Evolutionary ecology of plant defences against herbivores


  • Marc T. J. Johnson

    Corresponding author
      Correspondence author. Department of Biology, University of Toronto Mississauga, Mississauga, Ontario, L5L 1C6, Canada. E-mail: marc.johnson@utoronto.ca
    Search for more papers by this author

Correspondence author. Department of Biology, University of Toronto Mississauga, Mississauga, Ontario, L5L 1C6, Canada. E-mail: marc.johnson@utoronto.ca


1. Herbivores consume a large portion of the biomass produced by plants in virtually all ecosystems, which has dramatic effects on both the ecology and evolution of plants. In response to this threat, plants have evolved a diverse arsenal of direct and indirect defences to reduce herbivory and the impacts of damage on plant performance.

2. This special feature is a broad synthesis of the evolution and ecology of plant defences. The first objective of this special feature is to provide a review of what we have learned about plant defences against herbivores. The second objective is to stimulate debate and sow fresh ideas for the future research.

3. The 11 articles in this issue address three fundamental questions: (i) How do plants defend themselves against a diverse array of enemies? (ii) Why do plant species vary in defence? And (iii) What are the ecological and ecosystem-level consequences of plant defence? In addressing these questions the articles cover the interdisciplinary nature of plant–herbivore evolutionary ecology, from genes to global patterns.

4. The articles contained in the special feature question existing paradigms and provide new analyses of data. In some cases, influential hypotheses are firmly supported with new analyses (e.g. the Resource Availability Hypothesis), whereas in other instances conventional wisdom is called into question (e.g. the importance of secondary metabolites in the microevolution of resistance) and popular hypotheses are rejected (e.g. the Apparency Hypothesis, the Latitudinal Biotic Interaction Hypothesis).

5. This is an exciting time for research on the evolutionary ecology of plant defences. The articles in this special feature provide a guide to how we can move forward in resolving existing problems and tackling new questions.


The interactions between plants and herbivores are among the most important ecological interactions in nature. Approximately three-quarters of earth’s extant macroscopic biodiversity and biomass is a plant or an herbivorous arthropod, nematode, mollusk or vertebrate (Strong, Lawton & Southwood 1984; Wilson 2001) (Fig. 1). These herbivores consume over 15% of the biomass produced annually in temperate and tropical ecosystems making herbivory the major conduit by which energy enters food webs (Cyr & Pace 1993; Agrawal 2011). In response to the perennial threat and loss of productivity, plants have evolved an arsenal of defences during the ≈ 410 million years that they have been consumed by herbivores (Ehrlich & Raven 1964; Labandeira 2007). These defences play a critical role in shaping the interactions between plants and herbivores, such that understanding the evolution and ecology of plant defences is tantamount to understanding the origin and functioning of extant ecosystems. An additional major incentive to studying plant defences against herbivores is the fact that herbivory costs world economies billions of dollars each year in lost revenue related to agriculture, horticulture and forestry (Oerke & Dehne 2004).

Figure 1.

 The diversity of damage caused by insects. (a) Insect herbivory on leaves of Liquidambar styraciflua (photo credit: M. Johnson). (b) Root herbivory on milkweeds caused by Tetraopes larva (photo credit: S. Rasmann). (c) Black rhinoceros (photo credit: J. Bailey) and (d) flightless takahe feeding on herbaceous vegetation (photo credit: M. Johnson). (e) Lesion nematodes penetrating a root (photo credit: U. Zunke, Nemapix Vol. 1). (f) Effects of periwinkle snail (Littoraria irrorata) grazing on coastal saltmarsh vegetation; vegetation only remains within the snail exclosure. (Inset) A snail feeding on Spartina (photo credit: B. Silliman).

This special feature of Functional Ecology represents the first broad synthesis on the evolution and ecology of plant defences against herbivores in nearly two decades (Fig. 2). The previous synthesis was organized by Fritz & Simms (1992) and published in their edited book ‘Plant Resistance to Herbivores and Pathogens’. On a personal level, their book served as an invaluable resource to me as a beginning Ph.D. student, providing a crash course to the field of plant–herbivore interactions and a treasure trove of unexplored ideas. Although there have been influential books (e.g. Karban & Baldwin 1997; Herrera & Pellmyr 2002) and many specific reviews published in the intervening years (Table S1, Supporting information), there has been no recent review that covers the breadth of current research on the evolution and ecology of plant defences against herbivores. This is a large and important gap since much has happened since 1992, as evidenced by the ≈12,000 peer-reviewed articles published on plant–herbivore interactions in the intervening years (Fig. 2). Filling this gap is the first objective of this special feature.

