Niche Construction Theory: a Practical Guide for Ecologists Author(S): John Odling-Smee, Douglas H

Home , Niche construction

Niche Construction Theory: A Practical Guide for Ecologists Author(s): John Odling-Smee, Douglas H. Erwin, Eric P. Palkovacs, Marcus W. Feldman, Kevin N. Laland Source: The Quarterly Review of Biology, Vol. 88, No. 1 (March 2013), pp. 3-28 Published by: The University of Chicago Press Stable URL: http://www.jstor.org/stable/10.1086/669266 . Accessed: 01/04/2013 08:08

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to The Quarterly Review of Biology.

http://www.jstor.org

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions Volume 88, No. 1 March 2013 THE QUARTERLY REVIEW of Biology

NICHE CONSTRUCTION THEORY: A PRACTICAL GUIDE FOR ECOLOGISTS

John Odling-Smee Mansﬁeld College, University of Oxford Oxford OX1 3TF United Kingdom e-mail: [email protected]

Douglas H. Erwin Department of Paleobiology, Smithsonian Institution Washington, DC 20560 and Santa Fe Institute Santa Fe, New Mexico 87501 USA e-mail: [email protected]

Eric P. Palkovacs Department of Ecology and Evolutionary Biology, University of California Santa Cruz, California 95064 USA e-mail: [email protected]

Marcus W. Feldman Department of Biology, Stanford University Stanford, California 94305-5020 USA e-mail: [email protected]

Kevin N. Laland* School of Biology, University of St. Andrews St. Andrews, Fife KY16 9TS United Kingdom e-mail: [email protected] *Corresponding Author

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 4 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

keywords niche construction, ecological inheritance, ecosystem engineering, eco- evolutionary feedbacks, ecosystem ecology

abstract Niche construction theory (NCT) explicitly recognizes environmental modification by organisms (“niche construction”) and their legacy over time (“ecological inheritance”) to be evolutionary processes in their own right. Here we illustrate how niche construction theory provides useful conceptual tools and theoretical insights for integrating ecosystem ecology and evolutionary theory. We begin by briefly describing NCT, and illustrating how it differs from conventional evolutionary approaches. We then distinguish between two aspects of niche construction—environment alteration and subsequent evolution in response to constructed environments—equating the first of these with “ecosystem engineering.” We describe some of the ecological and evolutionary impacts on ecosystems of niche construction, ecosystem engineering, and ecological inheritance, and illustrate how these processes trigger ecological and evolutionary feedbacks and leave detectable ecological signatures that are open to investigation. Finally, we provide a practical guide to how NCT could be deployed by ecologists and evolutionary biologists to explore eco-evolutionary dynamics. We suggest that, by highlighting the ecological and evolutionary ramifications of changes that organisms bring about in ecosystems, NCT helps link ecosystem ecology to evolutionary biology, potentially leading to a deeper understanding of how ecosystems change over time.

Introduction nities, and focus on the interactions of pheno- COLOGY has long been portrayed as a types in different species (for instance, in food Ediscipline separated by the different ways webs). This means that many forms of environ- in which its two principal subfields, population- mental modification by organisms are left out community and ecosystem ecology, relate to of coevolutionary analyses, and partly accounts evolution (Ehrlich 1986; O’Neill et al. 1986; for the prevailing division between population- Jones and Lawton 1995; Likens 1995; Loreau community ecology and ecosystem ecology. As 2010; Schoener 2011). Why? a result: “The disciplinary links between ecosystem science and evolutionary biology are A primary concern of many ecologists is among the weakest in the biological sciences” to understand how energy and matter flow (Matthews et al. 2011:690). through organisms and their environments. In The need to overcome this weakness is now contrast, evolutionary biologists are principally becoming urgent, particularly with increasing concerned with information: that is, with the recognition that evolution and ecology can acquisition and inheritance across generations happen at the same pace (Kingsolver et al. of algorithmic information by organisms 2001; Hairston et al. 2005; Ellner et al. 2011) (Chaitin 1987), primarily in the form of the and must shape each other (Palkovacs and heritable DNA sequences that underpin their Hendry 2010). Schoener (2011) describes adaptations. Standard evolutionary theory “The Newest Synthesis” as “the emerging field (henceforth SET) permits evolutionary ecolo- of eco-evolutionary dynamics, whose major gists to integrate population-community ecol- precept is that both directions of effect— ogy and evolutionary biology (Whitham et al. ecology to evolution and evolution to ecolo- 2006; Rowntree et al. 2011), but at the price of gy—are substantial” (Schoener 2011:426). A restricting population-community ecology primary goal of this new field is to elucidate the largely to biota (O’Neill et al. 1986). SET rec- consequences of bidirectional eco-evo interac- ognizes abiota as sources of natural selection, tions and “eco-evolutionary feedbacks” in eco- but rarely considers the converse relationship, systems (Post and Palkovacs 2009). where organisms modify abiotic components The aim of this article is to illustrate how in environments, to be ecologically consequen- niche construction theory provides useful tial or evolutionarily generative. As ecosystems theoretical insights and practical tools that necessarily include abiota, evolutionary ecolo- contribute to the integration of ecosystem ec- gists frequently “edit out” abiota from commu- ology and evolutionary theory. The niche-

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 5 construction perspective explicitly recognizes Wright et al. 2002). For simplicity, we refer to environmental modification by organisms all such activities as ecosystem engineering, (“niche construction”), and its legacy over time although some of these activities are ex- (“ecological inheritance”), to be evolutionary cluded from more strict definitions of this processes: that is, they cause evolutionary change term. Ecological inheritance refers to lega- by acting as sources of modified selection, as cies of change, in both biota and abiota, well as of modified phenotypes (Lewontin bequeathed by niche-constructing organisms 1983; Odling-Smee et al. 2003). This stance to subsequent populations, which modify se- can be contrasted with the more tacit recogni- lection pressures on descendant organisms tion of organisms’ environmental impacts in (Odling-Smee et al. 2003); this can be re- standard accounts. The extension has pro- garded as a second general inheritance sys- duced a body of conceptual and formal theory, tem in evolution. known as “niche construction theory” (hence- NCT has generated a body of conceptual forth NCT), which explores the ecological and and formal theory that explores the ramifica- evolutionary ramifications of niche construc- tions of niche construction for evolutionary bi- tion. NCT has begun to be used as a vehicle ology (Odling-Smee 1988; Laland et al. 1996, for integrating ecosystem ecology and evolu- 1999; Odling-Smee et al. 2003; Laland and tion (Erwin 2008; Kylafis and Loreau 2008; Sterelny 2006; Silver and DiPaolo 2006; Leh- Krakauer et al. 2009; Post and Palkovacs mann 2008; Van Dyken and Wade 2012), and 2009; Loreau 2010; Van Dyken and Wade for related disciplines (Boni and Feldman 2012). We begin by summarizing these 2005; Laland et al. 2010; Kendal et al. 2011), findings. including ecology (Erwin 2008; Kylafis and NCT is derived from insights that were Loreau 2008; Krakauer et al. 2009; Post and first introduced to evolutionary biology in Palkovacs 2009; Loreau 2010), and the Earth the 1980s by Richard Lewontin (1982, 1983, sciences (Corenblit et al. 2009, 2011). Insights 2000). Niche construction refers to the modi- from mathematical evolutionary theory, sum- fication of both biotic and abiotic components marized in Table 1, provide unambiguous in environments via trophic interactions and evidence that niche construction is of consid- the informed (i.e., based on genetic or ac- erable ecological and evolutionary importance. quired information) physical “work” of organ- The significance of niche construction isms. It includes the metabolic, physiological, for evolution is threefold. First, niche con- and behavioral activities of organisms, as well as struction can influence spatial and tempo- their choices. For example, many species of ral patterns in the strength and direction animals manufacture nests, burrows, holes, of selection acting on the constructors them- webs, and pupal cases; algae and plants change selves (Odling-Smee et al. 1996, 2003). By levels of atmospheric redox states and influ- modifying their own environments, and by ence energy and matter flows by modifying legating modified selection pressures to nutrient cycles; fungi and bacteria decompose their descendents via ecological inheri- organic matter; and bacteria also fix nutrients tance, populations can influence the direc- and excrete compounds that alter environ- tion and the rate of their evolution; for ments. Niche-constructing species include instance, generating time-lagged responses several categories of species recognized in the to selection (Laland et al. 1999). Second, ecological literature, including ecosystem engi- niche construction can increase the abun- neers (i.e., species that modify their environ- dance of individuals within species by in- ments via nontrophic interactions), keystone creasing fecundity and/or extending the species (i.e., rare species with large effects on longevity of individuals. By influencing pop- communities and ecosystems disproportionate ulation structure, it may thus decrease the to their abundance, often via predation), dom- significance of drift and potentially increase inant species (common species with large ef- the longevity of species, independent of any fects on communities and ecosystems, often direct selection. Third, through ecological via competition), and foundation and facul- spillovers that occur in the process of modi- tative species (i.e., habitat-creating species; fying their own niches, organisms can also

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 6 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

TABLE 1 Twelve insights from niche construction theory Finding References Niche construction can: 1. Fix genes or phenotypes that would, under standard evolutionary theory, Laland et al. 1996, 1999, 2001; Kerr et al. be deleterious; support stable polymorphisms where none are expected 1999; Creanza et al. 2012 and eliminate polymorphisms that without niche construction would be stable. 2. Affect evolutionary rates, both speeding up and slowing down responses Laland et al. 1996, 1999, 2001; Silver and to selection under different conditions. Di Paolo 2006 3. Cause evolutionary time lags, generate momentum, inertia, and Laland et al. 1996, 1999, 2001; Kerr et al. autocatalytic effects. Interactions with evolving environments can produce 1999 catastrophic responses to selection, as well as cyclical dynamics. 4. Drive niche-constructing traits to fixation by creating statistical Silver and Di Paolo 2006; Rendell et al. associations with recipient traits. 2011 5. Influence the dynamics, competition, and diversity of meta-populations. Hui et al. 2004; Borenstein et al. 2006 6. Be favored, even when currently costly, because of the benefits that will Lehmann 2007, 2008 accrue to distant descendants. 7. Allow the persistence of organisms in currently inhospitable Kylafis and Loreau 2008 environmental conditions that would otherwise lead to their extinction; facilitate range expansion. 8. Regulate environmental states, keeping essential parameters within Laland et al. 1996, 1999; Kylafis and tolerable ranges. Loreau 2008 9. Facilitate the evolution of cooperative behavior. Lehmann 2007, 2008; Van Dyken and Wade 2012 10. Drive coevolutionary events, both exacerbate and ameliorate Krakauer et al. 2009; Kylafis and Loreau competition, and affect the likelihood of coexistence. 2011 11. Affect carrying capacities, species diversity and robustness, and Krakauer et al. 2009 macroevolutionary trends. 12. Affect long-term fitness (not just the number of offspring or grand- McNamara and Houston 2006; Lehmann offspring) by contributing to the long-term legacy of alleles, genotypes, 2007; Palmer and Feldman 2012 or phenotypes within a population.

