Lying on my side, I inspect each leaf of a four-year-old red maple (Acer rubrum) sapling. It is midsummer in central Minnesota, and, despite the hot weather, the leaves of this tree and most of its neighbors are covered in brown spots. These spots are a symptom of maple anthracnose, which is caused by infection with any of several ascomycete fungi. Especially in small trees, like the one I am inspecting, anthracnose can slow growth by impairing photosynthesis in damaged leaves and make them more vulnerable to other infections. I finish grading the infected leaves on my tree of interest, record the measurements on my clipboard, and roll over to check its nearest neighbor, also a red maple. Planted only 50 cm (about a foot and a half) apart from each other, these trees are part of the Forests and Biodiversity (FAB) experiment at the Cedar Creek Ecosystem Science Reserve, a long-term research site funded by the National Science Foundation. These trees are in a red maple monoculture plot, meaning each tree has only other red maples as neighbors. Many of these trees are covered in leaf anthracnose, probably because the fungal pathogens that cause it overwinter in the layer of dead leaves, or litter, of infected trees and emerge in the spring to reinfect new growth. Having so many red maples around seems to make any given tree more likely to develop anthracnose.
But when I stand up and walk only a few feet away from this maple monoculture, which only reaches my mid-thigh after several years of growth, I encounter a markedly different part of the FAB experiment. On entering a “biculture,” or two-species plot, of red maple and jack pine (Pinus banksiana), I am surrounded by pines that exceed my own height and maples that reach my chest. These trees were planted at the same time as the maples in the adjacent monocultural plot, but living with diverse neighbors has clearly made a difference to them. Whereas maples in monoculture tend to be short and stocky, with little space between leaf buds, maples that are forced to compete for light with faster-growing pine tend to come up long and spindly, with tons of space between each set of leaf buds. And, exposed to a lower level of accumulated leaf litter from other red maples, these trees have a much lower incidence of leaf anthracnose.
These two plots illustrate two extremes of the FAB experiment, which I comanaged while a graduate student at the University of Minnesota. Previously, as a young ecologist, I had devoured the reports of experiments designed to assess the importance of biodiversity coming out of Cedar Creek and applied to a doctoral program at Minnesota in the hopes that I would get to work there. I quickly heard back from Jeannine Cavender-Bares, a plant ecophysiologist and evolutionary ecologist who would become my doctoral advisor. At the time, she and a group of colleagues at Minnesota were in the final stages of planning a tree biodiversity experiment designed to expand on the foundation laid by grassland experiments at Cedar Creek and across Europe. This group was open to bringing in a new graduate student to help with the establishment of the new project and to do some preliminary research. Freshly returned from two years as a Peace Corps volunteer in subtropical Paraguay, I dug out my long underwear and moved to Minneapolis.
On the Origin of Biodiversity Research
But why go to the trouble of planting thousands of trees in various combinations of species, then take the time to make thousands of minute measurements of their every centimeter of growth and bout with an illness or pest? For me, as for a generation of researchers at Cedar Creek, experiments like FAB have emerged as a powerful approach for asking what role biodiversity, meaning the variety of life in a particular place or across the globe, might play in keeping the natural world working in the way we prefer it to do. To understand these experiments and the findings that come from them, though, we ought to take a step back and consider the history of what I would call the science of biodiversity.
When we speak of the environmental challenges of the current era, we can hardly avoid worrying about the erosion of biodiversity. While considerable disagreement persists over how biodiversity across the earth’s diverse landscapes has changed over the last ten thousand years, a period in which humans have exerted a growing influence over the biosphere, a consensus has emerged that, at a global level, our planet has entered into a period of precipitous biodiversity loss (Butchart et al., 2010). A case in point: a recent report from a group of Danish and Swedish authors predicts that the loss of mammal diversity that has taken place since the end of the last Ice Age will take 2.5 billion years—two-thirds of the time during which there has been life on the planet—to be replenished by natural evolutionary processes (Davis et al., 2018).
