CONCEPTS & SYNTHESIS
EMPHASIZING NEW IDEAS TO STIMULATE RESEARCH IN ECOLOGY
Ecological Monographs, 84(2), 2014, pp. 245–263
Ó 2014 by the Ecological Society of America
Under niche construction: an operational bridge between ecology,
evolution, and ecosystem science
BLAKE MATTHEWS,
1,8
LUC DE MEESTER,
2
CLIVE G. JONES,
3
BAS W. IBELINGS,
4
TJEERD J. BOUMA,
5
VISA NUUTINEN,
6
JOHAN VAN DE KOPPEL,
5
AND JOHN ODLING-SMEE
7
1
EAWAG, Aquatic Ecology Department, Center for Ecology, Evolution and Biogeochemistry, Kastanienbaum 6047 Switzerland
2
Laboratory of Ecology, Evolution and Conservation, University of Leuven, 3000 Leuven, Belgium
3
Cary Institute of Ecosystem Studies, P.O. Box AB, Millbrook, New York 12545 USA
4
University of Geneva, Institut FA Forel, 10 Route de Suisse, Versoix
5
Royal Netherlands Institute for Sea Research (NIOZ), Post Box 140, 4400 AC Yerseke, The Netherlands
6
MTT Agrifood Research Finland, FIN-31600 Jokioinen, Finland
7
Mansfield College, University of Oxford, Oxford OX1 3TF United Kingdom
Abstract. All living organisms modify their biotic and abiotic environment. Niche
construction theory posits that organism-mediated modifications to the environment can
change selection pressures and influence the evolutionary trajectories of natural populations.
While there is broad support for this proposition in general, there is considerable uncertainty
about how niche construction is related to other similar concepts in ecology and evolution.
Comparative studies dealing with certain aspects of niche construction are increasingly
common, but there is a troubling lack of experimental tests of the core concepts of niche
construction theory. Here, we propose an operational framework to evaluate comparative and
experimental evidence of the evolutionary consequences of niche construction, and suggest
how such research can improve our understanding of ecological and evolutionary dynamics in
ecosystems. We advocate for a shift toward explicit experimental tests of how organism-
mediated environmental change can influence the selection pressures underlying evolutionary
responses, as well as targeted field-based comparative research to identify the mode of
evolution by niche construction and assess its importance in natural populations.
Key words: alternative stable states; coevolution; diffuse coevolution; eco-evolutionary dynamics; eco-
evolutionary feedbacks; ecosystem engineering; niche construction; trophic interactions.
INTRODUCTION
The basic premise of niche construction theory is that
organisms can act as potent agents of natural selection
by modifying biotic and abiotic environmental condi-
tions (Lewontin 1983, Odling-Smee et al. 2003, 2013).
Previous research on niche construction has extensively
documented how living organisms, through their me-
tabolism, activities, and choices, can alter their sur-
rounding environment and by doing so influence
prevailing selection pressures (Odling-Smee et al. 1996,
2003). Animals, for example, dig burrows, build nests,
aerate soils, construct webs, and forage for prey, while
plants photosynthesize, weather rocks, produce soil, and
create shade (Odling-Smee et al. 2003). Such activities
can modify the selective environment of the organism
doing the environmental modification (Odling-Smee et
al. 1996) or of an unrelated population (Odling-Smee et
al. 2003, 2013). Organism-mediated environmental
modifications can also persist through time and affect
selection pressures experienced by future generations, a
process referred to as ecological inheritance (Odling-
Smee et al. 2003). Ecological inheritance is a key element
of niche construction theory that is increasingly being
integrated into evolutionary theory (Bonduriansky and
Day 2009, Danchin et al. 2011, Bonduriansky 2012)
When using the term niche construction (Odling-Smee
et al. 2003, 2013), niche refers to the sum of all natural
selection pressures experienced by a population and
construc tion refers to the modification of selection
pressures, either through physical modification of the
environment or through habitat choice (Odling-Smee et
Manuscript received 21 May 2013; revised 9 October 2013;
accepted 22 October 2013. Corresponding Editor: N. J. B.
Kraft.