change the niches of other species in an byproducts, acquired characters, and collective ecosystem. Where these spillovers are ef- activity in ecosystems, the broader conceptual- fectively coupled to other species they can ization offered by NCT compared to SET po- lead to coevolution. Thus, niche construc- tentially allows it to make useful contributions tion has the potential to percolate through to ecosystem ecology. The principal differ- ecosystems and precipitate multiple evolu- ences between SET and NCT are summarized tionary and coevolutionary events. In NCT, in Table 2. it is possible for one:many, many:one, and In the following sections, we ﬁrst distinguish many:many relationships to occur between between the environment-altering and subse- niche-constructing populations and other quent evolutionary aspects of niche construc- populations that coevolve as a result of the tion; we then consider how these effects can niche construction. Beavers provide a one: be detected by ecologists and paleoecologists many example because beaver dams alter and used to comprehend eco-evolutionary the environments of many populations (Naiman et al. 1988), while the soil envi- dynamics. ronment coconstructed by earthworms, arthropods, plants, and bacteria, among Niche Construction Theory and others, is a many:many case. By drawing Ecosystem-Level Ecology attention to the many important conse- Post and Palkovacs (2009) distinguish be- quences that ﬂow from time-lagged effects, tween two aspects of niche construction:

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 7

TABLE 2 Comparing standard evolutionary theory and niche construction theory Standard Evolutionary Theory Niche Construction Theory Focus: Organismic evolution in response to environments. Focus: The coevolution of organisms and environments. Causation: Primarily unidirectional, with autonomous selective Causation: Primarily reciprocal, with selective environments shaping organisms. Reciprocal causation is environments shaping organisms, and organisms recognized in some “special cases” where the source of shaping selective environments, either relative to selection is biotic (e.g., sexual selection, predator-prey themselves or other organisms. coevolution). Niche construction: Organisms acknowledged to change Niche construction: Treated as an evolutionary process in environmental states, but this is treated as the product of its own right. Focus is not exclusively on adaptations, natural selection and rarely as an evolutionary process in but includes changes in environments caused by the its own right. Focus is restricted to adaptations expressed byproducts of organisms (e.g., detritus), acquired outside the bodies of the organisms (e.g., extended characters (e.g., learned), or the collective phenotypes). metabolism or behaviors of multiple individuals/ species. Inheritance: Primarily genetic, although maternal, epigenetic, Inheritance: Genetic and ecological inheritance (i.e., cytoplasmic, and cultural inheritances recognized as legacies of selection pressures previously modified by “special cases.” niche construction). Genetic and ecological inheritance interact to form “niche inheritance.” Maternal, epigenetic, cytoplasmic, and cultural inheritances can be examples. Organism-environment complementarity (adaptation): The Organism-environment complementarity (adaptation): The product of natural selection. match between organism and environment results from dynamic interactions between niche construction and natural selection. environment alteration and the subsequent Glossary). Jones et al. (1997) describe allo- evolution of population(s) in an ecosystem in genic engineers that modify environments by response to natural selection pressures previ- mechanically changing materials from one ously altered by the niche constructor(s). form to another (e.g., nest-building wasps) Although environmental alteration may com- and autogenic engineers that modify environ- monly lead to subsequent evolution, some acts ments by changing themselves (e.g., trees of niche construction are dissipated, swamped, that provide living space for insects, birds, and or counteracted by other processes, such that mammals, or windbreaks). Researchers have their ecological and/or evolutionary ramifica- proposed various classes of ecosystem engi- tions are trivial and, in principle, ecologically neering, including structural engineers, bio- and evolutionarily negligible cases of niche turbators and bioconsolidators, chemical construction need not coincide (Post and engineers, light engineers, and wind attenua- Palkovacs 2009). This means that the eco- tors (Berke 2010; Jones et al. 2010). Although logical consequences of niche construction to some ecologists the term ecosystem engi- may sometimes usefully be studied without neering is restricted to nontrophic effects, it is considering evolutionary ramifications and important to recognize that important envi- vice versa. ronmental alterations can also occur through trophic interactions, which is why we consider them here. the environment-altering aspect of Jones et al. (1994, 1997) and Cuddington et niche construction al. (2007) drew attention to the paucity of eco- The environment-altering aspect of niche logical research dedicated to studying organ- construction is similar to ecosystem engineering isms that modulate the availability of resources in ecology (Jones et al. 1994, 1997; Wright and habitats in ecosystems. Many species influ- and Jones 2006; Cuddington et al. 2007; see ence energy flows, mass flows, and trophic

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 8 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

Glossary

Algorithmic information: Structural and functional “know-how” carried by organisms typically, but not exclusively, in their genomes (Chaitin 1987). Byproducts: Phenotypic effects that evolve as a consequence of selection on some other character, rather than being an adaptation directly favored by selection for its current role. Eco-evolutionary dynamics: Interactions between ecology and evolution occurring on overlapping time scales (Pelletier et al. 2009). The recognition that evolution can happen at the pace of ecology means that ecological changes can directly shape evolution and vice versa. Ecological inheritance: The inheritance, via an external environment, of one or more natural selection pressures previously modified by niche-constructing organisms (Odling-Smee et al. 2003). Ecological inheritance typically depends on organisms bequeathing altered selective environments to their descendants, but other organisms, including unrelated conspecifics and members of other species that share the same ecosystem may also be affected by this legacy. Where an act of niche construction leads to a change in the species composition of the local ecological community, this too is regarded as an aspect of the ecological inheritance. Ecosystem engineering: The creation, destruction, or modification of habitats and/or modulation of the availability of resources to other species by organisms (Jones et al. 1994). Ecosystem engineering can be equated with the environment-altering component of niche construction, although some definitions of ecosystem engineering are more restrictive. EMGAs (environmentally mediated genotypic associations): Indirect but specific connections between distinct genotypes mediated either by biotic or abiotic environmental components in the external environment (Odling-Smee et al. 2003). EMGAs may either associate different genes in a single population or they may associate different genes in different populations. Engineering webs: Webs of connectance in ecosystems, caused by species influencing energy and mass flows, and creating habitat and other resources for other species (Jones et al. 1994, 1997). Engineering webs contribute to the stability and dynamics of ecosystems, alongside webs of trophic interactions. Ecosystem engineers may influence and control energy and matter flows without themselves being part of those flows. Extended phenotypes: Phenotypic characters expressed outside the body of the organism (Dawkins 1982). Although byproducts are sometimes characterized as extended phenotypes, Dawkins is explicit in stating that this is “not profitable,” and restricts use of the term to adaptations. Thus, extended phenotypes correspond to that subset of niche-constructing activities that are biological adaptations. Niche construction: The process whereby organisms, through their metabolism, their activities, and their choices, modify their own and/or each other’s niches (Odling- Smee et al. 2003). Niche construction may result in changes in one or more natural selection pressures in the external environment of populations. Niche-constructing organisms may either alter the natural selection pressures of their own population, of other populations, or of both. Niche construction may occur through physical perturbation of the environment or through relocation to a new environment. Niche construction may have both positive and negative effects on the constructor’s fitness.

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 9

Niche inheritance: Combinations of genetic inheritance and ecological inheritance that descendant organisms inherit from ancestral organisms. Time lags: Delayed evolutionary response of characters exposed to selection pressures modified by niche construction. These time lags are generated by ecological inheritance that affects availability of a resource for multiple descendant generations, and can be orders of magnitude larger than the number of generations of niche constructing affecting that resource. Time lags may generate unusual evolutionary dynamics, including momentum, inertia, and autocatalytic effects, as well as opposite and catastrophic responses to selection. patterns in ecosystems nontrophically by gen- energy and matter resources, henceforth erating engineering webs based on mosaics labeled Rp, and semantic informational re- of connectivity among different species. Al- sources—e.g., “know-how” or algorithmic though the ecological consequences of information, such as is encoded by genes, niche construction should be the same as or is acquired through learning (Chaitin those of ecosystem engineering, NCT fo- 1987), labeled Ri (Figure 1). Abiota are cuses primarily on those environment- usually physical energy and matter re- altering activities of organisms that modify sources only (exclusively Rp), while biota natural selection in ways that affect the evo- always carry both physical Rp and algorith- lution of population(s) (e.g., Loreau 2010). mic informational Ri resources. All organ- As a discipline, evolutionary biology has fo- isms carry Ri (i.e., genetic information) in cused overwhelmingly on the passage of ge- their genomes in addition to the Rp in netic information via DNA, although there is their bodies, while animals also carry Ri increasing recognition that a broader notion (i.e., acquired knowledge) in their brains. of heredity may be required (Bonduriansky Artifacts primarily carry Rp, but it is possible and Day 2009; Danchin et al. 2011; Bonduri- for them to carry Ri too as, for instance, in ansky 2012). In contrast, ecosystem ecologists human artifacts such as computers or books. are concerned with the flux and flow of energy These distinctions are significant since the dif- and materials. If researchers are to bridge ferences between different kinds of R poten- these disciplines, they require a framework ca- tially lead to different ecological responses to pable of reconciling these differences. With niche construction. this in mind, we define an ecological variable, R, representing any resource or condition that signatures of niche construction/ is both a potential source of selection for a ecosystem engineering recipient population, and is modifiable by at least one niche-constructing population Let us assume that R is modified by (Laland et al. 1996, 1999). An increase in R can niche construction, such that R becomes ϩ⌬ ⌬ have either positive (e.g., R ϭ prey) or negative R R, where R could be positive or ⌬ (e.g., R ϭ predator) consequences for the fit- negative. Then R is a potential “ecologi- ness of organisms in the recipient population. cal signature” of niche construction (eco- Here we differentiate between three broad system engineering), and offers a useful categories of resource/condition that have way to measure the ecological change different properties, and respond to niche con- caused by organisms. Detection and quan- struction in different ways (Figure 1). R could tification of the impact of niche construc- represent an abiotic component of the envi- tion on an ecosystem therefore requires ronment (e.g., water, sediment), or a biotic that ⌬R must be differentiated from all component (e.g., another organism), or an ar- other potential sources of change in R, in- tifact or other construct built by an organism cluding those resulting from abiotic (physi- (e.g., a spider’s web). cal or chemical) processes or from other We further subdivide R into physical biota not of focal interest. This is the usual

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 10 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