Yet such sobering statistics beg a second question: does biodiversity loss really matter? Of course, to many of us—including, I imagine, most readers of Arnoldia—the diversity of earth’s species represents an irreplaceable gift. We sense a precious value, whether spiritual, emotional, or cultural, inherent to the diversity of life on earth. It is challenging, though, to convince others. And so, those of us who wish to protect biodiversity must ask ourselves whether there is an extrinsic value to diversity and, if so, how we can justify its conservation. The ecosystem services movement has answered this question by, at least to some extent, evaluating biodiversity in terms of dollars and cents. For instance, Canadian scholars Robin Naidoo and Wiktor Adamowicz (2006) estimate that the financial returns from visits by ecotourists to a Ugandan park rich in bird biodiversity far exceed the costs of maintaining the park. This approach, however, doesn’t fully capture a deeper question: does biodiversity support the vast array of ecosystem processes—or functions—that keep our biosphere working and, in doing so, sustain human life? In other words, are more-biodiverse ecosystems stronger and more resilient?
The history of this question is a long one, with origins prior to the formal elaboration of the concepts of biodiversity (by American conservation biologist Raymond F. Dasmann in 1968), of ecosystems (by English botanist Arthur Tansley in 1935), and of ecology itself (by German biologist Ernst Haeckel in 1866). Instead, the question of how biodiversity affects ecosystem function was posed first, at least within Western scientific discourse, by the founding mind of modern biology, Charles Darwin.
While Darwin is known foremost for his role in developing the concept of evolution by natural selection, his works also offer up a clairvoyant catalogue of research questions for contemporary biologists, one that we have yet to plumb fully some 130 years following his death. Through his lifetime, Darwin contributed important insights to the study of insect pollination; plant physiology; soil formation; the genetic origins of animal behavior; and the natural history of barnacles, coral reefs, and carnivorous plants; among other topics. Indeed, if we turn to Darwin’s On The Origin of Species, first published in 1859, we find a claim that, though peripheral to the broader case for adaptive evolution, constitutes the origin of an important field of biodiversity research: “If a plot of ground be sown with one species of grass, and a similar plot be sown with several distinct genera of grasses, a greater number of plants and a greater weight of dry herbage can thus be raised.” In this brief aside, Darwin argues that it was, at the time, well known that more-biodiverse systems—well, grasslands, at least—ought to be more productive than less-diverse ones.
Indeed, indigenous peoples, and especially farmers, have known for millennia that more-diverse ecosystems are more productive. For instance, the “three sisters” technique of growing diverse gardens of corn, beans, and squash developed in pre-Columbian central Mexico and subsequently radiated throughout the Americas. Contemporary studies have demonstrated that this system of polyculture—or growing multiple crop species together—boosts yield compared to monocultures of constituent species (Zhang et al., 2014). Experimental assessment of traditional Chinese polycultures consisting of varying mixtures of wheat, corn, and soybeans have revealed similar trends (Zhang and Li, 2003). Such traditional techniques have continued to evolve to this day, resulting in, among other practices, the contemporary interest in “companion planting” among home gardeners and farmers. For instance, many gardeners in North America are familiar with the practice of planting African marigolds, mints, and other aromatics in their gardens, both for their own aesthetic or culinary uses and, allegedly, to deter pests. Despite this mountain of traditional knowledge and practical evidence, the empirical reality of the link between biodiversity and ecosystem function went without formal evaluation for over a century before slowly climbing back into the crosshairs of experimental biologists.
An Ecological Reawakening
For much of the twentieth century, ecologists explored tantalizingly around the question of how biodiversity might shape ecosystems, often taking diversity to be a consequence of ecological conditions in a particular place rather than a cause of those same conditions (e.g., Connell and Orias, 1964). Eventually, as ecologists became more attuned to the ecological importance of stability—how much conditions in a forest or grassland might remain constant from season to season or year to year—they began to interrogate its relationship with biodiversity. At the center of this debate was the question of whether increasing the number of species in a community made that community more stable through beneficial effects such as symbiosis (Elton, 1958) or destabilized it by increasing the likelihood of, for instance, local numbers of a critical species crashing due to catastrophic disease (May, 1973).
