Introduction
AGROECOLOGY’S STRENGTH is in its interdisciplinary nature. Since its inception, agroecology
has embraced that multiple disciplines should engage with each other to holistically
design and analyze agricultural systems. Rather than putting plant pathology, weed
science, and soil management, for instance, into separate fields with little opportunity
for mutual influence, agroecology has integrated those different disciplines (e.g.,
Carroll et al. 1990). Importantly, agroecology also has a strong history of seeing the social dynamics
of agriculture as essential to understanding the biology. This reciprocal relationship
between humans and the environment is core to agroecology’s pursuits. Thus, agroecology
is not only a production system, but also a movement that encourages the questioning
of industrial agriculture and that helps shape the values of citizens who approach
agriculture ecologically (Wezel et al. 2009; Méndez et al. 2013).
The space that agroecology provides for study is inherently broad-from basic ecology
to social sciences and activism (Wezel et al. 2009; Méndez et al. 2013). Yet it is easy to name typical areas of study within agroecology where basic studies
are then applied to improve production and sustainability of the system. For instance,
within the biological realm, knowledge of plant physiological responses to aspects
of the abiotic environment (light, temperature, nutrients, etc.) can be applied to
optimize crop growth; the understanding of soil type and quality can be applied to
soil health and plant nutrition; exploration of plant competition and facilitation
can improve weed control and intercropping; a better understanding of the ecology
and population biology of pests can be applied to control them; consideration of the
flows of matter and energy within and between systems can be used to ascertain system
sustainability; and understanding of the role of biodiversity and biological interactions
in system can be used to improve ecosystem function. On the social side, understanding
social processes in rural and urban settings can help illuminate social determinants
of characteristics of our food and farming systems. Exploration of the politics and
role of agribusiness in shaping food and farming systems can identify points of leverage
for making change.
Many of these areas of research, especially those in the biological realm, mirror
important areas in the discipline of ecology. What differs is that they are applied
to agricultural systems and not natural ones. Additionally, biological work in agroecology
explicitly incorporates the human element, whereas not all ecological studies do (though
now they do more and more: Alberti et al. 2003). Nevertheless, what tends to be uncommon within agroecology is an evolutionary perspective.
Perhaps, agroecology will need to grow to encompass evolution as happened to ecology
(e.g., Bradshaw 1984; Antonovics 1992; Purugganan and Gibson 2003).
Yet evolutionary biology has been partner to ecology for a while. In fact, in academia,
many departments dominated by ecologists are called some version of “Ecology and Evolutionary
Biology,” giving equal weight to ecological and evolutionary lines of research. In
the past half a century, a field that bridges the two, namely evolutionary ecology,
has also emerged (the journal so-named had its first issue in 1987). In ecology and
evolutionary biology departments, genetics has become an important tool for investigating
ecological and evolutionary processes. Yet despite the importance of plant breeding
in agriculture and all the molecular and quantitative genetic work that has been performed
on our crops, genetics is less well-integrated into the holistic view of agroecology,
where methods tend to emphasize whole plant responses, soil microbial communities,
and fluxes and flows of nutrients, among others. This lack of engagement with genetics
exists even though many agroecology textbooks have a chapter on genetic resources
or genetics (e.g., Carroll et al. 1990; Gliessman 2006) that may mention concepts such as adaptation and genetic variation, especially as
applied to crops. Despite the inclusion of these evolutionary and genetic concepts
in seminal works, little of the literature on genetics or genetic resources claim
agroecology as its true home. Perhaps this may be due, in part, to a lack of interest
in evolutionary work by the agroecological community due to the perception that evolutionary
change requires very long periods of time that misalign with the annual management
timeline of agriculture.
Others have noted this lack of evolutionary emphasis in agroecology and multiple authors
have highlighted the value that an evolutionary perspective can bring to agriculture.
For instance, Denison et al. (2003) has argued that the only way for plant breeding to continue to enhance yield is to
focus on problems that natural selection has not already optimized. By taking this
view, problems such as increasing individual yield or photosynthetic efficiency may
seem unworthy of breeders’ time, while those such as increasing group fitness or resistance
to evolving pests are ripe for investment. Similarly, Weiner et al. (2010) argue that breeding for increased total yield at the expense of individual plant
performance can be combined with knowledge of ecological responses to crop planting
density to improve crop management systems-an approach they call evolutionary agroecology.