Figure 1. Types of Resource Subject to Niche Construction

$ If the Ri in other organisms can be modiﬁed by the separate Ri in niche-constructing organisms (e.g., genetic manipulation by parasites), then the Ri in other organisms becomes an ecological resource for the niche constructors. § Some organisms superimpose “signals,” such as scent marks, on abiota. If they do that, abiota may carry Ri as well as Rp. However, in that case, abiota are probably better described as Ri-carrying “quasi-artifacts“ rather than as abiota. strategy of researchers who study NCT, eco- that they carry in their genomes or brains, system engineering, and eco-evolutionary the ⌬R ecological signatures left by niche con- feedbacks (Jones et al. 1994, 1997; Odling- struction should reﬂect this information to the Smee et al. 2003; Cuddington et al. 2007; extent that an ecological signature of niche Post and Palkovacs 2009), with a variety of construction might stand out relative to observational, correlational, and experimen- changes in R resulting from other external pro- tal methods (for details see Jones and Lawton cesses. Below we spell out how abiota, biota, 1995; Odling-Smee et al. 2003; Hairston et al. and artifacts potentially offer different signa- 2005; Ellner et al. 2011). These include exper- tures of prior niche construction. imental manipulation of the presence, abundance, or potency of niche construction, Abiota common gardening experiments, selection experiments and experimental evolution, and of- As abiota carry only physical resources ten involve breaking a complex system down (Rp), niche-constructing organisms are into component parts, which involve single- only able to change abiota physically or species or species-pair effects. chemically. However, there are often clear As the niche-constructing activities of signatures of prior niche construction ⌬ many animals are shaped by the knowledge ( Rp) left in abiota; for example, in sedi-

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 11 ments, soils, oceans, and rocks (e.g., Erwin states of other organisms in ways long- 2008). Unlike other causes of change in abiota, studied by ecologists, for example, by provi- changes due to niche-construction must be sioning or protecting other organisms (e.g., due to the information (Ri) in those organ- offspring), supplying them with nutrients isms. Abiotic resources that have been affected (e.g., mutualism), eating them (e.g., prey), by multiple generations of niche construction or competing with them (e.g., for light). At may have been driven far from thermody- the population level, a niche-constructing namic equilibrium by the prior nonrandom population should therefore leave an ecolog- work of organisms, and may therefore occupy ical signature of its prior activities in biota ⌬ states that could not occur, or would be highly ( Rp). It might modify the demographics of unlikely to occur, on a “dead planet.” Exam- another population (e.g., the impact of gup- ples include soil states changed by niche- pies on algal growth), or its population struc- constructing invertebrates, ocean states ture (e.g., the impact of guppy predators on changed by Ediacaran organisms, and sedi- guppy size distributions), or its distribution ment or rock states changed by burrowing or- across an environment (e.g., the impact of ale- ganisms (Odling-Smee et al. 2003; Erwin 2008; wife populations on zooplankton communities Erwin and Tweedt 2012). Even when it is due in different lakes), or one of its engineering to byproducts, for instance, by their metabo- functions in an ecosystem (Post and Palkovacs lism or their detritus, any alteration to an abi- 2009). otic ecosystem component caused by niche An ecological signature in a recipient popu- construction will ultimately still be shaped by lation could also show signs of feedback to the genetic information. It may therefore be non- original niche-constructing population in the random and should frequently stand out rela- form of a modified biotic resource, generating ⌬ ⌬ tive to any change caused by nonliving agents another ( Rp and probably Ri) ecological sig- by virtue of its improbability (for instance, ele- nature in the original niche constructors. This vated rates of nutrient flux or improbable could result in demographic signatures among chemical composition). Such signatures may populations in ecosystems, potentially medi- well be produced by multiple individuals, per- ated by changes in intermediate abiotic ecosys- haps emanating from multiple species, and tem components, even in the absence of any possibly detectable over periods of time sub- genetic coevolution. Such interactions contrib- stantially longer than the lifetimes of the con- ute to many of the phenomena that are structors. The temporal and spatial patterns routinely studied by ecologists (for exam- associated with the signature often covary with ple, trophic relations, community assem- the density and duration of the constructor blies, energy/matter flows in ecosystems, and population(s) and the magnitude of their in- “engineering webs”). NCT predicts that these dividual effects (Jones et al. 1994, 1997), and phenomena should depend on the niche- may be characterized by long-term trends in constructing activities of organisms far more accumulation, depletion, or transformation of often than has previously been recognized (ex- resources over time. Such patterns are also cept by those ecologists who study ecosystem likely to represent an extremely broad range of engineering); it is therefore important that scales. niche-constructing effects be quantified. Niche-constructing organisms may also Biota change the information (Ri) carried by other The biotic components of ecosystems dif- organisms. For instance, horizontal gene trans- fer from abiota in potentially being able to fer can occur in bacteria while a parasite, such respond to prior niche construction by recip- as a virus or an insect, may affect the genome of rocating with further niche construction of its host by inserting a hostile “message” into its their own. Niche-constructing organisms can host’s genome. In animals, including inverte- change biota in two ways: either by modify- brates, useful acquired knowledge is often ing the physical (Rp) or informational (Ri) transmitted through social learning from con- states of other organisms (Figure 1). Niche- specifics and heterospecifics (Galef and La- constructing organisms modify the physical land 2005; Leadbeater and Chittka 2007).

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 12 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

Flack et al. (2006) describe how aggressive po- ple adaptation-byproduct dichotomy would licing by dominant pigtail macaques creates a imply that they are byproducts. However, hu- “social niche,” stabilizing and integrating ma- mans and other animals possess some very gen- caque societies. Similarly, Oh and Badyaev eral, knowledge-gaining adaptations, such as (2010) describe how male house finches ac- the ability to learn, which allow them to actively seek or create social environments with quire and store information in their brains, to properties that enhance their attractiveness or communicate this information to others, and conspicuousness to females (see also Cornwal- in humans to build on that shared knowledge ⌬ lis and Uller 2010). One of the strongest Ri base cumulatively (for instance, through tech- impacts of human niche construction is the nological advance). This means that some ar- selective modification of other species, both tifacts may possess a designed property in spite intentional (e.g., domestication) and uninten- of the fact that they are not adaptations. This tional (e.g., evolution in response to human ⌬ property should shape the ( Rp) ecological activities such as urbanization; Palkovacs et al. signatures found in artifacts, and potentially 2012). In such cases, “transmitting” organisms ⌬ facilitate their recognition. Finally, artifacts leave a Ri signature in the “receiving” organ- ⌬ may sometimes include ( Ri) signs or signals isms. such as the “advertising signals” in a bower- bird’s bower (Madden et al. 2012). Artifacts Niche-constructing organisms may also correlates of signatures of niche modify selection by building artifacts, includ- construction ing nests, burrows, webs, mounds, and dams, If a new niche-constructing activity evolves, it all the way up to the houses, cars, factories, and should start to generate a signature immedi- computers of contemporary humans. We in- ately, although probably only locally at first. clude in this category constructed resources Conversely, if a niche-constructing activity such as crop fields, hedgerows, terraces, canals, changes, or if the niche-constructing pop- artificial lakes, forest clearings, and so forth ulation goes extinct, then the signature it (Kendal et al. 2011). Artifacts are usually phys- previously generated should dissipate from ical abiotic resources, but can be biotic resources (e.g., crop fields, ant gardens) and are that time on. However, niche-constructed often underpinned by information stored in effects might be expected to accumulate the genome or brains of the constructor, as over time, before they become detectable, though they were “intelligently designed.” either because the per-capita impact of a single Although the most familiar artifacts (nests, act of niche construction is very small or be- mounds) are “extended phenotypes” (Dawk- cause low-impact effects are dissipated, coun- ins 1982), this concept is of limited utility here teracted, or swamped by other processes. as it is restricted to niche-constructing adapta- However, in searching for the covariation be- tions, yet artifacts can sometimes be built by the tween putative niche-constructing acts and byproducts of organisms. Examples include the putative ecological effects, we note that this dead coral substrate and sand of coral relationship may be time-lagged (Laland et al. reefs, which are constructed by interac- 1996, 1999; Jones et al. 2010). Analyses of such tions among multiple species over long pe- time lags have been based on genetic mod- riods of time, and the paths and routes els, but similar time lags occur in demo- created by habitual animals. Moreover, graphic and ecological models (Ihara and particularly in the case of those manufac- Feldman 2004; Jones et al. 2010), partly be- tured by humans, artifacts built through cause structure formation and decay and biotic the expression of acquired knowledge may responses to structural and abiotic change are still be informed and structured, even if rarely instantaneous (Jones et al. 2010), they are not themselves biological adapta- although the dynamics of gene-frequency tions. For instance, human buildings are not change may also contribute to time lags biological adaptations and, accordingly, a sim- (Laland et al. 1996).

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 13

Abiota Moreover, the density of a recipient popula- It should be possible to detect association may covary with the density of a tions between specific niche-constructing constructor population n time steps ago, populations and specific ecological signa- if their interaction occurs via a resource, tures in abiota. It may even also be possible manufactured by the constructor population to detect traces of such associations in re- that takes time to accumulate, or is gradually mains of extinct populations, the “ghosts” depleted through niche construction (La- of ancient niche constructors, still capable land et al. 1999). Indirect associations increase of affecting contemporary ecological and the likelihood that the effects are time-lagged. evolutionary processes (Odling-Smee et al. Although these observations might make spe- 2003). For instance, bog-forming Sphag- cies’ associations appear complicated, in prac- num mosses produce peat that can persist tice, it need not make the job of the ecologist for hundreds or thousands of years after the or evolutionary biologist any more difficult, death of the living moss (van Breemen 1995), and may often make it easier since established while the plants of the Carboniferous procedures and methods can address the produced the fossil fuels that with hu- problem and render systems more predictable man help are constructing the modern (Jones and Lawton 1995; Odling-Smee et al. climate of Earth. These kinds of traces 2003). The key to progress is to break down are particularly likely to be found in abi- complicated pathways into tractable compo- ota, which commonly persist for a longer nent pieces (e.g., Post and Palkovacs 2009; time than the populations that gave rise to see the section entitled Methods of Imple- them. Even if the information (Ri) expressed mentation). by an extinct niche-constructing population Ecological associations should typically oc- is now gone, it may still be possible to retrieve cur in the presence of the species expressing a proxy for it from a functionally equivalent the genetic or acquired knowledge responsi- contemporary population, thereby allowing ble for the niche construction, although its original biological functionality to be in- this association may be time-lagged and may ferred. For example, the individual organisms responsible for the oxygenation of the fade or disappear when the relevant niche- atmosphere are extinct, but many contem- constructing population disappears. It follows porary organisms exhibit photosynthesis, sug- that ecological signatures of the novel appear- gesting parallel characteristics in the original ance, the ongoing presence, or the subsequent oxygenators. Features related to the prevalence disappearance of niche constructors may af- or density of extinct or ancestral organisms, the fect inference about the structure, func- signatures of their niche-constructing activity, tion, and regulation of ecosystems. Here may still be recoverable in contemporary NCT can offer some useful predictions—for oceans, sediments, or in the atmosphere (for instance, concerning when invasions occur examples see Erwin 2008; Corenblit et al. and how they should change ecosystem func- 2011). tioning when they do occur (e.g., see Odling- Smee et al. 2003; Loreau 2010). Biota Reciprocal acts of niche construction Ecological associations between niche- by recipient biota may be repeated many constructing and recipient biota typically in- times, so the knock-on consequences of a volve ongoing interactions between currently single type of niche construction could living organisms, corresponding to familiar cascade through many compartments of ecological relationships, such as predator-prey, an ecosystem. Such causal chains, or cas- host-parasite, mutualisms, or competition for a cades, resulting from niche construction shared resource. However, NCT draws atten- may include evolutionary as well as ecolog- tion to the possibility that such interactions ical responses and trigger eco-evolutionary may also be mediated by any number of abiotic dynamics (Pelletier et al. 2009; Post and or biotic intermediate ecosystem components. Palkovacs 2009; Loreau 2010).