In their review of the field, American ecologist David Tilman and colleagues (2014) trace a “reawakening” in the study of biodiversity to incipient awareness of catastrophic global biodiversity loss during the 1980s, which culminated in a 1991 conference of ecologists in Bayreuth, Germany. The papers emerging from this meeting—which were ultimately collected in an edited volume, Biodiversity and Ecosystem Function, published in 1994—effectively launched the field of contemporary research on biodiversity-ecosystem functioning, otherwise known as BEF.
The observational findings and theories marshaled in the very early nineties, however, lacked the gold standard of ecological evidence: experimentation. This was not long in coming; two progenitor BEF experiments were already in development at the time of the Bayreuth Conference. At Imperial College London’s Centre for Population Biology, pilot testing of the futuristically named Ecotron facility began in 1991. The Ecotron, still operational today, consists of sixteen isolated rooms, each with its own light, temperature, and atmospheric control systems. Beginning in 1993, these rooms were assigned to one of three biodiversity treatments. The lowest diversity rooms contained boxes of soil enriched with two common British plant species (e.g., sow thistle), three plant-eating invertebrates (e.g., aphids), one predator (e.g., an aphid predator), and three decomposer species (e.g., earthworms). A second set of rooms contained extra species of each class, and the most diverse rooms contained sixteen plants, five herbivores, two predators, and eight decomposers. Environmental conditions were held constant, and during a two-hundred-day period, an international team of ecologists monitored a variety of ecosystem functions: how much organisms in each room respired, how quickly organic matter decomposed, to what extent nutrients and water ran off, and how productive plants were in each room. In the end, more-diverse communities of plants and animals consumed more carbon dioxide (respired more) and grew more than less-diverse ones (Naeem et al., 1994). Diversity supercharged the functionality of ecosystems with more species.
In a complement to the highly controlled approach of the Ecotron, Tilman and collaborators at Cedar Creek, in Minnesota, were simultaneously figuring out how to ask BEF questions in the field. Based on an observational study in which more-diverse grassland plots showed greater stability in biomass production than less-diverse plots following an extreme drought (Downing and Tilman, 1994), they established what came to be known as Cedar Creek’s “Little Biodiversity” experiment. In this seminal experiment, 147 plots, each nine meters square, were denuded of existing vegetation and seeded with one, two, four, six, eight, twelve, or twenty-four species of prairie plants. Echoing findings from the Ecotron, diverse plots (and especially any plot with twelve or twenty-four species) produced far more biomass than less-diverse plots. Furthermore, even after only two summers of growth, more-diverse plant communities in the Little Biodiversity experiment showed lower levels of soil nitrogen, suggesting that their roots more completely and efficiently utilized available nutrients (Tilman et al., 1996).
These findings have been borne out repeatedly through the expanded “Big Biodiversity” experiment, planted between 1994 and 1995. These plots are larger, more numerous, and contain as many as thirty-two species (Tilman et al., 1997). The assigned diversity levels of most of its original plots are still maintained through diligent weeding by an army of fresh-faced interns hired by Cedar Creek’s managers every summer. Now in its twenty-fifth year of growth, the Big Biodiversity experiment still serves as a critical platform for BEF research.