8
E-mail: blake.matthews@eawag.ch
245
al. 2003). At the outset, niche construction theory
focused on how organisms can modify their own
selective environments (Odling-Smee et al. 1996), and
so many classic examples of niche construction highlight
the importance of reciprocal interactions between
organisms and their own selective environment (Od-
ling-Smee et al. 2003). Leaf cutter ants, for example,
cultivate gardens of fungus upon which they are
obligately dependent (Mueller and Gerardo 2002),
and, in some cases, this has culminated in a loss of
genes associated with the acquisition of specific nutrients
(Ellers et al. 2012). Earthworms modify the structure of
their soil environment in a way that facilitates water
uptake into their bodies, thereby partially solving a
critical physiological problem associated with living in
terrestrial environments (Turner 2002). However, it is
increasingly evident that organism-mediated environ-
mental modifications can have a wide range of direct
and indirect evolutionary effects on multiple species in
natural communities (Odling-Smee et al. 2013, Walsh
2013). Odling-Smee et al. (2003) describe one type of
indirect evolutionary effect as an environmentally
mediated genotypic association (EMGA), which is an
association that develops between distinct genotypes in
the environment mediated by the effect of organisms on
biotic or abiotic conditions. For example, earthworms
might influence the selective environment experienced by
plants growing in the same soils, potentially leading to
covariance between the plant’s fitness and the worm’s
genes that underlie modifications to the soil environment
(Odling-Smee et al. 2003).
Clarifying the relationship between environment-
modifying activities of organisms and fitness variation
has been controversial throughout the development of
niche construction theory (Dawkins 2004, Laland and
Sterelny 2006). Dawkins (2004) argues that the buildup
of covariance between fitness and phenotype is much
more likely to occur within a gene pool, consistent with
the idea of an extended phenotype (Dawkins 1982),
rather than across gene pools (Dawkins 2004). In the
case of an extended phenotype, the phenotypic trait that
underlies the organism-mediated modifications of the
environment must vary within a population, have a
genetic basis, and be the target of the altered selection
regime caused by the environmental modifications
(Dawkins 2004, Brodie 2005). For example, genetically
based variation among gall wasps in their ability to
construct oak galls c an affect rate s of par asitoid
infection in the next generation of gall wasps, leading
to a covariance between gall forming traits and offspring
fitness (Bailey et al. 2009). While not disputing the
importance of extended phenotypes, niche construction
theory (Odling-Smee et al. 2003) argues that the traits
underlying specific environmental modifications neither
need to have a strong genetic basis (for example, they
can be acquired characters) nor need to be the same
traits that develop strong associations with fitness.
Hence, compared to Dawkins (2004), Odling-Smee et
al. (2003, 2013) consider a broader range of selective
agents that can potentially drive evolution, and suggest
that covariance between fitness and phenotype can
frequently build up across s pecies, resulting from
organism-mediated modifications to both biotic and
abiotic environmental conditions. While empirical data
and theoretical work are increasingly supporting this
view (Kerr et al. 1999, Odling-Sme e et al. 2003,
Krakauer et al. 2009, Laland and Boogert 2010, Kylafis
and Loreau 2011), the ongoing challenge is to determine
how much of the variance in fitness of one organism can
be explained by organism-mediated environmental
modifications compared to other agents of selection.
Since its inception, niche construction theory has
captured the attention of a wide range of evolutionary
biologists, ecologists, and philosophers (Erwin 2008,
Lehmann 2008, K rakauer et al. 2009, Post and
Palkovacs 2009, Loreau 2010, Kylafis and Loreau
2011, Van Dyken and Wade 2012), but has also
provoked considerable debate as to its novelty (Brodie
2005), scope (Okasha 2005, Kylafis and Loreau 2008),
and usefulness (Dawkins 2004). Niche construction has
been defined with a deliberately broad scope (Laland
and Sterelny 2006), and this has offered ecologists new
insights about how modifications to the environment by
organisms might persist o ver time (e.g., ecological
inheritance), result from byproducts and acquired
characters (Odling-Smee et al. 2003), and interact with
other environmental sources of selection so as to
influence evolutionary change in natural populations
(Odling-Smee et al. 2013).
