. However, our results highlight that different mechanisms underlie reductions in herbivory in response to plant genotypic versus species diversity, probably due to inherent differences in trait variation within versus among species. Ultimately, a mechanistic understanding of consumption dynamics such as outlined in this study increases our ability to predict how biodiversity affects patterns of resource use in food webs. rspb.royalsocietypublishing.org as the accumulation of biomass or production of fruits [46]. Thus, we hypothesized that increases in the production of fruits in polycultures may depend in part on reductions in herbivory. This hypothesis could be supported in two ways using our data. At the plant level, the hypothesis would be supported by a positive genotypic correlation between the decrease in herbivore damage and increase in fruit production that genotypes experienced in polycultures. At the plot level, the net biodiversity effect [17] measures the deviation of polyculture performance as predicted from monocultures. Thus, relating net biodiversity effects for reduced herbivore damage and increased fruit production in polycultures is analogous to the plant-level genotypic correlation described above. This plot-level analysis is particularly useful for our question because it captures the total response of the plot, which is the spatial level at which P. japonica interacted with plants in this experiment. Instead of positive relationships between reduced herbivore damage and increased plant productivity in polycultures, we found that a trade-off occurs via both plant-level genotypic correlations and plot-level net biodiversity effects (see figure 4; electronic supplementary material, figure S1). In other words, polyculture assemblages that were particularly good at increasing fruit production were also particularly bad at resisting herbivore damage. We suggest these negative relationships indicate that (i) reduced herbivory by P. japonica in polycultures does not feed back to increase plant productivity and (ii) costs in the complementary use of resources by plant genotypes may manifest across trophic levels. With regard to the first point, we have previously found that O. biennis is highly tolerant of herbivory from P. japonica [28], which may be due to mechanisms such as photosynthetic overcompensation [47], or the fact that folivory from P. japonica occurs naturally during the pre-reproductive and reproductive stages of O. biennis—typically the most tolerant life stages of plants [48]. Thus, it may not be surprising that differences in folivory by P. japonica in response to O. biennis genotypic diversity do not alter the strength of top-down control over productivity. Regarding the second point, we found that reductions in herbivory and increases in the production of fruits in polycultures were due primarily to the combination of plant genotypes (i.e. complementarity), whereas the effect of particular genotypes (i.e. selection) was less important. Furthermore, the trade-off between net biodiversity effects for reduced herbivory versus increased fruit production (figure 4) was driven by the negative Downloaded from http://rspb.royalsocietypublishing.org/ on June 18, 2017 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. M. 2007 From selection to complementarity: shifts in the causes of biodiversity –productivity relationships in a long-term biodiversity experiment. Proc. R. Soc. B 274, 871–876. (doi:10.1098/rspb. 2006.0351) Mraja A, Unsicker SB, Reichelt M, Gershenzon J, Roscher C. 2011 Plant community diversity influences allocation to direct chemical defence in Plantago lanceolata. PLoS ONE 6, e28055. (doi:10. 