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 14 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

Artifacts eco-evolutionary paths in ecosystems de- Ecological signatures left in artifacts by scribed by NCT. niche construction may be very different from those left in other kinds of abiota, from ecological signatures to since they often comprise the entire arti- modified selection fact and may possess a designed quality. To affect evolution, any change in any This should be reflected in particularly ecological variable caused by niche con- strong associations between the expressed struction (⌬R) must translate into at least ⌬ information (Ri) and the signature ( Rp), one modified natural selection pressure particularly in extended phenotypes such as for at least one population in an ecosystem. nests, webs, and mounds. They should also It cannot do that unless the ecological sig- clearly demonstrate their roles in contributing nature caused by the niche construction to the fitness of their constructors. Artifacts are becomes a component in the ecological also often highly organized, one consequence inheritance of at least one population. of which is that they may spontaneously dissi- There would be no evolutionary conse- pate or be deconstructed by other organisms, quences if the ecological changes caused leaving only transitory ecological signatures by NCT dissipated too rapidly, or if they unless they are constantly repaired and/or re- were swamped by other processes or by constructed. Moreover, since artifacts are often other agents, or if there were a constraint difficult and costly to construct and maintain, on the rate of evolution such as the ab- they, or their component parts, will frequently sence of genetic variation. In such cases, be valuable to other organisms and may re- niche construction would have only eco- quire defending. Artifacts can also feed back logical consequences. Niche construction to the constructor population to modify may have evolutionary consequences that the latter’s developmental environment. A are not easily detectable. For example, the good example is provided by Leca et al. evolutionary consequences of counterac- (2010) who describe how piles of stone tive niche construction may be stabilizing, tools left by macaques shape the learning and they may be difficult to detect until the “constructor” disappears from an ecosys- of tool-using traditions in these monkeys. tem. In practice, the criteria necessary for an The Evolutionary Consequences of ecological signature of niche construction Prior Niche Construction to be evolutionarily consequential are eas- We now turn to the second major aspect ily satisfied. Jones et al. (1994, 1997) de- of niche construction, namely, the subse- scribe the principal ecological factors that quent evolution of populations due to natural scale up the impact of “engineering” spe- selection modified by organisms. Note the cies in ecosystems: “(1) lifetime per capita niche-constructing population that causes the activity of individual engineering organ- ecological change does not necessarily have to isms; (2) population density; (3) the local be the same as the recipient population(s) and regional spatial distribution of the whose subsequent evolution is affected by the population; (4) the length of time the pop- selection that has been modified by that par- ulation has been at a site; (5) the type and ticular ecological change (Odling-Smee et al. formation of the constructs, artifacts, or 2003; Post and Palkovacs 2009). This means impacts, and their durability in the ab- that niche construction can potentially trigger sence of the engineers; and (6) the num- both “diffuse coevolution” (Strauss and Irwin ber and types of resources that are directly 2004), and coupled coevolutionary dynamics. or indirectly controlled, the ways these re- Below, we consider what it takes to convert an sources are controlled, and the number of (⌬R) ecological signature into an evolu- other organisms that depend on these re- tionarily significant modified natural selec- sources” (Jones et al. 1997:1952). All of tion pressure for one or more populations, these ecological factors should also make and then review the principal alternative any (⌬R) change caused by the niche con-

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 15 struction more likely to translate into an feeds back exclusively to that genotype in the evolutionarily significant, modified natural bearer responsible for building the extended selection pressure for at least one popula- phenotype and subsequently affects its fitness. tion in an ecosystem. The higher the value Path (1) is described by SET and NCT in the of each of these six variables, the more ecolog- same way. Path (2) is the path modeled in the ically potent the niche construction should be earliest niche construction models (Laland and the more likely it is that this will subse- et al. 1996, 1999) as well as some other mod- quently translate into evolutionarily significant eling frameworks (Bailey 2012). Genotype(s) ecological inheritance for at least one popula- in a niche-constructing population express tion in an ecosystem. It is not necessary for all an environment-altering trait, which modi- six of the above variables to be high simultane- fies an ecological variable. The (⌬R) signa- ously. Some organisms, such as beavers, are ture of change becomes a modified natural potent niche constructors because they pro- selection pressure for the same population. duce large, durable effects (Moore 2006). Unlike path (1), however, the modified selec- Others, such as photosynthesizing cyanobac- tion in this case does not solely feed back to the teria whose individual impacts on their en- allele(s) or genotype(s) responsible for the vironments are tiny, can have a vast impact environment-altering trait(s); rather it (also) if they are present in sufficiently high den- affects the fitness of different allele(s) or geno- sities over sufficiently large areas for suffi- type(s) in the same population. For example, cient time and may affect vast numbers of the beaver dam modifies selection on genes species (Odling-Smee 2010). expressed in beaver browsing or exposure An additional factor that facilitates the con- to parasites. Path (2) can be supported by version of an ecological signature of niche con- environment-altering byproducts as well as struction into a modified natural selection adaptations. pressure is genetic inheritance. Insofar as most Path (3) is similar to path (2), except organisms in a niche-constructing population that the organisms affected by modified share the same genes, and insofar as these natural selection are no longer in the same genes predispose them to niche construct in population as the organisms responsible the same way, generation after generation, for modifying the natural selection. For the collective consequences of their niche instance, beaver dam building and brows- construction are likely to accumulate over ing modify selection acting on genes ex- time to the point where the (⌬R) ecologi- pressed in the production of defensive cal signatures they generate do become chemicals by Populus trees (Martinsen et al. modified natural selection pressures on 1998). Note that path (3) can also repre- the population, which may cause further sent environment-altering traits that are ei- evolution of that or other populations in ther adaptations (e.g., dam building) or an ecosystem. byproducts (e.g., excretion). This “diffuse coevolution” is more accurately described as alternative eco-evolutionary “feed-forward” than “feedback,” and may be feedback paths direct or indirect. The latter would happen if Figure 2 describes the principal alter- an environment-altering population causes a native eco-evolutionary feedback paths change in the availability of an abiotic re- connecting the environment-altering and source, such as water or light, that indirectly selection-pressure-modifying aspects of niche affects a different recipient population. Path construction. (3) might also work via intermediate biota that The simplest path (1) corresponds to could serve as no more than a catalyst for a that emphasized in Dawkins’s extended subsequent evolutionary change in a recipient phenotype. A genotype in a population ex- population. For instance, an intermedi- presses an environment-altering adaptation, ate population might respond to a niche- for instance, the “houses” built by caddisfly lar- constructing population with a demographic vae (Dawkins 1982). This extended phenotype change, without itself evolving further. This then modifies a natural selection pressure that might, nevertheless, translate into a modified

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 16 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

Figure 2. Eco-Evolutionary Feedback Paths Alternative eco-evolutionary feedback paths that connect the environment-altering and selection pressure- modifying aspects of niche construction. Here shaded circles represent biota, white squares represent abiotic compartments in ecosystems, and shaded sections in white squares represent the ⌬R signature of niche construction that has modified the state of the abiota. The grey lines represent chromosomes, with the small white sections specific genes. Continuous arrows represent niche construction, and dashed lines natural selection. Path (1): A genotype in a population expresses an environment-altering trait that builds an extended phenotype, which modifies a natural selection pressure that feeds back exclusively to the genotype responsible for building the extended phenotype. Path (2): A genotype in a niche-constructing population expresses an environment-altering trait that modifies an ecological variable, thereby generating a modified natural selection pressure for the same niche-constructing population, but here affecting the fitness of different genotypes from those responsible for the niche construction. Path (3): Similar to path (2), except that the organisms affected by modified selection are no longer in the same population as the organisms responsible for modifying selection. Here the niche construction of the first population drives ecological change and coevolutionary episodes in the second population. Path (4): Resembles path (3), but with the additional property that the coevolution of two or more populations is facilitated via an additional feedback path. Here the second population modifies another abiotic resource through its own niche construction, which modifies selection acting back on the original population. natural selection pressure for a third popula- ical spillovers in influencing redox gradients tion that subsequently evolves further. Post and and other aspects of the Cambrian Explosion. Palkovacs (2009) give contemporary examples, Path (4) resembles path (3) but with the while Erwin (2008) discusses several cases of additional property that it promotes the ancient niche construction, including sedi- coevolution of two or more populations via ment perturbation and the persistence of shell an additional feedback path or paths. If a beds in marine ecosystems throughout the recipient population responds to modified Phanerozoic, as well as their possible macroev- selection induced by niche construction, olutionary consequences in deep time extend- by itself becoming an environment-changing ing as far back as the Cambrian. Erwin and niche-constructing population that subse- Tweedt (2012) discuss the role of such ecolog- quently modifies selection on the original

