Nutrient limitation of forest growth has been difficult to predict, and in temperate forests, long-term tests of single-nutrient versus multiple-element limitation are few. Nutrient co-limitation is the expected outcome of the ability of plants to adjust allocation to minimize limitation by any single resource. Nutrient limitation of productivity in northern hardwood forests was predicted by the Multiple Element Limitation (MEL) model to shift over time since harvest from single limitation by N to P at ~30 years and then, in mature forests, to co-limitation by N and P. Our work tested those predictions for tree growth in a fully factorial N and P addition experiment in 13 forest stands that we grouped in young (20-30 years), mid-age (40-50 years), and mature (>100 years old) age classes in New Hampshire, USA. Over 8 years of treatment, we found evidence of additive co-limitation of tree growth by N and P. We did not find evidence that limitation varied with time since disturbance. Our results suggest that processes contributing to co-limitation in these northern hardwood forests are effective across stands that vary widely in N status and are not sensitive to disturbance by forest harvest over time periods of several decades.

Ecological niche differentiation is a process that accompanies lineage diversification and community assembly. Traditionally, the degree of niche differentiation is estimated by contrasting niche hypervolumes of two taxa, reconstructed using ecologically relevant variables. These methods disregard the fact that niches can shift in different ways and directions. Without means of discriminating between different types of niche differentiation, important evolutionary and ecological patterns may go unrecognized. Herein, we introduce a new conceptual and methodological framework that allows quantification and classification of niche differentiation and divergence between taxa along single niche axis. This new method, the Niche Divergence Plane, is based on species' responses to an underlying environmental gradient, from which we derive a two-dimensional plane defined by two indices, niche exclusivity and niche dissimilarity. These two indices identify the proportion of the environmental gradient that is unique to each species, that is, how much of the environmental gradient species do not share (niche breadth exclusivity) and how different the species' responses are along the environmental gradient (niche dissimilarity). Thus, the latter can also be seen as a measure of the differences in niche preference or importance, even when there is significant overlap in niche breadth (i.e., low niche exclusivity). Based on the position of the two indices on the divergence plane, we can distinguish niche conservatism from four other general types of niche divergence: hard, soft, weighted, and nested. We demonstrate that the Niche Divergence Plane complements traditional measures of niche similarity (e.g., Schoener's D or Hellinger's I). Additionally, we show an empirical comparison using the Niche Divergence Plane framework on two Ambystoma salamanders. Overall, we demonstrate that the Niche Divergence Plane is a versatile tool that can be used to complement and expand previous methods of ecological niche comparisons and the study of ecological niche divergence.

Endemic plant conservation is important for maintaining regional diversity and in reducing homogenization. Research on rare endemics often focuses on genetic variability and pollen limitation. Less well studied is how other life history traits impact reproduction. We examined the effects of plant size on demographics of a threatened species, smooth coneflower (Echinacea laevigata), which is a narrow endemic in forest glade habitats in the southeastern USA. We tagged and recorded the size and reproductive status of plants in five plots (e.g., subpopulations) in Montgomery County, Virginia, USA, from 2014 to 2017. We used integral projection modeling to assess key aspects of the life cycle, estimate how changes in size and reproductive status impacted population growth, survival, and future reproduction. We found that very few plants reproduced each year (range: 1.2-18.2%). Plant size was a strong predictor of survival, reproduction, and growth probabilities the next year. We identified a population level cost of reproduction that reduced lambda by 2%. Reproductive plants experienced additional costs including a 5.7% decline in survival, 16% decline in growth, and a 10% lower chance of flowering in the next year compared to plants that were not reproductive. Overall, lambdas were above 1 indicating the population was growing in all years. This work demonstrated plant size had strong impacts on reproductive dynamics and reproduction can incur a cost on future population dynamics. Therefore, management of this species should focus on preserving environmental conditions that promote plant growth such as preventing woody encroachment.