1371/journal.pone.0028055) Raubenheimer D, Simpson SJ. 1993 The geometry of compensatory feeding in the locust. Anim. Behav. 45, 953 –964. (doi:10.1006/anbe.1993.1114) Simpson SJ, Raubenheimer D. 1993 A multi-level analysis of feeding behaviour: the geometry of nutritional decisions. Phil. Trans. R. Soc. Lond. B 342, 381–402. (doi:10.1098/rstb.1993.0166) Johnson MTJ, Agrawal AA, Maron JL, Salminen JP. 2009 Heritability, covariation and natural selection on 24 traits of common evening primrose (Oenothera biennis) from a field experiment. J. Evol. Biol. 22, 1295– 1307. (doi:10.1111/j.1420-9101. 2009.01747.x) Freeland WJ, Janzen DH. 1974 Strategies in herbivory by mammals: the role of plant secondary compounds. Am. Nat. 108, 269–289. (doi:10.1086/ 282907) Marsh K, Wallis I, Andrew R, Foley W. 2006 The detoxification limitation hypothesis: where did it come from and where is it going? J. Chem. Ecol. 32, 1247– 1266. (doi:10.1007/s10886-006-9082-3) Behmer ST, Simpson SJ, Raubenheimer D. 2002 Herbivore foraging in chemically heterogeneous environments: nutrients and secondary metabolites. Ecology 83, 2489– 2501. (doi:10.1890/00129658(2002)083[2489:HFICHE]2.0.CO;2) Steppuhn A, Baldwin IT. 2007 Resistance management in a native plant: nicotine prevents herbivores from compensating for plant protease inhibitors. Ecol. Lett. 10, 499 –511. (doi:10.1111/j. 1461-0248.2007.01045.x) Hambäck PA, Beckerman AP. 2003 Herbivory and plant resource competition: a review of two interacting interactions. Oikos 101, 26 –37. (doi:10. 1034/j.1600-0706.2003.12568.x) Morrison KD, Reekie EG. 1995 Pattern of defoliation and its effect on photosynthetic capacity in Oenothera biennis. J. Ecol. 83, 759. (doi:10.2307/ 2261412) Boege K, Marquis RJ. 2005 Facing herbivory as you grow up: the ontogeny of resistance in plants. Trends Ecol. Evol. 20, 441– 448. (doi:10.1016/j.tree. 2005.05.001) Carroll IT, Cardinale BJ, Nesbit RM. 2011 Niche and fitness differences relate the maintenance of diversity to ecosystem function. Ecology 92, 1157– 1165. (doi:10.1890/i0012-9658-92-5-1157) 8 Proc R Soc B 280: 20130639 24. Atsatt PR, O’Dowd DJ. 1976 Plant defense guilds. Science 193, 24– 29. (doi:10.1126/science. 193.4247.24) 25. Fritz RS, Simms EL. 1992 Plant resistance to herbivores and pathogens: ecology, evolution, and genetics, p. 590. Chicago, IL: University of Chicago Press. 26. Agrawal AA, Lau JA, Hamback PA. 2006 Community heterogeneity and the evolution of interactions between plants and insect herbivores. Q. Rev. Biol. 81, 349– 376. (doi:10.1086/511529) 27. Potter DA, Held DW. 2002 Biology and management of the Japanese beetle. Annu. Rev. Entomol. 47, 175– 205. (doi:10.1146/annurev.ento.47.091201.145153) 28. McArt SH, Halitschke R, Salminen JP, Thaler JS. 2013 Leaf herbivory increases plant fitness via induced resistance to seed predators. Ecology 94, 966 –975. (doi: 10.1890/12-1664.1) 29. Cleland RE. 1972 Oenothera: cytogenetics and evolution. New York, NY: Academic Press. 30. Johnson M. 2011 The contribution of evening primrose (Oenothera biennis) to a modern synthesis of evolutionary ecology. Popul. Ecol. 53, 9 –21. (doi:10.1007/s10144-010-0250-5) 31. Larson EL, Bogdanowicz SM, Agrawal AA, Johnson MTJ, Harrison RG. 2008 Isolation and characterization of polymorphic microsatellite loci in common evening primrose (Oenothera biennis). Mol. Ecol. Res. 8, 434 –436. (doi:10.1111/j.1471-8286. 