As mentioned above, the agroecological canon contains chapters on genetic resources
(Gliessman 2006) and ecological genetics (Carroll et al. 1990), though these perspectives rarely infuse agroecology.
Therefore, I argue that although agroecology’s very strength is its interdisciplinary
nature, nevertheless it can be seen as insufficiently so because evolutionary and
genetic perspectives receive little attention. In this paper, I aim to delve a bit
deeper into what encompasses evolutionary research, elaborate on the evolutionary
work that has been carried out in agriculture, and highlight some fruitful areas of
research that could broaden agroecology’s reach and import if embraced by the community.
In doing so, we can introduce evolutionary agroecology into the agroecological pantheon.
Areas of research in evolution
It can be helpful to clarify what areas of research are encompassed within evolutionary
biology as relevant to agricultural systems. These nodes of study can focus on long
or short term processes, encompass a range of scales, be oriented towards molecular
or quantitative genetic analyses, and explore ecological processes to differing degrees.
Macroevolutionary topics of research within evolution tend to incorporate long term
perspectives. Such work might explore the creation, extinction, and diversity of species
and their relationships to one another. For instance, some seek to understand the
relationships of species within and among plant families (especially those that include
important crop species), as well as variation within families for important traits
(Meyer et al. 2012). This long term perspective sometimes integrates an understanding of factors, such
as biogeography, that affect the generation of biodiversity.
Others studies tend to have a short term, microevolutionary perspective that can span
from millennia (which is short on an evolutionary scale) to rapid evolution that occurs
over even shorter periods of time. In fact, agriculture has become an important system
in which to study the effects of rapid evolution because intense management can drive
evolutionary processes (e.g., Gould 1991). In these microevolutionary studies, major forces of evolution are of explicit interest,
namely selection (and subsequent adaptation), gene flow, drift, and mutation (Silverton
and Lovett Doust 1993). Many researchers employ modeling techniques of various sorts
to understand the interplay between different evolutionary processes (e.g., Yeaman and Whitlock 2011).
Researchers may try to find evidence (phenotypic and/or molecular) of the effects
of these evolutionary processes in the resulting patterns of quantitative or molecular
genetic variation, including in crop species (Mercer et al. 2008; Samberg et al. 2013; Meyer et al. 2016). These patterns can provide clues into evolutionary forces that have shaped the
species. Molecular aspects of this work have been enhanced by on-going developments
in genotyping (e.g., genotyping-by-sequencing), and evolutionary researchers now also
attempt to identify genes involved in various traits, especially those involved in
important adaptations (e.g., Fournier et al. 2011; Lasky et al. 2015). Often they ask questions about the identity and number of genes involved, what
their effects are on traits and fitness, how interdependent they may be, and how they
evolve.
While selection is often heavily affected by the abiotic and biotic conditions surrounding
a population, there are areas of evolutionary research that are more or less explicit
about links to the ecological context. For instance, in evolutionary ecology, that
ecological context is paramount. Evolutionary forces such as selection are studied
as they operate in the context of interactions of species with their environment and
with each other (Antonovics 1992). For instance, many have studied the rapid evolution that can follow gene flow in
hybrid zones and try to explore its dependence on the ecological context (Wu and Campbell 2006). Additionally, there would then also be an emphasis on the mechanisms of adaptation
that organisms employ to tolerate biotic and abiotic factors they face (Angert 2006). However, other areas of study that fall into the intersection of evolution and
ecology may not be typical of evolutionary ecology. For instance, some researchers
explore the importance of the genetic variation found within and among species as
a mechanism of tolerance to environmental variation or even as a booster of ecosystem
function (Kettenring et al. 2014).