Different groups of taxa exhibit varying degree of climatic niche conservatism or divergence due to evolutionary constraints imposed on taxa and distributional relationships among them. Herein, we explore to what extent regional environmental conditions that taxa occupy affect climatic niche overlap between pairs of congeneric species of Peromyscus mice exhibiting allopatric, parapatric, or sympatric distributions. We used Bayesian generalized linear mixed models to identify environmental variables that best explain differences in climatic niche overlap between species. Our results suggest that regional environmental conditions explain 13–44% of variation in climatic niche overlap. Specifically, allopatric and parapatric species pairs are more likely to occupy similar climatic niches in areas that are topographically less complex but with more complex habitats. Sympatric species are more likely to occupy similar climatic niches in areas that promote local niche partitioning (topographically less complex, warmer winter temperatures, higher precipitation, and higher habitat complexity on a local scale). By understanding the relationship between regional environmental conditions and niche overlap, we highlight how differences in geography can contribute to shaping niches of congeneric species.

Aim Despite the large literature documenting the negative effects of invasive grasses, we lack an understanding of the drivers of their habitat suitability, especially for shade-tolerant species that do not respond positively to canopy disturbance. We aimed to understand the environmental niche and potential spatial distribution of a relatively new invasive species, wavyleaf basketgrass (Oplismenus undulatifolius(Ard.) Roem. & Schult, WLBG) by leveraging data available at two different spatial scales. Location Mid-Atlantic region of the United States. Methods Maximum entropy modeling (Maxent) was used to predict the habitat suitability of WLBG at the regional scale and the landscape scale. Following variable evaluation, model calibration, and model evaluation, final models were created using 1,000 replicates and projected to each study area. Results At the regional scale, our best models show that suitability for WLBG was driven by relatively high annual mean temperatures, low temperature seasonality and monthly range, low slope, and high cumulative Normalized Difference Vegetation Index (NDVI). At the landscape scale, suitability was highest near roads and streams, far from trails, at low elevations, in sandy, moist soil, and in areas with high NDVI. Main Conclusions We found that invasion potential of this relatively new invader appears high in productive, mesic habitats at low slope and elevations. At the regional scale, our model predicted areas of suitable habitat far outside areas where WLBG has been reported, including large portions of Virginia and West Virginia, suggests serious potential for spread. However, large portions of this area carry a high extrapolation risk and should therefore be interpreted with caution. In contrast, at the landscape level, the suitability of WLBG is largely restricted to areas near current presence points, suggesting that the expansion risk of this species within Shenandoah National Park is somewhat limited.

With the spread of a new invasive plant species, it is vital to determine the effectiveness of removal strategies as well as their advantages and disadvantages before attempting widespread removal. While thousands of dollars have been spent to curtail the spread of wavyleaf basketgrass [Oplismenus undulatifolius (Ard.) P. Beauv.], a relatively new invasive species, the lack of a cohesive management plan and funding has made controlling this species especially difficult. We assessed the efficacy of a variety of chemical control methods and hand weeding for this species and followed select methods over time. We also assessed the potential for ecosystem recovery following removal by measuring total and native species richness in response to treatments. Our pilot study revealed a wide breadth of responses to our eight herbicides, with fluazifop plus fenoxaprop, imazapic, quizalofop, and sulfometuron methyl being the least effective. In our follow-up experiments, hand weeding, glyphosate, and clethodim treatments were effective at reducing O. undulatifolius percent cover, density, and biomass, with an average reduction of at least 48% in the first year. However, we found substantial variation in the effectiveness of clethodim between our two experiments, which was likely driven by site differences. We also found that all three of these removal methods were effective at reducing the number of O. undulatifolius flowering stems and the height of those stems, which will likely reduce the spread of this species to new areas. Finally, we found that these methods have the potential to restore total and native species richness, but that glyphosate-treated plots did not fully recover until 2 yr after treatment.