2007.01983.x) 32. Johnson MTJ. 2007 Genotype-by-environment interactions lead to variable selection on life-history strategy in common evening primrose (Oenothera biennis). J. Evol. Biol. 20, 190 –200. (doi:10.1111/j. 1420-9101.2006.01201.x) 33. Barbosa P, Hines J, Kaplan I, Martinson H, Szczepaniec A, Szendrei Z. 2009 Associational resistance and associational susceptibility: having right or wrong neighbors. Annu. Rev. Ecol. Evol. Syst. 40, 1–20. (doi:10.1146/annurev.ecolsys. 110308.120242) 34. Cardinale BJ, Wright JP, Cadotte MW, Carroll IT, Hector A, Srivastava DS, Loreau M, Weis JJ. 2007 Impacts of plant diversity on biomass production increase through time because of species complementarity. Proc. Natl Acad. Sci. USA 104, 18 123– 18 128. (doi:10.1073/pnas.0709069104) 35. Tilman D, Wedin D, Knops J. 1996 Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature 379, 718– 720. (doi:10.1038/ 379718a0) 36. van Ruijven J, Berendse F. 2005 Diversity – productivity relationships: initial effects, long-term patterns, and underlying mechanisms. Proc. Natl Acad. Sci. USA 102, 695– 700. (doi:10.1073/pnas. 0407524102) 37. Fargione J, Tilman D, Dybzinski R, Lambers JHR, Clark C, Harpole WS, Knops JMH, Reich PB, Loreau rspb.royalsocietypublishing.org 10. Crutsinger GM, Collins MD, Fordyce JA, Gompert Z, Nice CC, Sanders NJ. 2006 Plant genotypic diversity predicts community structure and governs an ecosystem process. Science 313, 966 –968. (doi:10. 1126/science.1128326) 11. Reusch TBH, Ehlers A, Hammerli A, Worm B. 2005 Ecosystem recovery after climatic extremes enhanced by genotypic diversity. Proc. Natl Acad. Sci. USA 102, 2826 –2831. (doi:10.1073/pnas.0500008102) 12. Utsumi S, Ando Y, Craig TP, Ohgushi T. 2011 Plant genotypic diversity increases the population size of a herbivorous insect. Proc. R. Soc. B 278, 3108–3115. (doi:10.1098/rspb.2011.0239) 13. Hughes AR, Stachowicz JJ. 2004 Genetic diversity enhances the resistance of a seagrass ecosystem to disturbance. Proc. Natl Acad. Sci. USA 101, 8998–9002. (doi:10.1073/pnas.0402642101) 14. Parker JD, Salminen J-P, Agrawal AA. 2010 Herbivory enhances positive effects of plant genotypic diversity. Ecol. Lett. 13, 553–563. (doi:10.1111/j.1461-0248.2010.01452.x) 15. Duffy JE. 2002 Biodiversity and ecosystem function: the consumer connection. Oikos 99, 201 –219. (doi:10.1034/j.1600-0706.2002.990201.x) 16. Leibold MA. 1989 Resource edibility and the effects of predators and productivity on the outcome of trophic interactions. Am. Nat. 134, 922–949. (doi:10.1086/285022) 17. Loreau M, Hector A. 2001 Partitioning selection and complementarity in biodiversity experiments. Nature 412, 72 –76. (doi:10.1038/35083573) 18. Tilman D. 1982 Resource competition and community structure. Princeton, NJ: Princeton University Press. 19. DeMott WR. 1998 Utilization of a cyanobacterium and a phosphorus-deficient green alga as complementary resources by daphnids. Ecology 79, 2463 –2481. (doi:10.1890/00129658(1998)079[2463:UOACAA]2.0.CO;2) 20. Steiner CF. 2001 The effects of prey heterogeneity and consumer identity on the limitation of trophiclevel biomass. Ecology 82, 2495 –2506. (doi:10. 1890/0012-9658(2001)082[2495:TEOPHA]2.0.CO;2) 21. Haddad NM, Crutsinger GM, Gross K, Haarstad J, Knops JMH, Tilman D. 2009 Plant species loss decreases arthropod diversity and shifts trophic structure. Ecol. Lett. 12, 1029–1039. (doi:10.1111/j. 1461-0248.2009.01356.x) 22. Salazar D, Marquis RJ. 2012 Herbivore pressure increases toward the equator. Proc. Natl Acad. Sci. USA 109, 12 616–12 620. (doi:10.1073/pnas. 1202907109) 23. McArt S, Cook-Patton S, Thaler J. 2012 Relationships between arthropod richness, evenness, and diversity are altered by complementarity among plant genotypes. Oecologia 168, 1013–1021. (doi:10. 1007/s00442-011-2150-6)