Figure 2.

 The number of articles published on plant defences against herbivores as determined from ISI’s Web of Science (WoS) (a) “Title” and (b) “Topic” search engine. After trying multiple combinations of words, the combination “(defens* OR defenc* OR resistan*) AND plant* AND (herbivor* OR insect* OR arthropod* OR mollus* OR nematod* OR vertebrat* OR mammal*)”, returned the highest number of articles that directly related to plant defences against herbivores. The (a) “Title” search identified articles in which only the title contained the search words listed above. This figure allows for the assessment of how the relative number of publications relating to plant defences against herbivores has change during the past 110 years (1900–2010). For example, the title search suggests that there have been 1·8 X more publications on plant defences against herbivores since Fritz & Simms (1992), compared to the number of articles published on this topic between 1900 and 1992. The (b) “Topic” search is more comprehensive because it searches for the above word combination in titles, abstracts, “author keywords” and “keyword’s plus”, however, WoS seldom includes the latter three fields prior to 1991, such that only the last 20 years could reliably be compared.

The second objective of the special feature is to stimulate debate and future research. Since 1992, research has led to breakthroughs on conceptual issues, applied problems and methodological approaches. Some questions have been resolved during this time, such as: Whether plants use induced responses as adaptive defences against herbivores (Karban & Baldwin 1997; Agrawal 1998; Baldwin 1998); how plants recognize damage via plant hormone signaling (e.g. ethylene, salicylic acid and especially jasmonic acid) (Heil 2009; Wu & Baldwin 2009; Agrawal 2011); the conditions under which overcompensation can evolve in response to herbivory (Juenger & Lennartsson 2000) and the ecological consequences of genetic variation (Whitham et al. 2006; Genung et al. 2011; Hersch-Green, Turley & Johnson, in press). Other problems remain unresolved: Such as why plant species vary so dramatically in their levels and effectiveness of defences (Stamp 2003; Endara & Coley 2011); whether plant volatiles have evolved as indirect defensive mechanisms (Dicke & Baldwin 2010; Kessler & Heil 2011); and the relative roles of various plant traits (e.g. chemical resistance, tolerance, growth rate) in defence against herbivores (Agrawal 2011; Carmona, Lajeunesse & Johnson 2011). These issues remain important problems for future research. At the same time, developments in genetics, chemistry, phylogenetics, statistics and computing are enabling us to answer new questions that were intractable 20 or even 5 years ago. This is an exciting time for research on the evolutionary ecology of plant defences and the articles published in this issue provide a guide to how we can resolve existing problems and tackle new questions.

Scope of the special feature

Modern research on the evolutionary ecology of plant defences focuses on three general questions: (i) How do plants defend themselves, from the perception of damage to the production of defences, against the many generalist and specialist herbivores that attack them? (ii) Why do plant species vary in defence? and (iii) What are the ecological and ecosystem-level consequences of plant defence? In this issue, 11 leading research groups consider the ‘state-of-the-art’ on specific topics that address these questions. A deliberate effort was made to include reviews that spanned the interdisciplinary range of approaches used to study plant defence, from genetics and genomics, to chemistry and physiology, to community ecology, ecosystem sciences and global patterns of herbivory and defence. A further effort was made to review topics that are under active investigation and have lacked recent synthesis, or to include reviews that provide a novel perspective on cutting edge problems. Furthermore, because a major objective of this issue is to stimulate debate – and ultimately new research – authors were encouraged to be bold in their conclusions and opinions of ongoing scholarly debates and controversies. The papers contained in the special feature thus question existing paradigms and provide new analyses of data. In some cases, influential hypotheses are confirmed (Endara & Coley 2011), while in other instances popular hypotheses and conventional wisdom are called into question (Carmona, Lajeunesse & Johnson 2011; Salminen & Karonen 2011; Endara & Coley 2011; Moles et al. 2011). Below I summarize how the eleven articles contained in this issue provide a guide to the implementation of new methods, highlight unanswered questions, and offer new predictions to be tested.

How do plants defend themselves against a diversity of enemies?