Ancient human–landscape interactions have left their imprint on plant communities around the world. The growing body of literature on long-term legacies demonstrates that legacy effects in temperate and tropical systems are both common and long lasting. In eastern North America, legacies of extensive landscape management by pre-Columbian Native Americans are found in soil characteristics (Cook-Patton et al., 2014), species distributions (Warren, 2016), and species composition (Foster et al., 2003). Ross et al. (2014) argued that ancient Native American landscape management may have induced divergence from dioecy in an historically important fruit tree. More than 1500 years after the fall of the Roman Empire, effects of ancient agriculture on soil properties, and consequently plant species richness and composition, still reverberate in Europe (Dambrine et al., 2007). Throughout the Amazon, patches of rich, anthropogenic soils called Amazonian Dark Earths (locally, terra preta) highlight the landscape altering impacts of ancient indigenous land use (Lehmann, 2009; Schmidt et al., 2014). Similarly, the legacy of ancient cultures reverberates today in the soils and vegetation of Mesoamerican forests (Ross, 2011; Beach et al., 2018). These legacies pose exciting questions about the dynamics and stability of ecological communities where humans are part of, rather than apart from, ecosystems (Foster et al., 2003). Legacy effects are interspecific, long-term effects that persist after the causal species is gone or ends the causal activity (Cuddington, 2011). Legacies can have important implications for understanding current biodiversity patterns and processes. As we continue to uncover these legacy effects, it will be useful to place them in a general framework to help us understand key features that various legacies share, and key features that are unique. In this essay, we argue that dynamic systems theory would be such a framework. We propose that human landscape legacies can be more fully understood within a dynamic systems framework (Beisner et al., 2003 and references therein). A system is depicted intuitively as a ball rolling through a hilly landscape (Fig. 1). The horizontal position of the ball depicts the system state (e.g., abundance of a harvested species), and the low point in each basin represents a local stable equilibrium point. Just as a ball rolls downhill away from peaks and into basins, a system is repelled away from some states and attracted to others. The dynamic systems framework derives from physics, as a way to describe how physical systems change through time. It now helps us understand phenomena as diverse as evolution in adaptive fitness landscapes (Nowak, 2006) and the dynamics of coupled social-ecological systems (Carpenter et al., 2009). A dynamic system (Fig. 1A) comprises all state variables and external factors. State variables are those whose dynamics, or variation through time, we wish to study. State variables may be the abundance of a one or more species or an ecosystem variable. State variables are characterized by relatively fast reciprocal feedbacks and fast responses to external factors. The state of the system is the set of values of the state variables (x-axis, Fig. 1). External factors are those that are hypothesized to govern the state variables. External factors change relatively slowly, such as climate, or parent material of soil. External factors affect the state variables but are not affected by them, so the feedbacks are not reciprocal. Because they change so slowly, we treat them as constant parameters in a model of the state variables. They create the context within which the state variables change through time. We depict this context as the topography of the dynamic landscape (Fig. 1). A system state is said to be locally stable and self-sustaining if it returns to that state after a small perturbation. This state is depicted as a ball residing at the lowest point in a particular valley. A perturbation may be anything that shifts the state variables independently of the topography of the landscape (solid arrows, Fig. 1A, B), such as a forest fire, extreme weather event, or a timber harvest. A perturbation can also be something that affects the external factors and alters the parameters, thereby creating a new landscape topography (open arrow, Fig. 1C). An example of this may be severe erosion that eliminates upper soil horizons, setting back soil development for millennia. Amazonian Dark Earths (terra preta) are a well-known legacy of Pre-Columbian Native American land management that can fit readily within the dynamic systems framework (Lehmann, 2009; Schmidt et al., 2014). Radiocarbon dating of terra preta dates them to the mid-Holocene and earlier with many dating to the first millennium BCE (Schmidt et al., 2014). These fertile patches of charcoal-enriched soil have