While generally received sympathetically, the broad
scope of niche construction theory has nonetheless led to
some confusion and conflicts about how aspects of the
theory are positioned in relation to other closely related
ideas in both ecology and evolution. For example, the
concept of reciprocal interactions between organisms
and their selective environments is both fundamental to
niche construction theory and long-established in some
areas of standard evolutionary theory (Fisher 1930,
Roughgarden 1976, Crespi 2004, Frank 2009), particu-
larly in classic work on coevolution and diffuse
coevolution (Thompson 2005, Haloin and Strauss
2008). In ecology, there is also some uncertainty about
precisely what new insights niche construction theory
can offer. On the one hand, niche construction theory
has already made important contributions to emerging
syntheses between ecological and evolutionary dynamics
(Fussmann et al. 2007, Kokko and Lo
´
pez-Sepulcre 2007,
Post and Palkovacs 2009, Kylafis and Loreau 2011,
Matthews et al. 2011b, Schoener 2011). In particular,
niche construction research has documented a broad
range of organism-mediated environmental modifica-
tions that can influence selection pressures (Odling-Smee
et al. 2003). With the growing realization that ecological
and evolutionary timescales can be congruent (Hairston
et al. 2005, Ellner et al. 2011), such environmental
modifications might turn out to be more important
BLAKE MATTHEWS ET AL.246
Ecological Mo nographs
Vol. 84, No. 2
CONCEPTS &SYNTHESIS
agents of selection and drivers of evolutionary change
than previously thought (Odling-Smee et al. 2013). On
the other hand, the precise relationship between niche
construction theory and eco-evolutionary dynamics is
unclear, and t here is co nfus ion about h ow niche
construction is related to other ecological concepts in
general, and to ecosystem engineering in particular
(Odling-Smee et al. 2003, Boogert et al. 2006, Pearce
2011). Even though ecosystem engineering theory clearly
recognizes that the engineering effects of organisms can
have important evolutionary consequences (Jones et al.
1994), the strict definitions of ecosystem engineering
(Jones et al. 1994, 1997) and niche construction (Odling-
Smee et al. 1996, 2003) refer to distinct concepts.
In ou r view, niche construction th eory has the
potential to bridge many related concepts in ecology,
evolution, and ecosystem science. With the goal of
integration in mind, Odling-Smee et al. (2013) recently
distinguished between two important ‘‘aspects’’ of the
process of niche construction. The first aspect is the
environment-altering activities of organisms, and the
second is the subsequent modification of the selective
environment (Odling-Smee et al. 1996, 2003, 2013).
Niche construction is only present if both aspects occur,
as not all environmenta l modification s will a lter
selection pressures. Similarly, not all changes to
selection pressures will cause an evolutionary response,
meaning that niche construction can occur without
influencing evolution. In order to evaluate the impor-
tance of evolution by organism-mediated environmental
modification in natural populations, we need to
translate niche construction theory into empirical
practice (Odling-Smee et al. 2013). To do this, we
propose the following criteria to test for the presence of
niche construction (Criteria 1 and 2) and determine
when niche construction affects evolution (Criterion 3).
1) An organism (i.e., a candidate niche constructor)
must significantly modify environmental conditions.
2) The organism-mediated environmental modifications
must influence selection pressures on a recipient of
niche construction.
3) There must be a detectable evolutionary response in a
recipient of niche construction that is caused by the
environmental modification of the niche constructor.
Here, we refer to the environment in relation to both
biotic and abiotic characteristics, and the selective
environment as the environmental context in which
natural selection occurs. The first two criteria define the
term niche construction (Odling-Smee et al. 2013). The
organism changing the environmental conditions is only
classified as a niche constructor if Criterion 2 is satisfied.
The third criterion is a test of evolution by niche
construction, or in other words, evolution via selection
that is mediated by organismal modification of the
environment. We consider an evolutionary response as a
genetic change in a population that alters the relation-
ship between the phenotype distribution (including
mean, variance, and other moments of the distribution)
and fitness variation. We distinguish between a niche
constructor and a recipient of niche construction, but
explicitly recognize that both can refer to the same
organism. For example, in the case of an extended
phenotype, the niche constructor and recipient of niche
construction would be organisms within the same gene
pool, wherea s in t he ca se of a n envi ronmen tally
mediated genotypic association, the niche constructor
and recipient could be different species.