This is not to say that agricultural researchers have not taken an evolutionary perspective
on their work. In fact, some veins of agricultural research are ripe with evolutionary
thought-especially those that have had to confront the issues that evolution can cause
in managed systems. The study of the development and spread of pesticide resistance
has a strong evolutionary history (e.g., Tabashnik 1994). Researchers have documented a multitude of cases of pesticide resistance and their
mechanisms (Gould 1991), investigated the forms and strengths of selection pressures being exerted (Metcalfe et al. 2000), clarified correlated selection responses on other traits, modeled selection responses
based on varied ecological contexts (Mallet and Porter 1992), and much more. Within plant pathology, especially, this evolutionary perspective
has extended to documenting the diversity within pathogen populations with the evolutionary
potential this confers (McDonald and Linde 2002), considering how to manage pathogens with the use of disease-resistant varieties,
and studying the evolutionary dynamics of pathogens that are developing new resistance
mechanisms (Stahl and Bishop 2000). However, the agroecology community has generally not embraced these themes. From
where does this division between those using genetics to explore evolutionary concepts
and the ecologically-oriented agroecological community arise? Perhaps agroecologists
see the intense selection pressures of pest control in conventional agricultural systems
that drive evolution as not applying to more sustainable systems? It is hard to know
whether that is true, due to the paucity of studies in this area. Yet I would argue
that there could be exciting synergies created by integration of the evolutionary
and ecological perspectives to improve our understanding of agricultural systems,
their potential, and their dynamics.
How an evolutionary perspective can broaden agroecology’s impacts
There are a number of areas of evolutionary study that engage with agriculture. However,
further engagement would broaden agroecology’s impact and reach. I will focus here
on elaborating three areas of research that have been primary in my laboratory, though
many others could be discussed. Here I explain the premise of each and how evolutionary
and genetic perspectives help us investigate essential agroecological issues. First,
use of an evolutionary perspective can illuminate the adaptations that crops possess,
which can be utilized in production or breeding for sustainable systems. Second, viewing
genetic diversity within crops and weeds through an evolutionary lens may improve
agroecosystem function, such as primary productivity, as well as adaptive capacity.
Third, explorations of gene flow or hybridization between neighboring crop fields,
between crops and neighboring related weed species, and between populations of weeds
can illuminate, and help to eliminate, weed issues.
Adaptations for sustainability
Agroecology takes as one of its goals, the sustainable design of cropping systems.
From a biological perspective, consideration of soil, pest, crop, and water management
are all of primary part of that, as well as landscape connectivity and fluxes and
flows in the system. An important part of this is also the choice of genetic materials.
Choice of genetic materials can be done at the level of an individual farmer saving
seed from the year before, getting seed from his or her neighbor, or purchasing improved
varieties. Choice of genetic materials can also occur at the level of breeding programs
identifying traits that are important to agroecosystem sustainability, as well as
sources for those traits. For instance, given the increasing variability in precipitation
with climate change, identification of sources of drought-and flooding-tolerance for
farmers to utilize, or to use in breeding programs, would be beneficial. But could
an evolutionary perspective improve this endeavor?
The evolutionary forces that operate on crops result in patterns of genetic variation,
some of which is adaptive, or specifically beneficial to plant performance. The relatively
recent evolutionary processes affecting crops include natural and intentional selection
by the environment and humans, respectively, which have left their marks on crop diversity
and crop traits (Mercer and Perales 2010). For instance, human selection has resulted in the domestication of many crops,
and the fixation of domestication traits (Meyer and Purugganan 2013) that facilitated human use and cultivation has provided great use-value to human
societies (Jardón-Barbolla 2015; Jardón-Barbolla, in this issue). The environment, on the other hand, has selected
for traits that improve productivity under conditions experienced by crops in a given
geographic area or agroecozone (Mercer et al. 2008; Anderson et al. 2016). Crops whose seeds are saved yearly, typically landraces or traditional varieties,
as well as wild relatives of crops, have evolved over time in response to these geographically
variable biotic and abiotic factors (e.g., moisture availability), often resulting
in local adaptation (Zeven 1998). Mechanisms for adaptation can be due to phenology or timing of development (e.g.,
flowering time), morphology (e.g., leaf size), physiology (e.g., stomatal density),
or other traits.
To best choose seed lots or varieties of crops that will fit well into a sustainable
agroecosystem, the farmer should consider the traits s/he wants the crop to have.
Particular climatic adaptations, such as drought tolerance, may be essential for boosting
the resilience of the agroecosystem to the vagaries of weather. For farmers seeking
out such traits in available varieties or for those who would like to breed crops
to enhance those same traits, identification of relevant crop varieties or breeding
material is essential.