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 17 niche-constructing population, then the two ates evolutionary feedback to the construc- populations may coevolve. Once again, the tor, then it must be an adaptation. In fact, feedback paths that connect coevolving niche- recent theory suggests that this need not be the constructing populations to each other could case and niche-constructed byproducts can be be either direct or indirect. If they are indirect consequential for the constructor’s evolution. (path 4), particularly if the coevolution in- This occurs when byproducts precipitate bouts volves modified intermediate abiota such as a of selection in their own population by induc- mineral nutrient or soil pH, then the different ing selection on other traits in the same popu- physical dynamics of the abiotic change in- lation and hitchhiking to fixation on the back duced by niche construction may need to be of this selection they generate (Silver and Di taken into account. A possible example is pro- Paolo 2006; Rendell et al. 2011). In this in- vided by Goddard (2008) who shows exper- stance, spatial structure (local dispersal and imentally how niche construction by the yeast mating) may give rise to statistical associa- Saccharomyces cerevisiae allows it to outcompete tions between niche-constructing traits and other species. This species eventually domi- any genotypes favored in the constructed nates in fruit niches where it is naturally initially environments. Drawing on an important rare, by modifying the environment through distinction made by philosopher Elliot So- fermentation (the Crabtree effect) in ways that ber (1984), here there is selection of the extend beyond ethanol production. NCT also niche-constructing trait, but not selection for describes more complicated patterns of con- it, and only the latter meets the defini- nection between environmental alteration and tion of an adaptation (Williams 1966; So- selection-pressure modification. There can be ber 1984). Nonetheless, in such hitchhiking one:many, many:one,ormany:many relationships cases, there remains evolutionarily conse- between environment-altering populations quential feedback to a niche-constructing and subsequently evolving recipient popula- population stemming directly from its niche- tions via paths 3 and 4. All of these more com- constructing activities. plicated consequences of niche construction may also occur on multiple different scales in the role of acquired characters space and time (Loreau 2010). All of the aforementioned paths through which niche construction acts to trigger evolu- the role of byproducts tionary episodes remain relevant even if the A primary benefit of our broad concep- niche construction is the expression of learned tualization of niche construction is that it knowledge. This point is of particular rele- highlights the important roles that niche- vance for understanding human anthropo- constructing byproducts may play in eco- genic change within ecosystems; clearly it is systems. This is significant because such acquired knowledge that underpins urbaniza- roles are often unintuitive and easy to dis- tion, deforestation, agricultural practices, and miss. It is far more apparent that the bea- the majority of major human impacts on the ver dam may drive coevolutionary episodes environment. Such processes undoubtedly than beaver dung may, yet the latter is a precipitate evolutionary episodes in humans very real possibility. Numerous examples and other species. It would be unwise for have been documented of seemingly in- researchers to assume that human-modified se- consequential and inadvertent acts by or- lective environments can be treated as equiva- ganisms whose aggregate activity generates lent to such independent sources of selection an important ecological signature. For ex- as arise from geological or climatic change ample, consider Euchondrus snails whose since there may be selective feedback to the consumption of endolithic lichens inadver- constructing population of a form that influ- tently generates tonnes of soil, thereby ences its constructing behavior—for instance, a playing a vital role in fertilizing desert eco- dairy-farming niche created the conditions systems (Jones and Shachak 1990). that favored the spread of alleles facilitating Typically, evolutionary biologists assume adult lactase persistence, but the consumption that if a niche-constructing activity gener- of dairy products is more likely in individuals

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 18 THE QUARTERLY REVIEW OF BIOLOGY Volume 88 with the lactose tolerant genotype (Gerbault importance of preexisting “empty niches” et al. 2011). Similarly, deforestation for crop (Schluter 2000; Gavrilets and Losos 2009; planting has apparently inadvertently pro- Tebbich et al. 2010). Although the recent moted insect-propagated diseases (such as literature on adaptive radiations has em- malaria) in many human populations, gener- phasized lineage diversification without ating selection for resistant alleles (Durham any necessary ecological changes, trophic 1991; Laland et al. 2010) and triggering fur- novelty is a critical aspect of adaptive radi- ther human niche construction, such as the ation (Martin and Wainwright 2011). widespread use of chemical insecticides. Rather than lineages simply diversifying to Hence, anthropogenic change leads to “fill” available niches, niches themselves the type of eco-evolutionary dynamics recog- may be diversifying (Erwin 2005), a process nized by ecologists, but underpinned by cul- that Losos (2010) termed “self-propagating tural learning. adaptive radiation.” This process of differential niche construction and subsequent spatial structure evolutionary diversification has been dem- The significance of niche construction onstrated in several microbial laboratory for the evolution of recipient populations is systems (Rainey and Travisano 1998; Ha- further enhanced by spatial structure. Spatially bets et al. 2006) and appears to underlie explicit niche-constructing models have re- the divergence between landlocked and vealed hitherto unrecognized forms of dynam- anadromous alewife populations (Palko- ical feedback in ecosystems. For instance, vacs and Post 2008; Schielke et al. 2012). niche-constructing traits, even if costly, can What remains to be explored both empir- drive themselves to fixation through spatially ically and through theoretical studies is the enhanced statistical association with genotypes extent to which niche construction can drive that they inadvertently favor (Silver and Di the acquisition of new resources sufficiently Paolo 2006). Rather than passively hitchhiking, to enable adaptive radiation. these traits may actively create the conditions that promote their own spread by association temporal dynamics with recipient traits. Such dynamics not only Temporal dynamics are likely to be increase the amount of niche-constructing equally important here, since engineering activity across a population, driving increases occurs over a broad range of temporal scales in the magnitude of the niche construc- (Hastings et al. 2007) and many niche- tion, but have also been shown, under constructing activities accumulate over time, more restricted conditions, to trigger a sec- sometimes even “deep time” (Erwin 2008). ondary hitchhiking process in which ge- Formal theory suggests that time-lagged ef- netic variation at other loci that promote fects stemming from ecological inheritance the potency of the niche-constructing be- can generate rich evolutionary dynamics havior can also be favored (Rendell et al. (e.g., momentum effects, inertia, or cata- 2011). Local spatial effects are also likely to strophic responses; Laland et al. 1996, 1999). be important in cases in which a popula- They also raise some fundamental questions tion promotes its own range expansion for biologists, such as how fitness is measured through niche construction (Kylafis and in natural populations of organisms. If niche- Loreau 2008). An important aspect of constructing organisms can influence pat- spatial structure is its ability to promote terns of selection not just on the current and simultaneous genetic and ecological di- offspring generations, but for many descen- versification if the strength (or nature) dent populations, then an accurate estimate of feedbacks differs from patch to patch of fitness would require consideration of the (Habets et al. 2006). This perspective is costs and benefits of niche construction over important because it has the potential to multiple generations (e.g., McNamara and link eco-evolutionary feedbacks to the pro- Houston 2006; Lehmann 2008). Under such cess of adaptive radiation, which is usually at- circumstances, it may be better to replace tributed to an environmental template and the conventional fitness proxies (such as repro-

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 19 ductive success) with estimates of genotype- by a genotype in a population of earthworms, specific intrinsic growth rates (McNamara and causes an ecological change in an abiotic eco- Houston 2006) or survivability (Palmer and system component, the soil. Subsequently, that Feldman 2012). It may be unwise to ignore change in the soil translates into an evolution- such temporal dynamics since the fitness of arily significant, modified natural selection niche-constructing genotypes is a dynamical pressure in the ecological inheritance of a pop- property of resource concentrations, and ulation of plants. The soil mediates an indirect hence may change over ecological time- connection between whatever genotype is un- spans. Moreover, niche-constructing traits derpinning the niche-constructing activities of that might be considered too costly to the earthworms and an evolutionarily respon- evolve just on the basis of the number of sive genotype in the plants. A more compli- offspring can be favored because of the cated eco-evolutionary chain is shown in benefits that accrue to more distant de- Figure 3b, based on an example studied by scendants (Lehmann 2008; Palmer and Palkovacs et al. (2009). Here predators dif- Feldman 2012). These theoretical dynam- ferentially prey on two guppy populations ics potentially shed light on time-lagged (i and ii), generating different prey size effects observed in ecosystems (Hastings et distributions, which trigger different pat- al. 2007; Jones et al. 2010). Finally, the terns of excretion and consequently affect observation that there can be delayed feed- nutrient cycling rates, leading to differen- back is also relevant to our earlier point tial patterns of algal growth, which poten- that byproducts may be evolutionarily con- tially feed back to affect selection on the sequential. Byproducts can trigger time- guppies through exploitation of caroten- lagged evolutionary events in descendant oids. This sequence of EMGAs creates a populations of the same species, but this causal chain through the guppies’ ecosys- selective feedback may occur too late to tem. greatly influence the selection of the In any real ecosystem, vast numbers of niche-constructing trait. EMGAs must connect multiple environment- altering populations to multiple recipient pop- emgas and engineering control webs ulations, directly and indirectly, through both Jones et al. (1994, 1997) describe webs biotic and abiotic environmental components of connectance in ecosystems caused by in rich networks. We imagine that such net- species influencing energy and mass flows works will frequently resemble engineering and creating habitat and other resources for webs, although they need not be identical. Fig- other species. The same engineering processes ure 3c illustrates a miniature model ecosys- are potentially indirect sources of selection tem comprising both evolving populations originating from niche-constructing popula- and intermediate abiota. These ecosystem tions and acting on recipient populations and compartments are connected by many may underlie webs of diffuse and direct coevo- EMGAs. If the abiota were to be edited out, lution. We describe such indirect evolutionary then Figure 3c would reduce to a conven- interactions as environmentally mediated genotypic tional community of populations, such as a associations or EMGAs (Odling-Smee et al. food web (Figure 3d). However, it is impor- 2003). EMGAs are “[i]ndirect but specific con- tant to note that Figure 3d may sometimes nections between distinct genotypes mediated be a poor model of ecosystem dynamics be- either by biotic or abiotic environmental cause it omits the evolutionarily significant components in the external environment” changes in intermediate abiota caused by (Odling-Smee et al. 2003:419). The key feature niche-constructing organisms that are also of EMGAs is that they map sources of selection responsible for driving ecological and evolu- stemming from one population’s genes onto tionary changes in ecosystems. Their inclu- genotypes in another population that evolve in sion, while it increases complexity, creates response to those modified sources. Figure 3a the linkages between population-community illustrates a single EMGA in an ecosystem. A and ecosystem processes and would be justi- niche-constructing phenotype, influenced fied if it increased predictability.

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 20 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

The recognition of networks of EMGAs changes in recipient populations are the eco- has potentially important implications for un- logical changes caused by environment- derstanding ecosystem dynamics. In practice, altering populations, but the preceding cause the feedback and feed-forward ramiﬁcations of of evolutionary changes in recipient popula- niche-constructing activities are likely to go tions may well be the prior evolution of niche- well beyond isolated “eco-evo” links and pro- constructing populations and, therefore, of the duce interwoven causal chains that thread eco- niche-constructing traits that cause the ecolog- systems. The immediate drivers of evolutionary ical changes in the ﬁrst place. Thus, EMGAs

Figure 3

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 21 connect the prior evolution (“evo”) of geno- systems. Thus, a full understanding of ecosys- types in niche-constructing populations to the tem dynamics will require incorporating subsequent evolution of genotypes in recipient food webs and engineering webs into dynam- populations via the modification of either bi- ical systems that simultaneously consider otic (e.g., their phenotypes) or abiotic (e.g., trophic and competitive interactions along- soil chemistry) intermediate ecosystem (“eco”) side ecosystem engineering and niche con- components. In other words, EMGAs are not struction. just “eco-evo” links in ecosystems, but rather While, in theory, any such “evo-eco-evo” “evo-eco-evo” links. causal chain might be traced back ad infi- This argument can be extended further nitum, in practice it is useful to assume and, in some instances, the cause of selection that some event initiated a particular eco- for the initial evolutionary event in an “evo-eco- evolutionary cascade. This might start with evo” link is the prior niche construction of an a change in the environment, which pre- earlier population, while there may be further cipitates an evolutionary response in a downstream ecological or evolutionary conse- focal population, but this is not the only quences of a given act of niche construction. In possibility; the initiating event may best be this manner, causal chains underpinning eco- characterized as an act of niche construc- system dynamics can potentially be traced, as tion. For example, behavioral plasticity in a illustrated by Post and Palkovacs (2009). It population of animal niche constructors is clear that such interactions will influence may generate an innovation that propa- and to some degree “control” (Jones et al. gates through the population via learning 1997) ecosystem dynamics. However, it will (for instance, birds learning to peck open also be apparent that “ecosystem evolu- milk bottles or monkeys discovering new tion” need not resemble the simulated dy- food sources). In such cases, where pre- namics of the same system reduced to its sumably there is comparatively little advan- biotic components, as in classical model tage to considering the prior evolution of ecosystems (e.g., May 1973), since such mod- the population responsible for the niche els omit EMGAs mediated by abiota and construction, it makes sense to view the hence many important drivers of coevolu- causal chain as triggered by an ecological tionary episodes in ecosystems. The inclusion rather than an evolutionary episode (in of these additional connections could poten- this case, a learned behavior). tially greatly affect the stability of model eco- This is particularly apparent in the case