A FABulous Journey
The Forests and Biodiversity (FAB) project, which I was recruited to work on in 2012, would mimic past grassland experiments insofar as plots were planted with varying species diversity. Yet, in many other ways, the forest project departed from its progenitors. From a logistical standpoint, rather than weighing out and broadcasting consistent quantities of seed, we planted each tree on a grid (sixty-four trees per plot), where each tree was only half a meter from its nearest neighbors. Some plots were monocultures, consisting entirely of one of twelve species native to Minnesota: red (Pinus resinosa), jack (P. banksiana), or white pine (P. strobus); eastern red cedar (Juniperus virginiana); paper birch (Betula papyrifera); red (Quercus rubra), northern pin (Q. ellipsoidalis), bur (Q. macrocarpa), and white oak (Q. alba); basswood (T. americana); red maple (Acer rubrum); and box elder (A. negundo). Other plots (bicultures) contained thirty-two individuals of each of two species. Yet others were planted with five-species polycultures, and we threw the entire kitchen sink at a set of twelve-species plots. We started with two-year-old bareroot seedlings, planted in May of 2013, and over the next three years, we replanted dead trees and weeded woody invaders so that each plot truly corresponded to its assigned tree-diversity treatment. By the time I finished my doctorate five years later, I could easily conceal myself within their densely interlocking boughs—at least in plots dominated by fast-growing pines and birches.
Beyond logistical considerations, the design of FAB also expanded on past research by making it possible for us to ask which dimensions of biodiversity might be most important to supporting ecological function. For instance, the vaunted boost in productivity associated with higher-diversity plots in the Big Biodiversity grassland study appears not to be entirely due to species richness—the number of species in a plot. Instead, it seems that some of the diversity-related boost really stemmed from functional diversity, the variability in morphological and physiological traits associated with species in a community. In particular, it appears that more-diverse plots provided opportunities for nitrogen-fixing legumes and drought-tolerant grasses to interact synergistically, boosting the productivity of their community by sharing resources. Legumes fertilized nearby grasses, which, because they differ in their growth form and resource needs, did not outcompete their beneficial neighbors (Fargione et al., 2007). In this sense, it can sometimes be difficult to determine whether it is more important to have a lot of species present or for those species present to have a diversity of functions.
While functional diversity can be difficult to measure, phylogenetic diversity—corresponding to the evolutionary distance between members of a community—offers a useful proxy. Closely related species tend to share traits and interact with their environment in similar ways, but such similarities are lost as evolution progresses. Subtly then, FAB was designed so that bicultures—all plots with just two species—varied widely in their functional and phylogenetic diversity. Some two-species pairs, like white oak and bur oak, were both closely related and quite similar in their leaf shape, nutrient consumption, and responses to environmental stresses like drought and shade. Other pairs of relatively closely related species, like red maple and basswood, differed quite a bit in these traits. Yet other pairs, like basswood and eastern red cedar were both distantly related—remember that the split between flowering angiosperms like basswood and non-flowering gymnosperms like pine took place roughly three hundred million years ago—and functionally distinct. And finally, some pairs of distantly related species, like red oak and white pine, had relatively similar functional traits despite considerable evolutionary divergence. (This does happen from time to time: consider the functional similarities of bats and insect-eating birds.) The presence of these four types of bicultures in FAB allowed us to tease apart the role of functional and phylogenetic diversity in bolstering the ecological functionality of our newly planted “forest.”
We also wanted to understand whether diversity within a single species—genetic diversity—was as important as diversity among species. Though intraspecific diversity in other species is often invisible to humans, it has been well-documented that some plant traits vary just as much within a species as among related species. And copious evidence from epidemiology to conservation biology has shown that genetically diverse populations are more stable and better poised to cope with environmental stressors than homogenous ones. To assess this question, we designed and planted a second tree-diversity experiment. The eight-hundred-tree Biodiversity in Willows and Poplars (BiWaP) experiment included plots varying not just in species richness but also in genetic diversity. We took advantage of the fact that many species in the willow family (Salicaceae) can be easily propagated by cuttings to grow hundreds of identical clones of several quaking aspens (Populus tremuloides), white aspens (P. alba), and black willows (Salix nigra). We then planted these trees in the field such that some had as neighbors only genetic clones of themselves while others had as neighbors multiple genotypes each of several species. As such, the genetic diversity comprised another dimension of biodiversity whose role in supporting ecosystem function we planned to test.