There are a few types of evolutionary studies that might be useful where relevant
adaptations need to be identified in existing seed lots or in potential breeding materials.
In the first type of study, a researcher assesses the adaptation of diverse genetic
materials to relevant conditions. To do so, s/he will grow diverse genetic materials
under one or a range of common conditions to distinguish varieties that do well in
particular areas or under particular sets of conditions (e.g., Mercer et al. 2008; Orozco-Ramírez et al. 2014). However, relating the performance measured to other measured trait values allows
researchers to identify the traits conferring that adaptation (e.g., Etterson and Shaw 2001). These kinds of studies can be especially helpful to better understand the underlying
bases of climatic adaptation and to identify varieties that farmers might want to
use. For participatory or evolutionary breeding programs, this kind of information
can be helpful to evaluate different varieties, to choose varieties for crosses, and
to judge advanced materials (e.g., Bellon et al. 2003).
There is a second form of study that can identify crop materials with relevant adaptations
or find sources of genetic variants that may be beneficial to use in breeding programs
by using molecular genetic approaches. By genotyping crop materials at many or thousands
of loci throughout the genome, researchers can then look for relationships between
genetic variation and relevant phenotypes. For instance, to uncover loci that may
control drought tolerance, researchers could phenotype genetic materials for their
drought responses and then relate those phenotypes to genotypes (e.g., Lasky et al. 2015). The identified loci are then good candidates (or are linked to good candidates)
for influencing the traits of interest. Relating genotype to phenotype can be done
on a range of accessions using genome-wide association studies (GWAS) (Tiffin and Ross-Ibarra 2014; Meyer et al. 2016) or using populations generated from controlled crosses for quantitative trait locus
(QTL) analysis (Alimi et al. 2013a; Almeida et al. 2014). These kinds of analyses can also be done using, as phenotypes, the environmental
variables describing the locations where crop materials were sourced, such as those
describing temperatures and precipitation (Lasky et al. 2015; Anderson et al. 2016). By doing so, researchers can clarify the materials that possess genotypes likely
to produce phenotypes of interest. Both types of research can contribute to an understanding
of which crops are important within agroecosystems, how humans and the environment
have molded them, and what might be useful materials to deploy in sustainable cropping
systems. Combining the two may be our best bet.
Genetic diversity-ecosystem function
We rely on many ecosystem functions in agriculture. Many components of the agroecosystem
play important roles that improve the outcomes and properties of the system. While
primary production is of obvious interest, other functions such as pest resistance
and nutrient cycling are also essential. Within agroecology there has been an interest
in understanding how cultural practices and diversification of crop species can improve
ecosystem functions (Altieri 1999). For instance, many have explored how adding additional crop species into a system
can increase its resilience in the face of disturbances (Lin 2011; Jacobi et al. 2015; Liebman and Schulte 2015). Since agroecosystems in industrialized countries are often species poor, opportunities
for improvement abound in that context.
Yet crops utilized in industrialized settings (organic or sustainable farms included)
also tend to harbor low genetic diversity compared to their wild progenitors-a form
of diversity in plants that can also increase ecosystem function (Kettenring et al. 2014). Varieties of crops that farmers use have often been improved or bred intensively.
That breeding process itself, as well as the process of domestication that preceded
it, reduces genetic variation in a number of ways (Yamasaki et al. 2005). Domestication reduces genetic diversity (or the number of genetic variants or alleles)
at loci associated with domestication traits, such as lack of seed shattering (Meyer and Purugganan 2013). Thus, so- called domestication genes will often lack diversity in landraces or
modern varieties as compared to wild relatives of crops (Hufford et al. 2012). Modern plant breeding also reduces the diversity of alleles at loci associated
with improvement traits (Lam et al. 2010), such as those associated with yield or quality. Thus, improvement genes will be
less diverse in modern varieties than in landraces and be less diverse in landraces
than in wild relatives (Yamasaki et al. 2005). Transgenic varieties are an extreme of this since a new locus is “created” with
the insertion of a transgene and only one allele exists (the transgene), although
there is also the null allele (or lack of the transgene). Hybrid varieties, or varieties
that are created by crossing two inbred lines, each lacking diversity, also harbor
less genetic variation than the open-pollinated varieties that preceded them. Even
neutral genes, which are not selected upon, will likely lose some diversity during
domestication due to genetic bottlenecks imposed along the way. This comparative lack
of diversity throughout the genome of modern, hybrid (and sometimes transgenic) varieties
is unfortunate due to the benefits that genetic diversity can bring to ecosystems
(Kettenring et al. 2014). Nevertheless, open-pollinated varieties are still used by some in industrialized
countries (especially those who plan to save seed) and are common in the developing
world, especially in centers of crop origin where open-pollinated landraces are used.