Embracing the genomic revolution

Arguably the most important development in the study of plant defences against herbivores relates to advances in molecular biology. Molecular methods have revolutionized virtually all areas of basic and applied biology. For example, it is now relatively straight forward to sequence the genome of any organism, or to sequence every transcribed gene (i.e. RNA) within a tissue using second generation sequencing technologies (Hudson 2008; Morozova & Marra 2008). There have been equally impressive advances in the ability to genetically modify organisms or to transiently alter the expression of specific genes and pathways (Berenbaum & Zangerl 2008; Rasmann & Agrawal 2009). There also exist many tools that allow us to study natural variation in genes, their expression, and how they are related to the perception and deterrence of herbivores. Advances in molecular biology have been so large, and they have come so quickly, that many of us are at a loss of what it all means and how we can implement these techniques into our own research. In this issue, Anderson & Mitchell-Olds (2011) review and demystify these advances, providing a handbook for how molecular methods, bioinformatic tools and statistics can be used to study the evolutionary ecology of plant defences using the most recent molecular methods.

New approaches to test old questions in chemical ecology

Since the early pioneering work of chemical ecologists (Dunstan & Henry 1901; Dethier 1941; Fraenkel 1959; Hartmann 2008) (Fig. 2), it has been appreciated that plants employ a diverse array of chemical defences against herbivores. Tannins are among the most ubiquitous group of chemical defences produced by plants (Xie & Dixon 2005; Rausher 2006; Gross 2008). As such, the study of variation in tannins among plant species, as well as their effects on herbivores, has been incorporated into some of the most influential plant defence theories (Feeny 1976; Bryant, Chapin & Klein 1983; Coley, Bryant & Chapin 1985; Herms & Mattson 1992; Appel 1993). Salminen & Karonen (2011) review our understanding of the role of tannins in defence and they argue that the commonly proposed defensive function of tannins – precipitating proteins within herbivore guts – has little biological relevance to most arthropod herbivores. They review evidence for the oxidative stress hypothesis (Appel 1993) and argue that perhaps the most important defensive role of tannins relates to the oxidative capacity of a poorly studied class of tannins – ellagitannins. Unfortunately, tests of the oxidative stress hypothesis have been hindered by a lack of easily employable methods. To address this gap, Salminen & Karonen (2011) provide a new method that will enable researchers to examine the oxidative capacity of phenolic compounds (including all types of tannins) in many plant tissues. This method will allow virtually any laboratory to test the oxidative stress hypothesis (Appel 1993) and, combined with the recommended profiling of specific phenolic compounds (Salminen & Karonen 2011), the method has the ability to quickly transform our understanding of the role of tannins in plant defence.

Induced responses to herbivory

Although some plant defence traits are constitutively expressed, it is now recognized that plants can actively perceive damage and induce responses to reduce herbivory and its effects on plant fitness. Since Karban & Baldwin’s (1997) landmark book on induced responses to herbivory, ISI’s Web of Science identifies over 800 articles that have been published on the ecology, evolution and mechanistic basis for induced responses to herbivory. Here, Karban (2011) reviews progress since his influential book and highlights questions that remain unresolved. For example, he argues that we still do not understand the evolutionary forces that select for adaptive induced responses, the role of priming and volatile communication in mounting fast responses to damage, or whether induced responses prevent herbivores from evolving counter-defensive strategies.

Indirect defences

In addition to the many direct defences that plants use to protect themselves from herbivores, recruitment of predators and parasitoids (i.e. the third trophic-level) can act as indirect defences. However, the adaptive significance of indirect defences has been the subject of much debate (Van Der Meijden & Klinkhamer 2000; Allison & Hare 2009; Dicke & Baldwin 2010). In this issue, Kessler & Heil (2011) illustrate that this debate is partially resolved once we recognize that plants attract predators and parasitoids via two general mechanisms: (i) Resource based indirect defences rely on the plant’s production of food (e.g. sugars, fat bodies, nesting sites, etc.) to attract natural enemies of herbivores, whereas (ii) information based indirect defences involve production of volatiles following herbivory. After considering evidence for these two strategies, the authors conclude that resource based indirect defences are likely adaptive in most cases, whereas information based defences may rarely be adaptive.