The Complex Role of Biodiversity
But what goes into measuring the functionality of an ecosystem—even a highly simplified and orderly biodiversity experiment? At the end of each summer at Cedar Creek, a team of interns—led by me for the first several years of the experiment—measured the stem diameter and height of each tree in FAB. Standardized equations then allowed for easy conversions of these measurements into estimates of trunk biomass. Encouragingly, trunk growth from year to year was higher for trees with more-diverse neighbors compared to those in monoculture (Grossman et al., 2017), although we did not see an effect of either species or genetic diversity in the BiWaP experiment (Grossman and Cavender-Bares, 2019). Paralleling findings from Big Biodiversity, Ecotron, and other grassland experiments around the world, our documentation of a productivity boost in more-diverse plots contributed to the growing consensus that this BEF phenomenon is not only confined to grasslands. Indeed, meta-analysis of tree growth data both from global forests (Liang et al., 2016) and managed or experimental systems (Zhang et al., 2012) corroborates our findings. This pattern is perhaps of special note given that monocultural plantations dominate production forestry the world over. Polycultures are harder to maintain and harvest; yet recent experimental findings like ours raise the question of whether increases in yield might compensate for higher costs of maintenance and harvesting.
Going beyond my initial focus on productivity, I wanted to determine how tree biodiversity in these systems related to herbivore vulnerability and disease susceptibility. Since we planted FAB inside a massive fenced enclosure, I knew I would never be able to study, for instance, the role of diversity in preventing deer browsing. But I could measure damage by insects and fungal pathogens, like red maple anthracnose. Over three years, I spent a month each autumn painstakingly measuring leaves of hundreds of plants with a translucent grid: I would estimate the original size of a given leaf and the amount of this tissue that had been chewed up by insects or infected by fungi. I also counted galls (small tumors formed by insect larvae) and leaf mines (burrows in leaves created by other larval feeders). Finally, I surveyed damage across the experiment stemming from two fungal diseases, each specialized to a single species in the FAB and BiWaP experiments.
The story that emerged from these measurements is a complicated one (Grossman and Cavender-Bares, 2019; Grossman et al., 2018). Having diverse neighbors frequently affected how vulnerable a given tree was to insect or disease damage, but the direction and strength of this relationship varied based on the species of tree and type of damage in question. For instance, having diverse neighbors reduced the likelihood that an oak would be attacked by leaf miners but increased the risk of leaf miner attack for birches! And yes, red maples with more conspecific neighbors were more likely to experience intense anthracnose infection. Fascinatingly, it also seemed that very nearby neighbors (within a one-meter radius of a focal tree) had a bigger impact on that tree’s risk of pest attack or disease than did farther away neighbors. This spatial scale-dependence of vulnerability to damage was relatively consistent across tree and pest or disease identity. Generally, though, it appears that other factors, like climate, the presence of predators, and chance, might play a role equivalent to or greater than that of diversity in affecting the vulnerability of trees to pests and pathogens.
While pests and diseases constitute the most famous consumers of living plant tissue, an entire food chain unfolds once leaves and roots are shed and begin to decompose, and I also wanted to know how tree diversity affected this microbial universe. Focusing on the rich, plant-dependent microbial life of rotting leaves and the soil below them, I was interested in using the FAB experiment as a platform to assess whether more-diverse tree communities might beget more active and diverse soil microbial communities. In both cases, we found subtle biodiversity effects. We found that the most important factor shaping the microbial community was the proportion of trees in a plot that were gymnosperms (pines and junipers) versus angiosperms (oaks, maples, birch, and basswood). Interestingly, pines, and especially junipers, created a hostile environment for bacteria, perhaps due to antimicrobial properties exuded by these species. Yet, since I collected samples after only three years of tree growth, it is important to note that the microbial communities of the FAB experiment have probably not finished responding to the presence of different combinations of tree species. So, this story is only just beginning to unfold.