In such contexts, farmers may be able to harness two benefits of genetic diversity
within crops: improved ecosystem function through ecological means and greater adaptability
and resilience through evolutionary processes in the face of climatic fluctuations.
Unfortunately, we know little about how genetic diversity within varieties affects
agroecosystem function within monocultures. However, by marrying ecological and evolutionary
perspectives (Kettenring et al. 2014), we are coming to understand that, in natural ecosystems, genetic diversity can
augment function in species-poor natural ecosystems in a number of ways. Genetic diversity
can increase annual productivity (Crutsinger et al. 2006), reproductive success (Williams 2001), and arthropod diversity (Bangert et al. 2005), while also augmenting resistance and resilience to disturbance (Hughes et al. 2004; Elhers et al. 2008). Genetic diversity affects ecosystem function via additive effects, such as
sampling effects, and non-additive effects, such as complementarity (Kettenring et al. 2014). For instance, after experiencing climatic extremes, plots with higher levels
of eelgrass genotypes had greater performance and abundance of fauna due to complementarity
between genotypes (some did better in mixture than in monoculture) (Reusch et al. 2005). We would do well to investigate these issues in agroecosystems.
In addition to exploiting the effect of genetic diversity on ecosystem function, genetic
diversity can also enable adaptive evolution in the face of changing conditions. For
landraces or open-pollinated varieties where seed is saved from year to year by farmers,
selection of various sorts operates on crop populations. The ways that selection can
affect a response in a plant population is a central question in evolutionary biology
(Lynch and Walsh 1998). Response to selection (R) is directly related to the amount of genetic variation that the population has
in the traits being selected upon (h2) and the strength of selection (s) (R = h2s: the breeders’ equation). Thus, for crop populations to adapt in the face of climate
change, they require variation in traits that increase their performance under the
new conditions, e.g., reduced precipitation (Mercer and Perales 2010). For example, in landraces of pearl millet grown in the Sahel, a recent, long-term
drought has caused rapid evolution of earlier flowering time (Vigoroux et al. 2011). Gene flow may also be key to facilitating adaptation to new conditions since
it can introduce novel alleles that are beneficial under new conditions (Davis and Shaw 2001; Mercer and Perales 2010), which can be accomplished through hybridization or seed sharing among crop populations
(Bellon et al. 2011). Ultimately, taking an evolutionary perspective here can help us understand how
to increase agroecosystem resilience as conditions shift.
Many studies in these areas could help agroecology embrace the importance of genetic
diversity within crops. Research comparing ecosystem functionality or adaptive potential
when more or less genetically diverse crop populations are used could further clarify
how adaptive evolution and genetic diversity-ecosystem function effects might operate
in agroecosystems. Unfortunately, simply comparing open-pollinated (OP) to hybrid
varieties often confounds genetic diversity with degree of breeding, thereby reducing
the apparent benefits of diversity (Campbell et al. 2014). Since we would expect a range in the amounts and pattern of genetic variation typical
of different crops based on their mating systems (i.e., sexual vs asexual, outcrossing vs self-pollinating) and the structure of any farmer seed-sharing networks, studies
could investigate what this means for the size of ecological and evolutionary responses
possible. Different agroecosystem functions may respond differently to genetic diversity,
and different traits would be experiencing different selection pressures, so studies
could explore the range of related responses. If these kinds of studies clarify benefits
of diversity, it would be prudent to broaden the genetic base of crops to augment
the resilience of individual crop populations, agroecosystems, and even agricultural
landscapes (via metapopulation dynamics: van Heerwaarden et al. 2010). Participatory plant breeding may be well-suited to play this role.