Figure 3. Feed-forward and Feedback Processes in Ecosystems (a) Two populations connected indirectly by an EMGA. A genotype in a worm population expresses an environment-altering phenotype that modifies the soil. The (⌬R) ecological change in the soil modifies a selection pressure for a plant population. The recipient plants respond with genotypic evolutionary changes. The biota, worms, and plants are symbolized by shaded circles, and the intermediate abiotic environmental component, the soil, is symbolized by the square in the middle. The square is partitioned into shaded and white components, the former representing the (⌬R) ecological change in the soil caused by the earthworms’ niche construction, while the latter represents the extent to which the soil’s state is due to other agents. In principle, these components are measurable. (b) This figure symbolizes the guppy system described by Palkovacs et al. (2009). Differential predation in (i) and (ii) by two predator populations (the first shaded circle) lead to different size distributions in two guppy populations (the second circle), leading to differences in rates of excretion in the two populations (niche construction) leaving different (⌬R) signatures (shaded section of white square) affecting nitrogen cycling in the local environments (white square), affecting algal growth (the third circle), and feeding back to affect selection on the guppy populations. (c) An EMGA based “eco- evolutionary network” in a miniature model of an ecosystem. (d) The same ecosystem reduced to a food web, in a conventional population-community approach. Again, shaded circles represent biota, white squares represent abiotic compartments in ecosystems, and shaded sections in white squares represents the ⌬R signature of niche construction that has modified the state of the abiota. Ecosystem dynamics may be affected by engineering processes and their knock-on consequences, such that sometimes (c) cannot safely be reduced to (d).

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 22 THE QUARTERLY REVIEW OF BIOLOGY Volume 88 of anthropogenic change where, for in- researchers to go beyond the normal prac- stance, deforestation or agricultural practices tice of ecosystem ecology and ask, “What are obviously not triggered by evolution of evolutionary ramifications follow from spe- “tree-chopping” or “crop-planting genes,” but cies’ ecological impacts on biota and abi- nonetheless can initiate eco-evolutionary ota?” events in an ecosystem (Laland et al. 2010). Any eco-evolutionary system underpinned Although we see utility in considering gene by niche construction can be viewed as a networks operating across multiple species in chain of component links that either com- ecosystems (Whitham et al. 2006), such exam- prise the environment-altering (ecosystem ples serve to emphasize that the interplay be- engineering) aspect of niche construction tween ecological and evolutionary processes (or its downstream consequences) or a sub- cannot always be reduced to the genetic level sequent evolutionary response to prior niche and, sometimes, such as where acquired char- construction (Odling-Smee et al. 2003). The acteristics underpin niche construction, a study of eco-evolutionary dynamics will require more appropriate focus may be on “envi- researchers to utilize the methods of both ecol- ronmentally mediated phenotypic associations.” ogy and evolutionary biology. The key to prog- One advantage of a niche-construction per- ress is to break down complicated pathways in spective is that it emphasizes the value of think- eco-evolutionary networks into tractable com- ing of phenotypes as being reconstructed in ponent pieces (e.g., Post and Palkovacs 2009), each generation by different developmental re- subject each to analysis and then reconstruct sources (genetic and nongenetic), rather than the network, including the strength of in- as expression of genetically encoded informa- teractions and how they vary over time to tion only. The focus of some methods (e.g., gain a systems-level understanding. So- community genetics) on heritable phenotypic phisticated techniques to study network variation to the exclusion of nonheritable vari- structure and dynamics have been devel- ation (e.g., plasticity, epigenetics, population oped via statistical mechanics and applied structure) may miss many interesting eco- to a wide variety of problems, from social evolutionary processes (Uller 2008; Ellner et al. dynamics revealed by internet traffic to the 2011; Palkovacs et al. 2012). dynamics of power grids and food webs (Newman 2010). Methods of Implementation The guppy system studied by Palkovacs The sheer complexity of engineering et al. (2009) serves as a useful example webs and the challenge of tracing evo-eco- (Figure 3b). The first link in the chain is evo causal chains through ecosystems differential predation on two guppy popu- appears to be a daunting ordeal for the lations. This environment-altering aspect ecologist. In reality, the practical study of of niche construction by the guppy preda- ecology and evolution is not changed by tors leaves an ecological signature in the this perspective, as standard methods reg- form of different size distributions in two ularly and successfully deployed by ecolo- guppy populations. Characterizing this gists and evolutionary biologists can be link is, at least in principle, straightfor- combined to good effect. What is different ward: it requires that the predators be is the focus of investigation, which moves identified, the intensity of predation mea- from the study of the ecological impact or sured, and the guppy size distributions evolutionary response in a single taxon to sampled and quantified. Ideally, statistical the investigation of eco-evolutionary sys- analyses would link the intensity of preda- tems, pathways, or networks. This requires tion to the resulting prey distributions. It that researchers go beyond the normal is, of course, possible that there has been practice of evolutionary biology and ask, an evolutionary response to this predation “What causes the selection pressures lead- in the guppies with, for instance, more in- ing to a specific evolutionary response?” tense predation favoring smaller adult guppy rather than treating those selection pres- sizes, alternative trophic morphology, or di- sures as a starting point. It also requires etary preferences. This is highly plausible

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 23 given the evidence for rapid evolution in canceling or enhancing niche constructors’ guppies (Reznik et al. 1997) and in many current activity by removal and/or providing other species in which the response to se- them with the resources necessary for popula- lection has been measured (Endler 1986; tion growth (e.g., removing prey or feeding the Kingsolver et al. 2001; Ellner et al. 2011). guppies); canceling or enhancing niche con- Again, this can be investigated using stan- structors’ current activity by removing from or dard procedures (Endler 1986). supplementing the ecosystem with introduced The second link in the chain is the differen- members of the same guild (i.e., removing or tial excretion by the two populations and its adding a different species that engineers in the impact on nitrogen (N) and phosphorus (P) same manner, such as another small fish with levels in the water. This second environment- similar impact on nutrient flux); artificially re- altering component of niche construction moving or manufacturing and introducing the stems from the guppies and leaves an ecologi- products of the niche constructors (e.g., chem- cal signature in two abiotic environmental ically extracting or adding N and P to mimic components: the local aquatic N and P levels. the effects of excretion); and counteracting Palkovacs et al. (2009) were able to demon- negative effects and facilitating positive influ- strate this link experimentally with a mesocosm ences of both abiotic and biotic factors that experiment. The results showed that high- may affect niche constructors through trophic predation guppy populations contributed or nontrophic links (e.g., interfering with the approximately double the amount of N predation regimes). In all cases, the effects of and P to the total nutrient pool via excre- these manipulations on algal growth could be tion compared to low-predation popula- monitored. tions. In this way, differential excretion was The final link in this chain could be to test identified as a correlate of the signature of a hypothesized feedback to trait evolution niche construction (modified N and P). by connecting changes in algal biomass to Under conditions of equal biomass, a pop- guppy color patterns by demonstrating a re- ulation dominated by smaller individuals sponse to (natural or sexual) selection in this (the high-predation population) is expected character through sensitivity to the amount to drive higher nutrient fluxes than a pop- of algae-derived carotenoids available in the ulation dominated by larger individuals environment, again through standard proce- (low predation; Hall et al. 2007). dures for detecting selection (Endler 1986). The third link in the chain connects the As this example demonstrates, testing links excretion-mediated nutrient fluxes to algal in eco-evolutionary chains of causality re- growth in the streams. This link could be quires working at a variety of scales—from the demonstrated in the field by correlating whole-ecosystem scale to the small-scale exper- measures of guppy excretion rates (or, more iment. Small-scale experiments are needed to realistically, estimates of these based on as- isolate trait change as the driver of ecological says of population densities as a function of change and can be useful for running select- size) with rates of algal growth. It could also ion experiments under controlled conditions. be demonstrated experimentally, either un- However, to show that trait-driven ecosystem der controlled conditions in the laboratory dynamics and feedback effects are occurring or micro-/mesocosms or in the field, by can- also requires simultaneous studies at the whole- celing or enhancing the environment- ecosystem scale, where such dynamics play out altering behavior (excretion) or mimicking its in nature. niche-constructed effects (Odling-Smee et al. In practice, the above procedures are 2003). Relevant experimental manipulations likely to be complicated by many additional include: canceling or enhancing the niche challenges, including the aforementioned constructors’ current activity by removing indi- spatial effects, time-lagged responses and viduals and/or supplementing their numbers other convoluted temporal dynamics, slow with introduced members of the same species responses to selection in long-lived species, (e.g., decreasing/increasing guppy num- and so forth. Nonetheless, this example illus- bers or manipulating size distributions); trates the tractability of tracing at least some