Using these three criteria we can evaluate which sets
of ecological and evolutionary interactions describe
evolution by niche construction, and which do not. We
summarize this approach graphically in Fig. 1 where we
consider a wide range of scenarios in which organisms
are connected with their biotic and abiotic environment
via pathways of evolutionary (dashed arrows) and non-
evolutionary (solid arrows) effects. Evolutionary effects
are those cases where organisms cause an evolutionary
respo nse (e.g., Criterion 3), while non-evolutionary
effects include the effects o rganisms have on the
abundance, distribution, and behavior of interacting
biota (e.g., collectively referred to as ecological effects),
as well as effects on the physical (e.g., engineering
effects) and chemical state of their environment (Crite-
rion 1; Fig. 2A). For a particular scenario in Fig. 1 to
satisfy evolution by niche construction (i.e., the mini-
mum condition for satisfying Criterion 3), the pathway
of effects must start (from the left) with a niche
constructor, it must include at least two sequential
effects (i.e., connections in sequence along the pathway
of effects), and there must be an evolutionary effect
beyond the first effect. This last condition follows from
our second criterion, which requires selection pressures
to be mediated through some form of environmental
modification by an organism, including changes to either
abiotic or biotic conditions (Fig. 2). Evolution by niche
construction does not occur for scenarios where the
evolutionary response of an organism is caused solely by
the direct selection effects of another organism or by an
environmental condition that is unmodified by another
organism. Such scenarios are examples of evolution, but
not of evolution by niche construction ( Fig. 1).
Following our scheme, there are many simple modules
of ecological interactions that do not meet all three
criteria (Fig. 1, modules within the ecology box but
outside the evolution box). This highlights that there is
considerable scope for ecologists to use niche construc-
tion theory to help integrate evolution and ecosystem
ecology. To facilitate this, we clarify how niche
construction (Criteria 1 and 2) and evolution by niche
construction (Criterion 3) are related to several key
concepts, such as: ecosystem engineering, (diffuse) co-
evolution, and eco-evolutionary dynamics and feed-
backs.
Ecosystem engineering.—The distinction between eco-
system engineering and niche construction is currently
unclear in the literature (Boogert et al. 2006, Post and
May 2014 247UNDER NICHE CONSTRUCTION
CONCEPTS &SYNTHESIS
Palkovacs 2009, Pearce 2011, Odling-Smee et al. 2013).
Ecosystem engineers are organisms that modify their
physical surroundings (e.g., light environment, physical
habitat structure) so as to modulate the availability of
resources or energy fluxes in an ecosystem (Jones et al.
1994, 1997). By comparison, niche constructors are
organisms that alter selection pressures of a recipient
organism by modifying any aspect of the abiotic and
biotic environment (Fig. 2). Evidence of ecosystem
engineering would only satisfy our first criterion, and
would not provide evidence of niche construction.
Nevertheless, ecosystem engineers are excellent candi-
dates for being niche constructors because their effects
on the physical environment can propagate to influence
chemical fluxes and species interactions, and cause
ecosystem effects that are large, multidimensional, and
persistent (Wright and Jones 2006, Hastings et al. 2007,
Jones 2012). Ecosystem engineering is hence a putative
mechanism of niche construction, and further work
should focus on the how engineers might alter selection
pressures on themselves or on other species (Criterion 2).
Coevolution and diffuse coevolution.—Based on our
criteria and schematic (Fig. 1), all examples of pairwise
coevolution and diffuse coevolution are examples of
evolution by niche construction. Pairwise coevolution is
the situation where two interacting organisms are both
niche constructors and recipients of niche construction
(Fig. 2B) and they both drive reciprocal evolutionary
FIG. 1. A Venn diagram showing which modules of biotic (square) and abiotic (circles) entities, which are connected by
evolutionary (dashed lines) and non-evolutionary effects (solid lines), are associated with different major concepts in ecology and
evolution (bounded by labeled shaded boxes). Non-evolutionary effects include organism-mediated effects on both biotic and
abiotic conditions (e.g., ecological effects shown in Fig. 2A), and evolutionary effects include evolutionary responses to selection.