(Trans-)gene flow
With the advent of transgenic crops in the mid-1990s, the fact that crops have the
ability to hybridize or cross with their wild relatives became abundantly clear. The
new technologies raised concerns in the agricultural and ecological communities due
to the potential adverse effects that novel genes may have when out of place (Snow and Palma 1997). This issue galvanized the scientific and activist communities alike and spawned
a range of scientific studies that had a distinctly evolutionary perspective (reviewed
in Ellstrand 2004; Mercer et al. 2006; Campbell et al. 2007). It also highlighted the need to understand gene flow between crop fields (from
transgenic to non-transgenic crops) or among weeds (from those that received a transgene
from a crop to those that did not) (Beckie et al. 2003; Ellstrand 2004).
The issue of gene flow generally, and transgene flow as a specific case, fits within
evolutionary biology because gene flow is one of the fundamental evolutionary processes
shaping the genetics of populations. Gene flow can homogenize populations that hybridize
with one another, but it is also able to introduce novel genetic variants on which
selection can act (Lenormand 2002). Transgenes constitute a form of novel genetic variant (as opposed to a naturally
occurring mutation) and the planting of transgenic crops introduces to the landscape
a large source of crop alleles, as well as the novel transgene. In the case of gene
flow between transgenic crops and neighboring wild relatives, their hybridization
often results in crop-wild hybrids in the wild population (Burke et al. 2002). Thus, once gene flow has occurred, the questions around how the genes affect the
wild population become largely about selection: will the lineage initiated by this
crop-wild hybrid be maintained in the population, increase in frequency, or be lost?
Will selection, or just simply drift, be implicated? How will the particular ecological
context in which the wild population and its hybrids live affect this outcome (Mercer et al. 2014), especially given ecological differences between agricultural and wild environments?
The frequency of gene flow can also influence the outcomes (Ellstrand 2004). Here
it becomes clear that agroecology, which is intimately interested in the movement
of nutrients and pesticides off- farm, could encompass work that seeks to understand
the movement and effect of novel genes or crop variants of other genes in wild, weedy,
and crop contexts.
A number of methods are utilized to illuminate the gene flow and selection occurring
in crop-wild hybrid zones. While the possibility for gene flow can be estimated using
controlled hand-pollination, natural rates of gene flow can be documented at different
spatial scales (e.g., Arias et al. 1994; Watrud et al. 2004) using some indicator (molecular, phenotypic) to identify hybridization events. Hybrids
generated through gene flow can then be tested for their ability to survive and reproduce
to evaluate their fitness (e.g., Cummings et al. 2002; Mercer et al. 2006). When hybrid fitness is then compared to that of their wild counterparts, we can
begin to understand the selective dynamics affecting plants possessing crop alleles,
including hybrids of different advanced generations. Fitness at different parts of
the life cycle, such as the seed and seedling stages, can also clarify how selection
may change at different times (Weiss et al. 2013; Kost et al. 2014; Pace et al. 2015). In addition, selective dynamics can change under different ecological conditions,
affecting the likelihood of introgression of crop alleles. For instance, since competition
and herbicide application had been shown to affect the fitness of hybrids relative
to wilds (Mercer et al. 2007), studies manipulating different aspects of competition have elaborated on those
dynamics (Mercer et al. 2014). Thus, bringing an evolutionary perspective to bear on the question of gene flow
in transgenic crops has helped us address an important agroecological question concerning
the effects of a management tool (transgenic crops) on neighboring cultivated, wild,
and weedy neighbors via gene flow and selection.
Conclusions
Agroecology has not intentionally excluded evolutionary thinking, but it has not necessarily
embraced it. Both the agroecological and evolutionary communities alike must recognize
the people working at the intersection of the two. Better integration of an evolutionary
perspective into agroecology and greater representation of agriculture within evolution
is certain to inspire novel avenues for research, allowing for exciting synergies.
For instance, evolution of pest populations in the face of climate change is a major
issue facing agriculture. Agroecology can only gain if these evolutionary dimensions
of agroecosystem change are included when working to enhance the resilience of agroecosystems.
Hopefully a new generation of agroecologists will be inspired to work at this juncture,
cultivating the field we may come to call evolutionary agroecology.