structured forests since their abandonment hundreds to thousands of years ago. Amazonian soils are typically clay-dominated and nutrient poor, yet terra preta contain significantly higher levels of both labile and stable soil organic matter and increased cation exchange capacity (Glaser and Birk, 2012). These soil characteristics result in significantly different forest species composition in terra preta sites versus in the surrounding forest matrix (Junqueira et al., 2010). The high fertility and stable soil organic matter have led authors to suggest that terra preta can be important for both sustainable agriculture and long-term carbon sequestration in the tropics (Glaser and Birk, 2012). A dynamic systems approach to terra preta helps us understand the interactions between the fast dynamics of the forest plant community and the slow dynamics of terra preta soils. A conceptual ecosystem model of terra preta genesis (Glaser and Birk, 2012) captures key dynamic processes that drive both the fertility and the long-term persistence of these soils. Large amounts of charcoal in terra preta leads to long-term stability of soil organic matter, high nutrient mineralization rates, and recalcitrance to microbial degradation (Lehmann, 2009; Glaser and Birk, 2012). This charcoal has residence times of millennia (Lehmann, 2009) indicating that terra preta soils change so slowly that they are effectively permanent relative to rates of change in the biota. Thus, formation of terra preta changed the parameters defining the dynamic landscape (type III, Fig. 1C). It will be fascinating to see whether future research, that quantifies rates of critical soil nutrient and carbon cycles, can further evaluate this hypothesis for terra preta. Of course, historical human cultures can each exert multiple legacy effects. In Mesoamerica, the ancient Maya left behind vast erosive earthworks that permanently altered soils and hydrology, creating type III legacy effects (Turner and Sabloff, 2012; Beach et al., 2018). However, other historical patterns in the Maya forest have been identified that are not related to extensive soil alterations. Working in the forests around the ancient Maya city of El Pilar, Ross (2011) revealed a strong signature of ancient Maya residential forest gardens in the tree species composition of the modern forest. El Pilar was abandoned nearly 1000 years ago, yet these extant garden forests harbor surprisingly high abundance of woody species once used by the Maya for medicine, food, and construction. Ross (2011) identified little edaphic difference between Maya garden forest plots and the surrounding forest matrix, suggesting that these garden forests do not qualify as type III legacy effects. We hypothesize that Maya garden forests may constitute an example of an alternative basin of attraction legacy effect (type II, Fig. 1B). The resulting alternative stable states are notoriously difficult to demonstrate empirically (Petraitis, 2013). It requires showing that (1) variables external to the system are unchanged, while (2) variables within the system show feedbacks that create two or more basins with local stable equilibria. In our current research, we propose that alteration of tree species composition by the ancient Maya attracted high numbers of frugivores and seed dispersers, which then stabilized this community type such that it persists today. Analysis of tree dispersal and recruitment around potential hub tree species (Bascompte et al., 2003) should allow us to test whether Maya garden forests do, in fact, represent a type II legacy effect. We anticipate that any major civilization is likely to leave a variety of legacies. The strength of the dynamic systems approach is that it places all legacy effects, short and long term, with biological or physical causes, into a common theoretical framework. In general, theory helps organize, guide, and advance scientific knowledge (NRC, 2008). It helps us understand that the infinite variety of species and ecosystems share common principles (Lawton, 1992). Dynamic systems theory provides a way to describe, understand, and predict the change through time of any species or ecosystem using a common language and well-established tools. Here, it provides us with a basis for organizing legacy effects into three types of dynamics, each differing in their underlying mechanisms in testable ways. To date, research on human landscape legacies has revealed a tremendous diversity of patterns and mechanisms that are specific to each site and each historical culture. The framework presented here complements this and places the dynamics of these legacies alongside those of countless other species and ecosystems whose dynamics we seek to understand and manage. The authors thank P. K. Diggle and two anonymous reviewers whose thoughtful suggestions led to a clearer presentation of these ideas.