Comparing the relative importance of resistance traits

Identifying and understanding the plant traits involved in defence against herbivores has been the foundation to basic and applied research on plant defences throughout the history of the field. Although secondary metabolites have long been identified as the most important traits involved in defence against herbivores (Fraenkel 1959; Ehrlich & Raven 1964; Berenbaum & Zangerl 2008; Hartmann 2008), the relative importance of secondary metabolites compared to other types of traits (e.g. physical resistance traits) have not been evaluated. Carmona, Lajeunesse & Johnson (2011) ask the simple question: What plant traits predict resistance against herbivores? Their meta-analysis of over 70 quantitative genetic studies shows that the answer to this question is anything but simple. In a microevolutionary context (i.e. within species variation), secondary metabolites do not predict resistance against herbivores ‘on average’. Instead, variation in life-history, morphology and physical leaf traits explained the most variation in resistance. The authors offer a new conceptual model for the evolution of chemical defence – secondary metabolites evolve as important defences not because they play a major role in resistance, but because the evolutionary constraints acting on these traits are relatively weak compared to other plant traits. These authors argue that a more pluralistic approach is needed to understand the mechanisms of plant defence against herbivores.

Why do plant species vary in defence?

Understanding the ecological and evolutionary processes that lead to variation in defence among plant species has been the most difficult challenge in the study of plant defence evolution. Dozens of hypotheses have been proposed (Feeny 1976; Coley, Bryant & Chapin 1985; Bryant, Chapin & Klein 1983; Herms & Mattson 1992; Berenbaum 1995; Stamp 2003; Agrawal 2007; Johnson, Smith & Rausher 2009; Vannette & Hunter 2011), but few of these hypotheses have proven to be general across systems. This special feature considers some of the most promising and popular hypotheses thought to explain variation in defence (Endara & Coley 2011; Moles et al. 2011), as well as emerging areas of study that provide new hope in explaining patterns of defence (Rasmann et al. 2011).

Below-ground defences along successional gradients

Research on plant defences against herbivores has focused primarily on the defence of above-ground tissues, while defence against the rich below-ground fauna of herbivores has garnered attention only in recent years (Van Der Putten 2003; Wardle et al. 2004; Heil 2011). Because of this disparity, there exist large gaps in our knowledge about the ecology and evolution of the defence of roots and other below-ground plant structures. For example, we lack answers to even some of the most basic questions, such as how much below-ground tissue is consumed by herbivores? Are root defences regulated independently of above-ground defences? How much can plants tolerate below-ground herbivory? Rasmann et al. (2011) review recent progress in understanding below-ground defences and argue that environmental changes along successional gradients will drive important changes in herbivory and defences. They make specific predictions about how direct and indirect mechanisms of defence are expected to change along successional gradients. Testing these predictions could fruitfully consume multiple research programs for the next decade.

Defences and herbivory along latitudinal gradients

Several prominent hypotheses on the evolution of biodiversity along latitudinal gradients predict that the amount of herbivory and investment in plant defence should increase with decreasing latitude (Dobzhansky 1950; Coley & Aide 1991; Pennings & Silliman 2005). Despite the popularity of these hypotheses, and numerous studies of herbivory around the globe, these hypotheses have not been closely scrutinized. Moles et al. (2011) use meta-analyses to examine whether herbivory, and multiple direct (e.g. thorns, secondary metabolite concentrations) and indirect (e.g. extra-floral nectaries) resistance traits, vary with latitude. Contrary to predictions, they find that herbivory and defence do not strongly covary with latitude. These results reject existing theory proposed to explain latitudinal patterns of species interactions and plant defence, and the authors suggest new theories are required to explain ecological and evolutionary patterns in herbivory and defence at large biogeographical scales.

Revisting hypotheses of plant defence evolution

The two most influential hypotheses proposed to explain the evolution of variation in plant defence among species are Feeny’s (1976)‘Apparency Hypothesis’ and Coley, Bryant & Chapin (1985)‘Resource Availability Hypothesis’. Despite these papers being cited over 3000 times (c. 1500 citations each according to Google Scholar), there has been no synthesis of the evidence in support or against these hypotheses. Endara & Coley (2011) rectify this situation by performing a meta-analysis of data on variation in plant apparency, plant growth rate, levels of herbivory, investment in secondary metabolites, physical resistance traits and different physiological and life-history traits. They find strong support for multiple predictions of the Resource Availability Hypothesis across multiple temperate and tropical ecosystems. By contrast, they find little support for the Apparency Hypothesis because herbivory was highest on fast growing ‘unapparent’ plants and lowest on slow-growing ‘apparent’ plants; the Apparency Hypothesis predicts either no difference in herbivory among apparent and unapparent plants or the opposite pattern. Thus, Endara & Coley’s (2011) evaluation of the existing data suggest that plant apparency is no longer a tenable hypothesis, whereas the resource availability hypothesis consistently explains macroevolutionary patterns of plant defence.