Across all these projects, I was surprised to find that species richness—long the standard metric of biodiversity for biologists—emerged as a still-critical predictor of ecosystem function. In study after study, the number of tree species in a plot predicted ecosystem function as well as or better than more abstruse dimensions of biodiversity. In some cases, the diversity of particular functional traits within plots emerged as an important predictor of particular functions. But, generally speaking, I saw little evidence that continuing to measure diversity in terms of species richness might obscure important connections between biodiversity and ecosystem function.
From Local to Global to Local
Encouragingly, my findings—or anyone else’s—from the tree-diversity experiments at Cedar Creek don’t have to be the final word on BEF relationships in forests. On establishment, FAB was inducted into TreeDivNet, a network of twenty-five tree-diversity experiments distributed across the globe. The 1.1 million trees making up TreeDivNet have been planted in sites on six continents and range from boreal to Mediterranean and tropical climates. Though the design of these experiments varies from site to site, each includes some experimental manipulation of tree diversity, as in FAB. At most sites, investigators have made periodic measurements of tree survival and growth, and of damage inflicted upon trees by pests and pathogens (Grossman et al., 2018b). This riot of findings has already contributed to our understanding of how changes in tree biodiversity are likely to affect the way that forests function. And the BEF framework, though developed through experimental work, has now given credence to the idea that biodiversity changes the way ecosystems function. This premise has now been borne out through observational studies of non-experimental (e.g. naturally occurring) ecosystems (van der Plas, 2019).
I argue that this holistic view on the value of biodiversity needs to inform the way that we, as managers and users of natural resources, make local decisions. Though large-scale, systemic change will be required for humans to fully address the current biodiversity crisis, such change can be instigated and incubated on the smallest scales. For urbanites, this might mean turning more and more of our marginal spaces into biodiversity havens. Opportunities of this nature include pollinator-friendly prairie yards, urban gardens and food forests, and even no-mow zones such as those currently being put into place at the Arnold Arboretum. Communities can also make choices in our roles as consumers, advocating for less chemically intensive agriculture that protects the incidental biodiversity concomitant with farming prior to the widespread adoption of blanket glyphosate-spraying on row crops.
For me, working mere meters away from the Big Biodiversity plots and playing my own part in the establishment of new biodiversity experiments has also highlighted the importance of humility. Empirical evidence shows us that biodiversity plays critical, complex roles in mediating the way ecosystems function. Yet we are often not nor, I would argue, will we ever be able to fully understand and thus manage these BEF dynamics. Instead of assuming that we can figure out how to optimize global biodiversity to provide for the ecosystem functions that we want, it might make more sense to take a precautionary approach. In doing so, we should be highly conservative in both senses of the word, protecting biodiversity far more stringently than we think is necessary to sustain critical ecological functioning, especially in the face of ongoing challenges such as climate change. We would be foolish, I believe, to fail to conserve global biodiversity, which BEF research has shown us to be valuable beyond measure.
The author wishes to acknowledge his doctoral advisor, Jeannine Cavender-Bares, as well as the other co-PIs of the Forests and Biodiversity (FAB) experiment: Sarah Hobbie, Rebecca Montgomery, and Peter Reich. Work described here also stems from collaborations with Peter Kennedy, Jess Gutknecht, and several undergraduate collaborators and interns. Susan Barrott offered helpful feedback on an early draft of this article. Finally, the author is incredibly grateful for the opportunity to work at the Cedar Creek LTER site and to collaborate with TreeDivNet partners.
Butchart, S. H. M., M. Walpole, B. Collen, A. van Strien, J. P. W. Scharlemann,… R. Watson. 2010. Global biodiversity: Indicators of recent declines. Science, 328: 1164–1168.
Connell, J. H., and E. Orias. 1964. The ecological regulation of species diversity. The American Naturalist, 98: 399–414.
Darwin, C. 1859. On the origin of species by means of natural selection, or the preservation of favoured races in the struggle for life. London: John Murray.
Davis, M., S. Faurby, and J. Svenning. 2018. Mammal diversity will take millions of years to recover from the current biodiversity crisis. Proceedings of the National Academy of Sciences, 115: 11262–11267.