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 24 THE QUARTERLY REVIEW OF BIOLOGY Volume 88 causal pathways via eco-evolutionary feed- The starting point for an investigation backs and feed-forwards. Researchers can need not be the first link in the chain, as potentially go beyond merely identifying researchers can potentially work backward as causal influences to quantifying their mag- well as forward in identifying causal influ- nitude and duration. Other procedures that ences (Odling-Smee et al. 2003). We envis- can be deployed include experimental manip- age that the signatures of niche construction ulation of the potency of niche construction, highlighted in the sections Niche Construc- common gardening experiments, selection ex- tion Theory and Ecosystem-Level Ecology periments, and experimental evolution. Where and The Evolutionary Consequences of relevant data are available, statistical ap- Prior Niche Construction might frequently proaches such as structural equation modeling be the starting points for analysis, triggering and causal graphs can help to isolate or con- investigations of their downstream conse- firm putative causal influences and/or reject quences for other ecosystem compartments causal hypotheses that are inconsistent with the or, alternatively, exploration of the causes of data (Pearl 2000; Shipley 2000). Once com- the niche construction. peting causal hypotheses are translated into The boundary of the system, in terms of statistical models, researchers can assess the where the eco-evolutionary causal chain or evidence for each by standard statistical anal- web should start or stop, is a decision for yses or comparing competing hypotheses by researchers, given the nature of their re- using information theoretic or Bayesian cri- search interests and the resources available teria. The same methods allow the magni- to them. In principle, systems could range in tude of causal influences to be quantified scale from short-term responses in isolated and niche-constructing effects on variables to eco-evolutionary links to analysis of long- be distinguished from other processes. With term patterns of change in entire ecosys- time-series data, the duration of niche- tems. In the latter case, we note increasing constructing effects and the rate of response to reference to “ecosystem evolution” within selection can also be quantified. the ecological literature (e.g., Loreau et al. Ecosystems are threaded by a bewildering 2004), and point out that, to the extent that array of EMGAs, so a primary concern for ecosystems comprise not just biota but also researchers is how to constrain the system to abiota, the integrated changes in entire eco- a manageable size. This amounts to deciding systems over evolutionary scales must be me- which acts of niche construction and which diated by niche construction. Related to this responses to selection are too weak, tran- point, given that much niche construction is sient, or trivial to merit inclusion and, con- stabilizing in effect, by counteracting prior versely, which are too important to ignore. In changes in the environment (Odling-Smee many instances, likely or plausible key niche- et al. 2003), it follows that the absence of constructing processes can be identified a changes in entire ecosystems over evolution- priori by taking account of the factors de- ary scales in spite of evolutionary changes in scribed in the section entitled From Ecolog- composite species must be mediated by ical Signatures to Modified Selection that niche construction. scale up the impact of niche construction, Thus far we have said little about genes, yet supplemented by a knowledge of the basic our definition of EMGAs places emphasis on biology of the species (Jones et al. 1994, indirect associations between potentially dis- 1997). Ideally, however, researchers would go tant genotypes. Increasingly, researchers are beyond identifying a plausible key niche- pointing out the plausibility of community or constructing process to actually demonstrating ecosystem genetics, and experimental and sta- its importance through data collection and statistical tools are now available to identify the tistical analysis or experimental manipulations genes underlying a niche-constructing output and estimating its magnitude and duration. or a phenotype that responds to selection Likewise, in principle, putatively unimportant (Whitham et al. 2006). Although we welcome processes can be confirmed to be so through and indeed encourage such developments, we the same procedures. also stress that the tracing of causality through

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 25 mapping eco-evolutionary processes in ecosys- 2006; Crain and Bertness 2006). Niche con- tems can be done entirely at the phenotypic struction is a core component of such dynam- level. Indeed, in instances where acquired ics since it connects ecological to evolutionary characters underpin the focal niche construc- processes. We hope that the conceptual tools tion, which includes most anthropogenic that we have presented here will be of practical change, analysis at the genetic level would be use to ecologists, leading to a deeper under- misplaced. As a starting point, a reasonable standing of how ecosystems change over time focus may be on “environmentally mediated and resist change. phenotypic associations,” with isolating EMGAs a long-term objective. acknowledgments There is increasing evidence that eco- Research supported in part by an ERC Advanced evolutionary feedbacks play important roles in Grant to Kevin N. Laland, by NIH grant GM28016 to the dynamics of many ecosystems (Hairston et Marcus W. Feldman, and support from the NASA al. 2005; Pelletier et al. 2009; Post and Palko- Astrobiology Institute to Douglas H. Erwin. We are vacs 2009), with important implications for grateful to Tobias Uller and Clive Jones for helpful conservation and biodiversity (Boogert et al. comments on an earlier draft.

REFERENCES

Bailey N. W. 2012. Evolutionary models of extended concepts and current understandings. Earth-Science phenotypes. Trends in Ecology and Evolution 27:561– Reviews 106:307–331. 569. Cornwallis C. K., Uller T. 2010. Towards an evolution- Berke S. K. 2010. Functional groups of ecosystems ary ecology of sexual traits. Trends in Ecology and engineers: a proposed classiﬁcation with com- Evolution 25:145–152. ments on current issues. Integrative and Compara- Crain C. M., Bertness M. D. 2006. Ecosystem engi- tive Biology 50:147–157. neering across environmental gradients: implica- Bonduriansky R. 2012. Rethinking heredity, again. tions for conservation and management. BioScience Trends in Ecology and Evolution 27:330–336. 56:211–218. Bonduriansky R., Day T. 2009. Nongenetic inheri- Creanza N., Fogarty L., Feldman M. W. 2012. Models tance and its evolutionary implications. Annual of cultural niche construction with selection and Review of Ecology, Evolution, and Systematics 40:103– assertive mating. PLoS ONE 7:e42744. 125. Cuddington K., Byers J. E., Wilson W. G., Hastings A. Boni M. F., Feldman M. W. 2005. Evolution of antibi- 2007. Ecosystem Engineers: Plants to Protists. Boston otic resistance by human and bacterial niche con- (Massachusetts): Academic Press. struction. Evolution 59:477–491. Danchin E´ ., Charmantier A., Champagne F. A., Me- Boogert N. J., Paterson D. M., Laland K. N. 2006. The soudi A., Pujol B., Blanchet S. 2011. Beyond DNA: implications of niche construction and ecosystem integrating inclusive inheritance into an extended engineering for conservation biology. BioScience theory of evolution. Nature Reviews Genetics 12:475– 56:570–578. 486. Borenstein E., Kendal J., Feldman M. 2006. Cultural Dawkins R. 1982. The Extended Phenotype: The Gene as niche construction in a metapopulations. Theoret- the Unit of Selection. Oxford (United Kingdom): ical Population Biology 70:92–104. Freeman Press. Chaitin G. J. 1987. Algorithmic Information Theory. Cam- Durham W. H. 1991. Coevolution: Genes, Culture, and bridge (United Kingdom): Cambridge University Human Diversity. Stanford (California): Stanford Press. University Press. Corenblit D., Steiger J., Gurnell A. M., Naiman R. J. Ehrlich P. R. 1986. The Machinery of Nature. New York: 2009. Plants intertwine ﬂuvial landform dynamics Simon and Schuster. with ecological succession and natural selection: a Ellner S. P., Geber M. A., Hairston N. G. Jr. 2011. niche construction perspective for riparian sys- Does rapid evolution matter? Measuring the rate tems. Global Ecology and Biogeography 18:507–520. of contemporary evolution and its impacts on eco- Corenblit D., Baas A. C. W., Bornette G., Darrozes J., logical dynamics. Ecology Letters 14:603–614. Delmotte S., Francis R. A., Gurnell A. M., Julien F., Endler J. A. 1986. Natural Selection in the Wild. Prince- Naiman R. J., Steiger J. 2011. Feedbacks between ton (New Jersey): Princeton University Press. geomorphology and biota controlling Earth surface Erwin D. H. 2005. Seeds of diversity. Science 308:1752– processes and landforms: a review of foundation 1753.

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 26 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

Erwin D. H. 2008. Macroevolution of ecosystem engi- Jones C. G., Lawton J. H., Shachak M. 1997. Positive neering, niche construction and diversity. Trends and negative effects of organisms as physical eco- in Ecology and Evolution 23:304–310. system engineers. Ecology 78:1946–1957. Erwin D. H., Tweedt S. 2012. Ecological drivers of the Jones C. G., Gutie´rrez J. L., Byers J. E., Crooks J. A., Ediacaran-Cambrian diversification of Metazoa. Lambrinos J. G., Talley T. S. 2010. A framework Evolutionary Ecology 26:417–433. for understanding physical ecosystem engineering Flack J. C., Girvan M., de Waal F. B. M., Krakauer by organisms. Oikos 119:1862–1869. D. C. 2006. Policing stabilizes construction of so- Kendal J., Tehrani J. J., Odling-Smee J. 2011. Human cial niches in primates. Nature 439:426–429. niche construction in interdisciplinary focus. Phil- Galef B. G. Jr., Laland K. N. 2005. Social learning in osophical Transactions of the Royal Society B: Biological animals: empirical studies and theoretical models. Sciences 366:785–792. BioScience 55:489–499. Kerr B., Schwilk D., Bergman A., Feldman W. 1997. Gavrilets S., Losos J. B. 2009. Adaptive radiation: con- Rekindling an old flame: a haploid model for the trasting theory with data. Science 323:732–737. evolution and impact of flammability in resprout- Gerbault P., Liebert A., Itan Y., Powell A., Currat M., ing plants. Evolutionary Ecology Research 1:807–833. Burger J., Swallow D. M., Thomas M. G. 2011. Kingsolver J. G., Hoekstra H. E., Hoekstra J. M., Ber- Evolution of lactase persistence: an example of rigan D., Vignieri S. N., Hill C. E., Hoang A., human niche construction. Philosophical Transac- Gilbert P., Beerli P. 2001. The strength of pheno- tions of the Royal Society B: Biological Sciences 366: typic selection in natural populations. American 863–877. Naturalist 157:245–261. Goddard M. 2008. Quantifying the complexities of Krakauer D. C., Page K. M., Erwin D. H. 2009. Diver- Saccharomyces cerevisiae’s ecosystem engineering via sity, dilemmas, and monopolies of niche construc- fermentation. Ecology 89:2077–2082. tion. American Naturalist 173:26–40. Habets M. G. J. L., Rozen D. E., Hoekstra R. F., De Kylafis, G., Loreau M. 2008. Ecological and evolution- Visser J. A. G. M. 2006. The effect of population ary consequences of niche construction for its structure on the adaptive radiation of microbial agent. Ecology Letters 11:1072–1081. populations evolving in spatially structured envi- Kylafis G., Loreau M. 2011. Niche construction in the ronments. Ecology Letters 9:1041–1048. light of niche theory. Ecology Letters 14:82–90. Hairston N. G. Jr., Ellner S. P., Geber M. A., Yoshida Laland K. N., Sterelny K. 2006. Perspective: seven T., Fox J. A. 2005. Rapid evolution and the con- reasons (not) to neglect niche construction. Evo- vergence of ecological and evolutionary time. Ecol- lution 60:1751–1762. ogy Letters 8:1114–1127. Laland K. N., Odling-Smee F. J., Feldman M. W. 1996. Hall R. O. Jr, Koch B. J., Marshall M. C., Taylor B. W., The evolutionary consequences of niche construc- Tronstad L. M. 2007. How Body Size Mediates the tion: a theoretical investigation using two-locus Role of Animals in Nutrient Cycling in Aquatic theory. Journal of Evolutionary Biology 9:293–316. Ecosystems. Pages 286–305 in Body Size: The Struc- Laland K. N., Odling-Smee F. J., Feldman M. W. 1999. ture and Function of Aquatic Ecosystems, edited by A. Evolutionary consequences of niche construction Hildrew, D. Raffaelli, and R. Edmonds-Brown. and their implications for ecology. Proceedings of Cambridge (United Kingdom): Cambridge Uni- the National Academy of Sciences of the United States of versity Press. America 96:10242–10247. Hastings A., Byers J. E., Crooks J. A., Cuddington K., Laland K. N., Odling-Smee F. J., Feldman M. W. 2001. Jones C. G., Lambrinos J. G., Talley T. S., Wilson Cultural niche construction and human evolu- W. G. 2007. Ecosystem engineering in space and tion. Journal of Evolutionary Biology 14:22–33. time. Ecology Letters 10:153–164. Laland K. N., Odling-Smee F. J., Myles S. 2010. How Hui C., Li Z., Yue D.-X. 2004. Metapopulation dynam- culture has shaped the human genome: bringing ics and distribution, and environmental heteroge- genetics and the human sciences together. Nature neity induced by niche construction. Ecological Reviews Genetics 11:137–148. Modelling 177:107–118. Leadbeater E., Chittka L. 2007. Social learning in insects— Ihara Y., Feldman M. W. 2004. Cultural niche con- from miniature brains to consensus building. Current struction and the evolution of small family size. Biology 17:R703–R713. Theoretical Population Biology 65:105–111. Leca J.-B., Gunst N., Huffman M. A. 2010. Indirect Jones C. G., Lawton J. H. 1995. Linking Species and social influence in the maintenance of the stone- Ecosystems. New York: Chapman and Hall. handling tradition in Japanese macaques, Macaca Jones C. G., Shachak M. 1990. Fertilization of desert fuscata. Animal Behaviour 79:117–126. soil by rock-eating snails. Nature 346:839–841. Lehmann L. 2007. The evolution of trans-generational Jones C. G., Lawton J. H., Shachak M. 1994. Organ- altruism: kin selection meets niche construction. isms as ecosystem engineers. Oikos 69:373–386. Journal of Evolutionary Biology 20:181–189.