The stars denote effects on the physical state of the abiotic environment, to distinguish ecosystem engineering (yellow box) from
effects on other abiotic conditions (e.g., the chemical environment). The minimum condition for evolution by niche construction to
occur is to have a pathway that starts and ends with an organism (i.e., a niche constructor and a recipient of niche construction),
and has at least two connections with an evolutionary effect beyond the first connection. Starting from the left of each pathway the
red dashed arrow defines where evolution by niche construction has occurred.
BLAKE MATTHEWS ET AL.248
Ecological Mo nographs
Vol. 84, No. 2
CONCEPTS &SYNTHESIS
responses on one another. Diffuse coevolution is the
case where a niche constructor drives an evolutionary
response of a recipient that is a different species, and
where this response is mediated through the niche
constructors’s ecological or evolutionary effect on
another species that interacts with the recipient (Haloin
and Str auss 2008). Hence, diffuse co-evolution is
equivalent to evolution by niche construction where
the selective environment is modified by species interac-
tions in the community. In sum, compared to all forms
of coevolution, evolution by niche construction consid-
ers a broader range of potential agents of selection and
effect pathways that underlie evolutionary responses
(Fig. 1).
Eco-evolutionary dynamics.—The emerging field of
eco-evolutionary dynamics has a very broad focus that
includes both the ecological and evolutionary responses
of populations to interactions between organisms and
their environment (Fussmann et al. 2007, Urban et al.
2008, Post and Palkovacs 2009, Matthews et al. 2011b,
Schoener 2011). Eco-evolutionary dynamics grew out
the recognition that population dynamics and pheno-
typic evolution can occur on similar timescales, leading
to an important contemporary interplay between
evolutionary and ecological dynamics in natural popu-
lations (Thompson 1998, Hairston et al. 2005, Schoener
2011). Evolution by niche construction is closely related
to eco-evolutionary dynamics but the two concepts have
slightly different emphases and are distinguishable in
our schematic (Fig. 1). Although the distinction is often
likely to be subtle, it is useful to identify the minimum
conditions that constitute each process in order to
perform more targeted experimental tests of the specific
mechanisms. Eco-evolutionary dynamic scenarios must
include at least two organisms and at least one
evolutionary and one ecological effect (i.e., a non-
evolutionary effect terminating with a biotic recipient).
Neither of these two conditions is necessary for
evolution by niche construction.
Following our scheme, there are simple cases of
evolution by niche construction that do not constitute
eco-evolutionary dynamics, and vice versa (Fig. 1).
Unlike eco-evolutionary dynamics, evolution by niche
construction includes scenarios made up of entirely
evolutionary effects (Fig. 1), including linked chains of
evolutionary effects (e.g., evolutionary cascades) and
reciprocal evolutionary effects (e.g., coevolution). In
addition, evolution by niche constru ction include s
simple scenarios where an evolutionary effect follows
from an organism’s effect on abiotic environmental
conditions. In relation to Fig. 1, for example, worms
(square) can modify (solid arrow) the soil environment
(circle) and affect the evolution (dashed arrow) of plants
(square). Such chains of interactions where abiotic
modifications influence selection pressures are an
important emphasis of niche construction theory (Od-
ling-Smee et al. 2013) but in their simplest form can fall
outside the domain of eco-evolutionary dynamics (Fig.
1).
Eco-evolutionary dynamics scenarios can also occur
without evolution by niche construction. In relation to
Fig. 1, for example, a predator (square) may cause an
evolutionary response (dashed arrow) in the life history
of a prey population (circle) that subsequently changes
prey consumption rates (solid arrow) on a resource
(circle). This is illustrated by recent work showing that
alewives, a common planktivorous fish in freshwater
lakes of eastern North America, drive evolution in
Daphnia in a way that alters their grazing rates on
FIG. 2. (A) A partitioning of how organisms can modify
their biotic and abiotic environments. (B) An elaboration of
how organism mediated environmental modifications can affect
the fitness of another organism (e.g., potentially a recipient of
niche construction), through a variety of pathways (abbreviated
following Fig. 2A: P, physical; C, chemical; CR, consumer–
resource; NTD, non-trophic direct). Niche construction can
occur when organism-mediated environmental modifications
alter the evolutionary response of organisms relative to other
environmental drivers of selection (e.g., unmodifiable environ-
ment).
May 2014 249UNDER NICHE CONSTRUCTION
CONCEPTS &SYNTHESIS