Downing, J. A., and D. Tilman. 1994. Biodiversity and stability in grasslands. Nature, 367: 363–365.
Elton, C. S. 1958. The ecology of invasions by animals and plants. London: Methuen and Co.
Fargione, J., D. Tilman, R. Dybzinski, J. H. R. Lambers, C. Clark,… M. Loreau. 2007. From selection to complementarity: shifts in the causes of biodiversity-productivity relationships in a long-term biodiversity experiment. Proceedings of the Royal Society, B 274: 871–6.
Grossman, J. J., and J. Cavender-Bares. 2019. Consequences of biodiversity shift across phylogenetic scales for aspen and willow growth, survival, and herbivory. Journal of Vegetation Science, 30: 301–311
Grossman, J. J., J. Cavender-Bares, S. E. Hobbie, P. B. Reich, and R. A. Montgomery. 2017. Species richness and traits predict overyielding in stem growth in an early-successional tree diversity experiment. Ecology, 98: 2601–2614.
Grossman, J. J., J. Cavender-Bares, P. B. Reich, R. A. Montgomery, and S. E. Hobbie. 2018a. Neighborhood diversity simultaneously increased and decreased susceptibility to contrasting herbivores in an early stage forest diversity experiment. Journal of Ecology, 107: 1492–1505.
Grossman, J. J., M. Vanhellemont, N. Barsoum, J. Bauhus, H. Bruelheide,… K. Verheyen. 2018b. Synthesis and future research directions linking tree diversity to growth, survival, and damage in a global network of tree diversity experiments. Environmental and Experimental Botany,
Isbell, F., D. Tilman, S. Polasky, S. Binder, and P. Hawthorne. 2013. Low biodiversity state persists two decades after cessation of nutrient enrichment. Ecology Letters: 454–460.
Liang, J., T. W. Crowther, N. Picard, S. Wiser, M. Zhou,… P. B. Reich. 2016. Positive biodiversity-productivity relationship predominant in global forests. Science, 354: 196.
May, R. M. 1973. Qualitative stability in model ecosystems. Ecology, 54: 638–641.
Naeem, S., L. J. Thompson, S. P. Lawler, J. H. Lawton, and R. M. Woodfin. 1994. Declining biodiversity can alter the performance of ecosystems. Nature, 368: 734–737.
Naidoo, R., and W. L. Adamowicz. 2006. Modeling opportunity costs of conservation in transitional landscapes. Conservation Biology, 20: 490–500.
van der Plas, F. 2019. Biodiversity and ecosystem functioning in naturally assembled communities. Biological Reviews, brv.12499.
Schulze, E.-D., and H. Mooney. 1994. Biodiversity and ecosystem function. Berlin: Springer-Verlag.
Tilman, D., F. Isbell, and J. M. Cowles. 2014. Biodiversity and ecosystem functioning. Annual Review of Ecology and Systematics, 45: 471–493.
Tilman, D., J. Knops, D. Wedin, P. Reich, M. Ritchie, and E. Siemann. 1997. The influence of functional diversity and composition on ecosystem processes. Science, 277: 1300–1302.
Tilman, D., D. Wedin, and J. M. H. Knops. 1996. Productivity and sustainability influenced by biodiversity in grassland ecosystems. Nature, 379: 718–720.
Zhang, C., J. A. Postma, L. M. York, and J. P. Lynch. 2014. Root foraging elicits niche complementarity-dependent yield advantage in the ancient “three sisters” (maize/bean/squash) polyculture. Annals of Botany, 114: 1719–1733.
Zhang, F., and L. Li. 2003. Using competitive and facilitative interactions in intercropping systems enhances crop productivity and nutrient-use efficiency. Plant and Soil, 248: 305–312.
Zhang, Y., H. Y. H. Chen, and P. B. Reich. 2012. Forest productivity increases with evenness, species richness and trait variation: A global meta-analysis. Journal of Ecology, 100: 742–749.
Jake J. Grossman is a Putnam Fellow at the Arnold Arboretum.