Very little is known about how emerald ash borer will affect hardwood swamps, which often have a large ash component compared to upland forests. The aim of this research project was to investigate the current status of ash swamps that have become infested with emerald ash borer and make predictions about the future state of these forests. We surveyed three hardwood swamps in southern Michigan that varied in the amount of ash present (12.4%-58.8% basal area), which appears to affect how these communities will change as a result of emerald ash borer. Our analyses indicate that the two swamps with the least amount of ash will likely experience a proportional increase in surrounding trees, while structure of the site with the most ash will be altered in the future and may transition to a forest that is strongly influenced by shrub-like individuals as well as non-ash canopy trees.

Environmental conditions may affect individual physiological processes that influence short-term performance and ultimately growth, survival and reproduction. As such, habitats selected by animals must provide suitable and adequate resources. Ectothermic species are highly dependent on climatic conditions and ambient temperatures that dictate body temperature regulation and in turn physiological processes. We investigated the thermoregulatory performance, habitat selection, and movements of an ectothermic vertebrate, the Eastern box turtle (Terrapene carolina carolina) to assess the importance of thermoregulatory physiology in habitat selection. We evaluated the relationship between habitat selection and thermoregulatory performance in Southwest Ohio over two active seasons from May until October. We found that T. carolina selected shaded habitats, including evergreen and deciduous forests, as well as herbaceous grasslands, conformed to the ambient temperatures throughout the active season, although these habitats had temperatures below those expected based on thermal optima of box turtles. Further, we found that movement was not correlated with internal body temperature. Our study shows that thermal conditions are not paramount in habitat selection of box turtles, but that cooler temperatures do not have an effect on the extent of their locomotion.

An increase in the quantity of available resources is known to affect temporal variability of aggregate community properties. However, it is unclear how might fluctuations in resource availability alter community-level temporal variability. Here we conduct a microcosm experiment with laboratory protist community subjected to manipulated resource pulses that vary in intensity, duration and time of supply, and examine the impact of fluctuating resource availability on temporal variability of the recipient community. The results showed that the temporal variation of total protist abundance increased with the magnitude of resource pulses, as protist community receiving infrequent resource pulses (i.e., high-magnitude nutrients per pulse) was relatively more unstable than community receiving multiple resource pulses (i.e., low-magnitude nutrients per pulse), although the same total amounts of nutrients were added to each community. Meanwhile, the timing effect of fluctuating resources did not significantly alter community temporal variability. Further analysis showed that fluctuating resource availability increased community temporal variability by increasing the degree of community-wide species synchrony and decreasing the stabilizing effects of dominant species. Hence, the importance of fluctuating resource availability in influencing community stability and the regulatory mechanisms merit more attention, especially when global ecosystems are experiencing high rates of anthropogenic nutrient inputs.

Aquatic ecosystem management requires knowledge of the links among landscape-level anthropogenic disturbances and aquatic ecosystem properties. With large catchment area to surface area ratios (CA: SA), reservoirs often receive substantial terrestrial subsidies and can be particularly sensitive to eutrophication. Reservoir numbers and attendant management problems are increasing, and tools are needed to categorize their eutrophication status. We analyzed a dataset of 109 reservoirs in Ohio (USA) in an effort to classify eutrophication status using landscape-level features and reservoir morphometry. These predictor variables were selected because they are relatively stable and easily measured. We employed regression tree analysis and used a composite eutrophication variable as our response variable. Our regression tree analysis accurately divided 67% of Ohio reservoirs into 4 eutrophication status groups using 3 predictor variables: percentage of catchment area composed of agriculture versus forest; maximum reservoir depth; and CA: SA. We can infer that reservoirs with catchments containing >71% forest will likely be oligotrophic to mesotrophic. For reservoirs with <71% catchment forest, trophic status is determined by the relative extent of catchment row crops and either CA: SA or maximum depth. We applied our regression tree to a subset of reservoirs in the Environmental Protection Agency's National Lakes Assessment (NLA; n = 339 reservoirs). With a few exceptions, we categorized NLA reservoirs by eutrophication status despite their broad geographical range across the contiguous USA. Our results show that a few easily measured, stable parameters can classify reservoir eutrophication status. Models like ours may be useful for broad-scale management decisions.