What are the ecological and ecosystem-level consequences of plant defence?

A co-evolutionary approach to plant tolerance

Tolerance is the ability for plants to maintain fitness in the presence of herbivory, and it is recognized as one of the major mechanisms of defence against herbivores. Unlike resistance traits, however, tolerance is traditionally thought to have little effect on the performance and fitness of herbivores, and therefore it is expected to have little role in a co-evolutionary context. In light of recent micro and macroevolutionary empirical results (Stinchcombe 2002; Agrawal & Fishbein 2008; Utsumi, Ando & Ohgushi 2009), Fornoni (2011) questions this assumption and argues that traits related to tolerance can and do impose selection on herbivore traits. To understand the evolution of tolerance in plant defence, Fornoni (2011) argues that we must examine the wider ecological consequences of tolerance. He further predicts that tolerance against herbivory may play an important role in enabling plants to invade new environments, which can subsequently impact the structure and dynamics of large complex communities.

Feedbacks between the ecology and evolution of plant defence

In the capstone article of Fritz & Simms (1992), Janis Antonovics (1992) called for the formation of a new discipline –‘community genetics’. It was his vision that community genetics would encompass ‘the study of the genetics of species interactions and their ecological and evolutionary consequences’ (Antonovics 1992). Although his ideas were initially slow to take hold, there has been an explosion of interest in community genetics since 2003 (Agrawal 2003; Whitham et al. 2006; Johnson & Stinchcombe 2007; Wade 2007; Hughes et al. 2008; Whitham et al. 2008). Here, Genung et al. (2011) consider the progress that has been made in understanding the ecological and ecosystem-level consequences of genetic variation and evolution within populations, and they consider how these effects can feed back to drive the evolution of multiple populations of plants and herbivores within communities. They present a novel conceptual and experimental framework that future studies can use to understand the dynamic interplay between the ecology and evolution of plant–herbivore interactions and to address unanswered questions.

Towards a modern synthesis of the evolutionary ecology of plant defences against herbivores

As stated at the outset of this Editorial, the objectives of this special feature are to review progress that has been made in the last two decades and to chart a course for the next decade of research on plant defences against herbivores. In the capstone article of this feature, Agrawal (2011) ties together the breadth of topics covered to achieve both of these objectives. He considers the progress we have made toward answering the three fundamental questions of the field (see above), and he considers how future research can combine micro and macroevolutionary conceptual and methodological approaches to forge new ground in the study of plant–herbivore interactions. It is clear from Agrawal’s capstone article, as well as from the other articles in the special feature, that recent advances have removed many limitations to the problems we can solve on the evolutionary ecology of plant defence. Interdisciplinary collaborations, long-term experiments, and substantial investment from government funding agencies and the private sector will enable us to advance the field more in the next 10 years than were accomplished in the previous century.

A final note to students and researchers new to the study of plant–herbivore interactions: Although this issue includes a breadth of articles that span from genes to global patterns of defence and herbivory, it was impossible to adequately review all of the progress that has been made. The study of plant–herbivore interactions is now so large and interdisciplinary that we have only scratched the surface. However, thanks to other recent reviews and a newly published book (Walters 2011), a near comprehensive synthesis of the field is now in hand for the benefit of new students and established researchers. I have included a supplemental list of recent reviews pertaining to the concepts and methods used to address the three questions outlined above (Table S1). The articles published in the special feature, combined with the reviews outlined in Table S1, offer a roadmap of how the field has progressed and where the next generation of research can take us in understanding the evolutionary ecology of plant defences against herbivores.


I thank the contributors and reviewers who helped bring this special feature to fruition and to Liz Baker who provided editorial guidance. I am especially indebted to Chuck Fox, whose advice, guidance and editorial oversight has improved every aspect of the issue. I thank A. Agrawal, C. Fox, N. Haddad, A. Moles, J.-P. Salminen, M. Turcotte and N. Turley for providing constructive feedback on earlier versions of this Editorial. This work was partially supported by National Science Foundation grants DEB-0919869 and DEB-0950486.