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions March 2013 NCT: A PRACTICAL GUIDE FOR ECOLOGISTS 27

Lehmann L. 2008. The adaptive dynamics of niche Menge B. A., Berlow E. L., Blanchette C. A., Navarrete constructing traits in spatially subdivided popula- S. A., Yamada S. B. 1994. The keystone species tions: evolving posthumous extended phenotypes. concept: variation in interaction strength in a Evolution 62:549–566. rocky intertidal habitat. Ecological Monographs 64: Lewontin R. C. 1982. Organism and Environment. 249–286. Pages 151–170 in Learning, Development and Cul- Moore J. W. 2006. Animal ecosystem engineers in ture: Essays in Evolutionary Epistemiology, edited by streams. BioScience 56:237–246. H. C. Plotkin. New York: Wiley. Naiman R. J., Johnston C. A., Kelley J. C. 1988. Alter- Lewontin R. C. 1983. Gene, Organism and Environ- ation of North American streams by beaver. Bio- ment. Pages 273–285 in Evolution from Molecules to Science 38:753–762. Men, edited by D. S. Bendall. Cambridge (United Newman M. E. J. 2010. Networks: An Introduction. Ox- Kingdom): Cambridge University Press. ford (United Kingdom): Oxford University Press. Lewontin R. C. 2000. The Triple Helix: Gene, Organism, Odling-Smee F. J. 1988. Niche-Constructing Phenotypes. and Environment. Cambridge (Massachusetts): Pages 73–132 in The Role of Behavior in Evolution, edited Harvard University Press. by H. C. Plotkin. Cambridge (Massachusetts): MIT Likens G. E. 1995. Forward (sic). Pages ix–x in Linking Press. Species and Ecosystems, edited by C. G. Jones and Odling-Smee F. J. 2010. Niche Inheritance. Pages J. H. Lawton. New York: Chapman and Hall. 175–207 in Evolution: The Extended Synthesis, edited Loreau M. 2010. From Populations to Ecosystems: Theo- by M. Pigliucci and G. B. Mu¨ller. Cambridge (Mas- retical Foundations for a New Ecological Synthesis. sachusetts): MIT Press. Princeton (New Jersey): Princeton University Odling-Smee F. J., Laland K. N., Feldman M. W. 1996. Press. Niche construction. American Naturalist 147:641–648. Loreau M., de Mazancourt C., Holt R. D. 2004. Eco- Odling-Smee F. J., Laland K. N., Feldman M. W. system Evolution and Conservation. Pages 327– 2003. Niche Construction: The Neglected Process in 343 in Evolutionary Conservation Biology, edited by Evolution. Princeton (New Jersey): Princeton R. Ferrie´re, U. Dieckmann, and D. Couvert. Cam- University Press. bridge (United Kingdom): Cambridge University Oh K. P., Badyaev A. V. 2010. Structure of social Press. networks in a passerine bird: consequences for Losos J. B. 2010. Adaptive radiation, ecological oppor- sexual selection and the evolution of mating strat- tunity, and evolutionary determinism. American egies. American Naturalist 176:E80–E89. Naturalist 175:623–639. O’Neill, R. V., DeAngelis D. L., Waide J. B., Allen Madden J. R., Dingle C., Isden J., Sparfeld J., Goldizen T. F. H. 1986. A Hierarchical Concept of Ecosystems. A. W., Endler J. A. 2012. Male spotted bowerbirds Princeton (New Jersey): Princeton University propagate fruit for use in their sexual display. Press. Current Biology 22:R264–R265. Palkovacs E. P., Hendry A. P. 2010. Eco-evolutionary Martin C. H., Wainwright P. C. 2011. Trophic novelty dynamics: intertwining ecological and evolution- is linked to exceptional rates of morphological ary processes in contemporary time. F1000 Biology diversification in two adaptive radiations of Cypri- Reports 2:1. nodon pupfish. Evolution 65:2197–2212. Palkovacs E. P., Post D. M. 2008. Eco-evolutionary Martinsen G. D., Driebe E. M., Whitham T. G. 1998. interactions between predators and prey: can Indirect interactions mediated by changing plant predator-induced changes to prey communities chemistry: beaver browsing benefits beetles. Ecol- feed back to shape predator foraging traits? Evo- ogy 79:192–200. lutionary Ecology Research 10:699–720. Matthews B., Narwani A., Hausch S., Nonaka E., Palkovacs E. P., Marshall M. C., Lamphere B. A., Hannes P., Yamamichi M., Sullam K. E., Bird K. C., Lynch B. R., Weese D. J., Fraser D. F., Reznick Thomas M. K., Hanley T. C., Turner C. B. 2011. D. N., Pringle C. M., Kinnison M. T. 2009. Exper- Toward an integration of evolutionary biology and imental evaluation of evolution and coevolution as ecosystem science. Ecology Letters 14:690–701. agents of ecosystem change in Trinidadian streams. May R. M. 1973. Stability and Complexity in Model Eco- Philosophical Transactions of the Royal Society B: Biological systems. Princeton (New Jersey): Princeton Univer- Sciences 364:1617–1628. sity Press. Palkovacs E. P., Kinnison M. T., Correa C., Dalton McNamara J. M., Houston A. I. 2006. State and Value: C. M., Hendry A. P. 2012. Fates beyond traits: A Perspective from Behavioural Ecology. Pages ecological consequences of human-induced trait 59–88 in Social Information Transmission and Hu- change. Evolutionary Applications 5:183–191. man Biology, edited by J. C. K. Wells, S. Strickland, Palmer M. E., Feldman M. W. 2012. Survivability is and K. Laland. Boca Raton (Florida): Taylor and more fundamental than evolvability. PLoS ONE Francis. 7:e38025.

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions 28 THE QUARTERLY REVIEW OF BIOLOGY Volume 88

Pearl J. 2000. Causality: Models, Reasoning, and Infer- Silver M., Di Paolo E. 2006. Spatial effects favour the ence. Cambridge (United Kingdom): Cambridge evolution of niche construction. Theoretical Popu- University Press. lation Biology 70:387–400. Pelletier F., Garant D., Hendry A. P. 2009. Eco- Sober E. 1984. The Nature of Selection: Evolutionary The- evolutionary dynamics. Philosophical Transactions of ory in Philosophical Focus. Cambridge (Massachu- the Royal Society B: Biological Sciences 364:1483–1489. setts): MIT Press. Post D. M., Palkovacs E. P. 2009. Eco-evolutionary Strauss S. Y., Irwin R. E. 2004. Ecological and evolu- feedbacks in community and ecosystem ecology: tionary consequences of multispecies plant-animal interactions between the ecological theatre and interactions. Annual Review of Ecology, Evolution, the evolutionary play. Philosophical Transactions of and Systematics 35:435–466. the Royal Society B: Biological Sciences 364:1629– Tebbich S., Sterelny K., Tesche I. K. 2010. The tale of 1640. the finch: adaptive radiation and behavioural flex- Rainey P. B., Travisano M. 1998. Adaptive radiation in ibility. Philosophical Transactions of the Royal Society a heterogeneous environment. Nature 394:69–72. B: Biological Sciences 365:1099–1109. Rendell L., Fogarty L., Laland K. N. 2011. Runaway cultural niche construction. Philosophical Transac- Uller T. 2008. Developmental plasticity and the evo- tions of the Royal Society of London B: Biological Sci- lution of parental effects. Trends in Ecology and ences 366:823–835. Evolution 23:432–438. Reznick D. N., Shaw F. H., Rodd F. H., Shaw R. G. van Breemen N. 1995. How Sphagnum bogs down 1997. Evaluation of the rate of evolution in natural other plants. Trends in Ecology and Evolution 10: populations of guppies (Poecilia reticulata). Science 270–275. 275:1934–1937. Van Dyken J. D., Wade M. J. 2012. Origins of altruism Rowntree J. K., Shuker D. M., Preziosi R. F. 2011. diversity II: runaway coevolution of altruistic strat- Community genetics: at the crossroads of ecology egies via “reciprocal niche construction.” Evolution and evolutionary genetics. Philosophical Transac- 66:2498–2513. tions of the Royal Society B: Biological Sciences 366: Whitham T. G. et al. 2006. A framework for commu- 1322–1460. nity and ecosystem genetics: from genes to ecosys- Schielke E. G., Palkovacs E. P., Post D. M. 2012. In- tems. Nature Reviews Genetics 7:510–523. traspecific niche differences in the alewife re- Williams G. C. 1966. Adaptation and Natural Selection: A vealed by combining stable isotopes and direct Critique of Some Current Evolutionary Thought. Princeton diets. Biological Theory doi:10.1007/s13752-012- (New Jersey): Princeton University Press. 0031-9. Wright J. P., Jones C. G. 2006. The concept of Schluter D. 2000. The Ecology of Adaptive Radiation. organisms as ecosystem engineers ten years on: Oxford (United Kingdom): Oxford University progress, limitations, and challenges. BioScience Press. 56:203–209. Schoener T. W. 2011. The newest synthesis: under- Wright J. P., Jones C. G., Flecker A. S. 2002. An standing the interplay of evolutionary and ecolog- ecosystem engineer, the beaver, increases species ical dynamics. Science 331:426–429. Shipley B. 2000. Cause and Correlation in Biology: A richness at the landscape scale. Oecologia 132:96– User’s Guide to Path Analysis, Structural Equations 101. and Causal Inference. Cambridge (United King- dom): Cambridge University Press. Handling Editor: James Thomson

This content downloaded from 138.251.144.206 on Mon, 1 Apr 2013 08:08:43 AM All use subject to JSTOR Terms and Conditions