• Premise of the study: A growing body of literature now documents how ancient human management of the landscape echoes through to extant environments in eastern North America. Plant domestication is a major theme in the study of human–nature interactions. Long‐term ecological impacts of human selection may last for centuries after management ends, yet little work has focused on legacies in the evolution of historically used trees. Ecological data will be valuable in teasing apart myriad variables that confound questions of land‐use legacies. We discuss the potential for legacies of ancient human selection and present a preliminary case study for the approach of integrating ecological and historical data for Diospyros virginiana, the American persimmon.• Methods: Herbarium samples of D. virginiana (28 male and 40 female) from across the species range provided specimen localities for edaphic analysis. Soil and environmental data were analyzed using nonparametric ordination, Wilcoxon summed rank test, and permutational MANOVA.• Key results: Edaphic data demonstrated substantial variation among sites, but revealed no significant differences between sexes. Permutational MANOVA showed no difference in environmental preferences for the tested variables between male and female trees (R2 < 0.01, P = 0.8).• Conclusions: Extending our understanding of landscape history to the long‐term impacts of artificial selection at the species or population level would be valuable in both theoretical and applied botanical research. Multidisciplinary approaches integrating ecological data will be essential for investigation of the evolutionary implications of historical human selection in economic species and the potential for adaptive flexibility in reproductive systems of long‐lived perennials.

ABSTRACTDifferent disturbances in similar habitats can produce unique successional assemblages of plants. We collected plant species composition and cover data to investigate the effects of three common types of disturbances—fire, anthropogenic clearing (‘cleared’), and clearing followed by goat grazing (‘cleared‐and‐grazed’)—on early‐successional coppice (dry forest) community structure and development on Eleuthera, Bahamas. For each disturbance type, both the ground layer (<0.5 m height) and shrub layer (>0.5 m height) were sampled in eight patches (>1 ha) of varying age (1–28 yr) since large‐scale mature coppice disturbance. Overall, plant communities differed among disturbance types; several common species had significantly higher cover in the shrub layer of fire patches, and cleared‐and‐grazed patches exhibited higher woody ground cover. Total percent cover in the shrub layer increased in a similar linear fashion along the investigated chronosequence of each disturbance type; however, cover of the common tree species, Bursera simaruba, increased at a notably slower rate in cleared‐and‐grazed patches. The pattern of increase and subsequent decrease in cover of Lantana spp. and Zanthoxylum fagara in the shrub layer was characterized by longer persistence and higher covers, respectively, in cleared‐and‐grazed patches, which also exhibited low peak cover and fast decline of nonwoody ground cover. Our results suggest that goats may accelerate some aspects of succession (e.g., quickly removing nonwoody ground cover) and retard other aspects (e.g., inhibiting growth of tree species and maintaining early‐successional shrubs in the shrub layer). These effects may lead to different successional trajectories, and have important conservation implications.

This lab is a variant of the (revised/in revision) supplement by Bolker et al. of Bolker et al. (2009). Spatial variation in nutrient availability and herbivory is likely to cause population differentiation and maintain genetic diversity in plant populations. Here we measure the extent to which mouse-ear cress (Arabidopsis thaliana) exhibits population and genotypic variation in their responses to these important environmental factors. We are particularly interested in whether these populations exhibit nutrient mediated compensation, where higher nutrient levels allow individuals to better tolerate herbivory. We use GLMMs to estimate the effect of nutrient levels on tolerance to herbivory in Arabidopsis thaliana (fixed effects), and the extent to which populations vary in their responses (variance components)1. In this example, we step deliberately through the process of model building and analysis. The primary approach uses glmer from the lme4 package to analyze the data as a lognormal-Poisson hierarchical model, but we also make use of other approaches. Load packages:

The fluctuating resource hypothesis (FRH) proposes that fluctuations in resource supply can temporally reduce competitive pressure from resident species, thereby providing ephemeral opportunities for invading species. Although FRH has the potential to integrate many existing hypotheses regarding mechanisms of community invasibility, previous tests and evaluations of FRH were based on single trophic level, did not take the timing effect into account, and had difficulties in distinguishing the effects of resource pulses from other simultaneous processes. Here we test FRH in multi-trophic aquatic microcosms by creating resource pulses, by controlling resource quantity, propagule supply and pulse recurrence frequency, and by manipulating the timing of pulses relative to the timing of the arrival of new species (i.e. invaders) to local communities. The novelty of our work lies in that we directly manipulate resource pulse timing relative to invader introduction events and thus demonstrate the importance of this timing effect for community invasibility. Our study supports FRH in general: invasion success was positively related to resource pulses, and invaders had strong performance in treatments receiving coincident pulses, although not all invaders gained more benefit when resources were supplied at large-magnitude than supplied at continuous rates. Since many ecosystems worldwide are experiencing high rates of anthropogenic nutrient input and increasing rates of precipitation, these ecosystems are potentially more fragile and susceptible to invasion. More experiments across multiple ecosystem types are needed to help formulate a general theory of community invasibility.

Nutrient enrichment, ecosystem size, and richness each may directly affect the stability of both populations and communities. Alternatively, nutrient enrichment and ecosystem size each may directly affect richness, which in turn may affect stability. No previous studies, however, have tested empirically how these three factors interact and co-determine stability. We manipulated nutrient input and ecosystem size in replicate microcosms containing a diverse bacterial flora, and a range of green algae and heterotrophic protozoa, and used these manipulations and the resulting variation in species richness to measure their combined effects on temporal stability of both populations and communities. Results showed that nutrient enrichment and ecosystem size controlled protist richness, and their effects on stability could be mediated by richness. In addition, both community-level and population-level stability increased with protist richness. Furthermore, mean species evenness and mean species richness was negatively related. Effects of statistical averaging, overyielding, and component population stability were identified as possible mechanisms involved explaini ng the stabilizing effects of richness on community stability. Their relative strength in influencing stability, however, is likely to change as mean evenness decreased with increasing richness. This decrease in evenness would tend to weaken the strength of the statistic averaging effect, but increase the strength of the other two mechanisms due to relatively lower population variability (component population stability) and higher mean biovolumes of dominant protists (overyielding).

It is widely reported that the productivity-richness relation (PRR) is highly variable, and several field studies suggest that the PRR varies with observational scale. Here we provide the first experimental study to test whether the PRR is scale-dependent when all replicate ecosystems have similar initial conditions. We also test the relation between productivity and compositional dissimilarity, and whether the PRR varies with ecosystem size. Moderately complex replicated microcosms were assembled consisting of a range of protozoa, algae, and a diverse bacterial flora. We found that the PRR of protozoan and algal communities varied with observational scale, but was unrelated to ecosystem size. Specifically, protozoan and algal richness increased monotonically with productivity at the local scale, but became flattened at the regional scale. This varying PRR at different scales occurred because dissimilarity among replicates decreased with productivity. Thus, in this model system, our experimental approach found a different form of scale dependence than previous field research. We speculate that this difference results from different processes governing extinctions at low levels of productivity.

Tolerance to apical meristem damage (AMD) is a form of plant defense against herbivory. Theoretical models come to different conclusions about the effects of inorganic soil nutrient levels on tolerance to AMD, and different plants have shown different relationships between these variables. To assign some order to these disparate patterns and to resolve conflicts among the models, the ‘limiting resources model’ (LRM) was developed. However, we believe that the LRM is actually comprised of several different models, which we describe. Our study marks the first comprehensive and simultaneous test of the entire LRM framework, treating it explicitly as separate models, which also evaluates the models’ underlying assumptions. We studied tolerance to AMD in laboratory‐reared natural populations of Arabidopsis thaliana from three different regions of Europe, spanning a wide latitudinal gradient. We show that, in different populations of this species, basic responses to nutrients and damage are best described by different models, which are based on different assumptions and make different predictions. This demonstrates the need for complexity in our explanations, and suggests that no one existing model can account for all relationships between tolerance to AMD and nutrients. Our results also demonstrate that fruit production can provide a misleading approximation of fitness in A. thaliana, contrary to the common assumption in the literature.